APPLED MICROBIOLOGY, Mar. 1975, p. 405-413Copyright 0 1975 American Society for Microbiology

Vol. 29, No. 3Printed in U.SA.

Two-Dimensional Polyacrylamide Gel Electrophoresis ofEnvelope Proteins of Escherichia coliW. CHARLES JOHNSON, T. J. SILHAVY,I AND W. BOOSI*

Department of Biological Chemistry, Harvard Medical School, and the Biochemical Research Laboratory,Massachusetts General Hospital, Boston, Massachusetts 02114

Received for publication 16 October 1974

A method of separating envelope proteins by two-dimensional polyacrylamidegel electrophoresis is described. Escherichia coli envelopes (inner and outermembranes) were prepared by French pressing and washed by repeatedcentrifugation. Membrane proteins were solubilized with guanidine thiocyanateand were dialyzed against urea prior to two-dimensional electrophoretic analysis.The slab gel apparatus and conditions were similar to the technique developed byMetz and Bogorad (1974) for the separation of ribosomal proteins. Thisseparation occurs in 8 M urea for the first dimension and in 0.2% sodium dodecylsulfate for the second dimension. The technique separates about 70 differentmembrane proteins in a highly reproducible fashion according to both intrinsiccharge and molecular weight. Some examples of alterations in the membraneprotein pattern are demonstrated. These alterations are caused by a mutationaffecting a sugar transport system and by growth in the presence of D-fucose,inducer of the transport system. A further example of membrane protein changesintroduced by growth at the nonpermissive temperature of a temperature-sensi-tive cell division mutant is shown. Finally, it is demonstrated that the majorouter membrane component of Escherichia coli K-12 contains more than fourproteins of similar molecular weight.

In recent years, the use of polyacrylamide gelelectrophoresis in the presence of sodium do-decyl sulfate (SDS) for the analysis of mem-brane proteins has become quite popular (1,3-5, 17). This method is advantageous becausethe same solvent conditions are used in the-primary solubilization and subsequent electro-phoretic analysis of the proteins of interest.Furthermore, it allows their molecular weight tobe estimated (22). However, separation of pro-teins of similar molecular weight will, for thesame reason, be rather poor. Indeed, despiteexcellent separation of many proteins in theEscherichia coli cell envelope, the major outermembrane protein appearing as one band by thistechnique has only recently been shown to con-sist of four proteins in some strains (18). Con-sequently, we looked for other methods whichmight combine the separating power of SDS-polyacrylamide gel electrophoresis with theadditional capability of separating proteins ac-cording to their intrinsic charge under denatur-ing conditions. The two-dimensional polyacryl-amide gel electrophoresis introduced for theseparation of ribosomal proteins by Kaltschmidt

I Present address: Fachbereich Biologie, UniversitAtKonstanz, 775 Konstanz, West Germany.

and Wittmann (6), and later modified by Metzand Bogorad (10) as well as Subramanian (21),seemed useful for this purpose. However, themajor drawback of this approach is that mem-brane proteins are water insoluble. Detergentssuch as Triton or SDS used for solubilizing mem-brane proteins usually interfere with electropho-retic separation in urea (the system most oftenchosen for the expression of charge effects) andare difficult to remove. On the other hand, thepresence of urea in the protein sample does notinterfere with electrophoresis in the presence ofSDS. Therefore, the mandatory sequence to fol-low for a two-dimensional electrophoresis is 8 Murea in the first dimension for separation accord-ing to intrinsic charge, followed by SDS in thesecond dimension for separation according tomolecular weight. Membrane proteins are notreadily soluble in 6 M urea, but once they aresolubilized by another solvent a large portion ofthe proteins remain in solution when dialyzedagainst 6 M urea. We adopted this method andused the chaotropic reagent guanidine thiocya-nate as the primary dissolving reagent, which isas effective a membrane solvent as SDS but hasthe added property of easy removal by dialysisagainst urea (12). The present paper gives an

405

on August 18, 2019 by guest

http://aem.asm

.org/D

ownloaded from

JOHNSON, SILHAVY, AND BOOS

account of the two-dimensional separation ofthe membrane proteins so treated. In particu-lar, we applied this method to search for the stillunknown but supposedly membrane-boundcomponents of the periplasmic galactose-bind-ing, protein-mediated f3-methylgalactosidetransport system (2) of E. coli.(A preliminary account of this paper was

given at the Mosbach Symposium, Germany,1974.)

