New Regulatory Mechanisms for Cardiac and Smooth Muscle Thin

Filaments

Darren Brian Foster

A thesis submitted to the Department of Biochemistry in conformity with the requirements

for the degree of Doctor of Philosophy

Queen's University Kingston, Ontario, Canada

June 200 1 copyright Darren Bnan Foster 200 1

National Library 1+1 of Canada Bibliomue nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 Wellington Street 395. tue Wellington O(tawa ON KI A ON4 Otlawa ON KIA ON4 Canada Canada

The author has granted a non- exclusive licence ailowing the National Library of Canada to reproduce, loan, distribute or sel1 copies of this thesis in microfonn, paper or electronic formats.

The author retains ownership of the copyright in this thesis. Neither the thesis nor substantiai extracts ftom it rnay be prhted or otherwise reproduced without the author's permission.

L'auteur a accordé une licence non exclusive permettant a la Bibliothèque nationaie du Canada de reproduire, prêter, distriiuer ou vendre des copies de cette thèse sous la forme de microfiche/film, de reproduction sur papier ou sur format électronique.

L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.

Abstract

Poor coronary circulation (ischemia) is a leading cause of mortality and morbidity in

developed countries. Understanding the cellular processes that are affected during ischemia and

following adequate repehsion, will help the design of new treatments that might protect hearts of

patients at risk. It has been shown previously that myocardial shng, in the isoIated rat heart

model, is chamcterized by specitic TnI proteoiysis at its C-terminus, yielding tmcated TnT1-,,,.

(McDonough,J.L.,et al. ( 1999) Circ.Res. 84:9-20 ). Tmsgenic mice expressing Tn1,-,,, recapitulate

the hallmarks of stunning (Murphy et al. (2000) Science 287:48S-J9 1). To assess the molecular

basis of the contractile dystùnction, we have undertaken biochemical studies of human TnI,-,,2, the

homolog of mousdrat TnI,-,,,. Tnincated TnI binds TnT, TnC, actin and actin-tropomyosin with

comparabIe affinity to intact TnI, and is marginally iess effective at inhibiting actin-tropomyosin-

activated ATPase activity of myosin subfragment 1. Inhibition of actin-tropomyosin ATPase byT4-

,,? was more easiIy reversed by 'Troponin C. than was inhibition by fil1 Ien,d TnI. Troponin

reconstituted with TnI,-,,? displayed Ca"-sensitization of its Ca"-activated acrin-tropomyosin-

troponin-S 1 ATPase when cornpareci with wild type troponin. Furtherrnore, the mutant troponin

increased maximum CaL--activatecl ATPase by 80%. Rabbit ventricular cardiomyocytes, adenovirally

transfected with mouse Tnl,-,,, displayed increased ce11 shortenïng. Given the decreased maximum

isometric force observed in previous studies of stunning, truncation of the C-terminus of Tni may

cause force and ATPase to become uncoupled.

Abnonal smooth muscle contraction may contniute to diseases such as asthma and

hypertension. Alterations to myosin Iight chah kinase activity or phosphatase activity change the

phosphorylation level of the 20-kDa myosin regulatory light ch in (MRLC), increashg Ca'-

sensitivity and basai tone. One Rho family GTPase-dependent kinase, Rho-associated kinase (ROK

or pl60ROCK) can induce Ca2'-independent contraction of Triton-skinned smooth muscle by

phosphorylating MRLC andor myosin light chah phosphatase. We have shotvn that another Rho

family GTPase-dependent kinase, p21-activated protein kinase (PAK), induces Triton-skinned

smooth muscle contraction, independently of calcium, to 62 +/- 12% (n = 10) of the value observed

in presence of calcium. Remarkably, PAK and ROK use different molecular mechanisms to achieve

the Caz'-independent contraction. Like ROK and myosin light chah kinase, PAK phosphorylates

MRLC at serine 19 in vitro, However, PAK-induced contraction correlates with enhanced

phosphorylation of caldesmon and desmin but not MRLC- The level of MRLC phosphorylation

remains similar to that in relaxed muscIe fibers (absence of GST-mPAK3 and calcium), even as the

force induced by GST-mPAK3 increases h m 26 to 70%. Thus, PAK uncouples force generation

from MRLC phosphorylation, These data support a mode1 of PAK-induced contraction in which

myosin phosphorylation is at Ieast complemented through regdation of thin filament proteins.

Because ROK and PAK homologues are present in smooth muscle, they may work in parallel to

regulate smooth muscle contraction.

We have also sought to provide a sound biochemical basis for PAK-induced contraction by

undertaking biochemical studies of caldesmon phosphorylated by PAK. Mass spectroscopy data

showed that stoichiometric phosphorylation occurs at Ser657 and Ser687, abutting the calmodulin-

binding sites A and B of chicken gizzard caldesmon, respectively, Phosphorylation of Sa657 and

Ser687 has important t'unctional impacts on caldesmon. PAK-phosphorylation reduces binding of

caidesmon to calmodulin by about 10-fold wkle binding of calmoduiïn to cddesmon partially

inhibits PAKphosphorylation. Phosphorylated caldesmondisplays amodest reduction in affinity for

actin-tropomyosin but is significantly less effective than unphosphorylated caldesmon at inhibithg

actin-activated SI ATPase activity in the prwence of tropomyosin. We propose that CE?'-

independent phosphorylation of caldesmon at the calmodulin-binding sites, by PAK, parrially

converts caldesmon frorn an actin-myosin ATPase inhibitory 'off state to an non-inhïbitory 'on' state

which is tünctiondly sirnilar to the CaL'l caimodulin-caidesmon complex.

Acknowledgments

1 wouId like to thank rny supervisor, Dr. Jennifer E, Van Eyk, who provideci a nurturing and

stimulating environment in which it was a pleasure to work. Her confidence and encouragement has

allowed me to regain my strïde. My labmates, Kent Arrell, Jererny Simpson, Jason McDonough,

Lemy Organ, Irina Neverova, Raif Labugger, David Colantonio and Nina Buscemi were adeficto

family for over thtee years, with whom we experienced each others' triurnphs, tribulations and

neuroses. 1 wiil miss each of them. The counsel of Dr. John Strauss has proven invaluable to me.

1 wish to extend my thanks to my new supervisors, Dr. Albert Wang and Dr. William Lehman for

their continuing support and understanding whiIe 1 'played hooky' finishing this thesis. I am eterndly

gnteful to Rose Silva, for moving administrative mountains. Once again, 1 am t h a n h l for the

unconditional support of my parents, Ken and Diane Foster, who surely must have thought i had

decided to stay at Queen's until a new building was named in my honour. Finally, 1 am thankfiil for

the love and understanding of my wife, and bedrock. Diane Bovenkarnp.

Table of Contents

......*.....*.........*.... ..............................*..*...... Title ... i

.. .................................................................... Absiract u

............................................................... Table of Contents vi

............................................................... List of Figures ix

................................................................ Abbreviations x

...................................................... CHAPTER 1 : introduction 1

3 A . Overview of Muscle Contraction ................................................ ...................... 1 . Evolution of the Sliding Filament Mode1 of Contraction - 2

2 . Proteins of the Sarcomere: .............................................. - 5 ....................................................... 2.1 Myosin 5

........................................................ 2.2 Actin - 7 .......................... 3 . The Myosin-Actin Interaction, Crossbridge C ycling I O

............................................. B . Regdation of Crossbridge Cycling 13 1 . Ca2+ as an Initiator of Cardiac and Smooth Muscle Contraction ................. 13

..................... 2 . Regulation of the Crossbridge Cycle in Cardiac Muscle -14 ............................... 2.1. Cardiac Muscle Regulatory Proteins 14

........................................ 2.1.1. Tropomyosin -14 ................................ 2.1.2. The Troponin Complex -15

................................... 2.1.2.I.Tropon.inT 15

.................................. 2.1.2.2. Troponin C -16 .................................... 2.1.2.3. Troponin 1 17

............................... 2.2.1. The Steric Blocking Mode1 21 2.2.2. The 2-State Allostenc mode1 ............................ -22 2.2.3. Newer 3-state models reconciie steric and allosteric aspects of muscle

........................................... regdation 23 ................... 2.2.4 Summary of Sîriated Muscle Regulation -24

...................... . 3 Regdation of Crossbridge Cycling in Smooth Muscle -26 ..................................... 3.1 Thick Filament Regulation -26

...................................... 3.1.1 Myosin Revisited -26 ....................... 3.2 Thin Fiiament Candidate Regulatory Proteins -28

.......................................... 3.2.1 Tropomyosin 28

........................................... 3.2.2 Caldesmon 28 3. 3 Economical Force Maintenance: The 'Latch' State .................... 32

............................. 3.3.1 The Latchbridge Hypothesis -33

List of Figures . Figure 1- 1 . Structure of striated muscie ............................................. 4

. Figure 1.2 Structure of rnyosin ................................................... 8

. Figure 1.3 Structure of actin ......................... ,. ....................... - 9

. Figure 1.4 The crossbndge cycle ................................................ 11

. Figure 1.5 Troponin and tropomyosin on the thin filament ............................. 19

. Figure 1.6 Schematic diagram of the fhctional domains of Troponin 1 .................. 20 Figure 1.7.3.state models of striateci muscle regdation .............................. -25 Figure 1.8. CaZ+-dependent activation of myosin and smooth muscle contraction ............ 29

. Figure 1.9 Conformations of smooth muscle myosin in vitro ........................... 30 . Figure 1.10 Structure of caldesmon ............................................ 34, 35 . ....... Figure 1-1 1 Schematic representation of arterial smooth muscle response to agonist 35a . Figure 1.12 Mode1 of the pathogenesis of ischemia ceperfusion injury .................... 44

Figure 3.1 . Fragments of TnI used in this study ...................................... 78 Figure 3.2 . TnT-Sepharose chromatography ........................................ 85 Figure 3.3 . Ultracentrifiigation of actin and actin-tropomyosin ......................... 86 Figure 3.4 . Inhibition of actin-TM-S1 ATPase activity by T d and Tni,.,,. and its reversd by

the action of TnC .................................................... 87 Figure 3.5 . ca2'- sensitivity of reconstituted troponins ............................... 88a Figure 3.6 . Andysis of adenovirally-transfected rab bit cardiomyocytes ................... 90

Figure 4.1 . A PAK homologue is present in intact but not triton X-100 skinned smooth muscle ........................................................... IO7

Figure 4.2 . Constitutively active GST-mPAK3 induces Ca3-independent contraction of guinea pig taenia coli skinned muscle fibers without involving MLCK or myosin tight chab phosphatase ....................................................... 108

Figure 4.3 . Uncouphg between MRLC phosphorylation and force development in skinned muscle fibers even though isolateci myosin and MRLC cm be phosphorylated at serine 19 by GST-mPAK3 ..................................................... -111

Figure 4 4 . PAK and ROK phosphoryIate different proteins in s k i ~ e d muscle fibers ....... 112

Figure 5.1 . Phosphorylation of caldesmon by mPAK3 in the presence and absence of calmodulin. . 123 . Figure 5.2 Identification of phosphorylation sites in caidesmon incubated with mPAK3 .... 124 . Figure 5.3 Effect of phosphorylation on biiding of caldesmon to caimodulin ............ 126 . ...... Figure 5.4 Effect of phosphorylation on binding of cddesmon to actin.tropomyosin -127

Figure 5.5 . inhibition of actin-activated skeletal S1-ATPase activity by caldesmon and phosphorylated caldesmon ........................................ 129

Abbreviations

Da L mol M OC Ci cem x g h min s v k rn CI n

dalton litre mole molesflitre, molar degree celsius Curie counts per minute times the gravitational field hour minute second volt kilo milli micro nano

Chernicals / Reagents / Methods

ADP ATP ATPase DTT EDTA EGTA FPLC GTP HPLC IPTG LB MALDl MOPS MS MSMS PAGE PMSF PVDF QTOF SDS TBS TPCK Tris

adenosine 5'-diphosphate adenosine S'-triphosphate adenosine 5'-triphosphatase activity dithiothreitol, Cleland's Reagent ethylenediaminetetraacetic acid ethylene(oxyethylenenitrilo)te~raacetic acid fàst performance liquid chromatography guanosine 5'-triphosp hate high performance liquid chromatography isopropyl P-D-thiogalactopyranoside Luria-Bertani broth matrix-assisted laser desorption ionization 3-IN-morpholino]propanesul fonic acid mass spectrometry tandem mass spectrometry polyacrylamide gel electrophoresis phenylmethylsulfonyl Buonde polyvynilidene difluonde quadrupole time-of-flight sodium dodecyl sulfate Tris-buffered saline L-1 -tosyIamido-2-phenylethylchloromethyI ketone Tris(hydroxymethyl)arnino methane

Pro teins

CaM ELC GFP GST M MLCK MLCP MLC20 PAK RLC ROK S 1

TnT

Tris-buffered saiine containhg Tween-20 detergent 20-po1yoxyethylenesori1itan rnonolaurate

actin bovine serum alumin catdesmon a recombinant, bacterially expressed domain of human fibroblast 1-caldesmon spanning residues 244-536 a recombinant, bacterially expressed domain of human fibroblast 1-cddesmon spanning residues 1 - 1 52 calmodulin essential light chah green fluorescent protein glutathione-S-hmsterase myosin myosin Iight c h i n kinase myosin light chain phosphatase 20 kDa smooth muscle myosin light chain, regulatory Iight chain p2 f -activated kinase regdatory Iight chain rho-associated kinase myosin subfiagment 1, enzymatic head of rnyosin. derived by proteolysis of myosin with papain or chymotrypsin tropomyosin troponin troponin, reconstituted with Tni,-,, troponin 1, inhibitory subunit of troponin troponin 1, amino acids 1- 192(3) troponin Cl Cd!+-binding subunit of troponin troponin Tl tropornyosin binding subunit of troponin

CBAPTER 1: Introduction

The study of muscle is, after a fastiion, a study of al1 that animates us, and it has piqued the

curiosity of mankind throughout the centuries. In the third century B.C., the Greek, Erasistratus,

was the fIrst to proclaim that 'Spiritus Animalis' (the intrinsic property of living things) could be

attributed to muscle (1)- Nevdeiess, it was 500 years before the physician Galen (13 1-20 1 A.D.)

conducted the first systematic anatomical study of skeletat muscie, 'de Mottr hf~~sculorum'. As a

physician with the 'College of the Gladiators', it is nimored he had ample supply of fresh muscle

tissue',

As the age of reason hit Europe in the 16001s, scientific curiosity, largely dotmant since the

times of ancient Greece, prompted legitimate pursuits of anatomy and physiology- For the first time,

sirnilarities were observeci benveen the tissueattached to the skeieton and the tissue composition of

the hem as well as the blood vessels. History will remmber William Harvey(1575-16571, physician

to King James who, in his treatise "de Moni Cordism(3), detriiled how heart muscle was

responsible for pumping blood throughout the body. Some of Hmey's otherconclusions, (Le. that

the heart w u the sourceofspirits and that the heart was a apository for fementing blood) have been

best lefi in obscurity (4).

mile the function of skeletal muscle may have been obvious to the Greeks and the purpose

of the heart was (Iargely) deduced by Harvey, it took ionger to discem the fiinction of the smooth

muscle- Not only is it found in the vasculatm, but in the lining of the interna1 organs the ainvays

of the lungs. Smooth muscle, it would seem, is the unsung hero among muscles, since it ministers

to the tünction of our iife-sustainhg organs and the life-bearing female uterus, al1 in retative

I Creative Licence in the name oibumour. Most of Galen's studies were conducted on dogs. pi@ and apes. See (2)

obscurity - until recently.

in this chapter, 1 will begin by introducing the key concepts and mokcular basis of

contraction that are comrnon to boih smooth and cardiac muscle. 1 will descnïe how interactions

betweenactin and myosin, collectively known as thecrossbridge cycle, form the basis of conûaction.

The remainder of this chapter will serve to introduce cardiac muscle and smooth muscle, in tum. It

will become clear that, despite sharing a contractile mechanism, the two muscle types differ

dramatically in the marner by which that mechanism is regzilared. Regdation of the crossbridge

cycle is the central therne of this thesis. Specifically, this relates to questions of how smooth and

cardiac muscle contraction are iniriared and susrnined as well as how the muscles relar. Finaily,

while the primary moleculardeterminants of srnooth and cardiac muscle contraction have been well

elucidrited, new concepts are emerging about ancillaty mechanisms that serve to fine-tune muscle

regulation. [n this thesis, 1 present biochemicat evidence for two new mechanisrns of crossbridge

replation that directly alter the Ca"-sensitivify of contraction in cardiac and smooth muscle

respectively.

A. Overview of Muscle Contraction

1. Evolution of the Sliding Fiirment Model of Contraction.

The earliest detailed studies of muscle were matornical in nature and sought to compare and

classi@ muscle types. Heart muscle and skeletal muscle appeared morphologically simiIar, by Iight

microscopy, each bearing a pattern of alternating light and dark striations and were hence catled

striateci muscle (Figure l-L). By contras?, visceral and vascular muscle had a less organized

appearance and, using conventional techniques, stained more eveniy than striated muscle and was

therefore temed "smooth" muscle, Perhaps it was the regular, array-iike morphology of striated

muscle that captured researchers' interest, but there cm be no doubt that advances in determining

the mechanism of muscle contraction were driven primacily by work on striated muscle, and in

particular skeletal muscle. Under a polacizing microscope, skeletal muscle appeared to be made of

repeating subunits called sarcomeres (L. sarkos, flesh + menrs, unmixed). The limits of the

sarcomere were bounded by thin lines called the Z-Iines. The Z lines are flanked by areas that stain

only lightly, called the [-bands (isotropie) . Adjacent to the I-bands, are dark staining region called

the A-bands (anisotropic). Dunng contraction, the length ofthe sarcomeres, or the distance between

the Z Iines, diminished owing to the near disappearance of the 1 bands (Figure 1-1).

By the early 1940's the structure of muscle had been extensively studied by light rnicroscopy.

Understanding how muscle worked, however, would require new tools provided by the (then)

fledgling fields of biochemistry and biophysics. On the biochemistry front, Straub, working with

Albert Szent-Gyorgyi, discovered that the primary muscle protein 'myosin' identified by Kuhn in

1864, was in fact two proteins, which they renarned actin and myosin. Szent-Gyorgyi dernonstrated

the essential role of actin and myosin in adenosine triposphate-induced contraction using

giycerinatedmuscle fibers (discusseâ in (5)) . Work on the biophysics of muscle contraction was led

by H.E. Huxley, a hard core physicist by training, who had an interest in muscle. His speciality was

X-ray difilaction and he assiduously applied his talents by bombarding pieces of muscle tissue with

X-rays and looking at the equatorial diffraction patterns of the sarcomeres (Figure 1-1). Prior to

Huxley's work, it was believed that myosin was responsible for contraction of muscle and that it

operated as an ATP-dependent spcing. However, Huxley's X-ray diffraction showed that as muscles

were allowed to contract in buffer coutaining ATP, the dif ic t ion patterns changed, and were most

easily explained in tems of the interaction of intadigitating parailel rods or tilaments. Huxley

envisioned a contractiIe mechanism in whkh inextendde actin filaments floated behveen theZ-lines

and extensible or spring-like myosin filaments were attacheci to the Z-lines.

However, Hwdey's collaboration with Jean Hanson caused them to reevaluate his mode!.

Through a series of studies they realized that myosin, not actin, was the predorninant protein of the

A-band, and ttiat actin was the protein attached to the Z-lines (contrary to Huxley's original

assignment, (6)) . Furthermore, they determined that a sprïng-Iike structure for either myosin or actin

was not necessary to explain ATP-induced contraction. Sarcomere shortening could be explained

by the sliding of the filaments past each other. This could be accomplished by the repeated

attachrnent of myosin. movement of the myosin relative to the fixed actin, subsequent detachment

and so on (7;s). In the sections that follow, I will review our current understanding of the molecular

rnechanisrn that underpins the sliding filament theory. But let us first acquaint ourselves ~ 4 t h the cast

of molecules around which this drarna is based.

2. Proteins of the Sarcomere: Myosin and Actin

2.1 Myosin

MuscIe contraction requires energy. Specifically. chemical energy rnust be converted to

mechanical energy to perform work. In cardiac and smooth musclealike, this task falIs to the protein,

myosin. It has aptly been temed a mechanoenzyme, or molecular motor, by virtue of its ability to

harness the free energy of A ï T hydroIysis to drive its interaction with, and rnovement of, actin

filaments.

Muscle myosin, or myosin II (Figure 1-?A), is a hexameric protein composed of two long

polypeptide chains (caILed myosin heavy chains or MHCs), and two sets of two smallerpolypeptides

(called the regdatory and essential light chains respectively, RLC and ELC). The amino-terminal

(MI,-terminal) portion of the heavy chauis form globular structures about L 10 A long, that house

cataiytic sites for ATP hydrolysis and binding sites for actin. The carboxy-texminal (COOH-terminal)

ends of the heavy chains intertwine, to form a 13-11., extended coiled-coi1 rod , or tail, that

effectively provides scaffolding for the gIobuIar enzymatic heads. The region between the N-terminal

globular 'head' and C-terminal 'tail' is called the neck region, and consists of a long a -helix that is

stabilized though the binding light chairis, RLC and ELC (or DTNB chah).

The myosin head region cm be proteolytically cleaved from the rod region to yield a soluble

enzymatic subfragment of myosin called S 1. The two principal functions of myosin head are actin-

binding and ATP hydrolysis. The x-ray crystallographic structure of Sl has been determined at a

resolution of 2.8 A, and provides an excellent starting point from which to describe the salienr

features and functions of myosin Figure1 -2B (9). S 1 is highly asymmetrical, 165A long, 65 A wide

and with a thickness of about 40A . The actin binding site lies at the d i s d end of the head region.

Residues that are critical for the docking of myosin to actin are located in a glycine- and lysine-rich

flexibIe loop (Gly627-Phe646) that is not visible in the structure. The ATP binding site is Iocated

on the opposite site of the S 1 head from the actin binding site. This dornain is structurally similar

to the nucleotide-binding pockets of other proteins. including the p21" GTPase and adenyiate

kinase. Finally. the COOH-terminal residues of the head region tom a long helix (85 A) that

extends fiom the middle of the head outward, toward the upper rod or neck region of myosin. This

long helix would not otherwise exist if it were not for the stabilizing effects of the b o n d light

chahs.

Rayment et ai. proposeci that the actin-binding and ATP-binding sites communicate via

conformational changes through a hingeregion buried within the myosin head. They conjectured that

closure of the ATP-binding pocket, whiIst the S 1 head is attached to actin, might account for a 45"

change in the neck region, which has long been proposeci to act as a lever arm (1 0). The change in

the orientationof the Iever arm produces the power stroke that causes movement of myosin relative

to the aciin fiaments, as predicted by the sliding filament theory of contraction.

Since the publication of the structure, research groups have constnicted MHC and MLC

chimeras, in which actin-binding sites ftom cardiac muscle myosin have b e n substituted into

skeletai MHC, and vice-versa. By combining biochernical and x-ray structural analysis of these

constnrcts, researchers will learn how events such as actin binding and ATP hydroiysis are coupled

to the conformational changes in the myosin head that give rise to force. A detailed consideration

of these studies is beyond the scope of this text but the reader is referred to several excellent

examples (1 1 ; 12).

2.2 Actin

Actin has been conserved throughout evolution, from the lowest eukaryotes to human kind.

its f ic t ion is primordial. It does no less than provide the structural scaffolding for the cell. In

muscle it is the structure onto which myosin p p p l e s during contraction. In mammals, six isoelectric

variants have been isolated, each encoded by aseparate gene. Their arnino acid sequence reveais that

the sequences are 95% identical and sequence diversence is contined primarily to the NHrterminai

six residues of the protein. Muscle, whether skeletal or smooth, is composed primarily ofhvo actin

isofomis, a or y , each of which is equally able to perform its pnmary function during contraction -

activation of myosin ATPase (see (13) for a review). As a result, biochemists use skeletal and

srnooth muscle actin interchangeably.

In vivo, actin is preponderantly polymenc, forms Iong double stranded filaments, and has

thetefore been termed F-actin. In vitro, actin can assume two forms. Under conditions of Iow salt,

in the absence of Mg"', the actin is a monomeric 42 kDa globular protein, cailed G-actin. X-ray

crystaibgraphy has shown that G-actin is bi-lobed and has a cleft running d o m its middle, which

houses sites for stoichiomemcaily bound cal- and ADP (14) (Figure 1-3A)- Each Iobe or domain

LMM HMM l

motor domain

myosin head

Uppcr S0K I30mniii

Figure 1-2. Structure of myosin. A. Schemaîic depiction of the structure of myosin II and its domains. B. Crystai stmcture chicken skeletal muscIe myosin S 1 (9), as presented in (15).

Su bdomain Subdomain 4 2

Subdomain Subdomain 3 1

Figure 1-3. Structure of actin. A. Atomic structure of G-actin obtained by X-ray crystallography of Actin-DNAsel CO-crystals (modified fiom (14)). The position of the subdomains is shown. B (lefi) 3D-image reconstruction of F-actin fïiaments observed by etectron microscopy (Courtesy of Dr. Wiam Lehman). (right) By fitting atomic CO-ordimates fiom Panel A with the 3D-image reconstruction, this provides an atomic mode1 for F-actin (reprinted fiom (15)).

may be further subdivided into two subdornains. The smaller of the domains, the N-teminal half of

the molecule, houses subdomains t and 2 while the Iarger C-terminal domain houses subdomains

3 and 4. The N- and C-tennini are in close proximity in the region of subdomain t . This region is

a docking site for myosin.

Under conditions of high salt and M$', actin monomers polymerize to fom long doubie-

helical strands or filaments, as found in vivo. The actin strands twist around each other with a

periodicity of 36 nm. Along the cables, the N-terminal domain (i-e. subdomains I and 2) is

predominantly exposed to the solvent and is designated the "outer dornain", The C-terminal domain

(subdomains 3 and 4 ) is on the inside and forms in inter-strand contacts. In muscle, which is repiete

with divalent metal cations and whose ionic strength is L 70 mM. actin assumes the filamentous form

and there is little free monomeric actin (Figure 1-3B)-

3. The Myosk-Actin Interaction, Crossbridge Cycling

Ultimately, validation of the sliding firament mode1 proposeci by Huxley and Hanson would

require that one could: 1) demonstrate direct evidence for the interaction of the actin and myosin

filaments via a 'crossbridge' and 2) demonstrate that myosin heads could repeatedly interact with

actin dong its Iength in amannerthat was consistent with the energetics ofATP consumption, Direct

interaction of actin and myosin was first observed by eiectron microscopy (16), images that were

subsequently improved upon by methods of helical reconstniction (17). The nature of that binding

and its coupling to ATP hydroiysis began, and continues, to be reveaied through solution

biochemistry using kinetic assays of ATP hydroIysis, and protein bindiig assays. CoIIectively these

data descnie the "crossbridge cycle" that underpins the slidinp filament modeI.

A simpiified mode1 of the crossbndge cycIe is shown in Figure 1-4 (see(l8))- in step 1,

myosin (indicated as the S 1 head) dissociates h m actin in the presence of ATP, In step 2 of this

Figure 14. The crossbridge cycle. Step 1, dissociation of myosin-ATP. Step 2, hydrolysis of ATP. Step 3, isomerism. Step 4, release of inorganic phosphate. Step 5, displacement of actin. Force production stems from the interconversion of 'weakly- bound' states of actomyosin (e.g. prior to step 3, bottom left corner) and 'strongly-bound' states (e.g. foilowing step 3 and prior to step 1). Note that isomerism of myosin need not necessarily occur whilst bound to actin (omitted for simplicity). Whether isomerism is t d y the rate-determinhg step in this cycle continues to be debated (reprinted fiom (18))

cycle, myosin hydrolyzes the teminal phosphoanhydride bond of ATP yet retains each of the

products, ADP and inorganic phosphate (Pi). The myosin-ADP-Pi complex undergoes a slow

isomerism to a conformation that both has high affinity for actin and is conformationally strained

or poised (myosin*ADP-Pi, step3). The release of Pi (step 4) provides the energy needed to hel the

large confornational change in the myosin head that can move actin filaments up to 1 I nrn (step 5).

ADP can then dissociate, which allows ATP to bind and start the cycle anew. Crossbndge cycIing

is the molecular mechanism that underlies Huxley's sliding filament model of force production.

Biochemists use measurements of Pi, or H' release, resulting from actin-myosin adenosine

triphosphatase (ATPase) activity. as an index of crossbridge cycling in vitro.

Though the model conveys the basic tèatures of the crossbridge cycle, it is noteworthy that

each of the transitions depicted in Figure 1-4, is not strictly unidirectional, and there are additional

equilibtia that rnay factor into a complete picture of crossbridge cycle. Researchers have classified

the many states ofactomyosin into nvo broad categories, weakly bound, non-force-producing, and

strongly bound. force-ptoducing states. Myosin is considered to be in a non-force-producing state

if it is bound to ATP, or ADP-Pi.Though not depicted in Figure 1-4, myosin-ATP and myosin-ADP-

Pi complexes cm bind to actin, albeit weakiy. Myosin is considered to be in a force-producing state

when bound to ADP or when it assumes a myosin*-ADP-Pi conformation. These conformations

interact strongly with actin. Therefore, force production may be considered to be the resuit of the

interplay between populations ofweakly bound and strongly bound states of actomyosin (see (l9;ZO)

for detailed discussion).

Given our understanding of the crossbridge cycle mentioned above, one can appreciate that

force generation in muscIe is dependent on two parameters: I) the nurnber of attached crossbridges

and 2) the rate at which the crossbridges cycle between the weakIy- and strongiy-bond states.

Therefore, any mechanism that served to increase either parameter would favour force genention.

Conversely, any mechamanism that decrease either of the these parameters would favour muscle

relaxation. In the pages that follow, I will discuss how smooth and cardiac muscle have developed

different strategies to affect the extent and rate of crossbndge cycling.

B. Regulation of Crossbridge Cycling

Though the pmcess that underlies force genemtion, crossbridge cycling, is common to

cardiac and smoothmuscle, the processes that regulatecontraction and relaxation in each are distinct.

Crossbridge regulation, in cardiac and smooth muscle, is highly dynarnic and exquisitely tine tuned

to reflect the hnction of each muscle type. A heart m u t contract and relax quickly and repeatedly

to deliver blood to body tissues. Visceral and vascular smooth muscle rarely relaxes tùlly and must

sustain contractions in response to different hormonal signais. Nature has therefore provided

regulatory mechanisms that retlect the characteristics of each type of muscle contraction. Cardiac

muscle regulation bears passing resemblance to a molecularsivit~h~ while smooth muscle regulation

is more &in to a rheoscnr. Thou& imperfect, these analogies provide a usehl starting point for a

detailed discussion. in this section 1 will briefly outline the role oCC~!' as the intracellular trigger

for contraction in both cardiac and srnooth muscle. Cardiac and srnooth muscle reguIatory

mechanisms will be discussed in tum. As in the section on crossbridge cycling, regulatory proteins

will be introduced before descniing their proposed mode of regulation.

1. Ca2* as an Initiator of Cardiac and Smooth Muscle Contraction

in the body, each muscle type c m be induced to contract either by electrochemicai stimuli

(Le- action potentids) stndor phannacochemical stunuIi (the action of hormones). Within the

muscle, stimuIator-y signals are relayed to the proteins of the myofilaments through secondary

messengers, the most important of which is Ca2-.

Cardiac muscle contracts rhythrnicaiiy in response to action potentials. Depolarization ofthe

cardiomyocyte membrane causes an influx of Ca2- fiom the extracellular space. Increased

intracellular calcium, in tum, ûiggers tùrther Ca" release from the sarcoplasmic reticulum. The

resultant CaL' level is sufficient to ensure binding of Ca" to the troponin complex, which initiates

contraction that pumps blood out of the heart (systole). Levels of intracellular Cd' fa11 quickly

however, as it is sequestered in the sarcoplasmic reticulum via a Ca2'-~TPase pump, and extruded

through the plasma membrane by the NaTlCa'.' exchanger. This ensures that the muscle relaxes and

that the heart fills with blood (diastole).

Smooth muscle contraction can also be initiated from an action potential, and intracellular

Ca" can be elevated through the process of Ca2'-induced Ca" release fiom the sarcoplasmic

reticulum as outlined in the previous paragraph. However, smooth muscle contraction is highly

influenced by the action of hormones. For exarnple, hormone stimulation of G-protein-coupled

receptors can cause activation of phosphoiipase C. Phospholipase C catalyzes hydrolysis of

phosphatidyl inositol 4,5-bisphosphate (PIP,) into diacyl glycerol (DAG) and inositol 1.4,s-

trisphosphate (IP,). In hun, IP3 binds to an [P, receptor on the sarcoplasmic reticulum which causes

release of Ca" into the cytoplasm (See (21) for a review). Elevated Ca" leads to srnooth muscle

contraction by mggering myosin phosphorylation which wil1 be discussed in section B.3.1.1.

2. Regulation of the Crossbridge Cycle in Cardiac Muscle

2.1. Cardiac Muscle Regdatory Proteùis

2.1.1. Tropomyosùi.

The primary fiinction of tropomyosin (TM) in cardiac muscle is to regulate accessîbility of

myosin-binding sites on actin. Its precise position on the azimuthai axis of actin is controlled by

troponin complex in response to Ca" (discussed in detail in section B.2.1.2). Tropomyosin is a 40

nrn-long coiled coil a-helical protein that lies on thin filaments along the long pitch helical array of

actin monomers (i.e the actin "strandsn)(17). Tropomyosin molecules associate together as a strand,

binding end to end, overlapping slightly(2nm), in a head-to-tail manner, such that each tropomyosin

spans 7 actin monomers along the filaments (38 nm). Tropomyosin is a modular protein with 7

quasi-repeating motifs, each designed to bind successive actin monomers (22;23).

Tropomyosin is actua1ly a dimer of two 33 kDa peptides that form a coiled coil. Cardiac

tropomyosin exists predominantly in two forms: ua homodimers and ap heterodimers. The ratio of

aa to ap forms of tropomyosin varies between the species, ranging fiom up to 50% ap in large

mammals with slow heartbeats Iike canle, to nearly 100% a a i n small animals with faster hembeats.

Iike the rabbit. [nterestingly, the u and P chains migrate differently on SDS-PAGE despite the high

sequence homology and conserved length behveen the two peptide chains (24).

2.1.2. The Troponin Complex

Troponin controls tropornyosin rnovernent across the azimuthal axis of actin in response to

Ca2'. It is anchored to the thin fiIament via its interaction with a specific region of tropomyosin.

Consequently, troponin assumes the same periodicity as tropomyosin. binding the thin filament every

38 nm. The length of a single troponin complex (26.5 nm) is shorter than tropomyosin, thus each

troponin extends over about two thirds of each successive tropomyosin molecule. Troponin is

composed of three proteins; Troponin T (TnT), Troponin 1 (Tni) and Troponin C (TnC).

2.1.2.1. Troponin T

TnT is fiequently descnibed as the anchor that holds the troponin complex on the thin

filament. It has a molecularweight of 34.5 kDa and possesses an extended N-terminal (Tl) domain

that binds to tropomyosin, and a gIobular C-terminal domain (T2) that binds to troponin C and

troponin 1. At 18.5 nrn long, TnT spans the head to tail overlap region of tropomyosin, and extends

over the C-terminal half of tropomyosin (25). The TnT-TM interaction specifies the stoichiornetry

of the troponin complex on the thin filament. (1 Tn : 1 TM : 7 actin monomers).

TnT, however, is more than a static anchor for troponin; it has an important functional role.

Specifically, TnT is required for Tn to confer maximum Ca"-dependent ATPase activity. in other

words, TnT enhances ATPase above that of simple Actin-TM-S 1 in vitro (26). It is beIieved that

TnT imparts a conformation to actin, or a position to tropomyosin, that favors the interaction of the

rnyosin head with the filament (27). As such, TnT may viewed be as a crucial effector of TM

movement, or a relay protein that couples binding of Ca" to Tn with TM movement.

2.1.2.2. Troponin C

Reversible Ca2'-binding to TnC is the event that initiates cardiac muscle contraction. What

follows is a host of intra- and intermolecularevents that ultirnately alters the position uftropomyosin

on actin and facilitates crossbridge cycling.

TnC has amolecular weight of 16 kDa and is the most acidic of the three troponin subunits

(PI 3.5). Its overall structure in solution is similar to that of another Ca" binding protein, calmodulin,

and to a lesser extent, the myosin light chains that adorn the neck region of MHC (for cornparison

s e (9)). The structure of TnC has been extensively studied by both x-ray crystallography (28) and

NMR (29) . It is bi-lobed, each lobe separated by a flexible central helix. Skeletal TnC c m bind four

Ca" ions while cardiac TnC c m bind only three. Ca2+-binding is mediated by an EF hand (hetix-

loop-helix), motif. Specifically, caL' is coordinated by the oxygen atoms fiom carboxyi and

hydroxyl groups of amino acids at positions 1,3,5,7,9, and 12 of a 12 amino acid loop flanked by

a-helices. The C-terminal domain contains two high afbity ca"/Mg- binding sites, In vivo, these

sites are likely to be saturated with ~ g - . The N-terminal domain of skeletal TnC contains two low

afnnity Cab-binding sites, whereas cardiac TnC retains only the second site. The EF-hands of the

N-terminal domain display selectivity for C$, though M$' c m bind to these domains. The

selectivity and low affinity of the N-terminai domain for ~ a " , make the N-terminai domain of TnC

suited to a role in Ca"-dependent regulation of contraction. In particular, the importance of the 2nd

EF-hand in muscle regulation was addressed by site-specific mutagenesis of Glu 65 (which

participates in Ca2'-coordination). Mutation of Glu 65 abolished both Ca"-binding and TnC-

mediated regulation of contraction (30).