MATERIALS AND METHODSBacterial strains and genetic methods. The

strains employed in this study are listed in Table 1along with their origins. Details of the construction ofthese strains have been previously described (20).Media and growth conditions. Cells were grown to

stationary phase with vigorous aeration in 500-mlcultures. M188-777 and 1000 were grown in minimalmedia A plus 0.2% succinate and 20 ,g of histidine perml (20), and BUG-6 was grown in 0.2% D-glycerol at 35and 42 C, respectively. D-Fucose (1 mM) was used asan inducer where noted in the text.Membrane preparation. Cells from a 500-ml cul-

ture were harvested and washed once in 50 ml ofmembrane buffer [10 mM tris(hydroxymethyl)ami-nomethane (Tris)-hydrochloride, pH 8.5, containing 5mM ethylenediaminetetraacetate, 0.2 M KCl,and 5 mM 2-mercaptoethanol (10) 1, resuspendedin 35 ml of the same buffer, and broken by twopassages through a French pressure cell at 9,000 lb/in2at 4 C. Whole cells were then removed by centrifuga-tion at 1,000 x g for 10 min. The supernatant was thencentrifuged at 30,000 x g for 1 h to pellet themembrane fraction. The membranes were resus-pended in 15 ml of membrane buffer and separatedinto 0.5-ml aliquots and solubilized.The outer membrane fraction was prepared as

described by Schnaitman (18). The alcohol precipi-tate was resuspended in membrane buffer and furtherprocessed as described below.Membrane solubilization. (i) Guanidine thio-

cyanate. Solid guanidine thiocyanate was addedto one aliquot of washed membranes to a final

concentration of 6 M. This produced an almost clear,yellow solution which was stirred at room temperaturefor 0.5 h and then dialyzed overnight against onechange of 100 volumes of 6 M urea containing 0.1 MTris-hydrochloride (pH 8.5) and 5 mM 2-mercapto-ethanol. The dialysate was centrifuged at 12,000 x gfor 1 h, the supernatant was carefully removed, andprotein concentration was determined by the methodof Lowry et al. (9). This solution was then ready fortwo-dimensional electrophoretic analysis. The pelletwas saved for further use (see below).

(ii) SDS solubilization. Two fractions were solubi-lized by SDS: an aliquot of washed membranes andthe insoluble material remaining after guanidinethiocyanate solubilization and urea dialysis. To thesuspension of washed membranes, 20% (wt/vol) SDScontaining 0.625 M Tris-hydrochloride (pH 6.8) and0.05 M 2-mercaptoethanol was added to a finalconcentration of 2% SDS and 1 to 2 mg of protein perml. The urea-insoluble pellet was suspended in 1.0 mlof a solution of 2% SDS containing 0.0625 M Tris-hydrochloride (pH 6.8) and 5 mM 2-mercaptoethanol.In addition, the urea-soluble material was preparedfor one-dimensional SDS electrophoresis by adding20% SDS (wt/vol) containing 0.625 M Tris-hydrochlo-ride (pH 6.8) and 0.05 M 2-mercaptoethanol to a finalconcentration of 2% SDS and 1 to 2 mg of protein perml. All suspensions were then placed in a boilingwater bath for 10 min and dialyzed against 100volumes of sample buffer (2% SDS [wt/vol] contain-ing 0.0625 M Tris-hydrochloride [pH 6.8], 10% glyc-erol, and 5 mM 2-mercaptoethanol) (7) overnight atroom temperature. The samples were then cen-trifuged at 12,000 x g for 1 h. The supernatant wasremoved, and the protein concentration was deter-mined. The samples were then subjected to one-dimensional SDS electrophoretic analysis. The re-maining SDS-insoluble material was clear in both thewhole membrane fraction and the urea-soluble frac-tion. However, the pellet from the urea-insolublematerial retained the original caramel color exhibitedafter the urea dialysis.

Electrophoresis. (i) One-dimensional SDS. Themethod employed was that described by Laemmli (7).A 4% stacking gel was used over an 8% separating gel.

TABLE 1. Bacterial strainsa

Strain Parent procedure Genotype Origin Reference

W3092cy- F-galK lacY Wu 22LA002 W3092cy- EMS, penicillin F- galK lacY his

selectionLA021 LA002 NTG F- galK lacY his mgIABC Silhavy 20M188-444 F- malA glpD strA ptsF his galE Silhavy 20

lacYM188-777 M188-444 P1 transduction, F- malA glpD strA his galE lacY Silhavy 20

W3092cy- donorM188-1000 M188-444 P1 transduction, F- malA glpD strA his mglABC Silhavy 20

LA021 donor galE lacY lac- galBUG-6 lac -gal- Clark 16, 19

a All strains are E. coli K-12; EMS, ethyl methane sulfonate; NTG, N-methyl-N'-nitro-N-nitrosoguanidine.

406 APPL. MICROBIOL.

on August 18, 2019 by guest

http://aem.asm

.org/D

ownloaded from

ELECTROPHORESIS OF E. COLI ENVELOPE PROTEINS

(ii) Two-dimensional SDS. The method used inthese studies was first described by Metz and Bogorad(10) except that 15-cm gel lengths were used (Subra-manian, personal communication). The first dimen-sional gel (run in 5-mm [inner diameter] tubes) con-tained 8 M urea and was buffered with 0.1 MTris-hydrochloride (pH 8.5). The electrode buffer was25 mM Tris-glycine (pH 9.2). Usually 300 gg ofprotein in 50- to 150-,ul aliquots was applied to eachgel and electrophoresed with 2 mA/gel at 350 V,dropping to 1 mA/gel and 400 V at the end of the runfor 4.5 h at 18 C.The second dimension was run as described by

Metz and Bogorad (10), except that the length of theapparatus was 5 cm longer. The first dimensional gelswere removed from the tubes and cemented on top ofthe slab gels with a 4% acrylamide-cementing gel. Thesecond dimension was run at 32 V (25 mA/gel decreas-ing to 4 mA/gel) for 17 h at 18 C. The gels were stainedfor 8 h at 37 C with a 0.1% solution of Coomassiebrilliant blue containing 50% methanol and 7.5%acetic acid and destained with one change of 15%methanol containing 7.5% acetic acid.

Molecular weight markers were polymerized in thecementing gel along with the first dimensional gel.Markers used were bovine serum albumin, oval-bumin, and chymotrypsinogen purchased fromWorthington Biochemicals, and hemerythrin was pro-vided by A. R. Subramanian.

Assays. Galactose-binding protein was detected byOuchterlony immunodiffusion (8). Transport of galac-tose by the f,-methylgalactoside transport system wasdetermined as previously described (15). fl-Galactosi-dase activity was determined by the hydrolysis ofO-nitrophenyl-B-D-galactopyranoside in toluenizedcells or without toluene in cell-free suspension (13).

RESULTS

Preparation of membranes. Membraneswere prepared in three steps: (i) harvesting andwashing of the cells; (ii) breakage by Frenchpressure cell at a pressure of 9,000 lb/in2; and(iii) three to four washings of the membranes bycentrifugation at 30,000 x g for 30 min.The reproducibility of membrane prepara-

tions using this method, as determined bysubsequent two-dimensional electrophoresis,was compared with that of membrane prepara-

tions obtained by sonic oscillation of the cells,followed by repeated washing and centrifuga-tion at 100,000 x g or more for several hours,and was found to be superior. Membrane prepa-

rations obtained by French pressing containedapparently fewer cytoplasmic constituents thanthose obtained with sonic oscillation. Mem-branes of cells fully induced for ,B-galactosidaseand prepared by the former method containedno detectable enzyme activity, whereas mem-

branes obtained by the latter technique stillexhibited traces of enzymatic activity even aftereight washings.