Each of the troponin subunits binds extensively with the others; consequently, TnC interacts

with both TnI and TnT. Some of these interactions serve to maintain the integity of the troponin

complex and are unaffected by calt"' binding to TnC. Others are Ca2'-sensitive. Upon Ca"-binding

to TnC, coordinate changes in Ca"-sensitive interactions between subunits confer a global change

in the conformation of Tn complex, and affects the interaction between TnT and tropomyosin.

2.1.2.3. Troponin 1

TnI, the ATPase inhibitory subunit of the troponin complex, has a molecular weight of 24

kDa and has a high isoelectric point (PI- 10). There are three isoforms of TnI, though only one is

expressed in the adult human heart. The cardiac isoform o f T d is larger than its skeletai two cousins,

as it a bears an extension of -30 arnino acids in its N-terminus (3 1). The cardiac N-terminal

extension harbours two Ser residues which can be phosphorylated by protein kinase A ( P U ) in

response to sympathetic B-adrenergic stimulation of the heart (i.e. fight or flight response7(32)).

Despite the lack of global structurai data on troponin, dual pronged biophysicai and NMR

approaches have provided considerable intemation regarding the domain structure of TnI (Figure

1-6). Data has corne prîmarily h m work on skeletal muscle. though cardiac muscle Tni has also

been studied. Tni harbours several TnC-binding sites. The first is Iocated in the N-terminal at

residues 70-1 10 in cardiac TnI (1-40 in fast skeletal) and binds to C-terminal globular dornain of

TnC (33). The second, residues 126-146 (96-1 16 in fast skeletal Tni) can interact with the, N- and

C-terminai domains ofTnC (34;35). Finally, residues 149-168 (128-148 in fast skeletai) bind to the

N-temiinus of TnC (36). This interaction is likely to be the first perturbecl upon cal' -binding to the

N-terminus of TnC.

Interaction of TnI with actin has also received considerable scrutiny. Residues 126-147 of

TnI, the sarne region that binds to TnC, also binds to the N-terminus of actin at residues L -7 and 19-

44 (37)- Its interaction with actin blocks weak binding of myosin to the N-terminus, and thus

inhibits actomyosin ATPase (35). Though al1 arnino acids within the minimal sequence were

required for hl1 ATPase inhibition (39), Van Eyk and Hodges reported that Lys 105 and Arg 1 15

(Lys L35 and Arg 145 in cardiac TnI, respectively) were the largest determinants of ATPase

inhibition, since their substitution with Gly drarnatically reduced peptide inhibition in an Actin-TM-

S 1 ATPase assay (4 1 ). Finally, a peptide approach was also used to delineate a second actin-binding

domain in fast skeletal Tnl located at residues 125- 145 (1 57- 177), though its contacts on actin have

not been determined (42).

Isolated troponin I is able to bind to actin with a stoichiometry of 1 : 1, and binding inhibits

ATPase directly. However, in native thin filaments, which contain actin, TM and Tn, the

stoichiometry of bound TnI is dictated by its partnership with the other troponin subunits and

tropomyosin ( 1 Tn : 1 Tm : 7 actin monomers). The currently accepted role for Tni within a whole

troponin cornplex, is that it tethers troponin-tropomyosin in a position that interferes with myosin

docking to actin. When Ca" binds to TnC, Tnt-tethering is released, which removes the constraint

on tropomyosin position.

Actin

Head-to-tai l overlap

Troponin I

/ Troponin

Troponin T

Figure 1-5. Troponin and tropomyosin on the îhin fdament .

2.2. Putting It Al1 Together : Tn-TM Regulation by Caz' - Movements in Concert.

2.2.1. The Steric Blocking Model.

We have considered Ca", the proteins of the troponin cornplex, and tropomyosin

individually, But how do these proteins act coordinateiy to regdate the crossbridge cycle discussed

in section A.3? In the early 1970's researchers were passionate about tropomyosin. It was known that

TM was present in d l muscle. Efforts to visualize tropornyosin on the actin filaments by electron

rnicroscopyrevealed that under Ca''-fiee conditions, TM Lies on the outeraspect ofthe (subdomains

1 and 2) of the actin filament. In this position TM bIocks the binding of myosin to actin (17).

Continued use ofX-ray diffraction methods, applied to muscle and oriented actin gels, dernonstrated

that the intensity of one of the x-ray retlections (or layer lines), changed in response to Cg-. This

was interpreted as tropomyosin movement (43). Time-resolved X-ray rnethods showed that, upon

muscle stimulation. the change in layer line intensity preceded contraction (44). Ongoing

biochemical studies led to the discovery of troponin as the Cap-sensor that controlled activation of

muscle proteins. Taken toçether, these data suggested that regdation of contraction might be best

expiained by a stenc blocking mechanism. whereby tropomyosin, under the control of troponin,

regulated crossbridge attachrnent in response to CaL', Perhaps owing to its apparent simplicity, this

model is still prominent in undergraduate life science texts.

The dawn of the '80's spelied trouble for the stenc blocking rnodels as detailed biochemical

analysis of striated muscle proteins uncovered observations that were irreconcilable with a strict

steric blocking model. Diffraction methods yielded to the themdynamïc and kinetic measurement

of protein-protein interactions. In one of the ht, and arguably one of the most influentid, of these

studies, Greene and Eisenberg (45) dernonstrated that purifieci rnyosin heads bound to reconstituted

actin-TM-Tn filaments in both the presence and absence of Ca", an observation incongruent with

steric blocking of crossbridges by TM , Work by Morris and Lehrer (46) indicated that TM done

could have diverse effects on actin-activated S 1 activity. Skeletal tropornyosin inhibited ATPase

while smooth muscle tropomyosin potentiated activity. This observation clearly cannot be

accommodated by a steric blocking model which would predict that bare actin would be the best

activator of ATPase and that TM could oniy block ATPase. These studies are only two of the mmy

reports that confound the steric blocking model. The reader is directed to (19) for an excellent

review.

2.2.2. The ZState Allosteric model

In their study. Greene and Eisenberg demonstrated not only that myosin could bind to C i r

regulated filaments, but that binding was cooperative particularly in the absence of Ca" (45).

Cooperativity, rnost cornmonly identified wiîh the binding of Cl2 to hemoglobin, is a property of

Iigand-protein interactions whereby the binding properties of a ligand are influenced by pior binding

events. [n Iight of ernerging studies, crossbridge regdation \vas deemed to involve an ailosterk.

rather than a steric mechanism. T.L. Hill, a physicist with a predilection for biological problems.

developed a statistical rnechanical modei of the thin filament that could exphin the myosin SI-

binding behaviorobserved by Greene and Eisenberg(readers with steelyresoiveare directed to (47)).

Mathematics aside, the modei simply states that the thin filament can exist in either of two states,

designated active and inactive, which hrirbour different apparent affinities for myosin. That is,

where KI and K2 are the apparent equiiibrium constants for thin filament interactions with S 1, and

T is the transition constant between the active and inactive states of the thin filament that is

infiuenced by cal', In fact, each of the constants, K, and K2, is a composite of two parameters, a

cooperativi;y parameter, y; and an intrinsic binding constant, Ko. Specifically KI =y,Ko, and

KmKo, . In the absence of Ca2', the inactive state predominates, and S 1 interacts with a low

apparent affinity (overall low y,Ko,). This low affinity is characterized by low intrinsic afinity

(small Ko,) yet high cooperativity (high y,), However, in the presence of Ca2', the active state of

Actin/Rvl/Tn predominates, which has much higher apparent affinity for SI (high y2Ko,)

characterized by higher intrinsic affinity (large Ka, yet comparatively lower cooperativity (lower

y 3 than the inactive state.

2.2.3. Newer 3-state models reconcile steric and allosteric aspects of muscle regulation

The Hill 2-state model adequately described equilibrium binding of myosin to the thin

filament. Nevertheless, it was il1 equipped to account for stopped flow kinetic data of S 1 binding that

indicated that there was a defined lag period before S 1 could bind the thin tilament. The simplest

explanation ofthis behaviour held that there was a population ofthin filament sites that were initially

inaccessible to S1. This prompted McKillop and Geeves (48) to propose a three-state model of

contraction that extended the Hi11 modei by subdividing Hill's 'inactive' state into two states called

'bIocked' and 'ciosed' (figure 1 -7A). The brocked state, in which troponin holds tropomyosin over

the myosin binding sites of actin, was proposed to be populated in the absence of Ca?. in the

presence of Ca"', tropomyosin can adopt 'cIosed' or 'open' states. Myosin can bind the filament in

either the closed or open conformation, but can only isomerize to a strongly bound (force producing)

state when tropomyosin is in the 'open' conformation. McKiIIop and Geeves tested the mode1 and

found that it reconciled thermodynamics and kinetics of myosin binding to actin. Furthemore,

invokïng a blocked state reintegrated elements of the stenc blocking into the fold. The model

received criticai backing fiom the structural studies of Lehman's group (49), who, by 3D-image

reconstruction of electron of micrographs of thin filaments, were able to demonstrate the existence

of the Ca"-free, CazT-induced and myosin-induced conformations oftropomyosin (Figure 1-7B). The

3 state model is also consistent with muchof the physiological dataobtained from muscle fibers (see

(20) for a detailed review that encompasses structural, kinetic and physiological data).

The McKillop and Geeves three-state model is widely acclaimed but, like the Hill two-state

model, it rests on several assumptions that are not universally accepted. Both models assume

cooperativity is propagated over the length of 7 ricsin molecules through rigid or discrete

tropomyosin molecules. Yet it has been shown that tropomyosin molecules lying end to end along

actin, can communicate with each other. That is, movement of one tropomyosin is accompanied by

the movement of contiguous tropomyosins. Furthemore, Tobacman and coworkers have

demonstrated that Ca"-binding to cardiac thin filaments is highly cooperative irrespective of the

presence of myosin(50). His subsequent work on the thermodynamics of thin filament assembly has

laid the h e w o r k for a model that incorporates al1 of the principal elements of the McKilIop and

Geeves 3-state model, yet modernizes it to address issues of detailed thmodynamic balance, that

arose fiom their protein-binding studies. In their model, Ca3-mediated movement of troponin-

tropomyosin allows actin itselJI to assume an "active" conformation that favors myosin interaction-

This model differs tiom its predecessors by treathg actin cables as dynamic entities. The model

holds that actin, not tropomyosinperse, propagates Cas and myosin signals dong the thin filament

(5 1)-

2.2.4 Summary of Striated Muscle Regulation

Iso men statim to R state oriy

possible for cpen

Figure 1-7.3-state models of striated muscle regdation, A. kinetic/thermodynamic model of McKillop and Geeves (48). In this mode1 îhere two are transition equilibrium constants, K, describes the equilibrium between the 'blocked' and 'closed' and is influenced by Ca2+. KT describes the transition between the 'closed' and 'open' States. Myosin binding is descriied by the equilibrium constant, K,. Only in the 'open' state can myosin isomerïze to the force producing state, K, B. The structural 3-state model of Vibert et al. (49) Troponin- tropomyosin decorated filament were incubated in Gaz+-fiee solution, Ca>-replete solution or Caa solution containing myosin SI. The computer images are obtained by 3D helical image reconstmction fkom negatively stained electron micrographs. The Left panel shows a helical projection, or cross section of the thin f3ament.The positions of tropomyosin are shown in red (-Ca2+), magenta (+Ca3) and cyan (+Ca'f+Sl). The panel on the right shows how the position of tropomyosin moves across the face of actin.

The commonelement o f d l these models is that Ca2*-binding to TnC facilitates the transition

between the crossbridge-inhibitory and crossbridge-facilitatory states. How is this accomplished?

CaL', troponin and tropomyosin work together, as a functionai unit, to regulate contraction in

response to Ca". When Ca" binds to the N-terminus of TnC, TnC undergoes a conformational

change that weakens the interaction behveen TnI and actin. Specifically the actin-binding inhibitory

region of TnI, which also binds TnC, 'switches' fiom its binding site on actin to a newly exposed

hydrophobic pocket on TnC. The switching ofTn1 fiom an actin-bound state to a TnC-bound state

hsis hvo consequences. Firstly, TnI itself no longer directly inhibits actin-TM-myosin ATPase, but

more importantly, switching removes constraints on the position of Tn-TM such that TM c m roll

azimuthally across the face of actin, which exposes weak and strong myosin-binding sites on actin.

(See (40;52) for reviews of troponin function). Myosin crossbridges can then dock with the thin

filaments in ri weakly-bound state, then undergo a transition to a strongly-bound state that produces

force.

3. Regulrtion of Crossbridge Cycling in Smooth Muscle

3.1 Thick Filament Regulrtion

3.1.1 Myosin Revisited.

1 have already introduced myosin as the motor protein that is responsible for ATP hydmiysis

and force production, As we have seen, contraction in cardiac muscle is controlled by thin filament

proteins Tn and Tm in response to Ca", wtiich effectively control access of myosin to actin. Thus,

the ATPase activity in a cardiac systern is controlled by the thin filament. By contrast, smooth

muscle Iacks a troponin complex, and modulation of ATPase is conferred through the myosin

molecule itself.

The rate and the extent to which crossbridge c y c h g occurs, is highly regulated in smooth

muscle. When Ca" influxis tx-iggered, it binds the Ca2+-binding protein calmodulin (Figure 1-8). The

Ca2'-calmodulin cornplex serves as a co-factor for the enzyme, myosin light chah kinase (MLCK).

As its n m e implies, this kinase phosphoryiates the regdatory light chain of myosin LI specifically,

on residue Ser 19 (53). Myosin is converteci h m an inactive state to astate that is highly activatable

by its partner in contraction - actin. Briefly stated, phosphorylation increases crossbridge association

and actomyosin ATPase(54). Consequently, force development and the 'sfdinç' velocity of thin

fi1ament.s are both proportional to the phosphorylation state of rnyosin. Finally, rnyosin

phosphorylation is also reversible by specific myosin light chain phosphatases (MLCPs). The ratio

between MLCK and MLCP activities dictates the extent of LC, phosphorylation. and ultimateiy how

much force is produced in a contraction- When Ca" retums to basal IeveIs, MLCK is inactivated,

myosin becomes predominantly dephosphorylated, and muscle relaxes.

How does phosphorylation activate rnyosin? Electron microscopy of smooth muscle myosin

shows that unphosphorylated myosin adopts a folded conformationirt vitro, in which the neck region

is folded back against its tail, cdled the 10s conformation based on ultracentrifbgation studies.

Phosphorylation causes myosin to assemble into filaments that are enzymatically active. Interestingly

however, in differentiated srnooth muscle, myosin forms stable filaments regardless of its

phosphorylation state, though unphosphorylated filaments remain inactive. Whether in 10s or

filamentous 6s form, the structure of inactive myosin is such that the globdar heads are constrained

in contact with the myosin rod. Phosphorylation deases the constrauit, which allows the globuhr

myosin head to iateract freely with actin. Studies of truncated myosin have shown that the rod

region of myosui is required for phosphorylation dependent regulation of myosin ATPase activity

(55) (figure 1-9).

3.2 Thin Filament Candidate Regulatory Proteins

3.2.1 Tropomyosin.

Though the role of tropomyosin in skeietd muscle contraction is well defineci, its role in

smooth muscle is much Iess clear. Tropomyosin does provide structural support for the actin

filaments, The smooth muscle isoform, composed ofpolypeptides designated a and P, retains rnost

of the physical characteristics of its skeletal cousin. However, the smooth muscle isoform displays

a greater propensity for head to tail polymerization, and more cooperative binding to actin (reviewed

in (56)). More importantly, skeletai and smooth muscle tropomyosin appears to exert different effects

on the conformation of F-actin. Specifically, skeletal muscle tropomyosin acts as an uncompetitive

inhibitor of actomyosin ATPase activity while the smooth muscle isoform acts as an uncompetitive

activator of ATPase ( 1 9). L e h a n and colleagues have recently looked at the equilibrium positions

of skeletal, cardiac and smooth muscle tropomyosin on the thin filament (59). They showed that

in the absence of troponin. skeletal tropomyosin lies on the outer aspect of the actin subdomains 1

and 2, in a position similar to the blocked or Cd'-free state shown in tigure 1-6B. By contrast.

smooth muscle and cardiac muscle tropomyosins lie on the inner portion of those domains, in a

position that more closely resembles the Ca2=induced state. Remember however that in muscle, the

positions of the skeletal and cardiac tropomyosins are controlled by troponin, whereas smooth

muscle tropomyosin is likely to adopt its equilibrium position. Clearly, if tropomyosin plays a role

in smooth muscle regdation, it wili not function as it does in cardiac or skeletal muscle.

3.2.2 Caldesmon

Smooth muscle h-caldesmon has a Mr of 87,000 and was originally identified by its ability

to bind actin and calmodulin (caldesmon, derived tiom caimodulin and 'desrnos', Greek for binding,

(60)). On the bais of analyticai ultracentrifbgation, it is believed to be a highly extended moIecuIe

INACTIVE ACTIVE

e---- MLCK - pho.ph-

OEPHOSPHORYLATED FILAMENTS P ~ O S P ~ O R I L A ~ O FIUYE)CTS

b

"\ * # FOLOED MONOMER

INACTIVE

Figure 1-9. Conformations of smooth muscle myosin in vitro PaneIs A. EIectron micrographs of myosin molecules, in the folded (10s) and extended (6s) confonnations, respectively. Panel B. interrelation between the conformations of myosin, and the role of LC2, phosphorylation. Phosphorylation of myosin not oniy influences filament assembly in vitro, but is essential for its activation by actin. Myosin assembly may be a point of regdation in non-muscle cells. However, in smooth muscle, myosin filaments are stable. (reprinted fiom (58)).

whose length is 75 nrn (6 1). Caldesmon has three finctionally distinct domains (Figure 1 - 1 O): an

NHz-terminal dornain (residues 1-250), a middle domain (residues 251-300) and a COOH-terminal

domain (residues 400-756). The NK-terminal domain houses myosin- and tr~~omyosin-binding

sites (62;63), while the COOH-terminal domain intetacts with actin (64-66), tropomyosin (67)and

the CaL'-binding protein, calmodulin (65;68;69). The middle domain assumes a long a-helical

structure (70) but its function is unclear and it is absent Tom non-muscle low molecula. weight

isoforms (7 1 ;72).

Like tropomyosin, caldesmon binds lengthwise alonç F-actin tilaments, Its affinity for actin

(&-O. 1pM) is enhanced 3-foId in the presence of tropomyosin (73). Imm~noc~tochrmical evidence

indicates that caldesrnon is confined to contractile a or y actin tilments and is absent h m

cytostructural P actin filaments in smooth muscle (74). The stoichiomeby with which caldesrnon

interacts with thin filaments has been the focus of considerable study. Snidies in vitro have yielded

values fiom 1 caldesmon per 16-18 actin monomers (75) to I caldesmon per 7 actin monomers (73).

In 'native' fiIaments extracteci fiom smooth muscle, the estimated stoichiometry is 1 caldesmon

molecule per 14-1 6 actin moIecules (76). When these fiIaments are labeled withmonoclonal antibodies

to theNH,-and COOH-termini of caldesmon and subjected to electron rnicroscopy, caldesmon showed

a binding periodicity of 38 nm (77), corresponding to the approximate length of tropomyosin, and half

of the estimated length of caidesmon. A model that reconciies the 1 :2: 16 stoichiometry of caldesmon:

actin:tropomyosin with the lengths of the caldesmon and tropomyosui molecules is depicted in Figure

1 - 10. In this model, caldesrnon spans two tropomyosin molecules ninning dong the groove of the actin

double heiix. A second caldesmon molecuIe is staggered, the lengtho fone tropomyosin molecule with

respect to the first, dong the opposite groove of the actin.

Ngai and Walsh h t observed that thin filament-bound caldesmon could inhibit actin-activateci

myosin ATPase in vitro (78). Since then, mearchers have speculated that caldesrnon might hc t ion

as amodulator of srnooth muscle contractility. Subsequent study has shown that caldesmon effectively

inhibits the actin-activated ATPase activity (up to 80% or greater), regardless of the source of actin or

enzymatic myosin fiagrnent (73). Maximal inhibition occurs at a ratio of 1 caidesmon to 7 achn

monomers but is considerabiy more effective in the presence of smmth muscle tropomyosin (73). In

the absence of tropomyosin, caldesmon inhibits ATPase by cornpeting with myosin-ATP for actin

(64;79). However, in the presence of tropomyosin, caldesmon not only inhibits rictomyosin binding,

but also acts to reduce the Vm, of ATPase. These observations imply that caidesmon, in native thin

filaments, dso affects one of the kinetic transitions of the cross-bridge cycle, likely the rate of Pi release

(80). However, Brenner and Chalovich have show that addition ofexogenous caldesrnon to Triton-

skinned muscle fibers affects force development by influencing muscle stifiess nther than the rate of

tension redevelopment, L. (83). This means that in the muscle fibers, cddesrnon appears to inhibit

crossbridge cycling by interferhg (competing) with the docking of myosin on actin.

33 Economical Force Maintenance: The 'Latch' State

Calcium undoubtedly îriggers force production in smwth muscle. However, &er the onset of

contraction, the concentration of intracellular Ca" (hereatterdenoted [Ca2*]J drops. The initial elevation

in [Ca2-ji is transient? Iasting only seconds, yet force can be sustained for minutes or hours (84). An

idealized arterial smooth muscle contraction foiiowing agonist-induced stimulation is depicted in Figure

1-1 1. Stimulation induces rapid Ca" influx, foiiowed by activation of MLCK and force development

Following the omet of contraction, [Ca3Ii drops to an intennediate Ievel (- 0 2 IrM) that is insufficient

to activate MLCK fully- Consequently, the stoichiometry of myosin phosphoryiation drops by 50%

foiiowing the initial phase of the contraction (85), yet maximal force c m be sustained unaI Ca2' returns

to basal levels (estimateci at 0.1 yM, (86)). Force maintenance under conditions of low Ca" and

phosphorylation, is aiso characterized by low ATP consumption, and has been termed 'latch' (87).

33.1 The Latchbridge Bypothesis.

Efforts to explain the latch state have given rise to the '4-state latchbridge' hypothesis (87;8S)

The premise of the latchbridge hypothesis is that there is a discrete population of myosin molecules that

become dephosphorylated whilst in contact with actin. Dephosphoryiated actomyosin crossbridges, called

latchbridges, are postulated to detach h m actin much more slowly than phosphorylated crossbridges.

This extends the t h e myosin spends in a strongiy actin-bound state and slows its rate of cycling. n ie

model adequately addresses how low levels of myosin phosphorylation lead to tension maintenance and

atm provides an expianation for reduced ATP consumption. Advocates of the latchbridge hypothesis

contend that contractile responses (force, phosphorylation, Cà') can be faithiùlly modeled, pphically,

using anpirical data for the various rate constants in the 4-state model (88). Neverthelas, the hypothesis

rests on assurnptions that tend to oversimpli@ the mechanism of the crossbridge cycle. One assumption

is that actin-bound myosin (a crossbridge) must be a substrate for both MiCK and MiCP, which is

difficult to prove experimentdiy. The model M e r assumes that dephosphorylated myosin cannot

reassociate once it becomes unbound h m the thin filament. Hoivever, biochemical snidies indicate that

reassociation is possible, albeit at a reduced rate (84). Moreover, the mode1 is il1 equipped to explain

reports in which contraction has been uncoupled h m myosin phosphorylation (89). Specifically,

contraction has been reporteci in the absence of CaZ', without measurable change in myosin

phosphorylation, when rat uterine muscle is stimulateci with oxytocin (90). Geahoffer has shown that

myosin an be phosphorylated without appreciabIe force production, under conditions of low [Cal , in

response to serotonin and carbachol (91). Moreover, studies have shown that nitrovasodiiators (e.g.

nitroprusside) cm cause acterial muscIe relaxation, while myosin rernains phosphorylated (92).

Coldesrnon

Figure 1-10. Structure of caldesmon, A. (preceding page) Caldesriion domains, The C-terminal domain, containing the actin and tropomyosin binding domains, is critical for ATPase inhibition, The N-terminus binds myosin and may act as a tether between aciin and myosin or simply to preserve the geoinetry of the contractile apparatus; courtesy of Dr. Albert Wang. B. Model proposed by Lehmon et O/. (771, as dcpicted in (82), shows a caldesmon molecule running along each of the helical grooves of F-actin, one offset from thc other by the Icngth o f one tropomyosin molecule.

33.2 Cmperative Crossbridge Reattachment

An alternative expianation for Iatch has been proposed by Somlyo (93). In their studies, srnooth

muscle relaxation was triggered by flash-photolysis ofcaged ATP (-Ci?), and parameters suchas muscle

force, sti fhess (indices of crossbridge attachent) and myosin phosphorylation were measured. The rate

of rnyosin dephosphorylation exceeded the rate of relaxation, and the rate of reiaxation was best fit by a

two-parameter exponential decay function. Together, these observations indicate that dephosphorylated

myosin heads are capable of cooperative reattachent. Attachrnent ofdephosphorylated. slowly cycling,

myosin heads is consistent with the high economy of smooth muscle.

333 An Extra Intermediate in the Crossbridge Cycle

Recent work dso indicates that smooth muscle myosin undergoes an extra step in the crossbridge

cycle that cardiac and skeletal muscle myosins do not. Specificdly, cryoelection microscopy and 3D-

image reconstruction has revealed that actin filaments, decorated with srnooth muscle S 1. exhibit diffmt

conformations in MgADP h m those observed in rigor (a nucleotide îi-ee state). Srnooth muscle rnyosin

undergoes an additional conformational change, in which the lever arm (neck) region rotates 23" upon

ADP relerise. This indiwtes that smooth muscle contains an extra skiindependent step that is not pment

in skeletai or cardiac muscle, It is believed that under strain ADP release would be tùrther slowed, thereby

slowing ATP rebinding and crossbridge cycling ( 183).

33.4 Thin Filament Mechaaisms.

Srnwtbmuscie thin thaments may lack troponin but they do cxintain otheractin-binding prote&,

aidesmon and cdpnin, that inhiiit actomyosin ATPase in *o. As such, they are potential regulatory

proteins. Caldesmon is confineci to the contractile a-actin 6iamenîs whiIe caiponin is present on

cytostnictural B-achn filaments as welI, I will confine my discussion to the putative roIe of caldesmon

though the reader is directed to (94) for more information on cdponin,

How might a thin filament protein fiiment conûiiute to latch? Since caldesmon is capable of

simuitaneously binding to actin and to myosin, it had been suggested that caldesmon might contniute

to latch by crosslinking the myosin and actin filaments. Restated, amechanism whereby caldesrnoncouid

simultaneously inhiiit actornyosin crossbridge cychg yet "tether" rnyosin to actin seemed like a perfect

mechanism to explain latch. Evidence that caldesmon caused superprecipitation of actomyosin supporteci

such a model. Subsequent studies, however, indicated that the affinity of the cddesmon rnyosin

interaction was likely not m n g enough to sustain the tension loads observeci in Iatch.

0 th lines of enquiry have saught to detmine how the known fùnctions of caldesmon are

regulated. If caldesmon inhiiits crossbridge cycling, regdatory mechanisms might relieve this inhibition

and thereby contriiute to smooth muscle contraction. Caldesrnon was originally isoiatd on the bais of

its ability to bind calmodulin and actin (60). The study a h showed that caldesrnon-binding to actin

could be outcompeted with large amounts of Ca"&odulin, in what the authors called a 'tlipflop'

mechanism. Once caldesmon had been shown to irihiiit actornyosin ATPase. it was not altogether

surprising that Ca2'-calrnodulin was shown to reverse this inhibition (46). Nevertheless, these reports

provided evidence that a noveI, Ca2'dependent, potential reguiatory mechanism for contraction could

be reconstituted in vitro. Substantial work has nmived the cdmoduiin-binding sites on caldesmon to

two amino acid sequences within the extreme COOH-temiinal domain(95;96), specificaily to sequences

6MMWEKGNVF~a and 687SRINEWLT~695 (Figure 1-9). Recently, Graether et al. deterrnined that the

tryptophan residues within these sequences are the primary calmoduiin-binding deterrninants on

caldesmon (97).

Phosphorylation has also been ûinsidered as a mechanism for the regulation of caldesmon,

Studies within the last 15 years reveal that caldesmon is asubstrate for at least six kinases in vïîro: Protein

kinaseC (PKC (98-100)),Ca"-Calmodulindependent kinase U(Cah4 kinase II, (10 1; 102), caseinbase

Iï (CUI, (103; KM)), cyclic AMPdependent kinase (PKA, (105)), ce11 division cycle 2 (çdc2, ( L 06-108))

kinase and mitogen activated protein kinase (MAPK or ERK, (I09;L 10)). Protein kinase C

phosphorylates caldesmon to astoichiom~ofappmximately 1.9 mol phosphate1 mol protein, primarily

at Ser 587, but aIso at Ser 600 and Ser 726 (99). Phosphorylation at these sites within the COOH-terminal

domain is sufficient to weaken ddesmon interactions with bath actin and calmoduh. Consequently,

p hosphorylation by PKC signi ficantly decreases the cfficiency with which caidesmon inhibits actomyosin

ATPase (1 L 1). Phosphorylation of h-caldesmon by cdc2 kinase, prirnarily on Thr 673, and to a much

lesser extent, on Ser 582, Set- 667, Thr 696, and Sm 702 (log), also decreases its a f i t y for actin and

caimodulin (107). Casein kinase II exerts its influence on caldesmon by phosphorylating sites within its

NH,-terminal dornain. Spe~ificaiiy~ phosphoryiation at Ser 26 and Ser 73 weakens the interaction

between caldesmon and myosin (104)- Ser 73 is also the pnrnary t q e t of CaM kinase II, and its

phosphorylation has the m e functionai consequence (1 12). It is notervorthy that CaM kinase II c m

phosphorylatecaldesrnon M e r , up to 8 moYmol, atsites within the NH?- and COOH-terminal domains.

Phosphorylation by CaM kinase rvithin its COOH-terminai reIieves cddesrnon-mediateci inhiiition of

ATPase without influencing its alfinity for a c h (102). Finally, MAP kinase has k e n reporteci to

phosphotyiate h-caldesmon up to a stoichiometry of 2 moi 1 mol at many of the sarne sites as cdc2 kinase

(determined by tryptic peptide mapping), though the preferred phosphorylaiion sites for the two kinases

appear to be markedy different (1 IO).

Despite abundant evidence that caldesrnon impinges on crossbridges cyciing in &O, and ihat its

fùnction can be regdated W y , there has been conspicrwudy Me evidence for a caldesmon-mediated,

or aidesmon-assiste4 contraction in smooth muscie. Models invoking caidesmon in contraction bave

fden out of vogue. Thus, 20 years after its discovery, caldesmon remains an enigma

4. Fine hining of Cross-bridge Cychg: The Ca2'-sensitivity of Contraction

In the preceding sections on wdiac and smooth muscle regulation, we have considered the

mechanisrns by which increased intracellular Ca" leads Co force production. However, the relationship

between Ca2- and force (fiequentiy the force/pCa relation) is highly variable, within cardiac and smooth

muscle. ~actors that influence ca2--force dependency are said to affect the ~a?-sensitivity of contraction.

If muscle produces a given force but requires Iess Ce to achieve it, cal'-sensitivity has increased.

Conversely, if greater C c is required to achieve a particular contractile force, Ca2'-sensitivity has

decreased. Ca" sensitivity cm be studied in intact muscle fibers, triton-skinned fibers or in reconstituted

Ni vitro systerns.

4.1 Ca2+ -sensitivity in Cardiac Muscle

Studies have indicated that several factors influence 0'--sensitivity in cardiac muscle. They c m

be broadly classifieci as I ) factors that inthence ca2'-binding to Tn and 2) factors that influence the

propensity for myosin to bind the thin filament at a particular CS'.

Factors that interfère with Ca2'-bhding to Tn will tend to tàvour muscle relaxation while those

that promote Ca2' binding will lead to th fiIarnent activation. Consequently, increases in pH and/or

decreases in ionic strength tend to increase Ca2'-sensitivity. Within the beating hem cellular processes

influence cal+-sensitivity. As rnentioned previousIy, sympathetic B-adrenergic stimulation Ieads to

activation of Protein Kinase A ( P U ) in the myocardiurn. P m phosphorylation of troponin reduces its

afbi ty for CaL'(l13;I 14). Specificaiiy, PKA phosphorylates Tni on residues Ser22-23 and infiuences

the interaction behveen Tni and TnC.

in intact cardiac muscle, increased ÇaiCOmere length and decmed interfilment spacing within

the myosin-actin lattice both bring the rnyosin into closer proxünity to actin. This increases the effective

concentration ofmyosin heads around actin and le& to increased crossbridge association at a aven Ca?

concentration, and therefore causes Ca" sensitization. hdeed, any parameter that increases myosin

attinity for actin, e.g. reduced ionic strength, increased ADP, or decreased ATP, will cause C s -

sensitization. Decreasd Iattice spacing in wdiac muscle also increases Ca" afjïnity for TnC. Thus,

cardiac muscle harbours a form of feedback between crossbridge attachrnent and Ca" binding to Tn that

doesn't occur in skeletai muscle.

4.2 CaZ+-sensitivity in Smooth Muscle

In smooth muscle, the ca2--force relation is also a composite îùnction of two sepamte

relationships, each of which is a potential mget for regulation. R e d that Ca" initiates contraction by

causing W C phosphorylation and phosphorylated myosin undergoes rapid crossbridge cycling with

actin. Restated, myosin phosphorylation is coupled to [Ci-], through the activities of MLCK and

MLCP. Mechanisrns that alter the efficiency of one these enzymes relative to the other, at a particular

[Ca"], wouid influence the dependency between the level of myosin phosphorylation and [Ca?']i.

Investigations into the Ca2' sensitizing action of GTP analogs (discussed section D.1 ) has Ied to the

identification of new mechanisms that inhibit MLCP.

The second dependency that may regulate crossbridge cycling, is the coupling of force to myosin

phosphorylation. That is, to what extent does force necessarily require myosin phosphorylation?

Cooperative crossbridge reattachment of dephosphorylated crossbndges (previously discussed in section

B.333) is an example of how force couid be maintained at low levels of myosin phosphorylation.

Aiternatively, the force-phosphorylation relationship might be altered by thin filament proteins such as

caldesrnon. in fact Pfitzer et aL(lI5) have demonstrateci that addition of an exogenous C-terminal

fiagrnent of caldamon to Triton-skinueci smooth muscle tibers increases the degree of myosin

phosphorylation required to achieve a given level of force. If caldesmon reguiation c m be provenin vit'o,

it is Iikeiy that it will invoIve modulation of the phosphorylation-force relationship.

C. Myocardial Stuanùig.

1. Stunoing: An Affliction of Cardiac Myonlaments?

Poor coronary circulation (ischemia) is a Ieading cause of mortal@ and morbidity in deveioped

countries; its treatment incurs tremendous costs, both medically and in t m of lost productivity.

CLinically, the severity of myocardial ischemic injury vanes gmtly. in a rnild fonn, caiied myocardiai

stunning, patient. expaiencereversible contractite dysfwiction following a transient episode ofischemia

and subsequent reestabiishment of bIood fiow ( 1 16). With Longer, more severe ischemic episodes, the

myocardiun becornes irreversibly h a g e d , a condition d l e d myocardial intarct. Understanding the

ceIlular processes that are afected both during ischernia and subsequent to adequate repertiision, wili

help the design of new treatments that might pmtect the hearts of patients at risk.

Though m y o d i a i stunning is considemi to represent the early stageormild forrn ofischemia-

reperfhon (VR) injury, it is a serious condition. Stunning is defined specifidy as a period of post-

ischemic contractile deficit characterized by poor cardiac output and reduced systoiic and diastolic

fiuiction, despite adequate pertusion and absence of cellular necrosis ( 1 16). Though the injuy is

revasible, patients diagnosed with stunninç m a t risk for Mer complications for several days. Stunned

myocardium ananses pathoIogically as a d t of coronary thrombolysis, or relaxation of vasospasrn, but

is dso a fiquent side-effect of su rg ia i procedures, including coronary angioplasty and heart bypass

operations (1 16).

How does a heart becorne stunned? Researchinto the apathogenesis ofstunnuig has reveaIed that

injury may be as dependent on the resumption of properperfiision as it is on the ischemic periodperse.

in th& comprehensive ceview, BolIi and Marban (1 16) suggest that ischemia acts as a "primer" for

subsquent events ûiggered by the onset of reperfùsioa The effeçts of ischemia are manifold C e h are

starved of oxygen, vititl for ATP production, and intraceIlular pH decceasesecceaSeS Under these conditions,

inmellular Na' accumulates due to the failure of Na' extrusion pumps. N a etwation would normally

be replaced by CS' via the Na'/Caz* exchanger, yet the exchanger is inhibited by cellular acidosis. An

ischemic myocyte is therefore Nar-replete and oxygendeprivd in th& integrated mode1 of i/R injury,

Bolli and Marban suggest that the onset of reperfiision rnay be desûuctive in two ways. Firstly, rapid

revasal of cellular açidosis restores the tùnction of the Na/ CaZ+ exchanger, causing rapid influx of Ca"

into thecekl, or Ca2--overhad. Ca'T-overload, it is thought, activates the~a"dependent protease, caipain.

Aberrant calpain activation lads to proteolysis of certain myofilament proteins. The second mechanism

of injury rnay involve genaation ofreactive oxygen sgecies (ROS) owing to the reintroductiono foxygen

upon repertùsion. ROSS might cause inûacellular darnage by either ox id i ig pmteuis, thereby altering

their tùnction. or by contributing to Ca'-overload by dmaging the sarcolemma (Figure I-12).

Regardless of the specific triggers of stunning, a hallmark of the condition is that systolic heart

function is depresseci for several days following the ischnnic episode. Efforts to detmine the nature of

the lesion that causes the h a r t dysfunction have involved the use of models. An emerging view

is that contractile dyshction resuits fiom a lesion of the myofilaments. Speciticdy, in rat modek, when

hearts exhibit stunning zit the organ level (rduced cardiac output and contrrictility), ~ a ? - mnsients, and

membrane excitability of the myocyte are unaffectai (1 17), but select myofilament pmteins, such as Td,

have undergone proteolysis (1 174 19). The importance of the TnI modification is underscored by recent

work which demonstrates that targeted transgenic expression ofa proteolytic hgment o f f ni, Td,-,,, (the

earliest marker of VR injury in isolateci rat hearts) in mice hearts, is suffi:cient to recapitulate the

phenotype of the stunned myocardiun (1 20; 12 1).

2. h search of a "snuining' protein: The sîudy of Td,,,

Cardiac diseases are characterized by a host of changes to the celldar processes bat affect

contractility of the hwrt. One emerghg exphenta i approach to the mdy of cardiac disease, inciuding

the VR injury known as stuming, is to identify proteins that have k e n specifically altered or rnodified

in the diseased state. MinimaUy, the identification of modified proteins wiii provide new diagnostic

markers of myocyte injury. However, for a subset of proteins, disease-induced modification will

substantively affect molecular hct ion, and conûiiüte directly to cardiac dysfùnction. Identification of

these 'causal' proteins wiii ultimateiy prove beneficial to the design of new phamacological, genetic or

peptidomimetic therapies (l22). In this section, 1 will provide an overview of the literature that served

as the impetus for the project presented in Chapter 3.

One of the prefmed animai models of myocardial stunning involves the study of isolated rat

hearts perfused with an elaborate pumping systern (the Langendorfmethod}. In this experimental systern,

blood flow is mimicked by pumping bufferreplete with essential metabolites and minerais (Krebs buffer}

through the isolated hm, lschemia is induced by shutting off the pmp, Tuming on the purnp, aiter a

specified period, reperhes the kart. This rnoàel has bbeen used by several laboratories and ha yielded

considerable information about the pathophysiology of srunning.

Studies from Dr. Marban's group have shown ventricuIar tnbeculae h m ischemic rat hearts

displayed decreased steady-state isometric force deveiopment and decreased Cs'-sensitivity of force

development ( 123). Twitch kinetics of force developrnent were also perturbed. Furthermore, membrane

excitabiiity and regdation of myocellular C b m s i e n t s wre unaffecteci by stunnïng, indicating that

decreased Cap-responsiveness was not limited by C2+-availability. This suggested that a lesion of the

myotiiarnents was responsiile for stunning (124). Pardel work by Van Eyket al., usingthe sarne animal

rnodel, confirmeci this hypothesis. Specificaüy, they noted decreased isometric force, causeci by stunning,

inTnton-skinned trabeculae abers, in which celimembranes are removeci, leawig only the myofilarnents

(1 19). This technique allows the researcherto measure force in b u k with defined Ca" concentrations.

Van Eyk et al. ( 1 19) noted that the reduction in tibre force production correlatai with the extent of the

1 REVERSIBLE ISCHEMINREPERFUSION l3 Na+overload _,---* ? Sarcolemmal ----,

damage i

Oxidative stress \ 4

J Calcium overload

I , -'-----? SRdarnage ------' 1 C& activated

proteases

? Oxidative modification ? Proteolysis of contractile of contractile proteins apparatus

l \ (troponin 1, a-actinin) 1 I I t

I I 1 I \ / 1

Decreased Ca2+ responsiveness l

? Repair of ; Oxidative

1

I

: ? De novo orotein

I '

1 synthesis

. ,

*'---& SLOW BUT COMPLETE &---'*

RECOVERY

Figure 1-12. Mode1 of the pathogenesis of ischemia/reperfusion injury. This diagram illustrates how two prominent hypotheses, the Ca2+-overload hypothesis and the Reactive Oxygen Species ( 'OS) hypothesis, might be integrated. From ref. (1 16).

ischemic injury. Fwthermore, as the severity of the ischemic injury increased, they noted progressive

and selective proteolysis of the myofilament protein Tni.

Tni proteolysis was dso observeci by Marban ( 1 1 7; LX), who sugesteci that C*-overload, upon

reperfusion (discussed previously), might cause activation of the Ca'+dependent pmtease, calpain. This

hypothesis was supportai by experùnents hat Ïndidicated that Tni proteolysis in stunning codd be

prevented by reperfusion with a low Ca'Xow pH buffer ( I L 7). Furttiermore, addition of calpain to

skinned muscle libers yielded aproteolytic hgrnent o f f d with the same molecularweight as that which

was observed in stunned myomdium (1 17). In th& review, Bolli and Marban cited identification of the

precise lesion ofstunning as one of the buming issues in the field. Van Eyk's Lab subsequently purifiai

and characterized the specific proteolytic pmduct of Tni h m ischemic rat hearts. In stunning, Td is

cleaved at its C-temiinus to yield TnI,-,,), ( ! 18).

However, it was necessary to detemine whethcr Tni proteolysis wris a causal event in the

pathogenesis of stunning or rnerely a bypmduct of üR injury. To that end. Murphyet al. undertook the

constructionofa m g e r i c moue mode1 ofmyocardid stunning(120). She expressed Tni,-,,, under the

control of the u MKC prornoter in duit mouse hem, These mice expressed Tni at fevels 10-20% of

total celldar TnI. Expresion did not aIter regular sarcomeric structure, but did incur deamseci systolic

(dPmax/dt) and diastotic (dPmiddt) hction. Trabeculae demonmced decrased maximum force

development. Furthetmore, Ci? transients w a e normal in the transgenic hearts. Thus, the transgenic

mouse model faithfuiiy recapitulated the salient feature of shrnning observed in the nt hart model.

2.1 The Stunning Roject: Shîement of Objectives

Though the tmsgenic mode1 filustrates that TnI proteolysis is &dent to cause stunning, how

does it do it? Anmvering this question will not only advance our understaudhg of the pathophysioiogy

of ischemic hem disase but wodd aIso lay the h e w o r k for better preventive therapies. in the projeçt

presented in Chaptec 3 we have begun to investigate TI&,,. Our hvo principal hypotheses were that the

pathop hysiological manifestations ofstunning would be most easily explained by: 1) aberrant crossbridge

cyciing, which we would assess using in viîro measurements of actin-TM-S 1 ATPase, or 2) disnrptioo

of specific protein-protein interactions within the regulated thin fiIament that would mitigate troponin

response to Ca". As biochemist., we pursued a reductionist strategy in which each ofthe muscle proteins

was purified, so that they could be reconstituted to determine the effects of TnI tnuication on its

molecular fiinctions.

D. Snidies of the Ca2+-sensitivity of Smooth Muscle Contraction.

1. New Signal Transduction Pathways

In the last 10 years, considerable attention has b e n paid to the observation that GTP analogs

cause Ca1+-sensitization in smwth muscle revmily pemieabïIized with a-toxin (126). Efforts to

characterize the mechanisrn of cal'-sensitintion have impiicated the srnail rnonomeric GTPase, Rho.

Normally cytosolic, Rho c o d m sensitivity once it has been reçniited to the plasma membrane in

response to muscle stimulation (1 27). It exerts its influence through its e@ector, Rho-associateci kinase

(ROK). Additionof exogenous constitutively active ROK to Triton-skinned smoothmuscle fibm causes

Ca'' -sensitization directly. Though ROK phosphoryIates both myosin RLC (1 28) and MLCPin vitro.

its preferred substrate in the fih is MLCP. Phosphorylation of MLCP reduces its activity (129). This

promotes contraction by allowing the activity of MLCK to go unchecked. Recdl that MLCK activates

myosin by phosphorylation on Seri9 of its RLC (see ( 1 30) for a recent review on Rho-mediated C&'-

sensihzation).

ûther Ca"-sensitizing pathways continue to be elucidated and hotly debated. Controvcrsy has

raged over whether PKC plays a roIe in Ca"-sensitization. Muscle treatment with phorbol esters,

compounds known to activate PKC by rnirnicking their cofactor, diacylgIycerol, confer C&-sensitivity

(84). Moreover, PKC undergoes translocation to the membrane in response to treatrnent with phorboi

esters (13 1 ; 132). mer labs questioaed the importance off KC. ciîing the fact that phorbol esters, in their

hands, codd not induced contraction in CaLL-ûee solutions (133). Nor couId PKC induce contraction in

aTriton-skinned fier system, though at least on group thought this rnight be due to loss of a d i h ' b l e

&or dining the preparation of Triton-skinned fibers. Studies £rom Kitizawa showed in fact that a

diffiisiiIe factor was Iost. The were able to purify and characterize the diffiisible factor which they named

CPf-17. The protein, when phosphoryIated by PKC , binds and inhibits MLCP, thereby causing Ci?-

sensitization (134). Recently? they haveshownthat this pathway can be activated by extmceltularago~sts

(25).

These developments appear to indicate that Ca3- sensitivity is dictated by the phosphorylation

status of thick filament myosin, which in tum is dictateci by the relative activities of MLCK and MLCP,

at a given [Ca"],. The question remains, however. Do regulatory mechanisms exist that detennine the

extent to which force production necessariiy paralIeIs myosin phosphorylation? If so, these rnechmisms

might involve thin filament regulatory proteins.

2, Toward a Role for Caldesmon.

In section B.3.3.3,I revieived the evidence that caldesmon regulates actornyosinin vitro. Yet for

a long time there was no compdiing evidence that caldesmon might affect contraction within the muscle.

Promisingly, caldesrnon could be phosphoryiated in its C-terminus by MAP kinase in canine aortic

smooth muscle ( 135). In phorbol ester-stimulateci ferret aorta, W kinase even undergoes translocation

to the plasma membrane and is îhen redistniuted to the thin filaments where caldesmon resides ( 136).

However, subsequent studies showed that MAP kinase couid not induce contraction nor Ca"-sensitivity

of Tnton-skinned fibers, despite its phosphotyIation of caldesmon (137).

A role for caldesrnon in contraction continues to be elusive yet cannot be mled out. Pfitzer et ai.

have s h o w that caldesmon influences the dependence of force upon myosin phosphorylation ( 1 15).

Excitingly, Earley et al. (138) showed that partial knockout of caldesmon with anti-sense

oligodeoxynucieotides caused elwated vascular tone in c u I t d smooth muscle tissue. This indiwted

that cddesmon may serve as a reiaxing façtor hat sets basal muscle tone. This is noteworthy, since it is

estimateci that basal tone consfitutes 30-400h of total muscle tone in pen'pherd mfeSlstaace vessels. The

importance of intrinsic muscle tone has been previously underestirnated, and remains Iargeiy

underexpIoited by current pharmacoIogica[ therapies (139).

2.1 Srnooh Muscle Project: Statement of Objectives.

The objective of my studies has been to pursue new mechanisms by which smooth musck

contraction might be regulated in a Triton-skinned fiber system and further validate the models usuig the

t d s of protein biochemistry. The study presented in Chapter 3 was conducted to test the hrpothesis that

the CaL'-sensitiang actions of GTP anaiogs might by mediated by p2 1 activated kinase (PAK). PAK, as

its narne would hply, is regulated by Racl or Cdc42, members ofthe p2 lras family ofsmall monomeric

GTPases in vivo. Rad, Cdc42 and RhoA ail participate in cytoskeletal remodeilhg in nonmuscle cells

and it ivas postulated that PAK, like the RhoA effector, ROK, might influence muscle contnction. The

results ofChapter4 indicate that PAiCdoes cause Ca"-independent contraction of ïr i ton-shed smooth

muscle. Moreover, contraction was not mediated through myosin phosphorylation, hinting rit a novel

rnechanism of contraction, perhaps through regdation of caldesmon. To build a solid biochemicai

foundation for a model of contraction mediated through caidesmon, the study presented in Chapter 5

tested the hypothesis that PAK could affèct caldesmon regutation of the actin-myosin interaction.

CEAPïER 2: Materials and Methods

1. Materials and Apparatus

Human recombinant cardiac Tu1 and Tni(1-192) cDNA consrnicts were obtained h m Dr. Anne

Murphy (Johns Hopkins University, Baltimore, MD). Human recombinant cardiac TnT-containhg E.

Coli ce1 pellets were obtained h m Spectral Diagnostics inc.(Toronto, ON). Purified recombinant

cardiac TnC was also obtaùled h m Spectral Diagnostics. Recombinant murine GST-mPAK3 was

obtained h m Dr, S. Bagrodia (Corne11 University, Ithaca, NY). Al1 chemicals were obtained from

Sigma-Aldrich Canada (OakvilIe, ON) or [CN (Inine, CA) uniess specified otherwise. Proteases, trypsUi,

chymotrypsin and papain were also obtained tiom Sigma Antibodies were obtained h m several sources

whosenames are listed in thesection ihat describes immunoblot analysis. Colurnnçhromatopphy resins

were obtained h m Amersharn Pharmacia Biotech (Baie d'Urfé, QC). Uno FPLC colurnns were ob tained

Tom BioRad Labontories (Mississauga ON) and used with a DuoFlo FPLC system (BioRad). Reversed-

phase HPLC (high performance liquid chromsitopphy) was pertbrmed using Zorbax C-8 columns

obtained h m Chromatographie Speciaities (Brockville, ON). The HPLC systern h m Vruian

(Mississauga, ON) consisted of a mode1 9 100 autosarnpler, 90 12 solvent distniution system (pump) and

a 9065 Polychromator, al1 of which were connecteci to a personal computer ninning Varian Star

Chromatography software version 4.51. HPLC grade solvents were obtained h m Fisher Scientific

(Nepan, ON). Distiüed water was deionized and fïitered using a 'NANûpure Uitraiïitration Systern'

h m BamstearyDubuque, iA). Centnfùges and centrifùge rotors were h m Beckman (Mississauga,

ON). SDS-PAGE (Laemmli method) was perforrned using the 'Protean II' gel system using a 'Power Pac

1000' power suppty (Bio Rad L;iboraiories),ChemilMiu1escent reagents for immunodeteciion were

obtair!ed h m NEN-Mandel (Guelph, ON).

2. Methods

PIease note that al1 procedures and protein sarnple manipulation were conducted at MT unless

specified oiherwise. Al1 buffers were made in advance and chiiied to 04°C. When required, protease

inhi'bitors and reducing agents were added to the buffer immediately prior to use.

2.1. Protein Purification

2.1.1 Myosin, Smwth Muscle

Myosin was purifid h m fresh chicken gizzards. 100 g of muscle was rninced in a m a t ginder

and added to 500 mL of extraction buffer A, consisting of 10 mM MOPS pH 7.0.50 mM NaCI, 1 mM

MgCl,, 1 rnM EGTA, 5 mM DïT. 0.1 mM PMSF, 1 pg/mL leupeptin, 2 pM pepstatin, 0.5% (vh)

Triton X-LOO. The rninced tissue was homogenized in a blender tor 30 s and centrifilged at 9000 rpm

in a Beckman JA-10 rotor (14,000 x g) for 15 min. The pellet was resuspended, stirred wiîh buffer A

lacking Triton X- LOO (400 mL. buffer B) and recentrifuged, This washinç procedure was repeated thnx

tirna. The myosin was extracteci h m the washed myotibrils. by homogenizinç the pellet in 40 mM

MOPS pH 7.2,40 mM NaCI, 2 mM EGTA, I mM EDTA. 10 mM ATP, 5 mh.i DTT. 0.1 mM PMSF,

1 p g h L leupeptin and 2 pM pepstatin (600 mL). This homogenate was stirred for 45 min prior to

centrifugation (as before). The supernatant was made up to 600 mM NaCl, 20 mM MgSO, and an

additionai 5 mM ATP was added (while ensuring that pH remained at 7.0) before adding ammonium

sulfate up to 42% saturation. The solution was stirred for 30 min then cenirifigeci at 14,000 x g for i 5

min. The supernatant was adjusted to 60% saturation with ammonium sulfate, stirred and recentrifùged.

The pellet was dissolved in 100 mL of buffer wntaining 5 mM MOPS pH 7 2 , Z mM NaCl, 1 m M

MgCl, 0.1 mM EGTA, 0.1 mM PMSF, 55mM Dm. The solution was diaiyzed twice against 6 L ofthe

same buffer, causïng the crude myosin to precipitate over appro,uimately 3 hours. Complete precipitation

was ensured by diluting the sample 2-fold in cold distiiied, deionized water and adding 10 mM MgC&

Afterstirring for 30 minutes, the solution was centrifuged at 9500 rpm in a JA-10 rotor (16,000 x g) for

50 min. The crude myosin pellet was resuspend and washed in buffer B (200 mL). Once washed, the

myosin was resuspended in buffer B and homogenized with a sintered glass mortar and pestle. Sorty mg

of rnyosùl was resuspended in a minimal volume (10 mL) of buffer B and made up to 10 mM ATP, IO

mM MgCl, and 500 mM NaCl, and subjected to Sepharose CL-JB gel exclusion chromatography on a

colunin (2.5 cm x L rn) equilibrated in 15 m M Tris-HCI pH 7.4,500 mM NaCl, 0.1 mM EGTA, 0.01%

Na,, 1 rnM DTT and 0.1 rnM PMSF. The tlow rate was maintained between 20-30 mUh. Fractions

(approx. 1 O rnL) were analyzed by 12% SDS-PAGE- The leadhg 2 3 of the myosin peak was determineci

to be phosphatase-Eee. These fiactions were pooled and dialyzed aga& buffer B. Precipitated myosin

was concentrated by centrifugation and stored on ice for up to 2 weeks with minimal loss of ATPase

acrivity.

2.1.2 Myosin SI, Smooth iMuscle

A portion of the isolated crude myosin was set aside for the preparation ofthe soluble, 1 15 kDa

enzymatic head, rnyosin subhgment 1 (S 1). S 1 was prepared according to (140) by cle3vingmyosin with

the papaya latex protease, papain. Papain (Sigma, 14 unitslmg) was dissolved in 2 mM EDTA, 5 rnM

cysteine pH 6.0 by vortexing the sarnple for 20-30 min. The protease was activateci by ïncubating at 37

O C for 1 hr. Myosin (5- 15 mgIrni,) was then incubateci with papain (rnyosidpapain = 20001 1 (w/w)) for

IO minutes at m m temperature. Papain was inachvated by the addition of5 mM iodoacetic acid pH 7.0.

The sample was centrifuged at 20,000 x g for 30 min (25,000 rpm in a 70.1 Ti rotor) at 4°C to rernovc

most ofthe rwaining uncleaved myosin and the insolublerod domain. To ensure theircompteteremovai,

the supernatant sample was dialyzed against 10 mM imidazole pH 7.0,25 rnM NaCI, 10 mMMgCl?, O. 1

mM EGTA, L ÛiM D ï T and again centrifigeci. 1 rnM ATP was added to the sarnple supernatant before

it was apptied to a FPLC Mono Q column at room temperature (approx 22°C). S1 was eIuted with a

linear0-400 mM NaCl gradient (total volume 30 mi., 0.5 Wrnin) and 1 mL fractions were collected and

analyzed by 15% SDS-PAGE. Fractions containing S 1 were dialyzed against the appropnate assay buffer

(including 5 mM D m and stored on ice. The actin/tropomyosin-activated MgATPase activity that was

routhely obtained was 200 mol Pi hydrolyzed m i d mg-' (0.48 mol Pi mol-' S1 s-l) when papain

cleavage was carrieci out within 36 hr of tissue homogenization. The basal MgATPase activity of S 1 (but

not its activation by actin) ùicreases as it becomes oxidized. Consequently, the 'fold activation' of S 1

decreases if there is a substantial time lag (3-4 days) between the preparation of myosh and subsequent

preparation of S 1. in my hands, well prcpared S 1, characterized by low basal activity and high actin-

activated activity (at ieast 5-foId activation), ceases to be activated by actin/tropornyosin afierone week.

Smooth muscle rnyosin is exquisitely sensitive to cysteine residue oxidation. Failure to include DIT in -

any of the buffers will drarnz5calIy increase the probability that the preparation will be enzymatically

inactive regardless of its integrity as assessed by SDS-PAGE. The only exception to this rule is that only

1 rnM DïT be used in the papaidmyosin digest, since excess DTï will quench the iodoacetic acid

needed to inactivate the papain. Note that iodoacetic acid carboxymethylates reactive cysteine on S 1. The

effeçt is beneficial sinceurboxymethlyation prevents oxidation ofreactive cysteines that would otherwise

lead disulfide bond formation within the myosin had, and loss of enzymatic activity.

2.13 Myosin, Skeletal Muscle

300 g of f k h back musde h m New Zeaiand White rabbits was obtained. The muscle was

cleaned ofblood, fat and membranous tissue before was cubed on ice and rninced in a meat grinder. The

muscle was then added to 500 mL of 1 X cold Guba-Straub buffer(0.3 M KCI, 0.1 M KH,PO,. 0.05 M

K,HPO,, 2 5 mM MgCl,, 1 mM EGTA, pH 6.5 ) supplemented with 2 mM ATP. After each addition

of ATP, the pH was monitored to ensure that it had not changed substantively. The rninced meat was

blended for 15 seconds. The blended mixture was allowed to stir for 15 min. The resulting thick

homogenate was then diluted to 1.5 L with IX Guba-Stnub buffer and allowed to stir for 15 min more.

The solution was then centrifùged at 9000 rpm in a JA-IO centrifuge for 20 min (14 000 x g). The

supernatant was 6itered through a layer of cheesecloth into a 2 L graduated cylinder and the volume was

measured. The cloudy supematant mixture was diluted into a large vat containing 12 volumes ofdidled

and deionized water supplemented with 2 m M DïT, stirred briefly with agjass rod, and allowed to stand

for 3 hours. A white, myosin-rich precipitate fonned and settled slowly. After 3 hours, supernatant was

siphoned without aspirating too much of the precipitate. The precipitiite was collected by centrihgation

at 9000 rpm in a JA-10 centrifuge for 20 min (14 000 x g). The pellet was dissoIved in 100 mL of 5 X

Guba-Straub buffer (l.5 M KCl, 0.5 M KH,PO,, 035 M K2HP0,, 12.5 mM MgCl,, 5 mM EGTA, pH

6.5). The volume was brought up to 500 mL and the solution was stirred for 15 min. Once the myosin

had dissolved, ATP was added to hai concentration of 1 rnM ATP. AAer the addition of ATP, the pH

of the solution was adjusteci to 6.5 if necessary. The solution was stirred for 15 min, transferred to TI35

cmtritùge tubes, and centrifùged at 30 000 rpm in an LS-72M (1 10 000 x g) ultracentxifuge for 1 h. The

supernatant was tiltered through cheesecloth into a 1 L graduated cylinder. The volume was measured

a d the solution was diluted into 12 volumes of cold distilled, deionized water as before, and allowed to

stand in the cold roorn. M e r 15-30 min, the preparation was centrifiiged at 4200 rpm in the JS4.2 rotor

(3 500 x g) for 30 min. The pellet was then redissolved in 70 mL of 5 X myosin wash buffer ( 2.5 M KCI,

5 mM EDTA. 0.075 M KH,PO, pH 6.7). The solution was thendiiuted to a final volume of 350 mL with

distilled, deionized water, and stirred for 15 min. The pH was checked again to ensure it was between

656.7. The myosin was again precipitated by diIuting the solution with 12 volumes of distille&

deionized water. The solution was stirred bridy and aiiowed to stand for 15 min, Again, the precipitate

was coliected by centrifiigation at 4 2 0 rprn in a JS42 rotor for 30 min, The resulting pellet was again

redissolved in 5 X myosin wash bu& (50 mi.) and diluted to 250 mL and stirred for 15 min- Once

stimng, ATP was added to a h a i concentration of 1 mM. The pH of the myosin solution was verified

to be between 6.5-6.7. Mer 15 min of sîining, the solution was transferred to TI45 centrifùge tubes and

ceniritkges at 27000 rpm for 1.5 hours (LOO 000 x g), The supematant was iïitered through cheesecloth.

The myosin was fairly pure at this stage (>go%).

2.1.4 hlyosin SI, Skeletal Muscle

ivlyosin is Iargely insoluble in the conditions of most in vitro assays and it is much easier to use

the 1 I 5 kDa, soluble enzymatic head of myosin known as S 1 (myosin subhgment 1). This was prepared

by chymotryptic digestion of purified myosin by the method of (14 1). Detailed protocols may be found

in (142) . Puriîïed myosin ( h m the previous section) was diaiyzed against 10 L ofdigestion buffer(l20

mM NaCl, 13 mM N-HPO,, 8 mM NaH2P0,. 1 mM EDTA pH 7.0). The concentration was then

deterrnined by taking an aliquot to m m temperature and diluting it 20 foId in K*(EDTA) buffër(0.6 M

KC1,l mM EDTA, in which myosin is soluble) and measuring the O.D. at 280 nm. ELYO= 0.59 1 (for 0.1 %

Wh).

Myosin, in digest bufferat 18-20 mg/mL, was allowed to warm to room temperature in a 200 mL

beaker. Sufficient chymotrypsin (TLCK-treated, Sigma) was wnghed out and dissolve- in 2 mL ofdigest

buffer then added to the myosin solution such that the weight ratio of myosin to chymotrypsin was 200: 1

(w/w). The reaction was ailowed to proceed at m m temperature for 10 min before it was quenched by

the addition of PMSF to a final concentration of 2 rnM. A second aliquot of PMSF was added afterl5

min, The solution was then centrifuged at 40 000 rpm in a Ti60 rotor (1 50 000 x g) for 2 h to rernove

mdigested myosin and the insoluble LMM tail region. The supematant (crude S 1) was dialyzed against

50 rnM Tris-HCI, 1 mM EDTA, 0.01% NaN,, pH 8.0. Foilowing dialysis, the S 1 solution was loaded

ont0 a LTNO Q6 FPLC column equiliirated with the same buffër. The S 1 (100 mg) was eluted with a

0-1 50 rmM NaCI gradient. EIution was monitored by measuring the absorbante of hctions at 280 nm

and subsequently by subjecting aliquots h m each ûaction to 12% SDS-PAGE. Samples were pool&

For long tenn storage 20% glycerol was added. Under these conditions, S 1 remains active for > 1 year at

-20°C.

2.1.5 Actin

mirification of actin is a two-step process. The 6rst step involves making a muscle protein

acetone powder, which is suitable for long term storage at -20°C. The second step entails the specific

extractionof actin fiom the acetone powder. This procedure is described in detail by Pardee and Spudich

( t 43). Each step is described brietly, in turn.

Acetone powder

Fresh (or bzen) rabbit skeletal muscle (300 g, tiom the back and m e ) was sûipped of fat and

connective tissue, cut into small pieces and minced on ice. The rninced tissue was stirred for 10 min in

1 L of buffer consisting of 1 s'O mM hHPO, pH 63,100 rnM KCI. The suspension was stnined through

two layers of cheesecloth. The residual tissue pulp, retained in the cheesecloth, was stirred in 1 L of 50

mM NaHCO, for 1 O min., and strained through cheesecloth as before. Solids were stirred in 1 rn~M EDTA

at pH 7-0 for 10 min. then sûained again. in a similar manner, the solids were both stirred in 7 L of cold

distikd, deionized HzO, then strained tsvice for 10 min. The solid was stirred with 600-700 m i of ice-

cold acetone and sûained through cheescloth. This process was repeated 3 times. The protein powder

was allowed to dry on 3 MM papa in the b e h o o d for 1-2 h at m m temperature and stored at -20°C

indesnitely.

Earcction of actin

Extraçtion buffer(2 rnM Tris-HCl, O 2 rnM AIT and 0.5 mM DIT, O 2 mM CaCI,, (pH was not

adjusted) was chried on ice to 4°C. Acetone powder was hydrated using 20 mi. of buffer per dry gram

of starting mataial (final pH should be around 7.6). The s1un-y mixture was stirred with a glas rod and

placed on ice or in the cold room for 30 min. Gtoves were worn throughout the procedure. The mixture

was then iiitered in a Buchner fiinne1 containhg a piece of Whaûnan 4 filter paper at m m temperature.

The retentate was re-extracted for 30 min and subsequently refiltered. The filtrate containing

unpolymerized or globular actin (G-actin) was made up to 50 rnM KCI, 2 mM MgCl, and ailowed to

stand on ice in the cold room for 2 hours (allowing actin to polymerize). The solution was made up to

600 mM KCI, and stirred for 1.5 h on ice (which causes copurifjmg tropomyosin to dissociate). The

polymerized, or filamentous, acth (F-actin) was recoverd by centrifugation at 40 000 rpm in a Beckman

60 Ti rotor (100,000 x g) at 4°C for 1 hr- The F-a& pellet was tramferrecl to a srnail glas homogenizer

and dispersed in extraction buffer (gentiy to minimize air contact and formation of bubbles which results

in oxidation of the actin). The actin was then diaiyzed 3 times (2L) over 2 days. Following diaiysis, actin

is in its G-form. The sarnple was centnfiged at 100,000 x g h r 1 h to remove any residuai precipitate or

denatureci protein. The G-actin was filter-sterilized through a 0.45 pm syringe filter (Nalgene) and

0.0 1% NaN, (wh) was added. The G-actin was made in the days irnmediately before the experiment to

be pertbrmed. To make F-actin for actin- binding experhents or ATPase assays, the G-actin was dialyzed

at leas? three times agrùnst the final assay baer containing at least 1 rnM DïT (to rernove traces of ATP

that may accelerate depolymerization on stonge). Once F-actin is obtained, it should be stored in aliquots

onice and not disturbed. F-actin filament length is sensitive to shearstress and manipulation. Following

the polymerization >95% of the actin is in the fillamentous fon . For G-actin E28qO. l%)=l.l while for

F-actin E29O(O. 1%) = 0.69 when corrected for Rayleigh scattering at 320 nrn.

2.1.6 Tropomyosin, acq Fast Skeletai (Cardiac) Muscle

Purification of skeletal muscle tropornyosin also begins by making an acetone powder of the

muscle. However, a different protoc01 for the preparation of acetone powder is d e s c r i i betow.

incidentalty, actin can be pwïfied to equal purity from either acetone powder preparation. See referme

( 144).

Acetone Powder

15 L of ethanol, 8 L of acetone and 20 L ofdistillecl, deionized water, were pre-chilled to 4°C.

Back muscle (1 kg) was obtauied h m New Zealand White rabbits and fiozen at -80°C until needed. On

the eve of the prepmtion, the muscle was removed tiom the ûeezer and allowed to thaw overnight. The

next moming the tissue was minced with a meat grinder. The minced muscle was mixed with 1 L of cold

distilled, deionized water and allowed to sit for 20 min, before squeezing the mixture through four Iriyers

ofcheesecloth. The insoluble residue was subjected to the following series of extractions. One extraction

with L Lof%% ethanol, 3 extractions with 4 L 50% ethanol , 2 extractions with 4 L of 95% ethanol and

2 extractions with 4 L of ice cold acetone, Each extraction involved mWng the tissue with the solvent

for 5 minutes then squeezing the tissue through cheesecloth. Gloves were worn throughout the procedure.

After ihe final extraction, the tesidue was spread on 3MM paper and allowed to air dry at m m

tempenture. A typicai preparation yielded 100 to 150 g of acetone powder h m 1 kg of tissue. Acetone

powder can be stored indefinitely at -20°C.

Preparation of tropornyosin fiom acetone powder.

1 û0-L 20 g of acetone powder was mixed in a 4 L beaker with the extraction buffkr (l M KCI.

1 mM DTT pH 7.0) at a ratio of 10 mi. of buffer per gram of acetone powder. The solution was adjusted

to pH 7.0 with 1 N NaOH and lefi to stir at 4°C for 1 hour, The pH was adjusted to 75 and the soiution

was allowed to stir ovemight. The foliowing morning, the acetone powder slurry was squeezed through

cheesecloth, thevolumeofthe expeiied extraction solution was measured and the solution was kept. The

acetone powder was reextracted with an equal volume of fksh extraction buffer, The pH was adjusted

to 7.0 and stirred for 1 h as before. The extraction buffer was collected, as beforir, by squeezing the sIurry

through cheesecloth. The two (supernatant) extracts were pooled and centnfiiged at 4200 rpm in a JS42

rotor (3500 x g) at 4°C for 10 min to remove any remaining acetone powder. The supematant was

adjusted to pH 4.6 using 1 M HCI, and stirred for 20 min. and centnfuged at 4200 rpm in the JS-42

centrifuge for 20 min. The supematant was discarded and the pellet was dissolved in 800 mi, of

extraction buffa, adjusted to pH 7.0 and stirred for 20 min at 4°C. The pellet was easily redissolved by

mixing it with a little buffér then homogenizing with a hand-held glas homogenizer beibre adding the

test of the buffer. The sample was then centrifuged at 4200 rpm in the 1s-4.2 centrifuge for 20 min.

Again the supernatant was colkcted and the pellet was discarded. The pH was adjusted to 4.6 and the

solution was stirred for 20 min at 4°C before cenûifuging. The supematant was discarded and the pellet

was redissolved in 800 mL. of KCI extraction buffer. The pH was adjusted to 7.0 and allowed to stir

overnight. The next morning the solution was centritùged and the pellet \vas discarded. The supernatant

was adjusted to pH 4.3 and stirred at 4OC for 20 min. The solution was centritùged as before and the

pellet was collected. The pellec was dissolved in 1 L of 0.5 mM DTT, adjusted to pH 7.0, and s h e d at

3°C for 20 min. The solution was then subjected to 3 rounds of ammonium suifate precipitation. Brietly,

ammonium sulfate was added h m 0-35% saturation at 4OC and allotved to stir for 20 minutes prior to

centrifugation at 9000 rpm in a JA-LO rotor (14 000 x g). ï h e pellet was discarded and the supematant

was brought up to 55% ammonium sulfate saturation, stirred, and centrifiged as before- Again, the pellet

was discarded and ammonium sulfate was added to the supernatant to b ~ g it to 65% saturation. The

solution was stirredand centrihged as before. At this stage skeletai tropomyosin (bothuaand aB fom),

in the supernatant fiaction is Fd-ly pure (>85%) but may be contaminateci with troponin.

Hydroxyapatite ctiromatography was used to purifi tropomyosin h m troponin as welI as to

resolve the a u and ap fonns of tropomyosui. Tropomyosin obtaùied h m the 65% ammonium suIfate

pellets was dialyzed against 20 rnM KKPO, pH 7.0, 1 M KCI, 1 mM Dm. Meanwhile, an

hydroxyapatite coIurnn, 2 cm X 50 cm (though a t 00 cm is recommended for better resolution), was

equiliirated with the sarne buffér at a flow rate of 15 mllhour. Approximately 300 mg of crude

tropomyosin were applied to the column. The colurnn was washed with 200 mL of equilïbration buffër

before eluting the tropomyosin with a0- 250 mM KH,PO, gradient (400 mUside where A=equilibration

buffer and B=250 m . W2P04. 1M KC1 and I m M DTï) at a flow rate of 15 mi/ h. Elution was

monitored by measuring theabsorbanceofeach hctionat 280 nrn. Tropomyosineluted between 150-20

mM KH,PO,. Tropomyosin eiuted as a single p&. Fractions collected during peak elution were

anaiyzed by SDS PAGE/Coomassie staining to observe the relative elution profiles of aa and ap

tropomyosin. The leading edge of the peak contained pure au-wpomyosin while the lagging edge

contained pure ap tropomyosin. Fractions within the h n t and back halves of the peak (which contained

predominantly aa tropomyosin and ap tropomyosin respectiveIy) were pooled separately and were

rechromatographed to improve resolution of the tropomyosin isofoms.

2.1.7 Tropomyosin, ap, Smooth Muscle

Tropomyosin was purifieci simultaneousIy with caldesmon by the method of Bretscher (85).

Tropomyosin eluted h m a DE-52 anion exchange column (or UN0 Q6 column ) between400-500 rnM

NaCl (see caldesmon purification for buffer conditions). Tropomyosin was estimated to be >95% pure.

Fractions were pooled and dialyzed against WHCO, three Cimes (2 L), lyophilized and stored at -20°C.

2.1.8 Caldesmon

Chicken gizzard caldesmon was p&ed essentiaüy as desmieci by Bretscher (145). Smooth

muscle was obtained h m gizzards (fiozen or f k h ) stripped of fat, and was minced in a meat gtinder.