Solubilization of the membrane prepara-tion. Preliminary experiments performed toexplore the optimal conditions for a two-dimen-sional separation of membrane proteins basedon charge and molecular weight made threefacts obvious. (i) The presence of detergentssuch as 1% Triton X-100 or SDS (agents com-monly used for membrane solubilization) isundesirable when membrane proteins are elec-trophoresed in gels containing 8 M urea (12);such agents are difficult and time-consuming toremove by dialysis against urea, due to micellformation of the detergent. (ii) The additionalpresence of small amounts of urea in the proteinsample does not affect electrophoresis in gelscontaining SDS (12, 17). (iii) Membranes can-not be solubilized significantly by 6 M urea(12), but once solubilized, 6 M urea is able tokeep 70 to 80% of the proteins in solution. Thesefacts determined the sequence of the electro-phoretic separations consisting of initial electro-phoresis in 8 M urea and a subsequent run inthe presence of SDS. Moreover, the difficulty ofremoval of detergents by dialysis suggested theuse of a chaotropic agent. Consequently, weused 6 M guanidine thiocyanate, an agentintroduced by Moldow et al. (12), to initiallysolubilize the isolated membranes. Under theconditions used, this agent will solubilize 60 to90% of the membrane preparation when proteinis determined according to Lowry et al. (9).Guanidine thiocyanate had to be removed bydialysis against 6 M urea prior to electrophore-sis in the first dimension, since effective electro-phoresis in such a high salt concentration is notfeasible. The precipitation of 20 to 30% of thedetectable proteins during this dialysis is themajor drawback experienced in the two-dimen-sional analysis of membrane proteins. Table 2shows the loss of protein content during thesolubilization of the membrane proteins asdetermined by the Folin reagent. However, the

TABLE 2. Membrane yield and percentage recovery

Sample Amt (mg) % Recoverya

Whole membranesM188-777 + D-fucose 10.80 100M188-777 8.00 100

Urea pelletM188-777 + D-fucose 2.41 22M188-777 2.64 33

Urea solubleM188-777 + D-fucose 4.11 38M188-777 5.41 68

aTotal recovery for M188-777 plus D-fucose was60%; for M188-777 it was 101%.

VOL. 29, 1975 407

on August 18, 2019 by guest

http://aem.asm

.org/D

ownloaded from

JOHNSON, SILHAVY, AND BOOS

work of Moldow et al. (12) has indicated thatdialysis against urea apparently does not pre-cipitate a particular protein, but that the pre-cipitated protein in the urea dialysate is arepresentative sample of the proteins remainingin the sample. By analysis with one-dimen-sional polyacrylamide gel electrophoresis in thepresence of SDS of all three fractions mentionedin Table 2, we came to the same conclusion.Therefore, it seemed justified to use the proteinsfinally solubilized in 6 M urea as the majorrepresentatives of the proteins found in the cellenvelope, even though the selective loss of someproteins cannot be excluded.Two-dimensional polyacrylamide gel elec-

trophoresis. The three figures show stained(Coomassie brilliant blue) gel slabs obtained bytwo-dimensional analysis of envelope proteinsof typical E. coli strains. The electrophoresistechnique is essentially that described by Metzand Bogorad in their analysis of ribosomalproteins. Separation of the proteins in the firstdimension (from left to right) in 8 M urea occursprincipally according to the intrinsic charge ofthe denatured proteins at pH 8.5, whereas in thesecond dimension (top to bottom) the separa-tion due to 0.2% SDS occurs mainly accordingto their molecular weight (22). Numerous pro-teins of the same molecular weight that wouldnot be separated by the classical one-dimen-sional SDS-gel electrophoresis can now easily bedetected individually. This is particularly ap-parent for the major protein of the outer mem-brane of E. coli and other gram-negative bacte-ria, which appears to be one protein band ofapproximately 42,000 daltons on the usual SDS-gel electrophoresis (17) but has been shownrecently to consist of at least four differentcomponents (18).

Application of two-dimensional polyacryl-amide gel electrophoresis. It has recently beenshown by Schnaitman (18) that the major outermembrane component of E. coli can be sepa-rated into four distinct proteins of similarmolecular weight. The proteins of the K-12strain M188-777 solubilized from the outermembrane and prepared according to Schnait-man's procedure are shown in Fig. 1. As can beseen, the major protein components consist ofmore than four protein spots of very similarmolecular weight. This clearly demonstrates theadvantage of using a two-dimensional electro-phoretic technique where the separating featureis electrical charge in one dimension and molec-ular weight in the second dimension. The num-ber of the major outer membrane proteins varies

with different strains and growth conditions andmight be as high as 18 (see Fig. 3A, B).Membrane alterations can be detected in a