The minced tissue was added to 3 volumes ofcold buffercontaining 50 rnM imidazole pH 6,9,300 mM

KCI, 1 mM EGTA, 5 mM EDTA, 1 mM PMSF, 1 mM DîT, and homogenized in a blender. The

homogenate (600 mL) was transferred to a 1 L beaker with astir bar which was, in htm, placed in a 3.5L

beaker (on a heatingtstirplate) contairing 800 mL ofboiling water. The muscie homogenate was heated,

with stirring to > 90°C. The heating process took 15 min and the homogenate was kept at >90°C for 5

minutes. The homogmate was dowed to chi11 on ice for 30 min pnor to mtrihgation at 9000 rpm in

a Beckman JA-10 rotor (14 000 x g) for 40 min, Ammonium sulfate was hen added to the supematant

until the solution was 30% saturated. The solution was cenûifuged at rpm for40 min. Ammonium

sulfate was added to the supematant until the solution w u 50% saturated and recentrifuged. The

precipitate was dissolved in 5 rnL gel filtration buffer containing 20 mM MOPS pH 72,500 mM NaCl,

0.1 mM EDTA and 1 mM Dm. ï h e sample was diaiyzed ovemight against 1 L of the same buffer at

4°C. Prior to gel filtration, the caldesrnon-richsample was centrifugeci for 30 min in aTi60 rotor at 440000

rpm ( 150 000 x g) to remove any polymerized tropomyosin. The protein sampie (approx 100 mg) was

applied to an equilibrated Sepharose CL4B column (2 cm X 75 cm) and eiuted at a flow rate of 30 mU

h. Effluent ûactions (10 mL) were analyzed spectrophotomeûically at 230 nm and subsequently by 12%

SDS-PAGE, Caldesmon (apparent Mr of 140 kDa) elutes at a Iowa elution volume than the major

contaminant tropomyosin. Frictions were pooled to maxirnize purity ratherthan yield. PooIed caldesmon

was dialyzed against 10 mM imidazole pH 7.0,30 rnM NaCl, 0.1 mM DR. The solution was applied

to a pre-equiüirated DE42 anion exchange colurnn (2.5 x 10 cm) for conventional chromatography at

4"C, or aItematively applied to a LN0 Q6 column (BioRad) and subjected to FPLC at m m

temperature. in either case, caldesmon was eluted with a hear gradient of 3O-5OO mM NaCl (250 mL)

at a flow rate of 1 mUmin. CaIdesmon was identifieci by SDS-PAGE, pooled, dialyzed 3 times against

10 mM NH,HCO,. lyophilized and stored at -20°C. Since caidesmon is susceptible to oxidization, it

should be redissolved in b&ér containing 5 mM D ï T and centrifùged at LOO 000 x g, to remove

aggregates, prior to experimentation.

2.1 9. Recombinant Td, Tnl.,, and TnT

Tni and TaII-,,, TnT were expressed in E coli. Glyceml(SO% vlv) stocks of the E. coli sûain

BL-2 1 DE-3, containing the plasmids thiit encode the ûoponin subuniîs, were streaked on LB-agar plates

containhg 50 p g / d ampicillin . Single colonies were grown ovemight at 37°C in a shaking incubator

(10 mi, of ovemight culture per Mer to be cultured the following &y). Ovemight cultures were

centrifigeci in IO mL aliquots to pellet the E. coli and remove the Li3 media which has accumulateci P-

lactamase. Fresh LB (10 d) was added to the peIiets which were redispersed by vortexing. These 10

mL sarnples were used to hocdate up to 8 L of LB containing f m h ampicillin ( 100 m&). Ceils were

gmwn at 37°C in a shaking incubator until an O.D, reading of0.6-0.8 was reached. Protein expression

was induced by the addition of iPTG to a ha1 concentration of50 mgL and M e r cuItured for 1 h.

Cells were harvesteû by centrifirgation in 1 L botties for 20 min at 4000 rpm in a JS-4.2 rotor and stored

at -70°C. The foliowing day, cells were thawed and muspendeci in 20 mM HEPES, pH 7.0,350 m M

KCl, 5 mM EDTA. 1 rnM EGTA, I mM P-mercaptoethanol. 1 mM PMSF and 0.5 mgmL lysozyme.

The suspension was allowed to sit on ice for 70 min prior to son idon five times for 30 s at I-min

intervals. Sonicated cells were centritùged at 10 000 x g and the supematant was discarded. The pellet.

containing inciusion bodies (recombinant protein) as well as bacterial ce11 walVmembrane proteins, was

washed by homogeniang it in 30 mM Tris pH 7.0, I M urea, 5 mM EDTA, 1% (vh) Triton X-100,l-2

rnM D?T with a large glas homogenizer. When urea buffa were used, it kvas aiways the first

comportent added to the solvent. The dissolved urea was deionized by stimng in 1 tbsp of AG50 1-X8

mixed bed resin (BioRad) to remove cyanates, then atered through a0.45 pm membrane, before the rest

of the buffer componenîs were added. Tris is a mmmended buffer for urea solutions since Tris

quenches the formation of new cyanates on storage at 4°C. Once hornogenized, the solution was

centrifiiged at IO 000 x g. The supernatant was discarded and the petlet washed 2-3 more times or until

the pelIet was whitish gray and the supernatant was clear. The washed inclusion bodies contain pr imdy

recombinant protein. The pellet was dissolved in 25 mL of 50 m M Tns-HC1 pH 8.0,8 M ma, 5mM

EDTA, L mM DTT and centrifkged at 100 000 x g to remove high molecular weight proteins or protein

aggregates. The supernatant was then applied to a CM-cetlulose column 2.0 X 10 cm, equiliirated with

the sarne buffer and the recombinant protein was eluted with a 100 rnL linear gradient of 0-300 mM

NaCl at flow rate of 2 mUmin at r o m temperature. Column fiactions were assayed by 12% SDS-PAGE.

At this stage the proteins were >95% pure. Proteins were dialyzed once against 0.05% TFA (3 L) then

twice against water containhg 0.5 mM DTT, aliquotted and iyophilized.

2.1.10 Calmoduiin

Recombinant caimodulin was purified by hydrophobie interaction chromatogriphy. Recombinant

cdmodulinexpressed in BL2 1 -DE3 wascultured essentialiy as descnied for the recombinant caldesmon

üagments with the exception that ceils were cultured at 30°C. Induction with PTG was done at 37°C.

This was accomplished by warming % ofthe volume of LB to 55-6U'C and adding it to the 30°C culture

at the time of induction. Harvested E coli were resuspended in the calmodulin homogenization buKèr

(25 mM Tris-HC1 pH 7.5, 1 m M DTT and 0.02% NaN, ) and sonicated. Sonication was hIIowed by

centrifùgation at 15 000 x g and clarification by centnfùgation at 150 000 x g. CaClz (final concentration

5 mM) was added to the sample before it was loaded ont0 a Phenyl-Sepharose (Pharmacia) coIumn ( 1.5

cm x 10 cm) equiiiirated in the same buffer at room temperature. The column was washed with 10

volumes ofbuffer then with one column volume ofbuffer containing 500 mM NaCl. Finally, caimodulin

was eluted in buffercontaining both 500 mM NaCl and 10 cnM EGTA (at least 2 column volumes). The

purity of the calmodulin was assesseci by SDS-PAGE before it was dialyzed extensively (3 changes)

against NH,HCO,, lyophllized and stored at -20°C. Note that PhenyI-Sepharose chromatopphy must

be çamçamed out at room temperature, since caimodulin wül not bind to the çolumn at 4°C.

2.1.11 Recombinant GST-mPAK3

The plasmid constnict received h m Dr. S. Bagrodia contained the murine PAK3 in a pGM

vector such that the sequence encoded a GST-mPAK3 Fusion protein. Expressing GST-mPAK3 in

traditionai expression strains ofE. coli. such as BL-2 1, proved highiy problernaticdue to plasrnid toxicity

and excessive proteolysis of PAK. The vector was subsequently trrinsfomed into E. coli strain M l 1 0

which is normally not recommended since it is not reaIly an expression strain. Expression levels were low

(< 5 pg PAK per L culture) and cells grew slowly. Cells were induced with [PTG as in the previous

section. Induction time was typically 1 hr. Cells were havesteci by centrifigation in a JS-42 rotor at 4

000 rpm for 10 min, then stored at -80°C ovemight.

Cells were thawed in bufier( 15-20 mL. per L pellet) containing 50 mM Tris-HCI pH 7.5,50 mM

NaCI, 5 mM EDTA, 0.5 mg/mL lysozyme, I mM PMSF, 5 p g h d pepstatin, 5 yg/rnL leupeptin and 1

mM DTT. Cells were ailowed to sit on ice for 70 min for the lysozyme to be effective. Cells (total vol

50-75 mL) wre sonicatd 5 times for 30 s at 1-min intervais (during which tirne the solution is kept on

ice or an ice/salt/ethanol mixture). Triton X-LOO. finai concentration 0.1% (vtv), w;ts stirred into the

homogenate and the mixture was altowed to stand for IO min at 4°C prior to final sonication for 30 s,

The cell homogenate was uitracentritùged in a Ti45 rotor at 35 000 rpm ( 100 000 x g) for 30 min. The

supernatant was mixed batchwise with 2 mL of glutathione-Sepharose (Arnersham Phamacia) in a 50

mL conicai tube. The resinlsample was incubated (on a turnbling mixer) for 30 min. The resin was

aiiowed to settle. Alternativeiy the c o n i d tube was centrifuged at lûûû rpm in a JS4.2 rotor to speed

the settling of the resin. The supernatant was removed and the resin was washed three h e s with 1 O mi.

of Tris-saiine-EDTA buffer (20 mM Tris-HCI pH 7.5,150 m M NaCl, 1 m M EDTA). GST-mPAK3 was

eluted h m the glutathione-Sepharose resin with successive batchwise incubations (5 mL each for 10

min) with 50 mM Tris-HCI pH 8.0,s rnM reduced glutathione. The PAK samples were diaiyzed against

10 mM irnidazole pH 7.0 if ihe preparation was to be used in Triton-skinned fiber experiments.

OtIierwise, PAK was didyzed against 20 rnMTrk-HCl pH 7.5,s mM MgCI,, 1 mM DIT. Ahdidysis,

PAK was concentrateci fiom a volume of IO mL to 600-800 pL using Centriprep 50 spin columns

(Amicon). PAK could be stored in 50% glycerol, 0.01% NaN, (wh) at -20°C for up to six months with

minimal loss oflanase activity. Kinase activity was tested on a batch to batch bais against myelin basic

protein as in Childs et al. (1 10)

2.2 Characterizaiion of Protein Phosphorytation

23.1 In vitro Phosphorylation by PAK

Typically caldesmon, myosinor ML% (1-2 mg/mL) were incubated with GST-mPAK3 (approx

0.02 mglmL) in 20 mM Tris HCl pH 7 5 5 mM MgCb, 1 mEv1 Dm. 1 mM $'P-ATP (approx 500 000

cpdnmol) was then added to trigger the reaction. Phosphorylation reactions were routinely performed

at 30°C. Phosphate incorporation was monitored by spottirtg aliquots ont0 Whatman P8 1 papa (5 PL),

washing the paper four times with 75 mM H,PO,, drying the paperwith successive washes ofethanol and

rtcetone, foilowed by scintillation counting on a Beckman LS-80 using a program/channeI with an open

detection window suitabte for monitoring high energy missions. PAK activity !vas observai over a

broad range of NaCl concentrations (0-300 mM) and MgC12 concentrations (1-10 rnM) at pHs ranghg

h m 7.0 to 8.0.

232 Phosphoamiuo Acid Analysis

15 pL of radioactively phosphorylated caldesmon, myosin or ML-,, (appmx. 2 m m ) , were

added to 100 PL of6 N HCI degassed and hydrolyzed at 1 L O T for 2 h. The srimpie was &en diluteci with

10 volumes of distilled, deionized water, evaporated under vacuum and reconstituted with 15 pL of pH

1.8 eiectrophoresis buffer (0.08% v/v acetic acid, 0.02% v/v formic a d ) , The hydrolyzed sample, as

weli as phosphoserine and phosphothreonïne standards (2-5 mol, h m Sigma), were subjected to

electrophoresis on thin layer cellulose plates (Kodak) at 400 V for 2 h then ailowed to air *The intenial

standard phosphoserine and phosphothreonine were visualized by first spraying the cellulose plates using

an atomizer containhg 1% ninhydrin (wlv) dissolved in acetone, then heating the plate in an oven or with

a hair dryer. The plate was then subjected to autoradiography. The identity of the radiolabeled

phosphoatnino acid was determined by comparing its electrophoretic mobility (hm the autoradiogram)

with that of the internai standards (stained wvith ninhydrin on the plate).

2 2 3 Phosphopeptide Mapping

Radioactively phosphorylated caldesmon or MLC20 (approx 50pg) was digested with üypsin

(TPCK-treated, fiom Sigma) at an enzyme : substnte weight ratio of 150, overnight at 37°C in

NH,HCO, pH 8.0,l mM CaCI,. (CaCl? suppresses autolysis of trypsin). A second aliquot of trypsin was

added the following moming and the reaction was dlolowed to proceed for I h. The digested sarnple was

lyophilized and redissolved in pH 1.8 electrophoresis buffer (0.08% v/v acetic acid, 0.02% v/v fonnic

acid). The samples (approx 10 p ~ ) were then subjected to hvo dimensional mapping on Kodak thin layer

cehlose plates (20 X 20 cm). The first dimension consisted of thin layer electrophoresis at 400 V for 2

h in pH 1.8 buffer. The plates were ailowd to air dry before they were subjected to ascending

chromatography (perpendicular to the direction of electrophoresis) in l-butanol : pyridine: acetic acid

water ,40:50:10:40. M e r chromatography, the plates were dowed to air dry in a h e h o o d prior to

autoradiography.

2 3 Binding Assays

23.1 Afnnity Chromatography

Hman recombinant TnC and TnT were coupled to Wr-activateci Sepharose 1B (Phmacia)

as desc r i iby the rnanufàcturer. ïhe columns (1 mL) w m equilibnted with 10 mM imidazole pH 75,

20 mM NaCl, 1 m M EGTA, 1 mM DTT, 0.0 1% NaK. Proteins were eluted by dropwise application of

buffers whose composition is given in the figure legends. Eluted tiactions were analyzed by 12% SDS

PAGE or by RP-HPLC.

232 Measurement of Inlrinsic Tryptophan Fluorescence

Fluorescence measurements were performed with a Perkin-Elmer spectrophotometer. Details

regardhg the binding buffers, excitation wavelengths, mission tilters and protein concentrations used

in fluorescence studies are dacribed in the figure legends of the appropriate chapters.

Binding curves in chapter 5 were fitted to the following equation using a non-linear regsession routine,

~=(&, /r ) [~ - (~ '4nP,Q! ' ] /2nP,

where AI is the change in tluorescence intensity in the caldesrnon 6-a-ment induced by the ligand (e.ç.

calmodulin), Io is the fluorescence of the caldesmon fragment in the absence of ligand. &, is the

maximum change in caldesmon fiagrnent fluorescence when saturated with Iigand, A = nP f L+&, n =

the number of binding sites per mole of caldesmon hgrnent P, = the total caldesmon fnCment

concentration, L, = the total Ligand concentration. Y, = the dissociation constant.

2 3 3 Ultracentrifugation

Two nmoi of F-actin were mixed with 0.4 nmol of tropomyosin and O to 1.5 nmol

phosphorylated or non-phosphorylated caldesmon (or Tni or Tni mutant or peptide depending on the

study) in 200 pL of binding buffer (40 mM Tris pH 7.5, LOO mMNaCI, 5 mM MgCl,- and 1 mM DIT).

Protein mixtures were allowed to equiliibrate for 30 min at m m temperature prier to centrifùgation at

100,000 x g in a B e h a n TL-Iûû ulûacentrifùge at room temperature. This tesuited in 90-99% of the

F-actin being pellet& Pellets were rinsed with actin-binding bufferonce and dissolveci in 100 pL 0.05%

(vh) TFA in water. Relative arnounts of caldesmon or Tni, tropomyosin and a& were determineci by

reversed-phase HPLC- Briefiy, sarnples (pellets) were appiied to a Zorbax SB-C8 HPLC column. The

proteins were eluted at a flow rate of 1 mUrnin using a 2% B/min Iinear gradient where eluent A was

0.05%(v/v) aqueous TFA and eluent B was 0.05% TFA in acetonitrile. Elution was monitored at 210

nm. The area under the peaks within the resulting chromatograms were determined by integration. The

arnount of protein present could be determined by cornparison with a standard curve of integration uni&

vs. nmol ofprotein, Standard curves were obtained foreach protein or ûagment, Note it is possible to use

higher concentrations of actin when doing binding experiments ~4th peptides. In this case, one would

monitor actin at 2280 nrn and the peptide at 2210 m.

Alternative method.

üitracenûifuged samples wereaIso be maiyzed by densitometry of Coomassie Blue-stained gels.

Gels were stained for at least one hou in 1% Coomassie Blue R, 50% Methmol and 10% Acetic acid

to ensure that staining was complete and diminateci ambiguity caused by diffaenha1 staining rates of the

proteins. Gels can then be destauied ovemight. with at least 2 changes of stain. Shortcuts should be

avoided. Gels were malyzed by scaming the with an HP6000 scanner. Densitometric volume of the

bands was detennined using Scanplot v. 5.06 sothvare.

2.4 Kinetic msay

2.4.1 Determination of Actin-Activated SI-AïTase Activity

Actin-activatd myosin S 1 ATPase assays, in the presence and absence of tropornyosin, were

conducteci in a 96-weii ELISA plate ( Iûû pL assay volume). inorganic phosphate was r n A

colorirnetridy as descn'bed in (38). ATPase buffer containeci 40 mM Tris pH 7.5, 50 mM NaCl, 5 mM

MgCl, and 1 mM DTï. Myosin S L ATPase activity was deteminecl for reaction mixtures containhg O

to 4 ph4 caldesmon (or Tni, Tni mutant or Tn), 10 pM actin and 0.5 pM S 1 with or without 2 @A

tropomyosin for 10 minutes at 37' C. ATPase reactioas were initiated with 4 m M ATP and taminateci

by adding 100 pL of a solution containing 3% ascorbic acid, 0.5 M HCI, 4% SDS and 0.5% (wiv)

ammonium molybdate. Phosphate release was Linear within the t O-min reaction. Colour was allowed

to develop for six minutes prior to the addition of 100 pL of a solution of 2% sodium citrate, 2% sodium

m-assenite, 2% acetic acid, foiiowed by 10-min incubation at 37°C before absorbante at 650 nrn was

measured in a Molecular Dynamics E-max plate reader. Phosphate content was determined by

cornparison to a potassium phosphate standard c w e ranging fiom 0-100 nrnol. Less than 10% of the

ATP was hydrolyzed over the course of the reaction The ATPase activity of S 1 was subtracted h m the

actin-activated values,

Successtùl ATPase experiments require particuiar attention to several panmeters. The effective

shelf life of srnooth muscle myosin. and its subhgments, is limited, thecefore the enzymes shouId be

prepared immediately pnor to experimentation. By contrast. skeletal muscle S I may be stored for over

1 yr at -20°C in ATPase buffer containing 20 mM Tris-HC1, pH 7.5,50 mM NaCl, 0.1 mM EDTA, 5

mM DTT and 50% glycerol at a cancenûation > 10 mg/mL. Troponin 1 and troponin T. individually, are

not very soluble. They aggregate andor oxidize upon kze/thaw. At ionic strengths < 200 mM, the

solubility limit is around 0.5 rnglmL. Troponin 1 and Twill cause aggregation of actin-TM if added to

the solution at high concentrations. F-actin ceases to effectively activate ATPase if the filaments have

been sheared. Once F-actin has been prepared, it is stored on ice, in aliquots and not distilrbed unies

absolutely necessary. F-actin loses its cqacity to activate ATPase afkr 1-3 weeks dependmg on how

it had been treated. Since troponins (individdy or as a complex) tend to aggregate on successive

W d t h a w cycles, they must be as pure (protease-fiee as possible), and stored on ice with a bacteriostatic

agent (0.0 1% NaN3) in mail aliquots. The reproducibility of the ATPase assays was highiy dependent

on the accuracy and the amount of myosin dispensed in the reaction. Therefore, accurate 'positive-

displacement' pipettes or HamiIton sytinges are reconunended. F-actin should bedispensed with pipette

tips whoseends have been cut off, to prevent shearllig of the actin filamenk The activity of actornyosin,

or acto-S 1 inçreases several-fold when the assays are carcied out at 37OC (compared with m m

temperaîure). Activity is inversely proportionai to the ionic mngth of the buffer used in the AïPase

assays.

2.5 Gel Electrophoresis Techniques

2.5.1 2-D Gel Electrophoresis

Proteins, (e.g . radiolabelal phophddesmon in Chapter 4) were resolved by nvodimensional

gel electrophoresis according to astandard protocol (Bulletin 1 144) h m Bio-Rad. Proteins were resolved

in the h t dimension by isoelectric focusing on a mini-Protean U isoelectric focusing gel electrophoresis

apparatus (tubegels. Bio-Rad) using an arnpholyte mixture oF 10% pH 3.5 to 10,O and Wh pH4.0 to 6.5.

Protein molution in tfie second dimension was d e d out by 12.5% SDS-PAGE. The hvo-dimensiond

gels were stained with Coomassie Blue and dried. When requid. autoradiography was pertbrmed

directIy on t'ilm (X-ûmat Blue YB- 1. Kodak). TO determine the pH gradient. blank gels were resohi

in the k t dimension and cut into evenly sized slices. Each slice was immerssd in 1 mL of 10 mM

potassium chioride, vortexed, and incubated at m m temperature for 1 h, and then the pH of the solution

was detwmined. Linear regression of the pH and gel distance was unied out (y = 0.3386~+4375, where

y = pH and x = cm h m acidic end of gel) to detennine the actual pl of the various phosphoproteins.

Migration in the secund dimension was compared with bmad range protein markers (New England

Biolabs, Beverly, Ak).

25.2 Gel Overiay hsay

intact and skinnedTaenîacoli muscle samples were analyzed by 10% SDS-PAGE supplemented

with Lû% giyceroi, 5 mM MgCl,, and L mM dithiothreitol. FoUowuig transfer to nitrocellulose, the

proteins were denahired by incubating in a solution of6 M -*dine-HCl, 50 mM ZnCh, 5 rnM MgCl,,

25 mM MES, pH 6.5, and 0.05% Triton X-100 (30 min at 4°C). Protek were then renatured by

incubating three tirnes with 50 rnM NaCl, 2.5 mM Dr, 25 rnM MES, pH 6.5,1.25 rnM MgCl,, 50 mM

ZnCl,, 1% bovine senun albumin, and 0.05% Triton X-100 (2 h at 4°C). Purifieci recombinant hurnan

Cdc42 and Racl were labeled with ["SI-GTP-yS prior to probing of the nitrocellulose blots.

2 5 3 [mrnunoblot Analysis

Taenia coli muscle fibers (skinned or unskinned as described in Chapter 4) were homogenized

and subjected to 12% SDS-PAGE. Gels were then allowed to equilibrate in transfer buffer containing 20

mM Tris, 192 mM glycine and 0.02% SDS, 20 % v/v methanol for 15 min. Gels were rnounted in a

'Trans-Blot' apparatus (BioRad). Proteins were transfmed to Immobilon P (MiIlipore) by applying 200

rnA for 90 min. The blot was then incubated with 5% (wlv) Carnation instant milk powder dissolved in

Tris-buffered saline (TBS) solution for 2 hours at m m temperature or at 4°C ovemight. The blot vas

rinsed with TBS + 0.05% Tween 20 (TTBS). Blots were then incubated with antibodies toward PAK

(anti-mPAK3 NT3, polyclonal, fiom Kinetek Biotechnologies, Vancouver), caldesmon (hHCD.

monoclonal, Sigma), Desmin (DEU-IO, monoclonal, Sigma) and myosin regdatory Iight chains (my2 1,

monoclonai, Sigma) in a 1% solution of instant milk dissolved 'lTBS. Blots were incubated for 2 hours

at room ternperanire. Blots were washed 4 x 5 min in Tiï3S. Blots were incubated with horseradish

peroxidaseconjugated anti-mouse anhibody for up to 2 h. Blots were washed as before and developed

using Enhanceci Chemil&escence reagent (Amersharn) and exposed to X-ray film.

Chapter 3: Cardiac Troponin 1 iruncation associatecl wiih myocardial stunnhg demonstrates

supranormal function: implications for the mechanism of myocardial shuining

Foreword

This manuscript in preparation has five CO-authors. 1 am the îïrst author of this study, t performed

di protein purification, binding studies and ATPase assays, and wrote the manuscript. Dr. Heather Fraser

performed the adenovinis transfection studies and theu associated physiological measurements. Dr. Anne

Murphy kindly provideci the plasmid constntcts encoding human cardiac troponin 1 and its deletion

mutant troponin Dr. Eduardo Marbm was Dr. Fraser's postdoctod supervisor. As senior author,

Dr. Jennifèr Van Eyk conceived the project, and contnbuted to experhental design.

Abstract

It has been show previousiy that rnyowdial stunning, in the isolated rat heart modei, is

characterized by specific Td proteolysis at its C-taminus, yidding tnuicatedTd,~,,. (McDonough,J.L.,et

al. (1999) Cire-Res, 849-20 ). Transgenic mice that express Td,-,,, mpitulate the cluiicril haIlmarks

of stunning (Murphy et al (2000) Science 287,488-49 1). To assess themolecular basis of the contractile

dystwiction we have undertaken biochemid studies ofhurnanTni,,,, the homolog ofmousdrat Tni,.,,.

Truncateù TnI binds TnT, TnC, actin and actin-tropomyosin rvith comparable affinity to intact Td, and

is miirginally l e s effective at inhiMing actin-tmpmyosin-activated ATPase activity of myosin

subfiagment 1 . Inhibition of actin-tropomyosin ATPase by TnI 1-192,. when c o r n p d with inhibition

by hi1 length TnI, was more eady reversed by Troponin C. Troponin reconstituted with TI&,,

displayed Cs*-sensitization ofits Ca'=activated iictin-tropomyosin-koponin S I ATPase when compared

with wild type troponin. Furthmore, the mutant troponin conferred increased maximum Ca'-activated

ATPase by SP/D. Rabbit ventricular cYdiomyocytes adenovinliy mfec ted with mouse Ta-,,,

disphyed incrwed ce11 shortening. These results suggest that Ca?'-desensitization of isornetric force,

obsmed in îïbers h m stunned myocardium, is not amiutabIe to TnI proteolysis per se. Given the

deçreased maximum force observed in previous studies, truncation o f the C-terminus of TnI rnay cause

force and ATPase to become uncoupled.

Introduction

Myocardiai ischemiakperfkion (VR) injury is a leading cause of morbidity and mortality in

North America Patients may experience ischemia as a a i t of coronary ihrombosis or as a necessary

complication ofheart surgery. YR injury is a serious condition, even in instances of rnild ischemia Under

these conditions, themyocardium survives but exhiiits contractile dysfunction that is revmileover time.

This has been tenned myocardial stunning (1 16).

Recent work has demonstrated that reduced contractility during myocardiai stunning cannot be

attributed to poor myocyte excitability or a b e m t Cri--handIing, suggesting a tesion ofthemyofilaments

(123; 125). Arnong myofilament proteins, only cardiac Tni undergoes pmteolysis in response to 15 or20

min ischemid45 min repertùsion, which was not observed with ischemia alone. Furthemore. the extent

O tTni proteolysis correlateci with decreased mâ,uimurn forcedevelopment in skinned muscle fibers, (1 19)

and in intact tnbeculae, fiom stunned hearts, in an isoiated rat heart model (125). The VR-hduced

proteolytic ûagments were subsequently identified using imrnunologid analysis andmass speaometry.

Tni,.,,; (hurnan Td,-,,3 resulting h m specific proteolysis of 17 amino acid midues h m the C-

temiinus, was the preponderant fiagrnent in the myocardium of mt h- hat undenvent the stunning

protocol(118). This prompted Murphy et al. to constmct a transgenic mouse mode1 of stunning by

expressing Tni,~,,, under the conml of the a-EvMC promoter in adult mouse karts (120; 121). Mice

expressing the transgene at levels of 10-20% had dilatai hearts but sarçorneric strucnrre was preserved.

Transgenic mouse h e m displayed both systolic and diastoiic dysfùnction. Lefi ventricular tmbedae

h m the Tn1,-,,, mice developed las maximai isomeûic force than trabeculae hmcontrul mice, despite

nomal ~a"' availability. Thus the transgenic mouse model faitNùlly recapituiated the salient features of

the rat hart mode1 of myocardiai stunning.

Troponin I is one of the subunits of the troponin complex which regulates cardiac and skeletal

muscle contraction in response to Ca" (See (52; 146) for reviews), Reversiile binding of Ca" to the of

Troponin C (TnC) subunit causes the troponin cornplex to undergo a conformationai change that aiters

the position of tropomyosin on actin, exposing myosin binding sites. Myosin then binds actin, forming

crossbridges that lead to force production, The currently accepted roIe for Tni is thaf by virtue of its

interaction with actin, it confines troponin-tropomyosin in a position that interferes with binding of

myosin to the thin filament when Cd' is absent. in the presence of cal', Tni undergoes a conhmational

change such that the actin binding region 126-146 (fast sTn195-115) becomes bound to TnC, dleviating

constraints on the position of troponin-tropomyosin. Since these mino acids bind actin and TnC, TnI

is ofien regarded as a molecular switch. Given the pivotd roIe of Tni in contraction, its proteolysis is a

potential cause of the reduced force production observed in stunned myocardium.

To determine the molecuiar mechanisni by which mncated Tni contributes to the stunning

phenotype in transçenic murine hearts, ive have undertaken the smdy of mouse Tn1,-,,, and its human

homolog Tni,-191. using biochemical and myocyte transfection studies. (Portions of this material have

appeared previously in abstract form.)

Materials and Methods

Tu1 fragments used in this study

Figure3- 1 shows aschernatic representation ofthe constructs and synthetic peptides, used in this

study, within the primary sequence of human Tni. Tnl,,,, corresponds to the sequence of Tnt that is

cleaved in the stunned rat hm. Tni,,,,, was synthesized in an attempt to ensure that the behavior of

TnI,,,,, did not result fÏom the loss ofcrucial tùnctionai determinants in the irnrnediate vicinity within

the primary sequence.

Protein preparation

The synthetic Tni peptides,Tni,,, andTni,,,,, weresynthesized using standard Fmoc-peptide

synthesis. Peptide authentici~ was verifid by MALDI-TOF mass spectrometry. Peptides were purified

by reversed phase high perfomance liquid chrornatography as described previously (147). Human TnI

and and TnT were expresseci in E. coli as descnied in the Novagen online manual- Inclusion

bodies nere isolated and dissolved in 6M ureii. Tni and Tnl,-,,, were subsequently subjected to CM-

celluiosechromatoyaphy. Thc CM-cellulose coIumn (2.5 cm X 10 cm), was equilibrated in 10 niMTris

pH 8.0, 6 M urea, 5 mM EDTA, L mM DIT, TnI and T d ,,, were eluted in equilibntion buffer

containing 500 rnM NaCl (total volume 250 mL, 2 mllmin). Puity of eluted samples was assessed by

12% SDS-PAGE. Purified proteins (>95%) were pooled, diaiyzed against water containing I rnM DïT,

lyophilized and stored at -80°C. Human cardiac TnC was supplied in purified form h m Spectrai

Diagnostics (Toronto, ON). SkeIetal muscte actin was purifiai as d e s c r i i in (143). Skeletal muscte

myosin subfiagrnent 1 (S 1) was purified according to Weeds and Taylor (141). aa-tropomyosin was

p r e p d as descriied by Srnilie (144). Refolding of hurnan recombinant troponin (Tn) h m its

constituent subunits, was conducted as desm'bed previously ( 147). Thesynthetic peptides weredissolved

in appropriate buffer and the pH checked. S 1 and F-actin concentrations were determinai by extinction

coefficient with correction for Rayleigh scattering at 320 nm (147). The concentrations of ail other thin

filament proteins were determined by the method of Lowy (6).

A M t y chromatography

Human recombinant TnC (2 mg) and TUT (2 mg) were coupled to CNBr-activated Sepharose48

(Phmacia) as described by the manufacturer. The affinity columns (1 mL) were equilïùrated with

quilibration buffer consisting of 10 rnM irnidazole pH 7.5,20 rnM NaCl, 1 mM EGTA, 1 mM DTT,

0.01% Nd,. T d and Tni,-,,, (in a 1 : 1 weight ratio) of each synthetic peptide individually, was applied

to the affinity colurnns. Following sampleapplication, TnT-Sepharose was reequilibrated before applying

a gradient containing 1 M NaCl in buffer(tota1 volume 20 mL, 0.5 Wrnin). This was followed by an

urea gradient tiom 0-8 M urea in equilibration buffer + l M NaCl. The TnC-Sepharose columns were

equilibrated in the sarne manneras TnT-Sepharose except 2 rnM CaCl, had been added to ail buffm. The

TnC column wits then washed batchwise wiîh equilibration buffér containing, LOO, 500 and 1000 mM

NaCl, 1OOO mM NaCl + S M urea and finaily 1000 mM NaCl + S M urea i. 10 mM EDTA. ExpeRrnents

conducted with Tni ~rTni , - , ,~ were analyzed by 12% SDS PAGE. Experirnents conducted with peptides

Tni,,, and Tni,,,, were anaIyzed by subjecting 100 pL aliquots of each hction to RP-HPLC

(Varian) on a Zorbax 300SB C-8 column. Peptides were eluted with a linear AB gndient where buffér

A contained 0.05% aqueous tnfluoroacetic acid (TFA) and buffer B contained 0-05 % TFA in

actetonitriIe. Elution was conducted at 2% Blmin at a flow rate of 1 mUrnin. Peptides were eluted in a

2% actetonitrildmin gradient and the areas of the eiuted peaks were integnted.

Actin-bindiog studies

Two m o l o f f -acth were mixed with 0.4 m o l oftropomyosin and O to 1.5 nmoI of Td or Td

mutant in 200 pL of binding buffer (40 mM Tris pH 7.5, 100 mM NaCI, 5 rnM MgCIL and 1 mM

Dm. Protein mixtures were ailowed to equilibrate for 30 min pnor to centrihgation at L00,000 x g in

Figure 3-1. Fragments of TnI used in this study. Recombinant human cardiac Td,- and Td,-,, were expressed in E. d i and used in biochemical experïments. In the

cultured myocyte experiments, murine vectors bearing cDNA encoding cardiac TnI,_ zlo and Td,-,,, were transfected. Synthetic peptides conesponding to the last C- terminal 17 and 22 amino acids Tdt93-209 and Td188-m were synthesized and purified as desdeci in 'Materials and Methods'. The numbers in brackets refer to the equivalent amino acid positions within the primary sequence of the mouselrat isoforms of TnI.

a Beckman T L 4 0 ultracentnfùge at room temperature. This resulted in 9049% of the F-actin being

pelleted, Peiiets were rinsed with binding buffer and dissoived in 100 pL of reducing SDS sample buffer

(New England Biolabs). Samples were analyzed by L2.4 SDS-PAGE. Gels were stained with Coomassie

blue. Densitomeûic analysis was conducteci using Scanplot v5.06 (Cunningham Software).

Actin-activated myosin S1 ATPase

ATPase assays were conducteci in a 96-weU ELISA plate (200 pL reaction volume). inorganic

phosphate was determineci colorimetricaily as described by Chifflet et al. (148). ATPase buffa contained

40 mM Tris pH 7.8, 25 mM NaCl, 5 mM MgCl, and 1 mM D R . ATPase activity was detemiined for

reaction mixtures containing TnI. Tni mutant or Tn 10 pM actin and 0.5 pM S 1, in the presence or

absence of 2 pM tropomyosin for IO minutes at 22°C. Reactions were initiated with 4 rnM ATP and

terminated by adding 100 pL of a solution containing 3% ascorbic acid. 0.5 M HCI, 4% SDS and 0.5%

(wlv) ammonium molybdate. Colour (meamring inorganic phosphate) was allowed to develop for six

minutes pnor to the addition of Iûû pL of a solution containing 2% sodium citrate. 2% sodium m-

menite, 2% acetic acid. Sarnples w m incubated for 10 min at 37°C before absorbame at 650 nrn was

measured in a Molecular Dynarnics E-max plate reader. Phosphate content was determined by

cornparison to a potassium phosphate standard cuve ranging h m 0-100 nrnol. Control experiments

indicated that the rate of ATP hydrolysis was Linear over the course of the reaction .The basal activity of

S 1 was subtracted h m the actin-activated values.

Actin-tropomyosin-troponinSI ATPase

Troponin regdation of thin filament activation was conducted in the sarne ATPase buffér

conditions containing a h a 1 concentration of 0.5 mM EGTA. Standardïzed C~"/EGTA solution was

added. The fke Ca1- solution in the b u k was cdcdated using the program as descnied previously

(147).The ratio of actin:TM:Tn in these experùnents was 722. Control experiments indicated that the

stoichiomeûy of thin tiiarnent- bound troponin and mutant troponin was the same under the conditions

of these experirnenîs.

Viral Vecton

The troponin 1 wild type (wt) cDNA coduig for the hii length protein (rat 1-2 10) and the cDNA

d i n g for the ûuncated protein (rat 1-193) were cloned into viral vector pAdEGI. This polycistronic

adenovirus shuttle vector contains an inducible ecdysone promoter and green fluorescent protein (GFP),

Recombinant adenoviruses containhg the various Tni constnicts were genented by Cre-Iox

recombination.

CeM Isolation and Tissue Culture

Animal care and use conîbmed to The Guide for the Care and Use of Labontory Animals

published by the US National institutes of Health (NIH Publication No. 86-23, revised 1985). Studies

involving isolated nbbit venüicular myocytes utilized New Zealand White rabbits (1-2 kg). Rabbits

were euthanized rapidly and painlessly (30 mgkg intravenous pentobarbital) in accordance with current

euidelines established by the American Veterinary Medical Association Panel on Euthanasia, before - hearts were rapidly removed. Myocytes were isolated by conventional enzyrnatic dissociation, These

were cultured in M 199 media supplemented with 2% (v/v) fetal bovine s e m and 1% (wjv) penicillin

/sîreptomycin. Cells were transfected with either wt TnI or Tni,.,,, virus co-expressing the GFP as a

marker. These vinises were induced with 1 FM ponasterone A (Invitmgen, San Diego,CA) for up to 72

hrs and the medium was changed daily.

CeN Shortenhg

CeUs were added to a fieId Samuiation chamber (3 mm wide by 8 cm long) which contaïned two

plathum wires (one on either side of the chamber) and a perfiisate inflow iine and suction ouülow lie-

The chamber was attached to the platform of an inverted microscope and the platinum wires were

connected to a Grass stimulatorthat was t r i g g v Experirnents were performed at m m

temperature. AAer a brief perioci of time (5 min) to allow ceiis to attach to the cover slip on the

stimulation chamber, cells were perfused for a minimum of 5 min with Tyrodes solution containhg (10

mM HEPES, 5 mM KC1,2 mM CaCL, 140 mM NaCI, 1 mM MgCf,, and 10 rnM glucose pH 7.4).

Cells were stimulated repeatedly at a rate of 0.5 Hz for 10-msec duration. Ce11 shortening of stimulated

ceUs was assessed using an edge detector (Crescent Electronics) attached to the camera on the side-port

of the inverted microscope. Ce11 shortening parameters measured were percent shortening, tirne to peak

shortening, time to half relaxation, duration and ce11 size. Green cells wre selected by their green

tluorescence using a xenon arc lamp with a 488/S3O nm tilter. ïhe edgedetection system was calibrated

using a ceIl hemocytometer.

Action Potentials and Indo Ratios

Action potentials were generated in cells after 72-hours culture and expression of eitherthe wild-

type TnI,,,, or truncated Td,-,,, virus. Action potentids were initiated by short depolarizing curent

pulses ( 3 1 msec, 500-800 PA). indo- 1 ratiornetric dye (30 PM, Molecular Probes) was used as an index

of intraceiiular CaL* , foiiowing stimulation. Ratios were measured after subtraction of background

fluorescence obtained when sealecil. hdo ratios rn& h m the 4051495 emissions were caüirated

to calcium concentrations O and 5 mM in the presence of metabolic inhibitors BDM and cyanide and

inomycin. Ki,, was 0.4 and was 9.8 and b (signal to noise ratio) was 7.8. The pipette solution

contained in 80 ph4 indo pentapotassium, 5 mM M m , 15 mM NaHEPES , 1î5 mM K-glutamate

, 19 mM KCl, and 0.5 mM MgCl,, pH 7 3

Western Blot Analysis

Western irnmunobIots were perfomed on exûacts h m ceUs der 24? 48, and 72 hours of

infection for the troponin I fidl length and truncated (Tni,-,,,) protein. CelIs, k m a 35-mm culturedish,

were lysed and the pellet re-suspended. SDS-PAGE gels were run on 4- 12% gradient gels under reducing

conditions. Protein was transferred to nitroceUulose membranes using the NOVEX ünmunoblot system

and probed for protein expression using Sigma Fast alkaline phosphate detection system.

Statistics

Data are presented as mean k SE. Cornpaison between 3 groups was performed using I-way

ANOVA followed by Tukey's post-hoc test for differences between al1 groups. When only two groups

were compared, a two-tailed unpaired t-test was used. Differences ivith ~ 0 . 0 5 were deemeû signifiant.

Results

Affinity chromatography.

Table 3-1 srnarizes the results obtained h m binding of the TnI, Tt~l,-,,~ and the synthetic

peptides to the TnT or TnC afl3nity columns. As shown in Figure 3-2, TnI and TnII-,, were not eluted

h m the TnT column even in the presence of I M NaCi, but rather, required denaturation with urea and

NaCI. The interaction between Tnl and TnC, in the presence of C$*, wu sufficiently strong that TnI

fgled to elute h m the TnC-Sephmse in 1 M NaCl, or 8 M urea + 1 M NaCl. Elution required addition

of EDTA to chelate the Ca" (table 3-1). The identical behavior was observed for TnI,.,,I (Table 3-1).

Taken together. these data indicate that no major Tnl-TnC or TnLTnT interaction has been dismpted by

the deletion of the C-terminus, at Ieast under theseexperimentiil conditions. This is supporteci by the fact

that neither synthetic peptide, comprising the arnino acid squence h m the C-tenninus ofTnI, bound

to either the TnT or TnC affinity column (Table 3- 1 ).

Actin-Binding and effects on ATPase 3cîivity

n i e dh i ty of Td,-,, for F-actin. or F-actin-TM, was not aItered, when compared with intact

Td, as assessed by F-actin co-sedimentation upon ultracentrifugation (Figure 3-31. As with intact Tni,

the affinity of TnI,-,92 for actin increased in the presence of TM. Although under the buffer

conditions used in this assay (LOO mM NaCl), both Tni and Td,-,,precipitated at high mole ratios

with respect to F-actin (or F-actin-TM), which preduded determination of the maximum binding

stoichiometry-

[t is weU documented that binding of Tni to a&-TM inhibits bindiig of myosin to acictin and

prevents actïn-TM-activated hydrolysis of ATP by myosin S1 (38). Figure 3-4A depicts the

concentrationdependent inhibition of Actin-TMS 1 ATPase activity by Tni and Td,-,, .- Both Td and

TI&-,, inhiiited actin-TM4 1 ATPase to the same extent at a mol ratio of I Tni : 1 actin-

Table 3-1. Affinity chromatography of TnI and its fragments.

rose

High Binding in urea*

binding*

Yes Yes

Tnl

Tn11-192

; g 1 + EDTA

denotes that 2 mM CaCI, was present in the buffer. ND denotes that binding was not determined.

TnCSeph

Low NaCl binding*

Yes

Yes

TnTSepharose

Low NaCl binding

Yes

Yes

High NaCl binding

Yes

Yes

Binding in urea

No

NO

25mM 1M O Urea 8M NaCl

Fraction 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Figure 33. TnT-Sepharose chromatography. 50pg of Tni and 50 pg Tni,_,92 were applied to a TnT-Sepharose equilibrated as descriùed in Materials and Methods. The column was washed with equilibration before high sait buffer was applied. The bulk 090%) of the Tni and TnII.la remaineci bound to colurnn in sait and couid oniy be eiuted under denaturing conditions.

S P S P S P S P S P S P S P S P pl ;: * -

Densitometric Analysis

0 0 0 2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

nmol Tnl or Tnl,,,,

Figure 3-3. üitracentrifigation of actin and actin-TM Panel A. increasing concentrations of Tnl and Tn[1_192 were incubated with actin or actin tropomyosin prîor to ultracentrifugation Representative Coomassie blue-stained geIs showing supernatants (S) and pellets (P) are shown. Panel B. The gels in Panel A, and a duplicate set of gels, were anaiyzed by densitometry. Tni (a) and Td,-,, ( v) values are shown. Each data point is an average (and range) of the values obtained h m the two sets of gels.

nmol Tnl or Tnl,,,,

nmol TnC 1 nmol Tnl o r Tnll-192

Figure 3-4. Inbiiition of Actin-TM-S1 ATPase acîivity by TnI and TnI and its reversai by the adion of TnC, Panel A. Actin-TM-S1 activity was measured as a hction of increasing conceniration of Td (O) and %i1-192 (O). Panel B. TnI (O) and T X I I ~ _ , ~ ~ (O) were added to Actin-TM41 in sufficient quantity to &%it ATPase activily by 90 and 80% respectively. Reversal of tbis inhibition was monitored as a function increasing conceniration of TnC. The buffer used was the same as that used for panel A, except that CaCh was added to a finai concentration of 1 mM. Actin-TM41 ATPase activity is 1Wk S k d a d deviations wae caicuiated h m tripiicate data points. Where exor bars are not visible, the standard deviation is d and lies d e r the symbol.

Surprisingly, at intermediate concentrations, TnI,,, was a marginaily l e s effective inhibitor thanTui,

with IC, values of 0.32 nmol and 0.27 mol, respectively.

in the presence of Ca", TnC celeases the Tni-mediated inhi'bition of the acto-S 1-TM ATPase

activity (Figure 3-4B). Tni, or Ta-,, - was added in a 2:7:2 ratio relative to actin and TM in assay buffer

containing Ca", yieldig 90% and 8û% inhiiition respectively. increasing concentration of TnC, in the

presence of CaL', was able to celease both Tni and Tni,-l,l inhibition (Figure 3-43). However, the

concentration o f f nC required to achieve 50% reversal of actin-TM-S 1 ATPase inhibition b ~ T n I ~ - ~ q was

less than tiir intact Tni (RC, dues of 0.2 nmol and 0.4 nrnol , respectively),

Analysis of reconstihited Tn

In vivo, TnI is part of the ternary troponin complex. The Ca'-dependent troponin-mediated

regdation of ATPase is presented in Figure 3-5. The troponin complexes composed of either intact Tni

(wt Tn) or TnI,-,,(Tn,-,,J ivere mid with actin and TM in a mole ratio of 2:7:2. From Figure 3-5A

we note that at pCa 9. Tn,,, was not M y inhibitory. Surprisingly, Tn,-,,l was hyper-responsive to Ca

". Specifically, at pCa 4.5, Tn,-,,, confied supranormal maximum AlTase activity, reaching 225

nmol/rnin/mg relative to 125 for wt troponin (Figure 3-SA). From the plots nonaiized for the maximum

Ca2'dependent change in activity for each troponin (Figure 3-54 inset), T&,, has increased ~ a " -

sensitivity compared to intact Tn by - 0.1 pCa unit.. importantly, co-sedimentation experirnents under

these conditions indicate that Tn,-,, binds to actin-TM with the sarne stoichiomeûy as wt Tn (data not

shown)- Thus, poor mutant troponin assembly or stability on the thin fiIrnent cannot account for the

changes observed, Nor can incornplete complex formation since a 1 : 1 : 1 ratio o f f nT, TnC to Tni or Tni,-

,, was confirmed by HPLC (data not shown). To determine the fùnctionai impact of ttuncated Tni on

troponin hctionat levels pertùient to models of myocardial stunnùig, maximai Ca2-activated ~TPase

was assesseci for mixtures of wild type and mutant îroponin, the resutts of which are shown in Figure 3-

% of Tn present as Tni-192

Figure 3-5. Ca2+-sensitivity of reconstituted troponins. Panel A. Troponins were refolded fkom individual Tnl (or TnILl&, TILT and TnC subunits as describecl in Materials and Methods. The Ca2+dependence of the actin-TM-Tn-Sl activity was determitleci for reconstituted whole troponin (0) and troponin-bearing TnIL_i92 (O). inset: The same data were normalized with respect to the maximum change in ATPase activity in Ca2+-fiee and Ca2+-reptete conditions. Panel B. The maximum Ca2+-activated ATPase activity was determinml as a function of the composition of troponin (% by mol).

SB. The plot shows that actin-TM-Tn-S 1 ATPase activity in& lineariy with the increase percentage

of Tn,-,,,, At 20% Tn,-,,,#O wt Tn, the ATPae was elevated relative to the activity conferred by 100%

wt Tn.

Viral Expression in Cultured Rabbit Ventricular Myocytes

To determine whether the results observeci with recodtuted myofilament proteins in vitro could

be confirmeci in a mode1 that more closely resembles conditions in viio , anaiysis of cultured adult rabbit

venmcular mdiomyocytes, adenoviraily-transfected with murine Tni or murine TG-,,,, was carried out.

Murine Tni has a single amino acid addition (Ah 26) that is absent h m human Td. Therefore, mTni,-

,,, is the mouse quivalent sequence to human TnIl.l,2 used in the in viho ATPase assays. The transfècted

myocytes had visible striations and maintained their rectanguiar shape (some rounded edges) over the

tirne course of the experirnent. The expression of Tn[,-,,, \vas inferred by the appearance of green

tluorescence protein that was co-transfected with TnI as well as, c o n h e d by western blot anaiysis (data

not shown). Little or no Tdl-lq3 expression wrts observed in the cultured myocytes at 24 and 48 hours

pst-infection, h0weverTn.i expression increased to approximately 50% oftotal cellular T d at 72 h ps t -

infection (data not shown).

Ce11 shortening behavior was evaluated in3 groups ofmyocytes: irnùit'ècted, Tni,,,,-, andTni,-,,,-

infecteci cells after24,48 and 72 h of culture at 37°C. Percent cell shortening was altered oniy in the cells

expressing Tni,,,, (72-h). Compared to ail other groups of celis at 24,48 and 72 h, the percent ce11

shortenhg ofthe Tni,,,,transfected ceiis, at 72 h, increased approximately -3-fold. (Figure 3-6B shows

data at 24 and 72 h). Ce11 shortening traces h m a representative myocyte in each group at the various

time points are shown in Figure 36A. To ensure that differences in ce11 shortening were attriiutable to

myofilament fundon, myocyte excitabiIity and Ca--handling were assesseci. Stimulation of action

potentials in 724 cultured celis, &ected with Tnil-l,o or Tdl-193 resulted in similar action potential

Time (msec) 24 hours 72 hours

Mut diastolic systolic

T

Figure 3-6. Analysis of Adenovirally-transfeoted rabbit cardiomyocytes A. Representative traces of ce11 shortening. Percent relaxation measured in cells uniiifecied or infected with virus for GFP in combination wiih Tn1,.211 or Tnl l.193, aAer 72 Iiours of culture at 37"CôCell sliortening at'i,~ vitro-infected rabbit ventricular myocytes at 24 and 72-11 expression in unitifected and Tnll.210 and ï'nl,.,,,- infected cells, + pc0.05 conipared with al1 otlier groups, pC0.05 vs 24 h of culture. C. Representative action potential measurenients obtained from cells cultured for 72 h aAcr infection with either the Tnl,.,,,, (IFS) or Tnl,.,y3.(n=6). D, Measurement of indo-1 ratios (4051490 nm) for tlic diastolic and systolic Ca2' transient of Tnl ,, (1143) and Tnl (n=3) action potentials.

shape and APD, and APD,(Figure 3-6C). Indo fluoresecence ratios, representing Cr? transients, were

also similar between the two groups of myocytes (Figure MD). SimiIarly, there was no ciifference in the

maximal change (systolic - diastolic) in intraceiiular calcium observed during an action potential (Tni,,,,

= 168 * 95 nM; Tni ,-,,, = 162 * 93 nM).

Discussion

To elucidate the molecular mecbanisms that underlie contractile dystùnction observed in the rat

and moue models of myocardid stunning, we undertook the study of the human Tt&,,, the homolog

of rat and moue Tni,-,,,. Tni interacts with TnT, TnC and actin. Therefore, the effect of C-terminal

tnuication on its ability to bind these proteins was assessed, dong with its effect on actin-activated S 1

ATPase activity. The tùnctional domains of TnE have been miipped extensively, though characterization

of the extrerne C-terminus of Tni has begun only mntly . Rarick er al. ( 149) have shown that deletion

of 23 residues fiom the C-terminus, (i.e. Tni ,-,,,) increûsed Ca"-sensitivity. The truncation is close to

the TnC binding site identifiai by Tripet et QI. (471, in skeletal TnI, residues 1 16-13 1 (cardiac Tni

residues 144-158), a regon that is homogobus benveen the Tni isofoms.

The effect of C-terminal truncation of Tni on its interaction ivith TnC was minimal. Fimly. the

Tni and TnI,-,, bound equally well to the TnC aîhity column in the presence of Ca". Furthermore. the

synthetic peptide Tni I93.509 (or the longer Tnl ,,,,), did not bind to a TnC affinity column. TnC was

more effective at reversing inhiiition of the actui-TM-S 1 ATPase activity by TnIl-,y2, than it was at

reversing inhibition by fidl length Tnl. Taken together, these data suggest that there is a weak Tni-TnC

interaction within or near Tni,,,, that {vas too weak tu be detected by affinity chromatography. whose

loss may intluence myo6iament tlnction. Aiternatively, the increased TnC-responsiveness might be

attn'butable to a change in the confornation of Td, caused by inincarion, that influences other TnC-

binding sites on Td. Recentiy, a TnC-binding site has been identified to Lie ivithin Tni residues 19 1-2 10

(2 13). In that study, binding ofsynthetic peptides, h m 5 different regions of TnI, to TnC was monitored

by plasmon &ce resonance, in the presenceand absence of Ca?. This method is su£ïiciently sensitive

that it appears to have detected an interaction that was not observable by &ty chromatography.

Hoivever, since there ivas no disceniible ciifference in apparent afEnity of Tni and Td,-,, for TnC-

Sepharose, contacts between the C-terminus of Tni and TnC must be weak, and not contribute gmiiy

to o v d l Tni-TnC binding affhty.

TheC-terminus off nl also has a subtieefféct on the interaction betweenTd and &-TM. Tni,

,, was marginally less effective as an inhibitor of ATPase unda non-saturathg conditions, though

inhibition is maximai andequivalent to wild typeTnI at higher concentrations. Ultracenirifbgation studies

also indicated littlt difference in the affinity of TnI,-,,, compared to intact Td, for either actin or acîin-

tropomyosin. This may indicate that Tni truncation mitigates the inhibitoty tùnction of ach-bound Td,

though we m o t nile out that the difference in ATPase inhibition at subsaturating Tni concenûations

may reveal subtle a f i t y differences that are not detectable within the sensitivity limits of the

ultracentnhgation experiments. it dm m u t be noted that the centnfugationstudies were done ata higher

salt concentration than the actin-TM-S L ATPase studies, in order to eliminate non-specific binding of the

proteins io the centrifuge tubes ( 147). The hataseci sait might be sufficient to mask any mal1 difference

in affinity. In myohiril reconstitution experiments. Tn1,-,,, (1 49) was a Iess efktive inhibitor than Td, -

irz, though TnI,,, was as effective as full-Iength Tni,,,,. Taken together. these resuIts indicate that

residues 1%- 199 mark the boundary of a region that is necessay for tùll ATPase inhilition..

In vivo, T d , as part of the troponin cornplex, rnakes extensive contacts with TnT and TnC, that

d u e n c e its hction. Reconstituted Tn bomd the actin tïiament with thesame stoichiometry, regardles

of whether the mutant or fbii length TnI was used. Yet, the pCa -ATPase curve in Fig 3-5AreveaIs îhree

separate observations. Ta,-,, displayed 1) reduced ATPase inhriition at Low Ca?', 2) hcreased Ca"

sensïtivity and 3) higher maximum ATPase activity. These d t s are dnmatic when compared wïth

the study on TnI-TnC (Figure 3-48] in which d a r m a x i m m ATPase activities were observed. The

reduced ATPase inhibition suggests that Tn recondtuted withTd,-,, is l e s effective at rnaintaining the

thin filament in a tumed off state ('blockd state' by the nomenclature of McKillop and Geeves (33)).

The increased Ca'T-sensitization obsmed is consistent with the work of Rarick er al. (149). The

heightened maximal ATPase was unexpected and cannot be compared with previous work by Rarick et

al. since their data was preseated as relative ATPase iictivity and the maximum activity in expainen&

using Tt&-,,, and Tni ,-,, was not reported. One fùrther experimental difference was that Our study

used refolded Tn white their study used TnT to competively displace tropnin barn myofibrils,

fotlowed by reconstitution withTnI andTnC. Our observation is intriguingsince it implies that Td,-

192, while inhibitory in the absence of Ca", actively contributes to supranormal activity in the

presence of Ca2-. It is undear whether the heightened activity represents increased specific activity

per cycling SI head or whether increased activity is due to increased S 1 attachent, However.

similarsuperactivation was observed at Iow S 1 concentrations (data not shown), where S 1 heads are

less IikeIy to contribute to thin filament activation.

Mutations in several rnyotilament proteins observed in Familial Hypertrophie

Cardiomyopathy (FHC) act through a dominant negative mechanism whereby the poisoned

polypeptide mitigates the hnction of the wild type protein disproportionately (208). As a specific

example, Redwood et al, (209), showed that the effect ofTni FHC mutant, RI45G, on actin-TM-S 1

ATPase (without Ca") was biphasic, as its mole ratio with respect to wild type Tni was varieci, To

determine whether Tn containing Tni,_,, idueoced the b c t i o n of wiId type troponin, mixtures

of wt Tn and Tn,-,, were assessed with respect to maxima1 Ca"-ac tivated AlTase ~ax i rnum Ca2--

dependent activity in Figure 3-SB indicated that ATPase varied linearly with the amount of Tn,_,,.

This indicates that the effect of TnI-19z on crossbridge cyciing is conhed within a regdatory unit

containing a 7: 1 : 1 ratio of actin, TM and Tn, and is not pmpagated cooperatively to the neighboring

regdatory units.

Measurements of ce11 shortening were conducted to assess aspects of contractility that are

not easily measured in vitro. In vitro, the Ca2'-regulated step in the crossbridge cycle is proposed to

be the transition of the myosin head between the weakiy and strongly bound states. in shortening

fibers the geometry ofthe myofilaments and the presence of negatively strained crossbridges dictate

that the rate ofshortening is limited by the rate of crossbridge detachment. That myocytes containing

Td,-,,, shortened 3-fold tàster indicates that Td,-,,, increases the rate of crossbridge detachment.

This is particularly intriguing since the preponderance of evidence suggests that the primary role for

TnI is to inhibit crossbridge attachrnent. either directiy or indirectly (through its influence on

tropomyosin).

It is unclear how TnI tancation rnight confer increased crossbridgedetachment. While there

may be subtle differences in its interaction with other members of the troponin cornplex, the effects

~ f T n l ~ - ~ ~ ~ on crossbridge cycling will Iikely be conferred through control of tropomyosin's position

on actin, likely through its interaction with TnT. An increased detachment might be obsewed ifTn1,-

contèrred aTM position that destabilized one ofthe strongly bound actomyosin states, as has been

suggested previously (123). Altematively, Tn1,.l,2 may cause actin-TM to adopt a potentiated state

in which various steps of the crossbndge cycle are accelerated (See (146)). The latterscenario would

explain both the ATPase and unloaded shortening velocity data but is still speculative.

Can our observations be integrated with the pathophysiology of myocardiai stunning? in

Figure 3-SB when Tn,-,,2 is present at 20%, a Ievet pertinent to stunning (1 18), we still observai

supranormal ATPase activity. Yet ventricular trabecdae fiom ischernic rat hearts and tramgenic

rnouse hearts bearing TnI,_,,, yielded demonstrated decreased isomemc force production. This

suggests that tnincation of Tnl at its C-terminus engenders a thin fiIament state in which force is

uncoupled fiom ATPase acavity, interestingly myocardial stunning is not the oniy instance in which

a loss of contractility at the organ ievel is manifested by instances of hypercontractility at the

molecular level. This paradox has been observed in studies of FHC mutations of TnT. Sweeney et

al. have reporteci that the sarne FHC mutants of TnT that conferred decreased isomeûic force

(TnT,,,, and TnTWlQ), demonstrated increased unloaded shortenhg velocity (150).

The importance of delineating the mechanism by which the C-terminal tmcation of Tni

alters crossbridge cycling, is underscored by studies in which selective TnI proteolysis was observed

in patients undergoing heart bypass surgery (15 1). Proteolysis occurred at both N-and C-termini of

TnI, While the fragments in that study may ormay not correspond to TnI,-,,? exactly, we have shown

that removal of only 17 arnino acids fiom the C-terminus substantively alters molecular function.

Knowing the precise Tnl degradation profile, and the respective effects of these TnI tiagments on

Tnl tünction within ailing myocardium, may improve patient prognosis.

CBAPTER4:

Different Molecular Mechanisms for Rho Family GTPase-dependent, Ca2+-independent

Contraction of Smooth Muscle

Foreword

The work presented in this chapter was published in the Journal of Biologicai Chemistry in

1998. Dr. Jennifer Van Eyk was the first author of this study. She performed al1 smooth muscle fiber

contraction studies and wrote the manuscript. Mr. D. Kent Arrell and 1 contributed substantially to

the experimentation and weredeemed to be "co-2nd authors" for this paper. Specifically, ML Arrell

performed al1 1D IEF, 2D electrophoresis, prepared al1 skinned fires, pei-tormed fiber

phosphorylation and quantitahon of phospho-myosin light chain within the fibers. t pertbrmed al1

in vitro phosphorylation, phosphopeptide mapping, phosphoarnino acid analysis and quantitation of

phosphorylation of puritied proteins in vitro. I also purified al1 proteins necessq for these studies

including myosin, myosin light chain, caldesmon, myosin Iight chain kinase. actin, calmodulin.

tropomyosinand GST-rnPAK3. Dr-John Strauss helped to establish the smooth muscle tibersystem,

provided the 1 D IEF protocol and contributed substantially to the writing of the manuscript. Dr.

Emilia Furmaniak Kazmierczak performed the gel overIay experiment and PAK inhiiition by

wortmannin, Dr. Taisto Heinonen was an early contniutor to the project and was the first to show

caldesmon phosphoryIation by PAK in virro. Dr. Graham P. Cote kindly provided recombinant

Cdc42, Rac 1, GST-ROK and had editorial input. Dr. Alan Mak was the senior author for this study.

Abstract

Abnorrnal smooth muscle contraction may contribute to diseases such as asthma and

hypertension. Alterations to myosin light chah kinase or phosphatase change the phosphorylation

level of the 20-kDamyosin regulatory light chain (MRLC), increases Ca'-sensitivity and basal tone.

One Rho farnily GTPase-dependent kinase, Rho-associated kinase (ROK or pl60ROCK) cari induce

~a"-independent contraction of Triton-skinned smooth muscle by phosphorylating MRLC andor

myosin light chain phosphatase. We show that another Rho farnily GTPase-dependent kinase, p21-

activated protein kinase (PAK), induces Triton-skinned smooth muscle contraction, independently

of calcium, to 62 +/- 12% (n = 10) of the value obsented in presence of calcium. Remarkably, PAK

and ROK use different rnolecular rnechanisms to achieve the Ca2'-independent contraction. Like

ROK and myosin light chain kinase, PAK phosphorylates MRLC at serine 19 in vitro. However,

PAK-induced contraction correlates with enhancd phosphorylation of caldesmon and desrnin but

not MRLC. The level of MRLC phosphorylation remains similar to that in relaxed muscle fibers

(absence of GST-mP.W and calcium) even as the force induced by GST-mPAK3 increases tiom

26 to 70%. Thus, PAK uncouples force genemtion fiom hlRLC phosphorylation. These data support

a mode1 of PAK-induced contraction in which myosin phosphorylation is at least complemented

throughregulation of thin filament proteuis. Because ROK and P AK homologs are present in mooth

muscle, they may work in parallel to regulate smooth muscle contraction.

introduction

In smooth muscle cells, elevation of [Ca2']i in response to electncal or chemical stimulation

causes calrnodulin to activate myosin Iight chain kinase (MLCK). This highly specific serine/

threonine kinase phosphorylates the 20-kDa regulatov light chain of myosin (MRLC) at serine 19,

thereby increasing the actin-activated ATPase activity of myosin and inducing contraction. MRLC

is dephosphorylated by myosin Iight chain phosphatases (MLCP) resulting in muscle relaxation. The

properties and activities ofMLCK or MLCP cm be modified by phosphorylation, providing a means

to link smooth musclecontraction to other signaling pathways ( 152; 153). Of particular interest is the

recent linding that a Rho-associated kinase (ROK or p 16OROCK) can induce Ca2'-independent

contraction of smooth muscle (154) by directly phosphorylating MRLC on serine 19 (155) and by

phosphorylating and inhibiting MLCP (129). The serinelthreonine protein kinases PAK and ROK

are activated through interactions with the Rho superfamily of Ras-related low molecular tveight

GTPases (M. = 21,000) (for reviews see Rets. (L56)and (157)). ROK specifically binds RhoA.

whereas PAK (M, = 62-68,000) associates with both Cdc42 and Rac but not RhoA.

Five closely related PAK isozymes have been identifiai in rat brain (1 58), human placenta

and platelets ( 1 59), mouse fibroblast ( 160), and skeletal and vascular smooth muscle (1 59). In lower

eukaryotes, PAK homologues include Ste20 in yeast (16 1) and the single-headed myosin I heaw

chin kinases in Dictyostelium (162). PAK consists of two domains: an N-terminai regulatory

domain that contains a Cdc42/Rac bindingdomain and a C-terminal catalytic domain. The three

mammalian PAK isoforms, PAK1, PA=, and P M , share 70% identity in overall mino acid

sequence and over 90% identity within the kinase catalytic domain. Binding of GTP-Cdc42/Rac

Ieads to autophosphorylation of PAK and activation toward exogenous substrates such as myelin

basic protein ((160;163;164)). PAK and ROK have been ïmplicated in aiterations to the actin

cytoskeletonduring ce11 motility (for reviews see Ref. ( 1 56) and ( 1 5 7)), suggesting that these kinases

may have similar or overlapping modulatory roles.

There is substantial evidence linking activation andior translocation of RhoA to Ca2'-

sensitivity of smooth muscle in some ((127; 165-167)) but not al1 situations (I66;167). In addition,

inactivation of RhoA (168) or ROK by a selective inhibitor (169) has been shown to correct

hypertension in the spontaneous hypertensive n t model. However, the activation of RhoA and ROK

cannot explain the increase in basal tone or Ca2'-sensitization under al1 agonist-stimulated conditions

(166; 167), indicating that alternate signaling pathways are likely to be involved. Because the PM

kinases have been implicated in the control of motile events in nonmuscle cells (1 56) (157), PAK

is a potential modulator of smooth muscle contraction. In this manuscript. it is shown that like ROK

(GST-ROK). PAK (GST-mPAK3) causes Ca"- independent contraction of Triton-skinned smooth

muscle fibers. However, PAK acts via a different moIecular mechanism than ROK to induce Ca2'-

independent force in smooth muscle fibers.

Materials and Methods

Protein Preparations

Intact smooth muscle myosin, MRLC, MLCK, and caldesmon were purified fiom chicken

gizzard (53; l 6 ) , whereas P A U was isolated fiom rat bnin (170). Recombinant caldesmon

hgments (CaD39 and CaD40) were prepared and purified as described in Ref. (17 1). Recombinant

Cdc42 and Racl were expressed and purified as described in (1 58). Plasrnid pGST-mPAK3 cacrying

mouse tïbroblast mPAK3 tùsed to GST in pGEX-KG was expressed and purified according to

(1 60). Plasmid pDK-mPAK3 carrying an inactive mPAK3 mutant with lysine amino acid residue

297 mutated to arginine (mPAK3 cDNA) was tùsed to GST by subcloning a BamHl fragment of

pDK-mPAK3 into pGEX-4T3.2 The mutation of lysine 297 in the ATP binding site of PAK is

sufficient to tender G S T - ~ P A K ~ ~ ~ ' ~ unable to phosphorylate isolated MRLC or myelin basic

protein in vitro. The GST-mPAK3 tùsion proteins were expressed in E. coli and when tested together

weresirnultaneously expressed, purified, and dialyzed against the same buffers. The catalytic subunit

of recombinant ROK (Rho kinase) was expressed as a GST fusion protein in baculovirus (GST-

ROK; (129)). All recombinant kinases were purified on a glutathione-Sepharose afinity colurnn

(Amersfiam Pharmacia Biotech), concentrated in a Cenmprep 30 (Amicon), and dialyzed against 10

mM imidazole, pH 7.0. As reported previously (l60), GST-mPAK3 is susceptrile to degradation,

leading to different activities for each preparation. To ensure consistency, activities of the various

GST-mPAK3 preparations were standardid against myelin basic protein phosphorylation. GST-

mPAK was used in the skinned 6ber assay withïn days but could maintain sufficient activity forin

vitro phosphorylation analysis if stored fiozen in 50% glycerol. Various preparations of GST-

d A K 3 ( O S to 5 pg/mL of active GST-mPAK3) were able to induce Ca"-independent contractions

in s h e d smooth muscle fibers ranging fiom 26.1 to 80.9% of Ca2--dependent contraction.

Skinned Muscle Fiber Experiments

Triton X-100 skinned fibers tiom adult guinea pig taenia coli were prepared as describeci

previously in (172) and (173). The skinned fiber bundles were stored at -20°C in a solution

containing 50% glycerol (vlv), 4 rnM EGTA, L mM Na&, 7.5 mM ATP, and 20 mM imidazole, pH

6.7, and used within 1 rnonth. Thin fiber bundles (- 100 pm in diameter) were mounted on an AME

80 1 (SensoNor, Worten, Norway) force tmnsducet for analysis (172). "Relaxing" solution consisted

of 10 mM MgCl?, 1 mM NaN,, 7.5 mM disodium ATP, 4 mM EGTA, and 20 mM imidazole (pH

6,7), 10 mM sodium phosphocreatine, and 10 units/mL creatine phosphokinase (ionic strength of 1 10

mM, 2 mM t?ee Mg*, and 7.2 mhI MgATP). "Contracting" solution consisted of relaxing solution

supplemented to 0.1 pM calmodulin and 4.0 mM CaClz (pCa = 4.3). The solutions used to bathe the

fibers contained GST-mPAK3, GST- mPAK3E9'R. GST-ROK. or an equivalent amount of 10 mM

imidazole buffer used in the finai dialysis during preparation of PAK and ROK.

Gel Overlay Assay

The gel overlay assriy was perfomed as outiined in ( 174). Intact and skinned T. coli sampies

were analyzed by 10% SDS-PAGE supplernented with 10% glycerol, 5 mM MgCi,, and I mM DTT.

Following transfer to nitrocellulose, the proteins w r e denatured by incubating in a solution of 6 M

guanidine HC1,50 m M ZnCI,, 5 mM MgCl,, 25 rnM MES, pH 6.5, and 0.05% Triton X-100 (30 min

at 4°C). Proteins were thcn renatured by incubating tiuee tirnes with 50 mM NaCl, 2.5 m M DIT ,

25 mM MES, pH 6.5,1.25 mM MgCl,, 50 m M ZnC1,1% bovine serum albumin, and 0.05% Triton

X-100 (vlv) (2 h at 4°C). Puritied recombinant human Cdc42 and Racl were labeled with [35S]GTP-

yS pnor to probing of the nitrocellulose blots.

Western Blot Analysis

10 and 12.5% SDS-PAGE and western blot analysis were camed out as described (1 19).

Detection of the PAK homolog in smooth muscle was achieved using an antibody raised against a

synthetic peptide corresponding to 13 residues at the N-terminal end of rnouse fibroblast mPAK3

(NT3 poIyclond antibody, gifi from S. Pelech, Kinetek, Vancouver, Canada; diIution 1:200).

Caldesmon, desmin, and MRLC were detected using the followingantibodies: clone WCD (dilution

1 :2000), clone DEU- 1 O (dilution 1 : 1 OO), andclone MY-2 1 (dilution 1 200, Sigma), respectiveiy. For

quantification of MRLC phosphorylation, skinned muscle fibers (2-3 fibersArne) were subjected to

one-dimensional isoelectric focusing (1 72)- separating unphosphorylated MRLC fiom mono- and

di-phosphorylated MRLC. The ratio of phosphory lated to total W C was quanti fied by

densitomeûy (Sigma Gel. Jandel Scientific).

Protein Phosphorylation Experiments

In vitro phosphorylation of intact myosin and isolated MRLC, caldesrnon, and caldesmon

fragments were canied out at 30°C in 10 m M Tris HCI, pH 7.5,jO mM NaCI, 5 mM MgCI2. and

1 rnM [)'Pl-ATP (5 x 10scpm/nmol) or in the same buffer containing 150 mM NaCI.

Aliquots of the reaction mixture were analyzed for protein phosphorylation using Whatman

P8 1 paper and SDS-PAGE/autoradiography as described previously (1 10). Phosphorylated amino

acids were identified by thin-layer electrophoresis afier hydrolysis of the phosphorylated proteins in

6 N hydrochlorïc acid as described previously (1 07; 108; 1 10).Two-dimensionai tryptic peptidemaps

were produced as described previously (107; 10S;I 10).

GST-mPAK3 and GST-ROK Phosphorylation in Skinned Muscle Fibers

Skinned fiers mountcd on a U-shaped pin were incubateci at 25°C under various conditions

using the same conditions as in the s h e d tiber assays (relaxing, contracting, and retaxing plus

GST- mPAK or relaxing plus GST-ROK); except when required, assay buffers contained 1 rnM [y-

"PIATP (0.25 mCifmL) instead of 7.2 mM ATP. The GST-rnPAK3 and GST-ROK concentrations

induced 70% of maximum Ca"-dependent contraction. After 90 min of incubation, fibers were

submerged in ice-cold 15% trichloroacetic acid and 2 m M inorganic phosphate followed by acetone

to inactivate the kinaseslphosphatases. This ensures preservation of the phosphorylation levels.