particular strain after alteration of the growthconditions and by introducing a mutation af-fecting the fl-methylgalactoside transport sys-tem. Figures 2A and 2B illustrate the proteinpattern of membranes prepared from E. coliK-12 strain M188-777 grown in both the pres-ence and absence of D-fucose. This sugar isknown to induce the cytoplasmic enzymes of thegalactose operon as well as the f3-methylgalacto-side transport system. As can be seen by com-paring Fig. 2A with Fig. 2B, D-fucose causes theincrease or new appearance of four to fivecomponents (numbered). Moreover, the proteinpattern of strain M188-1000, an mglABC (14)mutant isogenic with the strain shown in Fig.2A and 2B and grown in the presence ofD-fucose, similarly fails to exhibit or show areduced amount of the proteins seen to beinducible by fucose in the wild-type strainM188-777 (Fig. 2C). Thus, it seems likely thatsome of the proteins appearing in the wild-typestrain after induction with D-fucose representthe hitherto unknown components of the fl-methylgalactoside transport system. To corre-late the particular spots with components ofthis transport system and to avoid the compli-cation caused by the D-fucose-dependent induc-tion of unrelated proteins, it will be necessary touse a constitutive (mglR) strain and introducedeletions or polar mutations into the knownstructural genes of mglA, B, and C (14). Onlythen will it be possible to coordinate the proteinpattern obtained to the yet unknown transportcomponents.

Figures 3A and 3B show the protein patternobtained from the envelope of strain BUG-6, atemperature-sensitive mutant in cell division(septum formation) (16). The cells were grownat the permissive (Fig. 3A) and the nonpermis-sive (Fig. 3B) temperature. Several interestingfeatures can be noted. (i) The envelope containsabout 18 protein spots which migrate to amolecular weight position similar to that of themajor outer membrane protein (see Fig. 1). (ii)One of the spots (designated 1) found to beinducible by D-fucose in strains M188-777 ap-pears to be temperature sensitive. It had previ-ously been shown (19) that the synthesis of the,B-methylgalactoside transport system and oneof its components, the galactose-binding proteinin this strain, is temperature sensitive. (iii) Incontrast to the large changes occurring in theprotein composition of the "periplasmic compo-

408 APPL. MICROBIOL.

on August 18, 2019 by guest

http://aem.asm

.org/D

ownloaded from

ELECTROPHORESIS OF E. COLI ENVELOPE PROTEINS

FIG. 1. Two-dimensional polyacrylamide gel electrophoresis of outer membrane proteins. First dimensionwas (left to right) separation in 8Murea, pH 8.5, for 4.5 h at 350 to 400 V; second dimension (top to bottom) wasseparation in 0.2%o SDS, pH 6.48, for 17 h at 32 V. The separating slab gel contained 10% acrylamide, 0.5%bisacrylamide, and the cementing gel contained 4% acrylamide. The marker proteins, molecular weightindicated on the left, are (from the top) bovine serum albumin, ovalbumin, chymotrypsinogen, andhemerythrin. The gel slab is stained with Coomassie blue. The outer membranes of strain M188-777 wereprepared according to Schnaitman (18) and were applied in 6 M urea (see ref. 10).

nents" upon temperature shift (19), the enve-lope fraction of this strain changes less dramati-cally. Several spots of lower molecular weightare present at the nonpermissive temperature,but not at the permissive temperature; only afew spots appear to be temperature sensitive.

DISCUSSION

The rationale in using the present electropho-retic system was to try to separate the mem-brane proteins with respect to more than oneparameter sequentially, thus at least in partreducing the probability of multifold spots. Twoobvious parameters were intrinsic charge in onedimension and molecular weight in the other.

(i) The procedure for membrane preparationhad to be simple and fast to enable the analysisof a great number of mutant strains. Themethod of choice was French pressing andwashing by centrifugation at relatively lowcentrifugal forces of 30,000 x g, thus avoidingtime-consuming ultracentrifugation. This

might result in a low yield of membranes,particularly of inner cytoplasmic membrane.However, sonication followed by repeated ultra-centrifugation at 200,000 x g gave qualitativelysimilar results. More washings were required toremove cytoplasmic constituents as measuredby fl-galactosidase activity. (ii) Detergents suchas SDS or Triton X-100 commonly used forsolubilization of membrane proteins greatly in-terfere with separation according to charge in amilieu of 8 M urea and are difficult to remove.Therefore, the membranes had to be solubilizedin such a fashion that the dissolving agent couldeasily be removed by dialysis against urea, themedium for separation in the first dimension.With this concept in mind, we employed the

basic procedure of Moldow et al. (12), using thechaotropic agent guanidine thiocyanate to ini-tially solubilize the membranes. The reagentwas removed by dialysis against 6 M urea,which left the majority of the proteins in solu-tion and ready for electrophoresis in 8 M urea.The method of Metz and Bogorad (10) estab-lished that the urea gels could then be used for