Fibers were stored at -20°C until analysis.

Samples of radioactively labeled proteins fiom skinned fibers were resolved by 12.5% (see

Fig. 4-4, A and B) and 10% (see Fig. 4-4, D and E) SDS-PAGE. Duplicate protein sarnples resolved

simuItaneousIy by 10% SDS-PAGE were transfemd to hvo nitroceIlulose blots, and one blot was

probed with anti-caldesmon antibody (see Fig- 4-4, D and E), and the other blot was probed with

anti-desmin antibody (see Fig. 44D: ROK data not shown) as described above. Alignrnent was

accomplished by pin-holing the nitrocellulose and autoradiopraphy film with [$'Pl-ATP.

Identification of proteins phosphorylated in the presence of either PAK (see Fig.4- 4D) or ROK (see

Fig. 4E) was then canied out by exposure ofthe nitroceIluiose blots to a "P PhosphorIrnager screen

(Screen GP, Kodak). This allowed not only for the accurate alignment between western blots and

the subsequent autoradiograph results, but also for detection of radioactive ernissions only, because

the PhosphorImager screen does not detect chemilurninescent emissions. Phosphorimages of the

blots were developed using the Storm Phosphorhager 520 (Moiecular Dynamics, Su~yva le , CA).

This series of experiments has been perfomed using four different guinea pig T.coli s k i ~ e d fiber

preparations and two GST-mPAK3 preparations each yielding the identical pattern and alignrnents

of autoradiography and western blots.

Proteins h m skinned fibers that were radioactively IabeIed in the presence of GST-mPAK3

as described above were also resolved by two-dimensional gel electrophoresis (see Fig. 4 4 )

according to a standard protocol (Bulletin 1 144) 6om Bio-Rad. Proteins were resotved in the first

dimension by isoelectric focusing on a mini-Protean ii isoelectrk focusing gel electrophoresis

apparatus (Bio-Rad) using an ampholyte mixture of 10% pH 3.5 to 10.0 and 90% pH 4.0 to 6.5.

Protein resolution in the second dimension was canied out by 12.5% SDS-PAGE. The two-

dimensional gels were stained with Coornassie Bhe and dried, and autoradiography was performed

directiy on fiim (X-Omat Blue YB- 1, Kodak). To deterrnine the pH gradient, blank gels resolved in

the first dimension were cut into evedy sized slices, each siice irnmersed in 1 mL of 10 m M

potassium chloride, vortexed, and incubated at room temperature for 1 h, and then the pH of the

solution was determined. Linear regression ofthe pH and gel distance was canied out (y = 0.3386~

+ 4.375, where y = pH and x = cm tiom acidic end of gel) to determine the actual pI of the various

phosphoproteins. Migration in the second dimension was compared with broad range protein

markers (New England Biolabs, Beveriy, AIA).

Results and discussion

Overlay assays with Cdc42 and Racl provide a sensitive means to detect PAK kinases in smooth

muscle. [3S~]G~~yS-Cdc42 bound to bands of 62 and 65 kDa in extracts of guinea pig T. coli

smooth muscle, whereas i3'S] GTPyS-Racl detected a single band of 62 kDa (Fig. 4-1, A and B).

An antibody raised against the N-terminal 13 mino acid residues of mouse fibroblast mPAK3

reacted with a protein of 62 kDa in guinea pig smooth muscle and a protein of the same molecular

mass in rat aorta (Fig. 4-LC). These results indicate that smooth muscle contains one. and possibly

two, PAK isoforms (Fig. 4- 1 A). PAK was absent From Triton-skimed smooth muscle fibers (Fig.

4-1, A and B), suggesting that, Iike ROK (154), PAK is either a cytoplasmic or membrane-bound

enzyme.

Triton-ski~ed guinea pig T. coli smooth muscle fibers were induced to contract in a Ca2'

-independent manner when incubated in the presence of constiiutively active GST-mPAK3 (Fig 4-

2A). The force induced by GST-rnPAK3, 5pglmL. 55 nM) in relaxing buffer (pCa < 8.0) reached

a maximum equivalent to 62%+i- 12% (n=lO) of that achieved by the addition of a calcium-

containing activation solution(pCa4.3) Under the same conditions, the inactive PAKmutant , GST-

m p ~ ~ 3 E97R , was unabIe to induce force in the absence of Ca2- (Fig. 4-28). In previous studies, Ca2'-

independent srnooth muscle contraction has been induced by addition of phosphatase inhibitors

(1 72; 175; 176), or most recently by a another Rho-fmily GTPase dependent kinase, ROK (1 54) In

al1 cases, the degreeof smooth muscle contraction correlates with an increase in the level of MRLC

phosphorylation. in the case of ROK, contraction is promoted by the direct phosphoqlation of

MRLC on Ser 19, in addition to the phosphorylation and inhiiition of MLCP (129)-

This dual effect of ROK was demonstrated by the use of wortmannin, which is a potent

inhibitor of MLCK but does not affëct the activity of either ROK (1 54) or PAK (Fig. 4-2C). ï h e

Figure 4-1. A PAK homologue is present in intact but not triton X-100 skinned smooth muscle. Recombinant cdc42 (A) or Racl (B) Ioaded with 35S-GTPyS was usai in overlay assays of purifieci brain d A K 3 (lanes 1) and either intact (lanes 2) or Triton X-100 skinned (lanes 3) guinea pig Taenia coli smooth muscle fitiers. C . shows Western blot of 12.5% SDS-PAGE of intact (lane 2) or skinned (lane3) guinea pig Taenia coli and intact rat aorta (Iane 4) using an antibody raisecl against a synthetic peptide corresponding to residues 1-1 3 of mouse fïbroblast mPAK.3 (clone 3NT). The methods are descn%ed under "Materials and Methods".

Figure 4-2. Constitutively active GST-mPAK3 induces Caff-independent contraction of guinea pig Taenia coü skinned muscle fibers witbout involving MLCK or myusin üght chah phospbatsse. Typicai isometric force tracing of Triton X-100 skinned guinea pig Taenia coli fiers, in the presence of comtitutively active GST-mPAK3 (A), or the inactive mutant G s T - ~ P A K ~ ~ ~ ~ ~ (B). Fibers were contracteci @Ca 5.6 andior 4.3) and reIaxed @Ca 8.3) subsequent to incubation with GST-mPAK3 (bar) at pCa 8.3. (C). Dernonstration of high specificity of wortmannin for MLCK by its ability to inhibit in vitro phosphorylation of MRLC by MLCK (m) without inhiiiting phospholylation by GST-mPAK3 (a). 100% equals maximum phosphorylation of MRLC (1 mol phosphatdmol of protein) by either kinase in the absence of wortmannin. (D) A typicai isometric force tracing of Triton X-100 sicinneci muscle fibers in the presence of 1 mM w o r t m ~ and constitutively active GST-mPAK.3. 1 mM wortrnannin was sufficient to eIiminate Ca2*dependent contraction (data not shown), and would inhriit any contribution of MLCK to the contraction iriduced by the presence of PAK The methods are as outheci under "Materials and Methods".

addition of the constitutively active GST-ROK (catalytic domain to the hiton-skinned smooth

muscle fibers ) produces a wortrnannin sensitivecontraction at pCa 6.5 (Ca2--dependent contraction)

as well as a wortmannin-insensitive contraction at pCa 8.0 (Ca2'-independent contraction)( 154). On

the other hand, wortrnannin at aconcentration of 1 m M had little effect on the contraction ofsrnooth

muscle induced by GST-PAU at Iow Ca" (Fig MD), even though this concentration is sufficient

to completely inhibit MLCK-dependent contraction at elevated Ca" (data not shown). These results

demonstrate that the caL'-independent contraction, promoted by PAK, occurs without requirement

of MLCK activity. Furthemore, it seems udikeIy that PAK promotes contraction by inhibiting

MLCP because Ca"-independent contractions achieved by phosphatase inhibitors are invariably

dependent on MLCK activity and are abolished by MLCK inhibitors. (e.g. ( 172: 175: 176)). Thus,

PAK niost likely works by direct phosphorylation of a contractile protein rather than altering the

MLCK or MLCP.

These results prompted ihe investigation into whether PAK directly phosphorylates MRLC.

thus achieving contraction in a traditional manner. In virro anaiysis shows that GST-mPAK3

phosphorylates intact chicken gizzard smooth musclemyosin to 2 mol ofphosphate/moI (Fig. 4-3A)

Phosphate is incorporated into a single Ser residue of MRLC fig 4-3A. Furthemore, MLCK was

unable to phosphorylate MRLC following phosphorylation by GST-mPAK3 (Fig. 3-3C). These

results predict that PAK like ROK promotes srnooth muscle force generation by increasing MRLC

phosphorylation levels. However, under the conditions where GST-mPAK3 induced Triton-ski~ed

smooth muscle f i e r s to contract wîth 70% of the maximal force obtained in the presence of Ca"',

no significant increase in the Ievel of MRLC phosphorylation was observed- (Fig. 43D, lane 4) even

as the force induced by PAK increases fiom 2670% (Fig. 4 3 D lanes 1-3). The uncoupling bettveen

MRLC phosphorylation and force generation irnplies diat Pm does not directly or indirectly

activate myosin but must ernploy an alternative and novel mechanisms to contract the skinned

muscle fibers.

To begin to define the molecular basis of PM-induced contraction, it is criticai to identiQ

the proteins phosphorylated by mPAK3 in the skinned smooth muscle fibers. Protein substrates for

mPAK3 were Iabeled with 3 2 ~ under conditions where GST-mPAK3 produces 70% of maximal

Cdr-dependent force (Fig. 4-4). One- and two-dimensional gel electrophoretic anaiyses of the

proteins labeled during a PAK-induced contraction were performed. With one-dimensionat SDS-

PAGE, hvo proteins with approximate molecular masses of 58 and 145 kDa are more highiy

phosphorylated in the presence than the absence of GST-mPAK3 (Fig. 4-4, A and D, compare lanes

1 and 2). Little if any, phosphorylation of MRLC is detected in the fibers contracted with GST-

mPAK3 (Figs 4-3D and 4-4A).

The 58- and 14.5-kDa proteins were identified by western blot analysis as desmin and

caldesmon, respectively (Fig. 4-4D). Furthemore, the pl of the 58- and 145-kDa proteins were

determined by hvo-dimensional gel electrophoresis using a pH gradient from 4.0 to 6.5 followed by

12.5% SDS-PAGE (Fig, 4 4 3 . The pI values for the 58-kDa protein are 5.59 * 0.04 and 5.37 * 0.04

for the mono- and diphosphorylated fonns, respectively, and dthough the arnino acid sequence of

guinea pig desmin is not known, these pt are in the range expected for desmin (human and chick

unphosphorylated desmin, Swiss P 17661 and P02542, pl of 5.2 1 and 5.45, respectively). The 145-

kDa protein, identified as caidesmon, has pI values of 5.63 * 0.03 and 5.38 * 0.04 for the mono- and

di-phosphoryiated forms, respectively. Again, the amino acid sequence of guinea pig caidesmon is

not kuown, but these observed pl values are in the range of p1 for human and chick unphosphorylated

h-caidesmon (PR JH0628 and A33430) of 5-62 and 5.56, respectively. irnportantly, there is no

known contractile protein other than caldesmon with a molecuiar mass

Figure 4-3. Uncoupüng between MRLC phosphorylation and force development in skimeci muscle fibers even though isolated myosin and MRLC con be phosphorylated nt serine 19 by GST-mPAK.3. A. T i e course of in vilro phosphorylation of intact chicken gizzard smooth muscle myosin by GST-mPAK3. The asterisk indicates the time at which SDS-PAGE and autoradiography samples were obtained. Myosin phosphorylated by GST-rnPAK3 was andyzed by autoradiography of 12.5% SDS-PAGE. Lefi inset, lane G, Coomassie-stained gel: Iane A corresponding autoradiograph. Right inset, phospho-amino acid adysis. B. T i e courses of MRLC phosphorylation in the presence or absence of GST-mPAK3 foiiowed by the addition of MLCK afkr 30 min. The asterisk indicates the time at which the samples anaiyted in C were obtained. C. Tryptic peptide maps (lefi panels) and phospho-amino acid d y s i s (right panel) of isolated MRLC phosphorylated by MLCK or GST mPAK3- D. Plots of W C phosphorylation % and the correspondhg relative force. (% force relative to pCa 4.3) obtained in the guinea pig taenia coli muscle libers pCa 4.3 (5, n=5) and pCa 8.3 in the absence (4, n=6) and presence of GST- mPAK.3 (2-3, n=5) aller a 90-min incubation at 25°C. Experirnents were done using five guinea pig skinned taententa coli p r e p d o n s and three different GST-mPAK3 preparations. Upper panel shows a typicai western blot (using anti-MRLC monoclonal antibody) of the one-dirnensional isoelectrïc fôcusing gel (2-3 fiberdane). The ratio of phosphoryIated (mono- and di- phosphorylated protein) to unphosphorylated MRLC was detetmined by densitomdry- Standard curves were done to ensure that the exposures of the westem blots were within a iiuear range. The methods are as outlined in "Materials and Methods"

Figure 4-4. PAK and ROK phosphorylate different proteins in skinned muscle fibers. The protein substrates of PAK (A, C, D) and ROK (B and E) in the skinned muscle fibers were identified by autoradiography and western blot analysis. Skinned guinea pig taenia coli muscle fibers were incubated in the presence of y3'P-ATP in the absence (pCa 8.3, lane 1) and presence of calcium (pCa 4.3, lane 4). As well, fibers were incubated at pCa 8.3 in the presence of either constitutively active GST- mPAK3 (lane 2) or GST- ROK (lane 3). In vitro-PAK-phosphorylated MRLC is included as a standard marka protein. A and B show autoradiographs of 12.5 %SDS- PAGE analysis , whaeas D and E show autoradiographs fiom 10% SDS-PAGE analysis genaated fiom the same blots used for western blotting with anti-caldesmon or anti-desmin antibodies. C shows an autoradiograph of 2-dimensional gel electrophoresis of radioactively labeled 48 and 145 kDa PAK mbstrates in the Triton-skinned muscle fibers. The proteins were separateci in the first dimension by a pH gradient produced by an arnpholyte mixture of 10% pH 3.5. to 10.0 and 90% pH 4.0-6.5 and in the second dimension by 12.5% SDS-PAGE. Phosphorylation of MRLC by GST-ROK in the absence of calcium is similar to that by MLCK at pCa 3.3 (B. compare lanes 3 and 4). Caldesrnon is the only protein phosphorylated by GST-mPAK3 that is not phosphorylated by GST-ROK in the skinned muscle fibers (compare D, lane 2 and E, lane 3). F shows the time course of in vitro phosphorylation of chicken gizzard h-caldesmon by GST-mPAK3 (a) or GST-ROK (O). GST-mPAK3 phosphorylates h-CaD to 2 mol phosphatdmol protein- G shows an autoradiograph of 12.5% SDS PAGE of MRLC, C-terminai fragments of human fibroblast caldesmon ( 0 3 9 ) and intact chicken gizzard h-caldesmon phosphorylated, in vitro, by GST-mPAK3. The methods are as outiined unda "Materials and Methods"

B

MRLC 1 2 3 4 4.38 pH Gradient 6.18 -

, . . . '. ..'.,i$ . . ' ,. . . '

tCaD . .. - . - . - . .. . -. : . - 212 kD CaD b :. ..* a.,.h% p. , . . .: ,

,:...:~~~,.:;-.- - 116 . . . ., ...., . ,-. .; .;:;;'-:.:; - 97

ri :. ,. -,. 7 ; ., :. 4 Des - >S.. -':..

, ,, - . : - 66 .- . ". ; ..??j' - C i -- . . . .?

.. . . . :, - 56 . <

. . . .. i'. - 43 .,T* " . .

- -37

Auioradiograph Western Western Autoradiograph Western

Time (Minutes)

between 70 and 150 kDa that is present in Triton-ski~ed muscle fibers with a PI close to 5.60. For

example, smoothmuscle MLCK(Swiss P 1 1799) and myosin light chainphosphatase PP2A 130-kDa

subunit have cdculated pl values of 4-72 and 5.09, respectiveiy.

GST-ROK, the GST fùsion protein of the constitutiveIy active catalytic domain of ROK (3),

causes Triton-skinned smooth muscle fibers to contract in a Ca"-independent manner with up to

80% of maximal force ((154); data not shown). Under these conditions, the major proteins

phosphorylated in the skinned fibers by GST-ROK are MRLC, desmin, and two proteins mi&ng

at positions greater than 158 kDa (Fig, 4-4, B and E). Clearly. neither of these high molecular mass

proteins are caidesmon (145-kDa phospho-protein. Fig. 4-4E). One is most likely the catalytic

domain of MLCP, which is known to be phosphorylated by ROK in vitro and has an approxirnate

molecular mass of 158 kDa by SDS-PAGE.

Comparison of the protein substrates for ROK and PAK under conditions where GST-ROK

and GST-mPAK3 induce similar amounts of ca2'-independent force (79.5 versus 7 1.1 %,

respectively), indicated that GST-ROK incorporated more phosphate into MRLC than did GST-

mPAK3 (Fig. 4-4, compare A and B). As well, GST-ROK did not phosphorylate caldesmon, which

is one of the main substrates for GST-mPAK3 (Fig. 4-4, D and E). In vih-O phosphorylation studies

confirm that chicken gizzard h-caldesmon is a better substrate for GST-mPAK3 than GST-ROK

(Fig. MF). In vitro, GST-mPAK3 phosphorylated h-caldesmon to 2 mol of phosphate/mol of

protein (Fig. MF). This explains the rnono-and diphosphorylated States ofcaldesmon found in two-

dimensionai gel electrophoresis of the PAK phosphorylated Tr i ton-shed muscle fibers (Fig. 4-

4C). Furthmore, the C-terminal domain of human fïbrobiast l-caIdesmon (corresponding to

chicken gizzard caidesmon amino acid residues 458-752) is a substrate for GST-mPAK3 (Fig. 4-44,

but no phosphorylation of the N-terminal caldesmon domain was observed (data not shown). The

C terminus of caldesrnon contains multiple binding sites for actin, tropomyosin, and calmodulin.

Caldesmon inhibits the actin-activated Mg-ATPase of myosin (review see (19)) and has been

suggested to provide a basal inhibition of vascular tone (177). The force of contraction of Triton-

skinned srnooth muscle fibers increases upon the partial extraction of caldesmon (1 78) or decreases

because of cornpetitive binding of a 20-kDa actin-binding fiagrnent of caldesmon (1 15). As well,

a synthetic peptide of an actin-bindingregion of caldesrnon increases force of 13% in escin skinned

arterial muscle fibers at low concentrations of Ca" (177), probably by competing with endogenous

caldesrnon for the actin filament. Taken together, these results suggest that reduction in caldesmon

interaction with actin would increase force generation, resulting in contraction. Phosphorylation of

the C-terminus of caldesmon by PAK could release caldesmon inhibition of the ATPase activity

resulting in augmented force development.

In conclusion, although hvo different Rho farnily-dependent kinases, PAKand ROK areable

to induce Ca2'-independent contractions in srnooth muscle, they do so via different molecular

rnechanisms. ROK increases the steady state level of MRLC phosphorylation. PAK, on the other

hand, uncouples force fiom MRLC phosphorylation and likely acts by phosphorylating caldesmon.

The data presented are consistent with a mode1 of PAK-induced contraction in which myosin

phosphorylation is at least compIemented through the regulation of thin filament proteins.

Acknowledgment

We thank S. Bagrodiaand R. A. Cerione for the gi% ofGST-rnPAK3 and G S T - ~ P A K ~ ~ ? ~ ~ ;

K. Kaibuchi for GST-ROK; S. Pelech for the anti-mPAK3; and X--P. Feng and L. Organ for

technical help.

Chapter 5:

Phosphorylation of Caldesmon by p2l-Activated Kinase: Impücations for the CaZ+-sensitivity

of Smooth Muscle Contraction.

Foreword

This manuscript was published in the Journal of Biologicai Chemistry in January 2000. 1 was

the first author. [ performed the majority of the experimentation, protein preparation and

experimental design associatd with this project (in conjunction with Dr. Van Eyk). This involved

preparation of actin tropomyosin, myosin S 1 (both smooth muscle and skeletal muscle variants),

actin-binding, calmodulin-binding studies, design of ATPase assays (a new method for us).

purification of PAK: and characterization of in v i m phosphocylation, including phosphoarnino acid

analysis. Li-hua Shen prepared f AK. calmodulin, actin and myosin S 1 preparations and pertbnned

ATPase assays, as weil as the phosphoryiation. Drs. John KeIly and Pierre Thibault detemiined

phosphoryiationsites by MALDi-QTOF bIS/MS analysis. Dr. Jenny Van Eyksupervisedthe project

from its inception and is the corresponding author. Dr. Alan Mak is Dr. Shen's supervisor, and the

senior author who wrote the manuscript.

Absîract

We have previously shown that p21-activated kinase, PAK, induces Ca"-independent

contraction of Triton-skinned smooth muscle with concomitant increase in phosphorylation of

cardesmon and desrnin but not rnyosin-regulatory light chah (Van Eyket al. (1998) J, Biol. Chem.

273 23433-23439). In this study, we provide biochemical evidence impiicating a role for PAK in

Ca"-independent contraction of smooth muscle via phosphorylation of caldesmon. Mass

spectroscopy data show that stoichiornetric phosphorylation occurs at Ser657 and Ser687, abutting

the calmodulin-binding sites A and B of chicken gizzard caldesmon, respectively. Phosphorylation

of Se1657 and Se687 has important functional impacts on caldesmon, PAK-phosphotylation

reduces binding of caldesmon to calmodulin by about 10-fold while binding of calmodulin to

caldesmon partially inhibits PAK phosphorylation. Phosphorylated caldesmon displays a modest

reduction in affinity for actin-tropomyosin but is significantly less effective in inhibiting actin-

activated S 1 ATPase activity in the presence of tropomyosin, We propose that Ca"-independent

phosphorylation of caldesmon at the calmodulin-binding sites by P A K partially converts caldesmon

ûom an actin-myosin ATPase inhibitory 'off state to a non-inhibitory 'on' state which is functionally

similar to the Ca'? calmodulin-caldesmon cornplex.

Introduction

Recent data strongly implicate the monomeric Rho family GTPascs, including Rho, Rac and

Cdc42, in modulatïng Ca1*-sensitivity of smooth muscle contraction (84; 1 79). RhoA-activated

kinase, Rho kinase, enhances the Ca'--sensitivity of contraction by phosphorylating the myosin-

binding subunit of smooth muscle rnyosin phosphatase I (SMPPL), resulting in inhibition of its

activity ( 129) and /or by phosphorylating Ser 19 of MLC directly ( 180). Rac i and Cdc42 have been

irnplicated in the cytoskeletal remodeling processes that accompany lamellapodia and filopodia

formation in many types of non-muscle ceils. One of the key downstrem effectors of Rac and cdc42

is the Serflhr kinase, PAK( 18 1). We have shown previously that infusion of Triton-skinned guinea

pig T. coli srnooth muscle fibres with constitutively active PAK3 induces CaL'-independent

contraction to about 60% of the force obtained in the presence of Ca2'lcalmodulin (182). PAK-

induced contraction is accompanied by an increase in the phosphorylation ofcaidesrnon and desmin

but not MLC even ihough PAK is able to phosphorylate MLC at Serl9 in vitro ( 1 83;184). These

results suggest that Rac and cdc4Z in contrzlst to Rho, inducesmooth musde contraction by altering

the properties of actin and/or intermediate filaments.

Smooth muscle caldesmon is speculated to be a thin filament regdatory protein by virtue of

its ability to inhibit actïn-tropomyosin activateci ATPase activity of myosin (78; !85;186). It is an 89

kDa protein that binds in an extended conformation dong filaments ofactin-tropomyosin. it houses

binding sites for myosin (l87), tropomyosin (l4), calmoduiin (97) and actin (96). In vim,

caldesrnon-mediated inhr'bition of actomyosin ATPase activity can be regulated by calcium binding

proteins, incIuding calmoduiin (I88;189) or caitropia (190).

Our hdingthat PAK induces Cap-independent contraction in s h e d smooth muscle fibres

(8) and the recent demonstration that an unlaiown kinase besides MAPK phosphorylates pizzard

caldesmon in vivo (1 9 1) suggest that PAK phosphorylation of caldesrnon may be involved in the

regdation of Ca"-sensitivity of smooih muscle contraction. Here, we report biochanical evidence

supporting a role for PAK in the Ca3-sensitivity of smooth muscle contraction via phosphorylation

of caldesmon. Specifically, we have identified the sites of phosphoryiation, and studied

phosphorylated caldesmon with respect to its affinity for actin-tropomyosin, cdmodulin and its

ability to inhibit actomyosin ATPase activity.

Materials and Methods

Protein Preparation

h-Caldesmon and aB tropomyosin were purified from chicken gizzards essentially as

described by Bretscher (145). Skeletal muscle actin was purified fiom rabbit muscle as outlined in

(143). Smooth muscle S 1 was prepared by papain cleavageofginard myosin (140). Rabbit skeletal

S 1(A 1) was prepared by cleavage with chymotrypsin as described in (142). Recombinant murine

P A U , was expressed from the plasmid pGST-mPW in E- coli JM 1 O 1 andor JM 1 10 ceIIs as

described before ( 182).

Phosphorylation of caldesmon and identification of phosphorylation sites.

Caldesmon (1-2 mg/mL) was phosphorytated by GST-mPAK3 (- 5 pg/mL). at 37°C for

60 minutes, in 20 mM Tris pH 7.5. 100 mM NaCl, 5 mM MgCl:. 1 mM [.t'Pl-ATP (1-5 .u IO5

cpdnmol), 0.5 m M DIT. Quantification ofphosphorylation and analysis ofphosphorylated arnino-

acids were performed as described before ( 1 10).

Approximately 700 pg caldesrnon was dissolved in 200 pL of 100 mM NH,HCO,, pH 7.9.

containing 10 pg of endoproteinase Glu-C. The digestion was carried out overnight at room

temperature. The digest solutions were evaporated to dryness and redissolved in 5% acetic acid (200

PL). For nanoelectrospray mass spectrometry analysis, 20- pL aliquots of the sample solutions were

desaIted using ZipTipThl C,, (MiIlipore, Bedford, MA). Approximately 1 - pL of the desalted

solutions were used for both precursor ion scanning and tandem mass spectrometry (MSMS)

analyses. Phosphopeptides were detecied by precursor ion scanning (precursors of mlz 79) in

negative ion mode on a API 3000 triple quadrupole mass spectrometer (Perkin Elmer/SCIEX).

Precursor ion spectra were acquired in multïpie channel acquisition mode, typically over a period

of 3 minutes ( d z 4OO-~OOO,OS m a s units step size, 5 msec dwell time). Argon was used as the

collision gas and the collision offset voltage was 80 V. Phosphopeptide sequencing was achieved

b y M S N S using a prototype quadruple time-O f- flight mass spectrometer (QqTOFMS, Perkin

EImerISCIEX, Concord, ON, Canada) equipped with ananoelectrospray ionization source. Product

ion spectra were canied out in positive ion mode using argon as the collision gas and a coilision

offset voltage was 30 eV (laboratory ûame of reference). MSMS spectra were typically acquired

every two seconds over a period of three minutes.

Calmodulin-caldesrnon interaction

Interaction between phosphorylated and non-phosphorylated caldesrnon. and calmodulin was

studied using intrinsic Trp fluorescence as described previously (97). The binding buffer was 20 mM

Tris-HCI, pH 72,O.j m M CaCI2, 100 mM NaCl, 1 mM DIT. The excitation wavelength was 295

nm with a dit width of 10 m. Intensity measurement was made with a 290 nm filter at 330 nm and

a slit width of 10 nm. Binding curves were fitted to a binding equation to obtain dissociation

constants as described before (97) except that binding stoichiometry was set to 1 moi caldesmon per

mol calmodulin as previously determined (97).

Actin-binding assays

Two nmol of actin was mixed with 0.4 m o l of tropomyosin and O to 1.5 m l

phosphorylated or non-phosphorylated caldesmon in 200 L of binding buffer (40 mM Tris pH 7.5,

LOO rnM NaCl, 5 mM MgC12. and 1 m M DïT). Protein mixtures were allowed to equilibrate for 30

min pnor to centrifugation at 100,000 x g in a Beckman TL- 100 ultracentrifuge. Pellets were rinsed

with actin-binding buffer once and dissolved in 100 L 0.05% (vfv) TFA in water. ReIative amounts

of caldesmon, tropomyosin and actin were determined by HPLC using a Zorbax SB-CS HPLC

column (147).

Acün activated S1-ATPase assays Actin activated myosin S 1 ATPase assays in the presence and

absenceoftropomyosin were conducted in a96-well ELISA plate (100 PL assay volume). Inorganic

phosphate was determined colorimeûïcaily as described in (148). ATPase buffer contained 40 rnM

Tris pH 7.5, 50 mM NaCI, 5 m M MgClz and t mM DR. Myosin S1 ATPase activity was

determined for reaction mixtures containing O to 4 p M caldesmon, 10 pM actin and 0.5 PM S 1 with

or without 2 M tropomyosin for I O minutes at 37°C. ATPase reaction was initiateci with 4 rnM

ATP and teminated by adding LOO pL of a solution containing 3% ascorbic acid, 0.5 M HCI, 4%

SDS and 0.5% (wiv) ammonium molybdate. Phosphate release was linear tvithin the 10 minute

reaction. Colour was allowed to devdop for six minutes pnor to the addition of 2% sodium citrate

and 2% sodium m-arsenite followed by 10 min incubation before absorbance at 650 nrn was

rneasured in a Molecular Dynamics E-rnax plate reader. Phosphate content was determined by

compatison to a potassium phosphate standard curve. Less than 10% of the ATP was hydrolyzed

over the course of the reaction.

Results

Chicken gizzard h-caldesmon was phosphorylated in vitro using a constitutively active

murine GST-PAK3 to a maximum of 2 mol of phosphate per mot of caldesmon as shown in Figure

5- 1 .Only phosphorylated Ser ivas recovered fiom a partial hydrolysate of "P-labeled caldesmon

(inset Figure 5- 1).

To locate the phosphorylation sites, caldesmon phosphorylated by GST-PAK to 2 mol

phosphate1 mol protein was subjected to digestion by endoproteinase Glu-C, which cleaves peptide

bonds on the COOH-sides of Glu residues. The resulting digest was analyzed for phosphorylated

peptides by precursor ion scmning as described in Materials and Methods. Two major doubly

deprotonated ions were detected (peaks A and B in Figure 5-2A) together with a number of minor

peaks representing minor sites of phosphorylation. MS/MS sequence analysis determined that peak

A ( d z 648.0) and peak B (m/z 754.0) correspond to hvo singly phosphorylated peptides,

G,j,VRNIKpSMWE, and T,,AGLKVGVSpSRM&,,, respectively (data not shown). Ser657,

being the only Serin peptide A, can be unambiguously assigned as the site of phosphorylation in this

peptide. There are two adjacent Ser residues in peptide B, however. MSMS sequence analysis

indicated that Ser687 is most Iikeiy the site ofphosphorylation (Figure W C ) since a b-type fragment

ion at m/z 8 13.5 corresponding to the unphosphorylated peptide, TAGLKVGVS,, was detected

(data not shown). The non-phosphorylated couterparts of peptides A and B were not detected

indicating that Sa657 and Ser687 were fdiy phosphorylated, largely accounting for the observed

stoichiometry of 2 mol phosphate 1 mol protein (Figure 5-L). The precursor ion scan of the

endoproteinase Glu-C digest of unphosphorylated caldesmon shows no trace of peak A ( d z 648.0)

orpeak B (m/z 754.0) indicating that Se1657 and Sa687 are genuine PAK-target sites (Figure 5-2B).

Time (min)

Figure 5-1. Phosphorylation of caldesmon by mPAK3 in the presence and absence of cahodulin. Caldesmon (1 mg/mL) was phosphorylated by recombinant, constitutively active, GST-mPAK3 in the presence (O) or absence (a) of Ca~lcalmodulin (0.4 mghi,), as outlined in Materiais and Methods. Inset: autoradiogram of 32P-Iabeled amino acid tecovered h m partial acid hydrolysis of phosphorylated caldesmon; Pi = iaorganic phosphate, P-Ser = phosphorylated Sc, O = ongin.

Figure I 2 . Identification of phosphoryhtion sites in caidesmon incubated with mPAK3. A. precusor ion scan of m/z 79 (-ve ion mode) of the endoproteinase Glu-C digest of phosphorylated chicken gizzard cddesmon- Peaks a (mh = 648.0) and b (mlz = 754.0) represent the two major doubly-deprotonated ions. Sequence analyses by tandem mass spectrometry deterrnined the sequences of peaks a and b to be G651VENKpSMWE660 and T6,,AGLKVGVSpSRINKE69t, respectively. B. precursor ion scan of m/z 79 of the endoproteinase Glu-C digest of unphosphorylated caidesmon. C. the functionai domains in subdomain 4 of caldesmon; the calmodulin-binding sites A and B are boxed, and the phosphorylation sites by PAK at Sa657 and Serf587 are indicated.

Minor peaks at mlz 529,543 and 558, which amount to less than [2% of peak B, were detected in

the unphosphoryIated sample but were not furthet analyzed.

According to the mode1 ofMarston and Redwood (26, and Figure 5-2C), Sec657 and Sa687

locate at the &O-terminal ends of calmoduIin-binding sites A and 8 (96;97) in subdomain 4. We

therefore examined whether Ser657 and Ser687 were accessible to PAK when caldesmon formed

a compkx witl-t Ca'h/cdmoduIin. GST-PAK3 phosphoryIated caldesrnon-calmoduIin cornplex at

a similar initial rate but mach a stoichiornetry of 1.2 mol phosphate / mol protein (Figure 5-1).

Calrnodulin was not phosphorylated by PAK, and Ca'-did not affect PAK activity under the same

conditions (data not shown). This result suggest that binding of calmodulin to sites A and 5 of

caldesmon renders Ser657 andor Se1687 less accessible to PAK.

To detemine whether introduction of phosphate groups to Serti57 and Ser687 at the

calmoduIin-binding sites A and B c m affect calmodulin-binding, we compared binding of

phosp horylated and non-phosphoryiated caldesmon to caImoduIin using intrinsic Trp fluorescence

measurements (Figure 5-3). Phosphorylated and non-phosphorylated caldesmon have similar

fluorescence spectra sharing a similar emission maximum ai 350 nm suggesting that the phosphate

groups do not cause significant changes in the environments surrounding the Trp raidues which are

majordeterminants for CaIrnodulin-bindimg(L 5). Bindingof Ca?*-calmoduIin, which contains no Trp,

increased the htrinsic Trp tiuorescence of caidemon by a maximum of about 70% (Fig. 5-3B) and

caused a blue shifiofthe ernission maximum fkom 350 to 338 nm (Fig. S-3A). On the other hand,

calmoduiin inmases the fluorescence intensity by less than 40% at saturation accompanied by a

smaiier shift in emission maximum fiom 350 io 345 run (Figure 5-38). As shown in Figure 5-3B.

phosphorylation reduces the aftinity of caldesmon for Ca2--catmodulin by about LO-foId, increasing

Kd h m 0.1 to 0.9 FM.

A, (nrn)

0.8 I

Figure 5-3. Effect of phospborylation on binding of caldesmon to crlmodulia. A, Trp fluorescence emission spectra of phosphorylated caldesmon with calmoduiin (-) and non- phosphorylated caidesmon with caimodulin (-). in the absence of calmodulin, the emission spectra of phosphorylated and non-phosphorylated caidesmon are Wtually identical as indicated by the m w , CaD = caldesmon, P-CaD = phosphorylated caldesmon. Excitation was at 295 nm. B. binding curves of phosphorylated (O) and non-phosphoqlated (e) caldesmon to caImodulU?, using intrinsic Trp fluorescence measurements. AI is the change in fluorescence in caidesmon at 330 nm induced by calrnodulin, I, is fluorescence of caldesmon in the absence of caimodulin. A = wavelength. Inset shows the integnty of the phosphorylated and non- phosphorylated caldesmon and caimodului at the end of the bindùig studies.

0.0 O. 1 0.2 0.3 0.4 0.5 0.6 0.7

mol caldesmon ! mol actin

Figure 51. Effect of phosphorylation on biiding of cddesmon to actin- tropomyosin. Bindiig curves of non-phosphorylated caldesmon (a) and phosphocyIated caldesmon (O) to actin-tropomyosin was anaiyzed by sedimentation.

We compared the ability of caidesmon and its phosphorylated counterpart to interact with actin-

tropomyosinand to inhibit actin-activated myosin ATPase activity since the caImodulin-binding site

A has been shown to bind actin (177) and site B is in the middle of the tropomyosin-linked actin-

bindiag and Uihibitory region in subdomain 4 (76). As s h o w in Figure 5-4, caidesrnon

phosphorylated to 2 mol phosphate 1 mol protein has a modest reduction in affinity for actin-

tropomyosin; Kd was increased by less than two-fold tiom 1.0 PM to L.7 PM, However,

phosphorylation of caldesmon induces a significant release of inhibition of actin-S L ATPase (Figure

5-5) in the presence or absence of tropomyosin. At 0.2 moVmol of caldesmon / actin, non-

phosphorylated caldesmon inhibits actin-activated skeletal SI ATPase activity by 80% in the

presence of tropomyosin whereas about 40% inhibition was observed by the sarne amount of

phosphorylated caldesmon. Similar results were obtained using smooth muscle S 1 (data not shown).