VOL. 29, 1975 409

on August 18, 2019 by guest

http://aem.asm

.org/D

ownloaded from

410 JOHNSON, SILHAVY, AND BOOS APPL. MICROBIOL.

4.iE it-'* .X.X-. t) tsu5,,s,0 t-FiVi- 00: S S i S S E | l | I | i _'iiC- *- j,3rA l l 1111 11i1 I l11-_ZaRl R R | | | l l ! ! R | .>stis,2

*00 *0' w:00+ - ! ! - J s v w, 1 V t! |s;s

\E 40.'4 '$ ..5iicy tsig-i+& -ESt.ts .Zl3Lt li5S8<>Zt5i tst: , /J.ws,

0 '. ,',A,<;X<XXXWf9gf w X r C,.j0vSX: z g'-, .: :,. w:W. id

- X, <ES,X 0 z t 0 ' . 8 A A7 . S i . ii 'i . b t1, i

,''S,;a{S.0. g}z.isgiJi.f E

., S s :: XWikX 0 a A j 0 Nbas ! w w_'siS+ka<.<-0-

,. M.--D ut. 'a: F:X, ^,;zt

..iR Ai.-d 0000 S 0 X f f,,. . riii. ........................ K.¢; aiee99+i+a=R*; 0 * 0 0 a- §, 0 00 A R t s se <. ,2;1

tV S. fS-bi- ;1;.-s E|0 ,.tAi.,S i,-

4*';-.8Z<t'':l tiN

0; + 6-'5#t j:iji 5ii 'i .' ' Cl#aiOW !vii0, $,s,,rrg ;si

0, ., ,05 .<.'0| | Ef *'@I;.*i0giis$ >0 < AD- 00 ia 5 ,Sn;

; -q>3 ii .;g9913 06, 5 D-8|:}s748S i35i

,,f, >S,$,,@ S; ili50 X . ^ /Z S L-:: lisif

;., ;' XiS > dS W: ; >; JiS ns .: ii'g s2.-f5k2.8_

< ;= ri

S.,")6H''SS S-005 Wi <&-s0_i_ .\.

it.X,FCkk20 >, , w z i | r, 0|S'S00"ii.S;'2''t u ;'.. W.X_ s-.

_0 " t-X- _ gE

P <gN8''l+"

_

.' t _ t 't' 0-..:W;C-<

* e ;0 ' X i ' ^ .g 0 . 7<tL w<jf=|, 'YX

E .3 ^'2

g-g'; 1X0,;3t0| sH sW-9'0u&;lL 2tj2t E5 00

t1 00y00 A0. :.__ __RXt.^ 005si 02si 0 5 0. stie E ^ S * w . j 0 0r K.,,ziS .d,0 V 0 g di , . , 0 C <___L dii AvSj E .. . -i-,_E* i >, 0S _ , Fo;.

0>-_:.0isvs0- 0; i0r 8_. aq, >-

S3Siil>d w

ftMS' f ) n lii0' <\. 4 P 3S.¢ >as k , , ' , i a _ __ _ _.l-BS: _Nf :Z flD

Sl? t-:XfiSCtSs_.xa.k*._ _ __

'- ___.s lila -t 3gy<___*0- 0 0 0 i'2- tff E XSf & 0 . i . , 0 . :0 0 0 0 '--',1'. r.' _i

'iS 4FW, We'"'l'X, ib _ iu'., s<.iis .> >t;.lgt E;<,3"X. _s; s . X ( i5i.=i . a. . 0 Li ' . 0 . g . .;.;FIG. 2. Two-dimensional polyacrylamide gel electrophoresis of total envelope proteins. (A) Strain M188-777

grown in the absence of D-fucose; (B) strain M188-777 grown in the presence of D-fucose, inducer of the f3-methylgalactoside transport system; (C) strain M188-1000 defective in mglABC but otherwise isogenic withM188-777. The mutant M188-1000 was grown in the presence of D-fucose. Details as in legend to Fig. 1.

on August 18, 2019 by guest

http://aem.asm

.org/D

ownloaded from

ELECTROPHORESIS OF E. COLI ENVELOPE PROTEINS

,r, d t w ,_ ,,~~~~~~

*<0Fff,8<99, *;R i

FIG. 2C

SDS-slab gel electrophoresis with no interfer-ence by the remaining urea and resulting inseparation according to molecular weight.The actual electrophoresis technique pre-

sented no problems with respect to good migra-tion in both dimensions. The major drawback inthe procedure was protein loss during dialysisagainst urea prior to electrophoresis. However,analysis of the proteins contained in the severalsteps of the solubilization procedure by conven-tional SDS electrophoresis shows no selectiveloss. The possibility of thiocarbamylation bythe thiocyanate anion of primary or secondaryamines present in the membrane protein duringsolubilization apparently presents little prob-lem. Moldow et al. (12) found 1 nmol of[14C ]-labeled thiocyanate bound per 590 nmol ofamino acid, or 0.17%. It is doubtful that thissmall amount would be stained. We thereforebelieve that the method as presented here isaccurate and gives a better resolution of mem-brane protein than conventional tube urea orSDS electrophoresis alone.The advantageous use of this two-dimen-

sional polyacrylamide gel electrophoresis was

demonstrated by the following examples. (i)Analysis of outer membrane proteins revealedmore major components of similar molecularweight than had been reported before (18). (ii)The analysis of the total envelope protein ofstrains induced and uninduced for a particularsugar transport system showed alteration of fourto five proteins, which might possibly containthe hitherto unknown components of the trans-port machinery. Based on genetic data, theexistence of at least three components had beenpostulated (Boris Rotman [personalcommunication] has further evidence that themglA and C gene products are necessary fortranslocation of substrate) (14). The electro-phoretic analysis of the envelope proteins ofclearly defined polar mutants or deletions willanswer this question. (iii) The analysis of theenvelope proteins of a strain, temperature sen-sitive in cell division, revealed some alterationsin the protein pattern. One of the spots missingwhen grown at the nonpermissive temperaturewas found to be fucose inducible in the wildtype. This correlates with a previous observa-tion that this mutant does not synthesize an in-

VOL. 29, 1975 411

on August 18, 2019 by guest

http://aem.asm

.org/D

ownloaded from

JOHNSON, SILHAVY, AND BOOS

FIG. 3. Two-dimensional polyacrylamide gel electrophoresis of total envelope proteins of BUG-6, tempera-ture sensitive in cell division. Cells were grown at 35 (A) and 42 C (B). Details as in legend to Fig. 1.

412 APPL. MICROBIOL.

on August 18, 2019 by guest

http://aem.asm

.org/D

ownloaded from

ELECTROPHORESIS OF E. COLI ENVELOPE PROTEINS

tact fl-methylgalactoside transport system andgalactose-binding protein when grown at thenonpermissive temperature.

ACKNOWLEDGMENTS

We wish to thank Herman M. Kalckar for his hospitalityand encouragement and Alap Subramanian for helpful sug-gestions. Also we are grateful to Jane Parnes, who partici-pated in the initial stages of this project.

This work was supported by Public Health Service grantGM-18498 from the National Institute of General MedicalSciences and a grant fromthe Milton Fund. One of us (W.B)was the recipient of the Solomon A. Berson Research andDevelopment Award of the American Diabetes Association.

LITERATURE CITED

1. Ames, Giovanna Ferro-Luzzi. 1974. Resolution of bacte-rial proteins by polyacrylamide gel electrophoresis onslabs. Membrane, soluble, and periplasmic fractions. J.Biol. Chem. 249:634-644.

2. Boos, W. 1974. Pro and contra carrier proteins; sugartransport via the periplasmic galactose-binding pro-tein. Curr. Top. Membr. Transp. 5:51-136.

3. Bragg, P. D., and C. Hou. 1972. Organization of proteinsin the native and reformed outer membrane of E. coli.Biochim. Biophys. Acta 274:478-488.