In the absence of tropomyosin (inset Figure 5-9, caldesmon is much less effective in inhibition as

reported by others (76; 192); 0.4-0.5 moVmol of non-phosphorylated caldesmon / actin is reyuired

to cause a 40% inhibition while similar arnounts of phosphorylated caldesmon inhibits by 20%.

mol Caldesmon/ mol Actin

B

0.0 8.1 02 0.3 0.4 0.5

mol Caldesmon/moI Actin

Figure 5-5. inhibition of actin-activated skeletal SI-ATPase activity by caldesmon and phosphorylated caidesmon. A, inhibition of actin-activated S 1 - ATPase activity by non-phosphorylated (a) and phosphorylated caldesmon (A) in the presence of tropumyosin, B, same as A, except that tropomyosin was absent,

Discussion

This study provides biochemicai evidence to support the hypothesis that phosphorylation of

caldesmon by PAK may play a rote in inducing Ca"-independent contraction in smooth muscle. The

strategic location of Ser657 and Sert387 in the calmodulin-binding sites A and B provides a crucial

clue to our understanding of how phosphorylation of these sites may affect the function of

caldesmon. The sequences around Se657 and Ser687 are conserved in chicken, mouse and human

caldesmon and these regions also forrn parts of the extended actin-binding regions in subdamian 4

(76) further underscoring their importance. Ser657 and Ser687 are not recognized by MAPK

( 109; 1 10), casein kinase 11 (103). Ca"-dependent calmodulin kinase 11 ( 193), and PKC (98). al1 of

which have been s h o w to phosphorylate caldesmon in vitro. The sequences surrounding Ser657

(RNIKS,,,MWE) and Ser687 (KVGVSS,,7RiN) in caldesmon, and Serl9 (QRATS,,NVF) in MLC

have a hydrophobic residue in the +2 position which agees with Brzeska et a1.(2 11) who showed

that a Tyr at the +2 position is strongly preferred by PAK1. As well, seven of the eight

autophosphorylation sites in PAKl have a hydrophobic residue in position +2 (215). However,

Tusizon et a1.(210), using a series of synthetic peptide substrates, identified the signature

detemiinants for PAKl phosphorylation as KRES which bears Iittle resemblance to the caldesmon

and MLC phosphorylation sites except for the presence of a basic residue within positions - 1 and -5.

It appears, therefore, that secondary structures and a hydrophobic amino acid at the +2 position are

equally important determinants for PAK recognition.

Not unexpectedly, we found that phosphorylation of Ser657 and Ser687 interféred wiîh

interaction between calmodulùi and caldesmon. We have shown previously that aithough Trp659

and Trp692 in sites A and B, respectivety, are major determinants for caldesrnon-calmodulin

interaction, amino acid residues surrounding the Trp residues aIso contniute to optimal binding (97).

NMR data showed that sites A and B in synthetic peptides simultaneously bind to the two

hydrophobicregions ofcahodulin affecting al1 S Met residues in the Met puddIes (2 1 1 ) and become

a-helical upon binding to calmodulin (194;195). Furthemore, the helix fonned by site A is

arnphiphilic such that Ser657 is located on the polar surface (194). Introduction of phosphate groups

at these sites would interfere with contact between the polar surface of site A and calmodulin but

should have a minor impact on hydrophobic interactions since phosphorylation at these sites do not

appear to induce significant changes in the environments surrounding Trp residues in sites A and B

based on fluorescence data (Figure 5-3). This is also consistent with our tinding that binding of

caImodulin to sites Aand B attenuates subsequent phosphoryIation ofcaldesmon by PAK. indicating

that Ser657 and/or Ser687 become less accessible to PAK (Figure 5- 1). The actin-binding sites span

an extended region in subdomain 4ofcaldesmon (76). Introduction of two phosphates at Ser657 and

Sert587 is unlikely to induce extensive disruption in the actin-binding regions, thus abrogating

interaction, This may account for the modest reduction in amnity of phosphorylated caldesmon for

actin-tropomyosin (Figure 54): which is a reflection of perhaps a teorientation of the caldesmon

molecule dong the actin-tropomyosin tilarnent towards forming a non-inhibitory state. This

intetpretation is consistent with ow finding that phosphorylation significantly reduces the ability of

caldesmon to inhibit rnyosin S 1 -ATPase. A synthetic peptide spanning fiom Gly65 1 to Ser667

containing site A has been shown to bind actin and enhance contraction in saponin-treated single

hyperpermeable srnooth muscle cells of fmet aorta and portal vein ( 196), whereas Ser687 and site

B are situateci in the midst of a tropomyosin-dependent actin-binding region (residues 669-710),

which is beIieved to be involved in tropomyosin-linked caidesmon inhibition of actomyosin ATPase

activity(76). It is conceivable thatphosphorylation of Sera87 and Sa657 are responsible for altering

the tropomyosin-dependent and tropomyosin-independcrnt inhibition, respectively.

As shown in Figure 5-5, PAK-phosphorylation attenuates (-50%), but does not abolish,

caldesmon inhibition of actin-activated myosin ATPase activity invoking a phosphorylation

rnechanism by which caldesmon function c m be modulated independentiy of Cd'/calmodulin. it is

possible that introduction of phosphate groups to the calmodulin-binding region may partially

simulate a conformation repmsentative of the non-inhibitory caldesmon/~a"'-calmodulin complex,

It appears that caldesmon c m exist in a number of states with different inhibitory activities

depending on its phosphorylation status and binding to Ca"/calmodulin. One of these states is

generated by Ca"-independent phosphorylation of Ser657 and Ser687 with PAK, which possesses

intermediate inhibitory activity compared to the fully inhibitory non-phosphorylated caldesmon and

the non-inhibitos, Ca2-lcalmodulin-caldesmon complex. Formation of a complex benveen

phosphoryhted caldesrnon and calrnodulin, however, appears unfavounble in view of results

showing that the affinity of phospborylated caldesmon for Cd'/ calrnodulin is reduced 10-fold, and

Ser657 and/or Ser687 in the caldesrnon-calmodulin complex are iess accessible to PAK (Figure 5-31,

Taken together, data h m this study and others suggest that RadCdc47 and Rho GTPases act in

concert targeting thin- and thick-tiiament, respectively, in modulating Ca? sensitivity of srnooth

muscle contraction.

AcknowIedgments

We thank S. Bagrodia and R. A. Cerione for ihe @fis of GST-mPAK3. We extend thanks to Nina

Buscemi and Lenny Organ for technical assistance and to Dr. irena Neverova for helpfiii suggestions.

CBAPTER 6: General Discussion

The role of TnE in myocardial stunning

The results of Chapter 3 indicate that the behavior of Td,_,,, appears contrary to what we

rnight expect given the decreased contractility that defines stuming. Specifically, we expected the

maximum Ca2'-activated ATPase conferred by mutant TnI,_,, -bearing troponin, to be lower than

the ATPase measured ushg wild type troponin, in accordance with the observation of decreased

maximum force production of the stunned trabeculae (125). skinned muscle fibers from stunned

myocardium (isolated rat heart) (1 L 9; 1 X), or trabeculae €rom the transgenic rnice expressing murine

Td,-,,, (120). However, thestudies in Chapter 3 have shown that Tni,-,,l-bearingTn acmally confers

superactivation of actin-TM-TncS I ATPase achvity, With respect to the role of Tnl proteolysis in

myocardial stuming, we musr consider two possible interpretations: 1) Tnl proteolysis causes

stunning but the mechanism is comptex; 2) Tni proteolysis is a by-product of stunning and its effects

may be benign or compensatory.

To distinguish between the hvo possibitities, it is necessary to critically evaluate both the

evidence that implicates a role for TnI protedysis in the literature, and the strengths and limitations

ofthe studies of Chapter 3. This story is one that begins with physiological studies ofthe whole heart

and has nmowed as the evidence warranted. The stunned physiological phenotype is comrnon to al1

forrns of stunning regardless of experimental or animal model. Minimdly the molecular mechanism

must be reconded with these observations.

1 have previously highlighted iiterature, with respect to the a vivo rat model, in Chapters 1

and 3. This focus is deliberate. The rat mode1 is experimentally simpIe and has been studied

systematically by more than one laboratory. Work fiom the Iaboratory of Dr. Marbk has been

seminal. They arrived at their conclusion that TnI proteolysis is the lesion of stunning, only afier

other key parameters such as membrane excitability and Ca "-handling had been tested (1 23)-

Within the rat model, the lesion must lie with the myofilaments. Subsequent work indicating that Tni

proteolysis accompanied stunning were informative and suggestive of an invohement of TnI

proteolysis, though these studies were correlative and did not prove causality (1 17-1 19).

Importantly, expression of Td,-,,, in transgenic mice reproduced the systolic and diastoiic

dysfùnction and decreased the Ca2'-responsiveness of the myocardiurn (120). However, the

transgenic model of stunning is subject to al1 the criticisms that normaily befall transgenic studies.

For instance, the transgenic mouse mode1 of stunning is a chronic model, though medicaily,

stunning is an acute affliction. To what degree is the transgenic phenotype a retlection of altered or

compensatory expression ofcardiac genes? While skeptics are neverentirely satistied, it is clear that

regulation of Ca2--handling is preserved in the transgenic mice. Furthennore, mRNA Ievels of MHC

and phospholarnban were unaltered, indicating that their studies were not confounded by induced

cardiac hypertrophy. Therefore, subject to the usual caveats of transgenic models, the study of

Murphy et al. (120) does indicate that TnI,.,,, is sufficient to cause the stunning phenotype.

The study presented in Chapter 3 represents the first aaempt to characterize the rnolecuiar

lesion of stunning using biochemical methods. The reductionist approach offm the advantage hat

dysfùnctioncan be dissecteâ by controlling the protein composition within the experiment. Restated,

this approach dlows us to study the properties that are intrinsic to stumed troponin, and to

distinguish these properties fiom the effects of extrinsic factors like differences in the composition

andlor post-translational status of the other myofilarnents during stunning. However, the underlying

assurnption was that the force and acto-myosin ATPase activïty correlate. This now rnust be called

into question.

The Ca2--sensitization is at odds with with one pubIished report which studied stunned

Triton-skimed fibers (125) though it is consistent with the work of Van Eyk et al. (1 19). It is also

consistent with other biochemical studies of the C-terminus of Tni using reconstituted myofibrils

(149). This implies that the Ca "-desensitization, of both twitch force and steady-state isometnc

force, observed in other studies, is confened by a process other than Tni proteolysisperse. A critical

process that could affect the ~a"-sensitivity in the heart, but which has not been addressed in the

transgenic mouse or isolated rat heart models of stunning, is myofilarnent phosphorylation,

Specifically, Tni phosphorylation affects the Cg'-sensitivity of contractionand has been particularly

well documented (JO). As mentioned in Chapter 3 previously. phosphorylation of the N-terminal

serine 22-23 by PKA decreases the Ca "-sensitivity by influencing the Tni-TnC interaction in a

manner that decreases the afinity between ~ a " and TnC ( 1 13). Furthemore, Tni phosphorylation

by PKC, at sites at Ser43 and Ser45, decreases maximum Ca "-activated ATPase activity ( 197).

Interestingly, both PKC and PKG show overlapping speciticity with PKA and are able to

phosphorylate Tni at PKA site in vitro. though neither has been implicated in Tni phosphorylation

in vivo (197).

Differences between observations obtained fiom in vitro biochemistry of stunned troponin

and the physiology of stunned fibers may be reconciled by determining the efiects of myofilament

phosphorylation. if we grant that the ca2'-overload that accompanies ceperfusion is suficient to

activate aberrant Cap-activated proteolysis, there is no reason to suspect that aberrant Ca2'-

dependent phosphoryIation~dephosphorylation is any less likely. One reason this hypothesis has not

been tested is that traditional gel analysis of proteins is one dimensionai. That is, proteins are sorted

by molecuiar weight. Consequently, this type of analysis is well suited to the study of proteolysis,

since proteolysis is easily discernible as a change in molecular weight. The re-emergence of 2D gel

eIectrophoresis. in which proteins are separated by both molecular weight and charge, w i U enabIe

researchers to assess the phosphorylation status of the myofilament proteins. This will establish

whether Ca2'-dependent phosphorylation/dephosphorylation contributes to the stunning phenotype.

Given the eIevated Ca Ir-activated ATPase caused by Tn,-,, (Chapter 3), we m u t consider

the possibility that force and ATPase, two end products of crossbridge cycling, become uncoupkd

within the sninned myocardiurn. In the crossbridge cycle, myosin binds to actin, ATP is hydr&ed

and Pi is released (this is what is measured in Our ATPase experiments). Release of Pi is highiy

exergonic (AG is negative) due to an increase in entropy rather than a change in enthalpy. Pi reiease

thus provides the energy that is coupled to force generation as it drives the conformational change

in myosin-ADP, that rresults in ADP release and force production. It should be stated that there is

evidence that Pi release and force production are coincidental, while others believe that force

production is a function of ADP release (20).

Observations that Tni proteolysis causes increased Cà'-activated ATPase yet decreased Cazh-

activated force, suggests that the fiee energy of Pi release is not being harnessed to produce force.

Restated, Does TRI proteolysis engendera thin filament state that impingeson the abilityofmyosin

to convert ftee energy into work? If so, by what mechanism? 1 suggest that Tn containhg Tni,-,,?

rnight cause troponin-tropomyosin to adopt aconformation that destabilizes attachent of the strong

crossbridges. This can be tested in several ways.

Firstly, ATPase and force have never been rneasured simultaneousIy in a skinned muscle

fiber tiom s t u ~ e d myocardium. This is experirnentally difficult, yet important. Biochemically,

detded assessrnent of the affinity of myosin heads for regulated thin filaments should be tested.

Specifically, by performing binding experiments in the presence of either ATP or ATPyS (which

favors weak myosin binding) and ADP (which favors strong binding), crossbridge destabilization

cm be addressed directly. An uncoupling mode1 predicts that myosin-ADP would display reduced

a E t y for actin-tropomyosin bearing Tn,-,,.

If force production is ûuly uncoupled fiom ATPase, then certain crossbridge parameters

should be determined to delineate the mechanism. Do filaments decorated with Tn,-,,, affect the

force generated by a single myosin head? Does the lifetime of a force-bearing crossbridge differ

when Tn is replaced with Tn,-,,,? These types of parameters can be measured using new laser trap

technologies that measure forces of single molecules by stringing actin filaments between spectnn-

coated beads. When myosin interacts with the filaments. the deflection of beads is measured with

a photodiode,

If Tn proteolysis does cause TM to adopt a position thrit destabilizes myosin binding, or

othenvise accelerates myosin dissociation, can such a thin tilament conformation be visualized

directly? The process of 3D-image reconstruction of electron microgaphs has been used successfdly

to determine the tropomyosin position on actin in the presence of Ca" and myosin heads (49). Most

recently, the technique has allowed direct visualization of the troponin complex (198). These

techniques may be equally valuable for determining whether Tni proteolysis affects the position of

tropomyosin in response to Cz? and myosin.

The data presented in chapter 3 will be controversial, since they imply an uncoupling of

between force and ATPase activity, and beg M e r experimentation, Detailed investigation of

myocardiai shinnîng at the molecular level wili be seminai in the development of new treatments.

Finally parallel investigation of molecular stunning, heart failure and familial hypertrophie

cardiomyopathy will uhimately reveal whether different foms of heart disease share any cornmon

mechanisms.

The role of PAK and caldesmon in smooth muscle contraciion.

PAK exists in at least four isoforms: PAKf (a P U ) , P A U (y PAK), PAK3 (B PAK), and

PAK4, They are regulated by the Raci and Cdc42 members of the family of small monomenc

GTPases. The N-terminai regulatory dornain of PAK houses overlapping GTPase-binding and

autoinhibitory domains (209). Structure/function studies of site-specific mutations within this

domain show that mutations within the autoinhibitory domain fiequently diminish binding to the

Cdc42. Therefore, it has been suggested that PAK activity is suppressed by an intramolecular

interaction behveen the C-terminal catalytic domain and N-terminal regulatory domain. Binding

of RacJCdc42 to PAK causes activation by perhirbing this interaction (200).

Like its upstrearn binding partners. Rac 1 and CdcJ2. PAK also plays a role in reorganization

of the actin cytoskeleton, Recall that Racl is involved with actin reorganization associated with

Iamellapodia/filapodia formation, whereas Cdc42 is required for the formation of tiny microspikes.

A key issue that is currently being addressed by several labs. is whether. and under what

circumstances. PAEC is involved in these processes, Mutations in the PAK-binding domain of Rac1

appear to have little effect on filapodia formation, which would rippear to argue against the necessity

of P M in these processes (2 10). Paradoxicaily, however, a mutant of PAK that cannot bind Racl,

does promote formation of membrane ruffles, though its intrinsic kinase activity is not required

(202). PAK's kinase activity is essential, however, for disassernbly of actin stress fibers and focal

adhesion complexes (I99), and efforts to identiQ the physiologicai substrates for PAK have begun,

Changes in ceIl morphology and motility require more than assembiy or disassembly of a

cell's scaffolding; it aiso requires concerted movement. [n smooth muscle and non muscle celIs aiike,

movement is a result of actin-myosin interactions. Recently, Sanders et al. have shown that MLCK

is a substrate for recombinant PAK, and that phosphorylation of MLCK inhibits its activity toward

MLC,, in vitro (203). They also showed that over-expression of PAKI decreased MLCK activity

and rnyosin phosphorylation in BHK2 1 cells. MLCK phosphorylation correlated with inhibition of

ceIl spreading and dissociation of stress fibers. The authors propose that the actions of PAK and

ROK constitute opposing pathways that regulate ceIl spreadinglrounding by controlling the

phosphorylation state of myosin.

At the outset of the study presented in Chapter 4, the effects of Rho GTPase on ceIl

morphology had been studied extensively. In the srnooth muscle community, it had been known

since 1988 that GTP analogs could cause Ca2'-sensitization of muscle contraction ( 126) . Work tiom

the Somlyo laboratory has shown that the effect of GTRS is mediated through RhoA. Once a

downstream et'fector of Rho was identified (ROK), several groups sought to determine its role in

contraction. Early reports were contradictory. One Iaboratory reported that ROK phosphorylated

MRLC on Ser 19, whicli was sufticient to activate actomyosin ATPase activity ( 180). Furthemore.

ROK likely causes contraction inskinned smooth muscle by direct phosphorylation of ML-, (204).

In intact muscle, however, there was evidence that GTPyS caused Ca2--sensitization through

inhibition of myosin light chain phosphatase, MLCP (205). Subsequent work in intact muscle

concurred with studies in skinned fibers, that ROK did mediate Ca2--sensitization, but through

phosphorylation and inhibition of MLCP rathathan direct phosphorylation of MLG, (180). Given

the involvement of the Rho kinase pathway, might other relatively newly discovered GTPase-

regulated kinases, such as PAK, confer Ca"-sensitivity?

The study in Chapter 4 constituted the first investigation of PAK in a differentiated muscle

systern. Since its publication (l83), the results have been reproduced using ainvay smooth muscIe

fibers (McFawn & Fisher, unpubiished results). The data clearly indicate that force production cm

not resuit fiom traditional mechanisms, sincemyosin phosphoryIationwas undtered in PAK-induced

contractions, and contractions were wortmannin-independent. Our best efforts to determine the

substrates in these fibers implicated desmin and caidesmon. Consequently, this study constitutes

some of the most persuasive evidence in favour of thin filament-assisted contraction and impiicates

PAK as a potential mediator of cal'-sensitization in intact srnooth muscle.

A role for thin filament proteins in the regulation of contraction has been regarded with

skepticisrn by some (84). Therefore, it was incurnbent upon us to provide a solid biochemical

foundation for such a proposai. The results in Chapter 5 indicate that caldesmon is a substrate for

PAK in vitro, that phosphorylation mitigates ATPase inhibition, and h a only a marginal influence

on its affinity for actin. importantly, cddesmon is a substrate for PAK when it is pre-bound to the

thin filament (data not shown). These results suggest that phosphorylation of caldesmon, by PAK,

causes a change in thin filament conformation that exposes myosin-binding sites on actin. Restated,

crossbridge cycling is enhanced, regardless of the degree of myosin phosphorylation. We propose

that PAK contributes to CaL'-sensitization by changing the myosin phosphoryiation-force

dependency. This hypothesis is testable by directly rneasuring the force as a tùnction of myosin

phos$mj4ation in skimed smooth muscle fibers, as has been described previously (94). A tùrther

prediction of such a mode1 is that caldesmon phosphorylation in smooth muscle fibers would enable

cooperative reattachment of unphosphorylated crossbndges, as assessed by Somlyo (93).

A valid criticism of a mode: of PAK-mediated Ca2'-sensitization is that there is currently no

physiological evidence for it. This must be addressed. Currently, much of the work on PAK

activation cornes from non-muscle cells. A notable exception is the report that PAKcan be activated

in smooth muscle ceIls in response to Angiotensin 11 (206). Assuming that new physiological triggers

of PAK can be found in differentiated smooth muscle, it wiI1 then be necessary to determine the

relative roles of myosin and caidesmon phosphorylation,

Caldesmon phosphorylation by PAK has been overshadowed, to a degree, by reports that

PAK also phosphorylates and inhibits MLCK (203). What are the imptications for smooth muscie

contraction and for our newly proposai model? PM-mediateci inhibition of MLCK alone would

decreme Ca2.-seasitivity, yet 1 have juçt described how P A K effects on caldesmon would increuse

CaL'- sensitivity. 1s there a contradiction between the studies? It shouid first be noted that Sanders

etaL(203) conducted their studies in BHK2 1 (hamster kidney cells) and HeLa cells (an irnmortal ceII

Iine), whereas our studies were conducted on differentiated srnaoth muscle. The authors studied the

effects oFPAICon cell morphology which is dependent on assembly and disassembly of actornyosin

stress fibers. P A K inhibition of MLCK decremes MLC,, phosphorylation, which in turn, causes

stress fiber disassembly. Howevei in differentiated srnooth muscle. used in our studies.

myo filaments are not subject to phosphorylation-dependent assembly. Nevertheless. if physiological

activationut'PAK tveredemonstrated in srnooh muscle, which mechanism (MLCK phosphorylation

or caldesmon phosphorylation) would dominate? Firstly, it wouId be important to assess the

substrate specificity of PAK for MLCK and caIdesmon- If the catalytic et'ficiency (kcat/Km) for the

two PAK substrates differed by an order of magitude or greater, the matter might be academic.

However, assuming that MLCK and caldesmon are both legitimate candidate substrates in

differentiated smooth muscie, the implications are intriguing. PAK-mediated phosphorylation of

caldesmon would facilitate crossbridge cycling, while phosphorylation of MLCK would decrease

myosin activation (keeping ATPase Iow). That is, force wodd be sustained by faditating cychg

of unphosphorytated myosin. Such a mechanism would provide an eIegaat explmation for the high

economy of smooth muscle contraction, Le. the maintenance of force with littie ATP consumption.

in our study, we aiso demonserated that PAK dso phospborylates desmin in the Ttiton-

skinned smooth muscle fiers- Desrnin is a stmctural protein that comprises the intermediate

(cytosûuctural) filaments. Recently, Ohtakara et al. (207) published a report demonstrating that

phosphorylation ofmonomeric desmin decreases its propensity for polymenzation, ï h e identification

of both conh~ctile and cytostmctural substrates for PAK is timely, given that recent advances in

the field of smooth muscle contraction suggest the cytoskeleton is an integral component of force

regulation. Work by Mehta & Gunst (16) has shown that force production in canine tracheal strips,

stimulateci with acetylcholine, was associated with increased actin polymerization, and conversely,

that inhibition of actin polymenzation by Latrunculin A decreased force deveioprnent. Traditiondly,

a distinction has been made between the contractile and cytoskeletal elements of srnooth muscle.

That distinction is slowly fading, and the merger of these fields may provide the catalyst necessary

for the discovery of the precise role of PAK in smooth muscle.

Perspectives: Muscle contraction in the new millennium

The scope of this thesis is extremely broad, s p m i n g cardiac and smooth muscle systems.

and 1 am richer for it, Thin tilament reylation has been the overarching theme. In their

comprehensive review on striated muscle contraction Gordon et al. (20) suggest that our

understanding of muscle contraction has advanced to the point where structural, biochemical and

physiologicai models are beginning to be reconded. This understanding stems from painstaking

work of physiologists, biochemists and X- ray crystallographers.

Regardless of the scientific field, the central dogma of the scientific method has been 1)

hrmulate a hypothesis, 2) design an experiment to test the hypothesis. 3) observe the behaviour of

the system, and 4) evaluate whether the data fits the hypothesis. This method will continue to

provide the foundation for scientific understanding, but on its own, it is il1 equipped to handle what

will surely be the greatest hurdle to bioIogicaI science in the next century - complexity. By

complexity, 1 rnean the ùiterrelationships between the thousands of proteins, lipids and carbohydrates

that make up our bodies and their relationships with our environment. Natural complexity is

compounded by experimental complexity, whereby scientific 'tmths' may Vary depending on the

approach used to study thern. The shortcoming of the scientific method is that it lirnits study to that

which can only be known a priori. which brings me to my point. The greatest breakthroughs will

come h m laboratories that balance broad scope screening and hypothesis-dnven approaches.

This paradigm has been embraced by many. Many of the hottest discoveries of the Iast 20

y e m have corne fiom the screening of genetic mutations in yeast, D. melanoguster, C. elegans and

zebrafish arnong others. Furthemore, screening for new protein-protein interactions using the yeast

twa-hybrid methodology has contributed greatly to the study of signal transduction. However, these

techniques only begin to scratch the surface of complexity, because they don? address the processes

that tine tune the intetrelationships between proteins - post-tmslational moditication.

Proteomics, the study of the status and complement of proteins within a cell, promises to

address the complexity of post-translational modification and has been much heralded. In this

approach, 2-dimensional electrophoresis is hquently used to resolve proteins (and their post-

transiationd modifications), which can then be identified by mass spectrometry or by sequence

analysis. One of the advantages of the proteomic approach is that it is generally fiee ofbias. though

the properties of certain proteins do confound analysis (e.g, solubility, hi& pI, and ceMar

abundance). In tirne however, just as the field of genomics has driven the deveiopment of new

techniques in molecular bioIogy, the study of the proteome will spur advances in protein separation

and charaçterization technologies. Finally, the proteomic approach will prove powerful for the study

and documentation of the subtle and complex changes that occur dong the continuum of cardiac

injury.

References

Holmes,K.C. 1998. Muscle coamction. ~VotnrtLs.Found.~yn~p. 213:76-89.

University ofToronto and Faculty of Medicine. A Look Back at MuscIe. 2001, Ref Type: lnternet Communicatioo

Harvey,W. 1628, Erercitatio Anafomica de Mortr Cordis er Sungrtink in Animalibus (Anatomical L~ercise on the Motion of the Hem and Blood in Animais). Frankfun.

Burke, James. Not What it Seems. S c i h . (July), 110-1 L 1.2000.

Hmley,H.E. 1996. X personai view of muscle and motility mechanisms. Ann~r .RmPl~~s io l58: 1-19.

Huxley,H.E. and Hanson J. 1953. The smctunI ba i s of the cross-striations in muscle. Xutlirr 17230- 32.

tl~x1eyH.E. and HansonJ. 1954. Changes in the cross-smations of ~nuscle during contraction and stretch and lheir structunl inierpreration. Nutrrre l73:973-976.

Huxleyh. and Simmom,R. 1954. Mechanism of force genention in striated muscle. iVuture 17397 1 - 973.

Rayment.I., Rypniewski,W.R, Schmidt-Base.K., Srnith-R., Tomchick,D.R., Benningb[SrL, Winkelmann.DA,, WesenbergG., and KoIden.tI.;Zl. 1993. Th=-dimensional srructure of myosin subfngment- 1: a molecular motor; Science 26 150-58.

Raymcnt.I., Holden.H.hl.. Whittaker.M.. Yohn.C.B.. Lorenzht.. Holmes.K.C.. and Mi1ligan.R.A. 1993. Structure of the actin-myosh complex and itç implications For muscle contraction . Science 26 I:jS-65.

II. Sweeney,EI.L. and Ho1zbaur.E.L. 1996. Mucational andysis of motor proteins. =Innu.Rev.Phjsiol58:75 1- 792.

Lowey,S. and Trybus,K_M. 1995. Rote ofskeletai and smoath muscle myosin Iight ch*. BiophvsJ. 68: 120s-126s.

Lehmitn,W.. Vibert,P., Craiga, and BannyM. 1996. Actin and the structure of smooth muscle thin filaments. [n Biochemisuy of Smooth Muscle Conaction. M.Barany, editor. Academic Press, 47-60.

Kabsch,W., Mannherr,H.G., SuckD., P a i f 3-, and HohesJLC. 1990. Atomic smcnue of the actin:DNase L complex. Nature 3 4 7 3 7 4 .

GeevsMA. and HoimesJLC. 1999. Strucninl mechanism afrnuscle contraction. hnri.RevBiochem 68~687-728.

Hux1eyJ.E. 1957. The double anay of filaments in cross-smated muscle. J.Bioph~~.BiocfiemC~'tof~ 3 ~ 6 3 1-648-

MooreJ'B., HuxIeyHE., and DeRosier,D.J. 1970- Ilme-dimensionai reconsmction of F-actin, thin filaments and deconted thia filaments. JNoLBioL 50:279-295.

18. SpudichJA. 1994. How moIecular motors work Nahrre 3725 15-5 18.

Chalovich,J.M. 1992. Actin mediated regulation of musck contraction. Phannacol-Ther. W95-148.

Gordons\.M., Homsher,E., and RegnierM. 2000. Regdation of contraction in smated muscle. Physiol Rev. 80:853-924.

HorowitzA., MeniceCB., Lapone,R.. and Morgan,ELG. 1996. Mechanisms of smooth muscle contraction. Physiol Rev. 76:967-1003.

McLach1anA.D. and Stewart&L 1975. Tmpomyosin coiled-coi1 interactions: evidence for an unstaggered structure. J.Mol. Biol. 98:293-304.

h1cLachlanA.D. and Stewart,h.I. 1976. The 14-fold periodicity in alpha-tropomyosin and the interaction with actin. I.!Llol.Biol. 103:271-298.

Smi1lie.L.B. 1982. Preparation and identification of a. ,I.fe~horls En_?.mol. 85 Pt B:?3J-Xl.

Flicker.P.F., Phi1iips.G.N.. Jr., and C0hen.C. 1982. Troponin and its intemction with tropomyosin. An electron microscopie study. J.MoLBio1. 162 :493-50 1 .

Fanh.C.S.. Miyamoto,CA., Ramos.C.H.. da Si1vaA.C.. Quaggi0.R.B.. Fujimori.K.. Smillie.L.B., and Reinach2.C. 1994. Suuctuml and regdatory fùnctions of the Mi2- and COOH-teminal regions of skelrtal muscle troponin 1. J.Biol. Chm. 169:5230-5240.

Tobacman.L.S. 1996. ïh in filament-mediated regulation of cardiac contraction .-innu.Rev.Ph~siol 58:4474S 1.

Herzberg.0. and James,M,N. 1983. RetTned crysial suucrure ortroponin C from turkey skeletril muscle at 1.0 A resolution. J~Lfol.Bi01. 303:76 1-779.

Slupsky,CBI. and Sykes,B.D. 1995. NMR solution structure of calcium-satunted skeletal muscle troponin C. Biochemisrty 34: 15953-15964.

SweeneyJLL., Brito,RM., Rosevear.P.R, and PutkeyJA. 1990. The Iow-affmity Cd(+)-buiding sites in cardiaclslow skeletal muscle troponin C perfom distinct fiinctions: site [ done cannot mgger contraction. ProcNatlJcad.Sci.I/.SA 87:W 8-9542.

Wiikinson LM and Grand RJA. 1978. Comparkon of amino acid sequence of wponin I h m different smated muscles. Nature 271:3 1-35.

Moir,A.J., SolaroBJ., and Pemy,S.V. 1980. The site of phosphorylation of mponin 1 in the perfused rabbit h e m The effect of adrenaiine. BiochemJ- 185505-513-

Ngai,S.M. and Hodges&S. 1992, BiologicalIy important interactions between syntheuc peptides of the N-temiinal region oftroponin 1 and troponin C. JBioLChem. 267: 157 15- tS72O.

LesyzkJ., GrabarekZ., GergeIyJ., and CoIliosJH 1990. Characterization of zero-Iength cross-links between nbbit skeletal muscle troponin C and troponin 1: Evidence for direct interaction between the inhibitory region of Tni and the NH2-, regdatory domain of TnC BiochemLsyv 29 299-304.

Ngai,S.M., SonnichsettJ.D., and HodgesJLS. 1994. Photochernicd cross-linkuig between native rabbit skeletal troponin C and benzoylberizoyl-troponin i inhibitory peptide, residues 10.1-1 15. J.BioLCheni. 269:2 165-2 172.

Kobayashi-T., Tao,T., Gergely,J., and Col1insJ.H. 1994. Strucrure of the troponin cornplex, Implications of photocross-linking of troponin 1 to troponin C thiol mutants. JJfioLChem. 2695725-5729.

Levine,BA., Mou,A.J., and PerrySV. 1988. The Urtenction of troponin-1 with the N-terminal region of actin. Eiir.J.Biochem. 172:389-397.

Syska.H., Wi1kinsonJ.M.. GrandJtJ.. and Perry,S.V. 1976. The relationship between biological activity and prümy structure of mponin 1 h m white skeIed muscle of the rabbit Bi0chem.J. 153:375-387.

Talbot,J-4- and H0dgesJ.S. 198 1 . Synthetic studies on the inhibitory region of nbbit skeletal troponin 1. Relationship of mino acid sequence to biologicai acrivicy. J.Biol.Chem. 256:279S-2802.

So1aro.R.J. and Van EykJ. 1996. Altered interactions among thin Elment proteins modulate cardiac Function. J.Mof.Ce11 Cardiol. 2 8 2 17-230.

Van EykJ.E. and Hodges.R.S. 1988. The biological importance of wch amino acid residue of the uoponin 1 inhibitory sequence 103-1 LS in the interaction with uoponin C and uopomyosin-actin. J.Biol.Chem, 263: 1726- 1732.

Tripet&.. Van EykJE., and Hodgcs.R.S. 1997. ktapping of a second actin-tmparnysin and a second uoponin C binding site ivithin the C terminus of troponin 1. and their importance in the Cd+-dependent regulation of muscle concnction. 3.Mol.Biol. 27 1:72S-750.

Hm1ey.H.E. 1972. Structural changes in the myosin containhg filaments filmenu during çoncnction Cold Spriny Hurb.~mp.Qriant.Bioi. 27:36 1-376.

Kress.M.. Hu..ley,H.E.. Famqi4.R. and HendnxJ. 1986. Smictunl changes during activation of frog muscle studied by the-resolved X-ny d ihc t ion J.Mol.Biol. L38:325-342.

GreeneL-E, and Eisenberg,E. 1980. Cooperative binding of myosin subfragrnent-1 to the actin-troponin- tropomyosin cornplen ProcNatl,.lcadSci. U S A 7726 16-2620.

Lehrer,S.S. and Morris,ES. 1982. Dual elfects of tropomyosin and troponin-tropomyosin on actomyosin subhgment 1 ATPase. J-BioLChm 257:8073-8080.

HiII,T.L., Eisenberg,E., and Greenei.. L 980, Theoretical mode1 for the cooperative equilibrium binding of myosin subhgment I ro the actin-troponin-cropomyosin complex Proc-Natl-4cadSci.USd 772 186- 3 190.

bicWop,D.F. and Geeves,M.A. 1993. ReguIation of the interaction between acM and myosin subhgment 1: evidence for thrre States of the thin filament. BiophysJ 65:693-70 1.

VibertS., Craig,R, and L e h a W . 1997. Steric-mode1 for activation of muscle thin flaments. JMol.BioI. 1663-1 4.

Tobacman,L.S. and SawyerD. 1990. Calcium binds cooperativeIy to the regdatory sites of the cardiac thin filament, J.BioLChem. 26533 1-939.

5 1. Tobacman,L.S. and Buners,CA 2000. A new mode1 of cooperative myosin-thin filament binding J.Biol.Chem. 275:27587-27593.

FiIatov,V.L., KatrukhaA-G., Bulargina,T.V., and Gusev.N.B. 1999. Troponui: structure, properties, and mechanism of îünctioning. Biochernistry (hfosc.) 64:969-985.

IkebeM. and Hartshome,D.J. 1985. Phosphorylation of smooth muscle myosin at two distinct sites by myosin light chah kinase. J.BiuLCkem. 260: 10027- 1003 1.

SellersJ.R., PatoM.D., and AdelsteinJ1S. 198 1. Reverçible phosphorylation ofsmooth muscle myosin. heavy meromyosin, and platelet myosin JJtiuLChem. 256: 13 137-13 142.

Trybus,K.M., Freyzon.Y.. FaustLZ., and Sweeney.H.L. 1997. Spare the rod. spoil the regulation: necessity for a myosin rod Proc.Narl.Acad.Sci. LU.-l 94~48-52.

Smil1ie.L.B. 1996. Tropomyosin. In Biochemistry of Smooth Muscle Contraction. M.Barany, editor. Academic Press. 63-75.

SobieszekA. and Smal1.J.V. 1977. Regulation of the actin-myosin interaction in vertebrate smooth muscle: activation via a myosin light chah kinase and the effect of tropornyosin. J.Mol.Bio1. 1 12559-576,

Trybus.K.M, 1996. h4yosin regulation and assembly. In Biochemistry of Smooth Muscle Contraction, M-Barany, editor. Academic Press. 3746 .

Lehman,W., Hatch,V., Korman.V.. RosolM. . Ihomas.L,, Maynim.R.. GeevesMA.. Van EykJ.E.. Tobacman,L.S.. and Craig-R. 2000. Tropomyosin and actin isoforms modulate the localization of uopomyosin srrands on m i n filaments. J.:CloLBiol. 303593-606.

S0bue.K.. Muramoto.Y., Fujita-M.. and Kakiuchi.S. 198 1. Purification of a calmodulin-binding proicin from chicken g b a r d diat intencts with F-min. ProciVurl. ..LcodSci.LT.S.II. 78:5652-3655.

GncrffiP., Wang,C.L.. and Stafford,W.F. t988. Caldesmon. ~Ioleculiw weight and subunit composition by analytical ultracentrifugation J. Bid C'hem. 263:14 196-14202.

Li.Y., Zhuang,S., Guo.H., MabuchiX., LufLC.. and Wang,CA. 2000. The major myosin-binding site of caldesmon resides near its Ktenninal extreme- LBioLChem. 275: 10989-10994.

Watson&IM., Kuhn.A.E., Novy,R.E., iii.I., and Mak.A.S. 1990. Caldesmon-binding sites on tropomyosin, JJiol.Chem. 265: 18860- 18866.

ChalovichJM., BryanJ., Bensot$.E., and VeiazL. 1992. Localkation and characterization of a 73- kDa region of caldesmon which reversibly inhibits actomyosin ATPase activity. LBiol-Chem. 267: 16644 16650.

BartegiA., FanoumA., DerancourtJ., and KassabJL 1990. Characterhtion of the carboxyt-tecminaII0- kDa cyanogen bromide kgment ofcaidesmon as an act&czImoduIin-binding region J.ioKhcm. 265: 1523 1-15238.

WigarcLCJ., P a u l U , and MurphyRA. 1997. Energetic cost of activation processes during conmcuon O f swine artenal smooth muscle- J.PhysioLKond.) 50 112 13-733.

67. Fujii,T., Ozawa,J., Ogoma,Y., and Kondo,Y. 1988. Interaction benveen chicken g k d caldesmon and tropomyosin J.Biochem.(To&o.) lO.I:73 4-737.

Wang,C,L., Wang,L.W., Xu.SA., Lu,R.C., Saavedra-Alanis,V,, and BryanJ. 199 1. Locaiization of the caimoduh- and the actin-binding sites of caldesmon. J.Biol.Chem. 2669 166-9 172.

MezgueldiM-, Derancourt,J., Calas,B., Kassab,R, and FlittoumA 1994. Precise identification of the reylatory F-actin- and caImoduIin- binding sequences in the 10-kDa carboxyl-terminai domain of caldesmon. JBiol. Chrm. 269: 12824- 12832.

Wang,C.L., Chaiovich.l.M., Gnceffa,P., Lu,RC., MabuchCK-, and Stafford.W.F. 199 1. A Iong heliv h m the central region ofsmooth muscle caldesmon J.Biol.Chrm. 266: 13958-13963.

Hayashi,K., Fujio,Y., Kato,I., and SobueJC. 199 1. Stnictunl and knctional relationships between h- and 1-caldesmons. J.Bio1.Chem. 266:355-36 1.

Bal1.E.H. and KovalaT. 1988. Mapping of caIdesmon: relationship benveen the high and low molecuIar weight forms. Biochrmisr~ 276093-6095.

VeIaz.L., Hemnc.bl.E.. Bens0n.C.E.. and ChalovichJM. 1989. The binding of caldesmon to actin and its effect on the ATPase activity of soluble myosin subhgments in the presence and absence oftropomyosin. J Biol.Chem. 161:9602-96 10.

Furst.D.O., Cross.R.A.. De Mey.J., and Smal1J.V. 1986. Caldesmon is an tlongated flexible molecul: localized in the actomyosin domains of smooth muscle. EMBOJ. 5 2 5 1-257.

Smith.C.W., Pritchard.K.. and MarstonS.B. 1987. The mechanism of C S - regulation ofvascular smooth muscle thin filaments by caldesmon and calmodulin J.Biol.Chem. 262: 1 16-1 22.

Marston.S. and Huber.P.A. 1996. Caldesmon. In Biochemistry of Smooth Muscle Contraction. M-Barany, editor. Academic Press, 77-90.

LehrnqW.. Cnig,R., LuiJ., and Moody,C. 1989. Caidesmon and the structure of s m m h muscIe thin filaments: UnmunoIocalization of caldesrnon on thin filaments. JXucle Res. Cell hforiL 10: O 1-1 12.

Ngai,P.K, and Walsh,M.P. 1984. Inhibition ofsmooth muscle actin-activated myosin Mg2+-ATPase activity by caldesmon J.Biol.Chern. 259: 13656-13659.

Velaz,i., Chen,Y.D., and Chaiovich J.M. 1993. Characterization o f a caidesmon fragment titat cornpetes with myosin-ATP binding to actin. Bioph-vs.L 65:892-898.

HonuchiJCY., SamueI,M., and Chacko,S. 199 1. Mechanism for the inhibition of acto-heavy meromyosin ATPase by the actin/calmoduIin binding domain of caldesmon. Biochemistry 30:7 12-7 17.

Lehman,W, Moody,C., and Craig,R 1990. Caidesmon and the structure of vertebrate smooth muscle thin filaments. A minireview. AnnJLYAcad.Sci 599:75-84.

Watson,M.H., Kuhn,A.E., and Mak,A.S. 1990. Caldesmon, calmodulin and mpomyosm interactions. BiochÏmBiopl~~ilctu 1054: 103-1 13.

Brenner& Yu,L.C., and ChaiovïchJM. 199 1. Paralle1 Ïnhiiihon of active Force and reIaxed fiber

stifiess in skeletal muscle by caldesmon: implications for the pathway to force genention. PracAa1l-4cadSci.(I.SA~ 885739-57.13.

SomIyo,A.P. and Som1y0,A-V. 1994. Signal transduction and regulation in smooth muscle [published erratum appears in Nature 1994 Dec 22-29;372(6505):8 121. Nan<re 37223 1-236.

BamnyN. and BaranyX. 1996. In Biochemistry of Smooth Muscle Contraction. M-Barany, editor. Academic Press, 321-339.

Rembol$CM. and Murphy,R.A. 1988. Myoplasmic [Cd+] determines myosin phosphorylation in agonist- stimulated swine artenal smooth muscle. Circ.Res. 63593-603-

DiI1on.P.F.. Aksoy.M.O., Driska,S.P., and Murphy,R.A. 198 1. Myosin phosphorylation and the cross- bridge cycle in anenal smooth muscIe. Science 21 1:495497.

Hai,C.M. and Murphy.R.A. 1988. Cross-bridge phosphorylation and regulation of latch state in smooth muscle. ..lm.J.Physiol. m C 9 9 - 106.

StraussdD. and Murphy.RA. 1996. In Biochemistry of Smooth Muscle Contraction. Academic Press.

OishiJC, Takano-Ohmuro,H., Minakawa-Matsuo,N., Suga,O,, Karibe,H., KohamaK., and Uchida.M.K. 199 1. Oxytacin contncts nt uterine mooth muscle in C d ( + ) - k e medium without any phosphorylation of myosin light c h a h Biochem.Biopl~~s. Res. Commun. 176: 122- 128.

Gerthoffer,W.T., Murphey,K.A., and Gunst,S.J. 1989. Aequorin luminescence, myosin phosphorylation. and active stress in tracheal smooth muscle. Am.J.Pl~ysiol 157:C1062-C 1068.

McDanie1N.L.. CheeXL., Singer.H.X.. Murphy.R.~.. and Rembo1d.C.M. 1992. Nitrovasodilators relm menal smooth muscle by decreasing [Cd-li and uncoupling stress tiom myosin phosphorylation ,.lmJ.Physiol. 263:C46 1 -C467.

Somlyo,A.V., GoIdman,Y.E., FujimorLT., BondM., Trentham,DK, and Som1yoA.P. 1988. Cross- bridge kinetics, cooperativity, and negatively stnined cross- bridges in vertebrate smooth muscle. A laser- flash photolysis snidy. J.Gen.Physio1. 9 1: 165-192.

Winder,S.J. and Waish,MS. 1996. Calponin. Curr.Top.Cel1 Regil 34:33-61.

Zhan,Q.Q., Wong,S.S., and Wang,C.L. 199 1. A calmodulin-binding peptide of caldesmon. J.Biol.Chem. 266~21810-218 14.

Mmton,S.B., Fraser,LD., Huber,PA., Pritchard& GusevJB., and TomkK. 1994. Location of w o contact sites between human smooth muscle caldesmon and Ca(;?+)-calmodulin, J.ioLChem 269:s 134- 8 139.

Gneiher,SS., Heinonen,T.Y., Raharjo,W.H., Jin,JS., and Mak,A.S. 1997. Tryptophan residues in cddesmon are major determinam for calmodulin binding, Biochemisny 36:364-369.

Tanaka,T3 OhtaH., KandaX., HidakaJi., and SobueJL 1990. Phosphorylation of hi&-Mrcaldesmon by protein kinase C modulates the regulatory Function of this protein on the interaction between rtctin and myosin. Eur.J.Biochem. 188:495500.

99. 1kebe.M. and HornickT. 199 1. Determination of the phosphorylation sites of smooth muscle cddesmon by protein kinase C. Arch.Biochm.Biophvs. 288:538-542.

Vorotnikov,A.V., Gusev,N.B., HuaS., ColIins.JH., Redwood,C.S., and Marston,S.B. 1994. Phosphorylation of aorta caldesmon by endogenous proteolytic hgments of protein kinase C. J.hiit.de ResCell Motil. 1537-48.

ikebe&C. and Reardon,S. 1990. Phosphorylation of srnooth muscle caldesmon by calmodulin-dependent protein kinase 11, Identification of the phosphorylation sites. J.3iol.Chem. 265: 17607-176 i2.

NgaiJ.K. and Walsh,M.P. 1987. The effects of phosphorylation of smooth-muscle ca ldesmo~ Bi0chem.J. 24:J 1 7425.

Sutherland,C., Renaux.B.S.. McKay,DJ., and Wa1sh.M.P. 1994. Phosphorylation ofcaldesmon by smooth-muscle casein kinase II. J~tIuscIe Res.Cef1 Motil. L5:440-456.

Bogatcheva.N.V.. VorotnikovA-V., Birukov,K.G., Shirinsky,V.P., and Gusev.N.B. 1993. Phosphorylation by casein kinase II affects the interaction O € caldesrnon with smooth muscle myosin and tropomyosin. 3iochem.J. 290:137-442.

Hettasch.J.M. and Se1lersJ.R 199 1. Caldesrnon phosphorylation in intact human platelets by CAMP- dependent protein kinase and protein kinase C. J.Biol.Chem. 266: 1 1576-1 188 1.

Yamashiro-S., Yamakita.Y.. HosoyaJL. and Matsumun.F. 199 1. Phosphorylation of non-muscle caldesrnon by p34cdc2 b a s e duiing mitosis. :Vuntre 349: 169- 172.

Mak4.S.. WatsonM-H., Litwin.C.M.. and Wang.l.H. 199 1. Phosphorylation of caldesmon by cdcZ kinase. J.Biol.Chem. 266:667S-668 1.

MakAS.. Carpenter-M., Smillie.L.B., and WangJ.H. 199 1. Phosphorylation of caldesmon by p34cdc2 kinase. [dentitication of phosphorylation sites. J-BioLChem. 266: 1997 1-19975.

A h , L . P . and HathûwayS.R 1993. Identification of mitogen-activated protein kinase phosphotyIation sequences in mammaiian h-Caidrsmon FE3S Leu. 32256-60.

Childs,T.J., Watso&I.H., Sanghera.l.S., Campbell,D.L., PelechS-L., and Maks\.S. 1992, Phosphorylation of smooth muscle caldesmon by mitogen-activated protein ( M M ) kinase and expression of MAP kinase in differentiated smooth muscle cells. LBidChem. 2672853-22859,

Tanaka,T., OhtaJi., KandaK., H i d a k a . . and S0bue.K. 1990. Phosphorylation of high-Mr caidesmon by protein kinase C moddates the regdatory fùnction of this protein on the interaction between actin and myosin. Eur.1Biochem. I88:495-500.

ikebe,M. and Reardon,S. 1988. Binduig of cddesmon ta smooth muscle myosin J.Biol.Chem, 2633055- 3058-

Robertson,S.P., Johnson,.r.D., Holroyde,MJ.. Kranias,E.G., PotterJD., and Solaro-R.1. 1982. The effect of troponin 1 phosphorylation on the Cd+-binding propenies of the Cd+-regulatory site of bovine cudiac troponin. J.Biol.Cliem- 257:260-263-

SoIaroJU., MouAl., and Perry,S,V, L976. Phosphorylation of troponin i and the inotropic effect of

adrenaiine in the perfused nbbit hem. rVarure 262:6 15-6 17.

Pfitzer,G., Zeugner,C., TroschkM., and Chdovich,J.M. 1993. Cddesmon and a 20-kDa actin-binding hgment of caldesrnon inhibit tension development in skinned gizzard muscle fiber bundles. Proc.Na11-ricad.Sci.U.S.A 90:5904-5908.

B0lli.R and M&m,E. 1999. Molecular and celluIar mechanisms of myocardial stunning. Pliy~iol Rrv. 79:609-634.

Gao,W.D., Atar,D., Liu,Y., PerqN-G., MurphyAM., and Marban,E. 1997. Role of troponin I proteolysis in the parhogenesis ofsnimed myocardiun Circ.Res. 801393-399.

McDonoughJ.L., Arrel1.D.K.. and Van Eyk,J.E. 1999. Troponin t degradation and covalent complex formation accompanies myocardial ischemidreperfusion injury. CircRes. 849-20.

Van Eyk.J.E., Powers,F., Law,W., Larue,C., Hodges.R.S.. and Solaro.M. 1998. Breakdown and release of myofilament pmteins during ischemia and ischernialreperfusion in n t h e m : identification of degradaiion productç and effects on the @-force relation. Cire-Rex 3226 1-271.

MurphyAM.. Kog1er.H.. Gcorgakopoulos.D,. McDonough.1.L.. K;iss.DA.. Van Eyk.J.E.. and Marban.E. 2000. Transgenic mouse model of srunned rnyocardium Science 287:JSS-49 1.

Foster.0.B. and Van Eyk,J.E. 1999. in searcfi of the proteins that cause myocardial stunning. CircRes. S5:370-473.

Ga0,W.D.. Aiar.D., 3ackx.P.H.. and Muban-E. 1995. Relationship berween intncrhIar mlcium and contractile force in s m e d rnyocardiwn. Direct evidrnce for decreased myotilament C d + responsiveness and aitered diasioiic tirnction in intact venoricular muscle, CircRes. 76: IO36-IO48.

MarbanE. and Gao,W.D. 1995. S m e d myocariiium: a disease ofthe myofilatnents'? Basic Re~Carrliol. 90969-272,

Gao,W.D., Liu.Y., MeIlgrenJL, and Marban.E, 1996. incrinsic myofilament aiterations underiyùlg the decreased contractiIity of srunned myocardium. A consequence of Cd+-dependent pmteolysis? Circles. 781455-465.

NishimwaJ., Koiber,M., and van Breerne~C. 1988- Norepinephrine and GTP-gmma-S Încrease rnyofihnent Cd+ sensitivicy in dpha-toxin permeabilized arterïd smooth muscIe Biochem~iop~s.Res.Cornm~~n~ 157577-683.

Gong,M-C., FujiharaK, SomIyoAV, and Som1yoA.P- 1997. Translocation of rhoA m c i a t e d with Ca2+ sensitization ofsmooth muscie. JBioLChem. 272: 10704-10709.

Amano,M, Itofl., KimuraX, Fuhn,Y., ChiharaK-, Nakano,T., Mat.mm,Y.. and KaÏbuctiiX. 1996. Phosphoryiation and activation of myosin by Rhc-associated kinase ( R h - kinase), LBioLChern. 27 1 :20246-20249.

OkawaX. d al. 1996. Regdation ofmyosin phosphatase by Rho and Rheassociated kinase (Rho- kinase). Science 773:245-248.

Som1yoA.P. and Som1yoA.V. 2000. Signal transduccian by G-proteins, rho-kinase and protein phosphatase to smooth musde md non-muscle myosin IL dPhysiol522 Pt 2: 177-185.

K h d i l ~ . and M0rgan.K.G. 1992. Phenylephrine-ùiduced tmnslacation of protein kinase C and shortenhg oftwo spes of vaxular cells of the ferret. J.Physiol455:585-599.

KhaliLRA., Lajoie,C., Resnick,hI.S., and MorganKG. 1992. Ca(2t)-independent isofomis of pmtein kinase C differentially uanslocate in srnooh musde. Am.dPhysiol263:C? 14X7 19.

Walker,L.A.. Gaiily.P., 1ensenJ.E.. SomIyoAV., and SornlyoA.P. 1998. The unimpanrince of behg (protein kinase C) epsilon. FASEB J. i2:8 13-82 1.

Kitazaw~T., TakizawaN., Ikebebl., and EtoM. 1999. Reconstitution of pmtein kinase C-induced contractile C d + sensitizanon in mron X 100-demembmnated nbbit anenal smooih muscle. J.Phy.siol520 Pt i:139-152.

Adarn.L.P., Gapinski.C.1.. and Hathaway.D.R, 1992. PhosphoryIation sequences in h-caIdesmon h m phorbol ester-stimulated canine aonas. FEBSLe!i. 302:223-226.

Khalil.R.h. and M0rgan.K.G. 1993, PKC-mediated redistribution of mitogen-activated protein kinase during smooth musde ce11 activation ..lnrJPh~ioL 26j:C.IO6-C411.

Nix0n.G.F.. IizukaK. Haystead.C.M.. HaysteabTA., SomlyoAP.. and SomIyoA.V. 1995. Phosphorylation oFcaldesmon by mitopen-activcited protein kinase with no e f f e t on Cd+ sensitiviry in rabbit smooth musde. J.Piiysiol.(Lonll.I 487283-259,

EarleyJJ., SuX.. and MoreIandRS. 1998. Caldesmon inhibits active crossbridges in unstimulared vascular smooth muscle: an antisense oIigodeoxynucIeotide approach CircRes. 33:66 1-667.

JainM., Dai,H.B., Carrozza,l.P., Seilke.F.W., and MorganKG. 1996. intrinsic tone as patentid vriscular reserve in conductance and resistrtnce vessels. Circulation 94: 1083-1088-

Greene,L.E., Sellers,J.R, Eisenberg,E.. and AdelsteinJLS. 1983. Binding of giPard smooth muscle myosin subhgmeut 1 to actin in the presence and absence of adenovine S-mphosphate. Biochrm&ry 22530-535.

WeedsAG. and TaylorJLS. L975. Separation of subhgment-1 isoenzymes h m nbbit skeleai muscle myosin. Nature 25759-56.

MargossianSS. and Lowey,S. 1982. P q m t i o n of myosin and its subfngments h m rabbit skeletal musde. Metho& E ~ y o l . 85 Pt 855-7 1.

PardeçJ-D. and SpudicbJA 1982 f i f ica t ion of muscle actin . Methoris En,?o l . 85 Pt B: 164-8 1:164- L8L.

Smi1iie.L.B. 1 982. Preparation and identifkation of a, Medo& E e m o i . 85 Pt B:ZA-24 1.

BretscherA. 1984, Srnooth muscle caidesmoa. Rapid purification and F-actin cross-linking pmperties.

Tobacman,L.S. 1996. Thin filament-mediated regdation of cardiac contraction Anmr.Rev.Physio1 58:447-48 1.

Van Eyk,J.E., Thomas,L.T., Tripe*., Wiesner,RJ., PearIstoneJ.R, Farah,C.S., Reinach,F.C.. and HodgesRS. 1997. Distinct regions of troponin 1 regulate Cd+-dependent activation and C d + sensitivicy of the acto-S 1-TM ATPase activity of the thin filament. J.ioLCkem. 272: 10529-10537.

ChifflegS., Tomglia,A., C h i e s a , and Tolosa% 1988. A method for the detemination of inorganic phosphate in the presence of labile organic phosphate and hi& concentrations of protein: application to lens ATPases, Anal-Biochem. 168: 1-4.

Rarick,H.M., Tu,KH.. Solaro,R.J., and hlartinA.F. 1997. The C terminus of cardiac troponin 1 is essential for full inhibitory activiv and Cd+ sensitiviy of nt myofibrils. LBioLChem. 27226887- 26892.

Sweeney,H.L., Feng,H.S., YangZ., and Watkins.H. 1998. Functional analyses of troponin T mutations that cause hypemophic cardiomyopathy: insights into disease pathogenesis and troponin hnction. Proc.Natl.=lcurl.Sc.i. U.S.4 95: 1406- 144 10.

McDonoughJ.L., LabuggerB., PickettW.. TseA4.Y.. MacKrnzie.S., Pang.S.C.. A1ar.D.. R0pchan.G.. and Van Eyk.J.E. 200 1. Cardiac troponin I is modified in the myocardium of bypass patients. Circirluiion 10358-64.

Tan.J.L., Ravid.S., and Spudich J.A. 1992, Control oFnonmuscIe myosins by phosphorylation .4nnir.Rev.Biochem. 6 1373 1-59:721-759.

SniIlJ.T., Ga1lagher.P.J.. CIerring,B.P., and Kamm.K.E. 199 1. Vascularsmooth muscle contractile elements. Cellular regulation. Kvprrtemion 17:723-732.

Kureishi-Y., Kobayashi,S., h a n o M . , KimuraXI, KanaideH., Nakano,T. , Kaibuchi,K., and 1to.M. 1997. Rho-associated kinase directIy induces smooth muscle contraction through myosin light chah phosphorylation. J.BÏol.Chm 272: 12257-12260.

AmanoM., ItoM., K i m q K , FukamY., ChharaK., Nakano.T., MatsuuraY., and KaibuchiK. 1996. Phosphorylation and activation of myosin by Rho-associateci kinase (Rho- kinase). LBiol-Chern. 27 1 :20246-20249.

Zigmond,S.H. 1996. Signal nansduction and actin filament organization. Curr-OpinCe11 BioL 8:66-73.

Rid1eyA.J. 1996. Rho: iheme and variations. Cir=Biof. 6: 1256- 1264.

Manser,E., Leung,T., Salihuddin$I., ZhaoZS., and Lim,L. 1994, A brain serinelthreonîne protein kinase activated by Cdc42 and Rac 1- Nanrre 367:40-46.

Marcin,Gh., Bollag,G., McCocmickI., and AboA. 1995. Anovel serine kinase activated by racliCDC42Hsdependent autophosphoryhtion is related to P U 6 5 and STUO. EMBOJ. 14:4385.

Bapdia,S., Taylor,SJ., CreasylC.L., Chernoff,., and CenoneAA. 1995. Identification of a mousr p2 ICdc42iRac activated kinase. J.BiolClrem. 270:2273 1-22737.

Leberer,E., DignubD., Hougan,L., ThomasD.Y., and Whiteway,M. 1993. Dominant-negative mutants o f a yemt G-protein beta subunit identifv two functional regions involved in pheromone signallinp ~~0 J. 11:4805-4813.

Wu,C., Lee,S.F., Furmani&-Kazmierczak,E., Cote,G.P., ïhomasp.Y., and Leberers. 1996. Activation of myosin-I by members of the Ste20p protein kinase fmily. J.Biol.Chem. 271:3 1787-3 1790.

Manser,E., Chong,C., ZhaoZ.S., Leung,T., Michae1.G.. HaU,C., and Lim,L. 1995. Molecular cloning of a new member of the p2 1-Cdc42IRac-activated kinase (PAK) family. IBioLChem. 270:25070-25078.

JakobiP., Chen,C.J., Tuazon,P.T., and TnughJA. 1996. Molecularcloning and sequencing of the cytostatic G protein-activated protein kinase P.4K I. JBioLChem. 27 l:6206-62lL

HinaK., KikuchiA.~ Sasaki.T., Kur0da.S.. Kaibuchi,K., MatsuumY., Seki,H., Saidas., and Takai.Y. 1992. Involvement of rho p2 1 in the GTP-enhanced calcium ion sensitivity of smooth muscle conaction

J,Biol.Chem. 267237 19-5722.

Gong,M.C., 1inika.K.. Niu0n.G.. BrowneJ.P.. Halls\.. Eccleston.3.F.. Sugai&[., KobayashiS.. SomIyoA.V., and Som1yoA.P. 1996, Role of guanine nucleotide-binding proteins-ras-farnily or mmenc proteins or both-in C d + sensitization of smooth muscle. Proc.Nutl.Acud.Sci. U.S.A. 93: 1340- 1345.

Ono.B., SteusloffA., Just.1.. Aktories,K., and Pt3zer.G. 1996. Role of Rho proteins in carbachol-induced contractions in intact and permeabilized guinea-pig intestinal smooth muscle XPhvsiol496 ( Pt 2):3 17- 329.

Satoh.S., KreutzR., Wi1m.C.. GantenS. . and Pfitzer.G. 1994. Augmented agonist-induced Ca(?+)- sensitization ofcomnary anrry contraction in genetically hypertensivt nts. Evidence For altered signal transduction in the coronary smooth muscle cells. J.Clin.lntest. 94: 1397-1403.

Uehata.M., 1shizaki.T.. Sat0h.H.. Ono.T., KawahanT.. h4orishita.T. . Tarnakavdi.. Yamagami.K, InuiJ.. M a e k a w a , et al. 1997. Calcium sensitization of smooth muscle mediated by a Rho-associated protein kinase in hypertension. Naiirre 389:990-994.

Wu.C., Lee,S.F., Fumaniak-Kazmierczak,Ev Cote.G.P., Thomas,D.Y., and leberer,E. 1996. Activation of myosin4 by members of the Ste?Op protein kinase fmily. LBioLChem 2713 1787-3 1790.

NovyRE., Liu,L.F., Lin,C.S., Helfman,D.M., and Lin,J.J. 1993. Expression of smooth muscle and nonmuscle tropomyosins in Escherichia coli and characterizarion of bacterially produced tropomyosins, Biochim.Biophysdcta 1 162255265,

StraussJ.D., de LmemUeJ'., and PauIRJ. 1992. Effects of myosin kinase inhibithg peptide on contractility and LCZO phosphorylation in skinned smooth muscle ilrn.J.Physiol262:C1437-C 145.

Snauss.JD., BarthL., Van EykJ-E., and RueggJC. 1993. Methocls: A Compwion to Metho& in Enymology 5:28 1-290.

Mamer&., LeungT.. and Lim,L. 1995. Purification and assay of kinases that intenct with RadCdc42. ~bferliodrr Enzzol . X6:2 15-227.

PaulW., St~aussJD., and Krisanda..J.M. 1987. Tùe effects of cdcium on smooth muscle mechanics and energetics. Prog.ClinBioLRes. 245:3 19-332.

Truikle-MuIcahy,L, IchikawaX., HartçhorneDJ., Siegman,M.L, and Butier.T.M. 1995. Thiophosphorylation of the 130-Da subunit is associated with a decceascd activity of myosin light chah phosphatase in alpha-toxin-permeabilized smooth muscle. JJioI-Chem. 270: 18 19 1-18 194.

Katsuyama,H., Wang,C.L., and MorganJCG. 1992. Regdation ofvascular smooth muscle tone by caldesmon. J.Biol.Chem. 267: 14555- 14558,

MalmqvistJJ., Amer& Makuch,R, and Dabrowska,R L996. The effects of caldesmon exmction on mechanical propenies of skinned smooth muscie t3re preparations. Pjhgers .&ch. 432:24 1-247.

Pfiaer,G. and AmerA. 1998. involvement of small GTPases in the regulation of smooth muscle conmction [in Process Citation]. Acta Pi~usioLScand. 164:449-456.

AmanoM., ItoLI., Kimura,K., Fukata,Y., ChihwK., N h o . T . , MatsuumY., and KaibuchiX. 1996. Phosphoryhtion and activation of myosin by Rho-associared kinase (Rho-kinase). J. Biol.Chem. 27 1 :20%6-?0249.

Bagrodia,S., DerilardB., Davis.R.L. and Cen0ne.R.A. 1995. Cdc4Z and PAK-mediated signding leeds to Jun kinase and p38 mitogrn- activated pmtein kinase activation J.Biol.Chrm. 270:77995-27998.

Van Eyk.J.E., .4rrell.D.K., Foster.D.B., StnussJ.D., Hein0nen.T.Y.. Fumaniak-KaunirrczkE.. Cote,G.P.. and b1dc.A.S. 1998. Different moleculiw mechmisms for Rho family GTPase-dependent. C d + - independent contxaction of smooth muscle. J. Biol.Chem. 27323433-73439.

Whittaker M. WiIson-Kubalek EM, Smith JE. Faust L, Milligan RA. Sweenry HL. 1995 A 35-A movement of smooth muscle myosin on ADP releaseNaturr. 378:748-5 1.

Chew.T.L.. MasancchiaRA.. GoeckelerLM.. and Wysolmerski.R.B. 1998. Phosphorylation of non- muscle myosin II regulatory light chah by p21-actimted kinase (gamma-PAK) J.ACl~~scle Res.Cril :Cloril. 19:839-854.

LashJ.A., Sellers J.R, and Hathaway,D.R. 1986. The effects of caldesmon on smooth muscle heavy actomeromyosin AïPase activity and binding of heaty meromyosin to actin LBioLChem. 26 1: 16 155- 16160.

ChalovichJ.M., CorneIiusS., and Bens0n.C.E. 1987. Caidemon inhibits skeletal actomyosin subfhgment-1 AïTase activity and the binding of myosin subfngment-1 to actin J.BioLChem. 262:j7 1 1- 5716.

Hemric,M.E. and Cha1ovichJ.M. 1990. Characterïzation of caldesrnon binding to myosin. J.BioLChem 265: 19672-19678.

Chacko,S., Miyata,H., and HoriuchiKY. 1987. ModuIation of actomyosin ATPase by thin 6Iament- associated protek. Prog-ClinBioLRes. 245243-58: 143- 158.

PritchardK. and Marston,S.B. 1989. Cd+-calmodulin bmding to caldesmon and the caldesmon-actin- mpomyosin cornplex Its role in Cd+ regulation of the activity of synthetic smooth-muscle thin EiIaments Biochem-J. 257:839-843.

ManiRS., McCubbhW-D., and Kay,C.M. 1992, Calcium-dependent regulation of caldesmon by an 11- kDa smooth muscle calcium-binding pmtein, caltropin BiochemLrt-. 3 1 : 1 1896-1 190 1.

Krymsky,M.A., Chibalina,M.V., Shirinsky,VS., Mmton,S.B., and VoromikovAV. 1999. Evidence against the regulation of caldesmon inhibitory activity by pltlpl-lerk mitogen-activated prorein kinase in vitro and demonstration ofanother caldesmon hase in intact gizzard smooth muscle. FEBS Leu. 352:254-258.

Velaz.L., IngrahamRH., and Chalovich,J.M. 1990. Dissociation of the effect of caldesmon on the ATE'ase activity and on the binding of smooth heavy meromyosin to actin by partial digestion of caldesmon. J.iool.Chmn. 265:2929-2934.

IkebeJl., Reardon,S., Scott-Woo,G.C., ZhouL., and Koda,Y. 1990. Purification and chancterization of calmodulindependent multihnctional protein kinase h m smooth muscle: isolation of caldesmon kinase. Biochemkt~ 29: 1 1242-1 1248.

ZhangJl. and Voge1,H.J. 1994. The calmodulin-binding domain of caldesmon binds to calmodulin in an alpha-helical conformation. Biochemiïny 33: 1 163-1 171.

Zh0u.N.. YumT.. MakA.S.. and Voge1.H.J. 1997. NMR studies of caldesmon-calmodulin interactions. Biochemkrp 3 6 2 8 17-3825.

Katsuyama.tI.. Wang,C.L., and b1organ.K.G. 1992. Regulation of vascular smooth muscle tone by caldesmon. J. Biol.Chem. 767: 14555- 14558.

No1and.T-4.. Jr.. Guo& Raynor-R.L., 1ideamaN.M.. Averyhart-Fullard,V., Solaro.R.J., and Ku0J.F. 1995. Cardiac troponin 1 mutants. Phosphorylation by pmtein kinases C and A and regulation of Ca(2+)- stimulateci MgATE'ase of reconstituted actomyosin S- 1, J.Bio1.Chrm. 270:2%45-25454.

Lehman.W.. RosolM., Tobacman.L.S.. and Craig.R. 300 1. Troponin organization on relaxrd and activated thin tilaments revealed by electron rnicroscopy and three-dimensional reconstruction J,Mol.Biol. 307~739-744.

FrostJA., KhokhlalchevA., Stippec.S.. LVhiteM.A.. and Cobb.hI.H. 1998. Differential effects of PAKI-activating mutations reveal activity- dependent and -independent effects on cytoskeletal regulation LBioLChem. 273:28 19 1-28 198.

BagrodiaS. and CenoneSA. 1999. P A to the Future. Trends Ce11 Biol. 9:350-355.

Joneson+T., McDonoughM., Bar-SagiP., and Van AelstL. 1996. M C regulation ofactin p o l y m e ~ t i o n and prolifention by a pathwily distinct tiom Sun kinase [published emtum appears in Science 1997 Apr 11;276(53 10):185]. Science 274: 1374-136.

Sells,.MA, Knaus,ll.G., Bagrodia,S., Ambrose,D.M., Bokoch,GM., and ChemoffJ. 1997. Human p2I- activated kinase (Pak 1) regdates actin organization in munmaliui ce lk Curr.Biol. 7:202-2 10.

SandenL-C., Mat~umura,F.~ Bokoch,G.M., and de Lanerollep. 1999. inhibition of myosin light chah kinase by p21-activated k i ï ï e . Science 283:2083-2085.

Kureishi,Y., Kobayashi,S., AmanoM., KimuraX, KanaideJi-, Nakano,T-, KaibuchiX., and Ito,M. 1997. Rho-associated h a s e directly induces smooth muscle contraction thmugh myosin light chah p hosphorylation. JBiol. Chem. 273: 12257- 12260,

Kitanwa,T., ManioM., and SomIyo&P. 199 1. G pmtein-mediated inhiïition of myasin light-chain

phosphatase in vascular srnooth muscle. Pmc.NarlJlcad.Sci. U S A R8:9307-93 10.

Schmitz.U., khi&T., Ishida,M., SunpisitchatJ., HashamM.I., PeIechS., and Berk5.C. 1998. Angiotensin iI stimulates p21-activated kinase in vascuiar smooth muscle cells: role in activation of M. C k R m 82: 1272-1278.

Ohrakan,K., Inad@., Goto,H., Taki.W., ManserE., Lim,L, Izawa,L, and LnagakiM. 2000. p2 l- activated kinase PAK phosphorylates desmin at sites diflerent h m those for Rho-associated kinase. Biochem.Biophyx. Res.Cammun. 2727 12-7 16.

BottineIli R CovieIlo DA, Redwood CS. Pellegrino MA. Maron BJ, Spirim P, Watkins H, Reggiani C. 1998. A mutant tropomyosin thst causes hypemophic cardiomyopathy is expressed in vivo and associated with an incrgased calcium sensitivie. Circ Rrs.82: 106-15.

ElIioct K. Watkins H. and Redrvood CS.2000 Altered replatory propefties of human cardiac uoponin 1 mutants thnt cause hypemophic cardiomyapathy. J Biol CIim. 275:22069-74.

Tuazon PT. Spanos WC, Gump EL, 5fonnig CA. and Tnugh SA. 1997. Determinants for subsmte phospitorylation by p2 1-activated protein kinase (gamma-PM)Biochemk~.36: 16059-M.

O'Neil KT and DeGndo W. 1990. How calrnodufin binds its rargets: sequencr independent recognition of amphiphilic aIpha-helices. Trends Biocliem Sci. 1554-64.

Chapter 6. Muscle. 1994. In Color Atlas oFHistology. 2" ed. Garner L.P. and J.L. Hiatt editors. WiIliarns and Wilkins. p. 102.

Ferrieres G. Pugniere M. Mani JC. Villard S. Laprade hl. Doum P. Pau B. Grnier C. 7000 Systernatic mapping of regions ofhuman cardiac mponin I involveci in binding tocardiac troponin C: N- and C-terminal low affmity contributing regions. FEBS L m . -17939- 105.

Brzevka H, Knaus UG. Wang ZY+ Bokoch GM, Korn ED. I997 pZ 1-activated kinase has substrate specificity similm to Acanifiamoeba rnyosin Ihemy chah kinase and activates Acanrhamoeba myosin 19roc Nad Acad Sci U S A. 94: 1092-5.

Gaai A. Fiuang Z. T w o a PT.1999 Tnugh IA.Multisite autophosphoryIation of p21-activated protein kinase ~amma-PAK as a function of activation J Biol Chem.274:8022-8.