4. Fairbanks, G., T. L. Steck, and D. F. H. Wallach. 1971.Electrophoretic analysis of the major polypeptides ofthe human erythrocyte membrane. Biochemistry10:2606-2617.

5. Inouye, M., and M.-L. Yee. 1973. Homogeneity of enve-lope proteins of Escherichia coli separated by gelelectrophoresis in sodium dodecyl sulfate. J. Bacteriol.113:304-312.

6. Kaltschmidt, E., and H. G. Wittmann. 1970. Ribosomalproteins. VII. Two-dimensional polyacrylamide gelelectrophoresis for fingerprinting of ribosomal proteins.Anal. Biochem. 36:401-412.

7. Laemmli, V. K. 1970. Cleavage of structural proteinsduring the assembly of the head of bacteriophage T4.Nature (London) 227:680-685.

8. Lengeler, J., K. 0. Herman, H. J. Unsold, and W. Boos.1971. The regulation of the P-methylgalactoside trans-port system and of the galactose-binding protein ofEscherichia coli K12. Eur. J. Biochem. 19:457-470.

9. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1951. Protein measurement with the Folinphenol reagent. J. Biol. Chem. 193:265-275.

10. Metz, L. J., and L. Bogorad. 1974. Two-dimensional

polyacrylamide gel electrophoresis: an improvedmethod for ribosomal proteins. Anal. Biochem.57:200-210.

11. Miller, J. H. 1972. Experiments in molecular genetics.Cold Spring Harbor Laboratory, Cold Spring Harbor,N.Y.

12. Moldow, C., J. Robertson, and L. Rothfield. 1972. Purifi-cation of bacterial membrane proteins. The use ofguanidinium thiocyanate and urea. J. Membr. Biol.10:137-152.

13. Neu, H. C., and L. A. Heppel. 1965. The release ofenzymes from Escherichia coli by osmotic shock andduring the formation of spheroplasts. J. Biol. Chem.240:3685-3692.

14. Ordal, G. W., and J. Adler. 1974. Isolation and com-plementation of mutants in galactose taxis and trans-port. J. Bacteriol. 117:509-516.

15. Parnes,-J. R., and W. Boos. 1973. Unidirectional trans-port activity of the ft-methylgalactoside transport sys-tem. J. Biol. Chem. 248:4436-4445.

16. Reeve, J. N., D. J. Groves, and D. J. Clark. 1970.Regulation of cell division in Escherichia coli: charac-terization of temperature-sensitive division mutants. J.Bacteriol. 104:1052-1064.

17. Schnaitman, C. A. 1970. Protein composition of the cellwall and cytoplasmic membrane of Escherichia coli. J.Bacteriol. 104:890-901.

18. Schnaitman, C. A. 1974. Outer membrane proteins ofEscherichia coli. HI. Evidence that the major protein ofEscherichia coli 0111 outer membrane consists of fourdistinct polypeptide species. J. Bacteriol. 118:442-453.

19. Shen, B. H. P., and W. Boos. 1973. Regulation of theB-methylgalactoside transport system and the galac-tose-binding protein by the cell cycle of Escherichiacoli. Proc. Natl. Acad. Sci. U.S.A. 70:1481-1485.

20. Silhavy, T. J., and W. Boos. 1974. Selection procedure formutants in the fi-methylgalactoside transport systemof Escherichia coli utilizing the compound 2R-glyceryl-f-D-galactopyranoside. J. Bacteriol. 120:424-432.

21. Subramanian, A. R. 1974. Sensitive separation procedurefor Escherichia coli ribosomal proteins and the resolu-tion of high-molecular-weight components. Eur. J.Biochem. 45:541-546.

22. Weber, K., and M. Osborn. 1969. The reliability ofmolecular weight determination by dodecyl sulfate-polyacrylamide gel electrophoresis. J. Biol. Chem.244:4406-4412.

23. Wu, H. C. P., W. Boos, and H. M. Kalckar. 1969. Role ofthe galactose transport system in the retention ofintracellular galactose in Escherichia coli. J. Mol. Biol.41:109-120.

VOL. 29, 1975 413

on August 18, 2019 by guest

http://aem.asm

.org/D

ownloaded from