Galactokinase Is a Novel Modi er of Calcineurin-Induced ... · wing vein phenotype induced by the...

INVESTIGATION

Galactokinase Is a Novel Modifier ofCalcineurin-Induced Cardiomyopathy in Drosophila

Teresa E. Lee,* Lin Yu,† Matthew J. Wolf,†,‡ and Howard A. Rockman*,†,‡,1

*Department of Cell Biology, †Department of Medicine, and ‡Department of Molecular Genetics and Microbiology,Duke University, Durham, North Carolina 27710

ABSTRACT Activated/uninhibited calcineurin is both necessary and sufficient to induce cardiac hypertrophy, a condition that oftenleads to dilated cardiomyopathy, heart failure, and sudden cardiac death. We expressed constitutively active calcineurin in the adultheart of Drosophila melanogaster and identified enlarged cardiac chamber dimensions and reduced cardiac contractility. In addition,expressing constitutively active calcineurin in the fly heart using the Gal4/UAS system induced an increase in heart wall thickness. Weperformed a targeted genetic screen for modifiers of calcineurin-induced cardiac enlargement based on previous calcineurin studies inthe fly and identified galactokinase as a novel modifier of calcineurin-induced cardiomyopathy. Genomic deficiencies spanning thegalactokinase locus, transposable elements that disrupt galactokinase, and cardiac-specific RNAi knockdown of galactokinase sup-pressed constitutively active calcineurin-induced cardiomyopathy. In addition, in flies expressing constitutively active calcineurin usingthe Gal4/UAS system, a transposable element in galactokinase suppressed the increase in heart wall thickness. Finally, geneticdisruption of galactokinase suppressed calcineurin-induced wing vein abnormalities. Collectively, we generated a model for discoveringnovel modifiers of calcineurin-induced cardiac enlargement in the fly and identified galactokinase as a previously unknown regulator ofcalcineurin-induced cardiomyopathy in adult Drosophila.

ACTIVATED/uninhibited calcineurin is both necessaryand sufficient to induce cardiac hypertrophy (Molkentin

et al. 1998; Wilkins and Molkentin 2002; Van Berlo et al.2013). Transgenic mice expressing constitutively active cal-cineurin (CanAact) display cardiac hypertrophy (Molkentinet al. 1998) and the genetic or pharmacological inhibitionof calcineurin suppresses agonist and pressure overload-induced cardiac hypertrophy (Sussman et al. 1998; Taigenet al. 2000; Wilkins and Molkentin 2002; Van Berlo et al.2013). Prolonged cardiac hypertrophy is a known risk factorfor dilated cardiomyopathy, heart failure, and sudden death(Levy et al. 1990; Messerli and Ketelhut 1991; Drazner et al.2004; George 2013; Grossman and Paulus 2013). In con-trast, cardiac hypertrophy stimulated by exercise is physio-logical, is not typically associated with abnormal cardiac

function, and does not stimulate calcineurin/nuclear factorof activated T cells (NFAT) signaling (Wilkins et al. 2004),supporting the concept that calcineurin promotes patholog-ical cardiac hypertrophy.

Calcineurin acts as a calcium/calmodulin-dependent pro-tein phosphatase that consists of two subunits: a large CanAsubunit (60 kDa) and a small CanB subunit (19 kDa). In themouse, there are three CanA genes (Ppp3ca, Ppp3cb, andPpp3cc) and two CanB genes (Ppp3r1 and Ppp3r2); in thefly, there are three CanA genes (CanA1, CanA-14F, andPp2B-14D) and two CanB genes (CanB and CanB2) (NCBIGene, http://www.ncbi.nlm.nih.gov/gene). The large CanAsubunit has phosphatase activity and consists of severaldomains: the catalytic domain, which regulates protein de-phosphorylation (Klee et al. 1979); the CanB binding do-main (Klee et al. 1988); the calcium/calmodulin bindingdomain; and the autoinhibitory domain (Shibasaki et al.2002). In the inactive state, the autoinhibitory domaininhibits the catalytic domain. Binding of calcium/calmodulinactivates calcineurin by alleviating this autoinhibition. Aconstitutively active calcineurin (CanAact) is generated byeliminating the autoinhibitory domain and has been usedto investigate calcineurin signaling (Molkentin et al. 1998;

Copyright © 2014 by the Genetics Society of Americadoi: 10.1534/genetics.114.166777Manuscript received June 17, 2014; accepted for publication July 17, 2014; publishedEarly Online July 31, 2014.Supporting information is available online at http://www.genetics.org/lookup/suppl/doi:10.1534/genetics.114.166777/-/DC1.1Corresponding author: Department of Medicine, Duke University Medical Center,DUMC 3104, 226 CARL Bldg., Research Dr., Durham, NC 27710.E-mail: [email protected]

Genetics, Vol. 198, 591–603 October 2014 591

http://flybase.org/reports/FBgn0010015.html





http://www.ncbi.nlm.nih.gov/gene

http://www.genetics.org/lookup/suppl/doi:10.1534/genetics.114.166777/-/DC1


mailto:[email protected]

Sullivan and Rubin 2002; Gajewski et al. 2003). Expressionof CanAact has been found to induce cardiac hypertrophy(Molkentin et al. 1998), skeletal muscle hypertrophy (Musaroet al. 1999; Semsarian et al. 1999), and slow twitch skeletalmuscle specification (Chin et al. 1998; Wu et al. 2000). Thesmaller subunit (CanB) is constitutively bound to CanA andis required for maintaining calcineurin expression (Watanabeet al. 1995; Klee et al. 1998; Parsons et al. 2004). Defi-ciency of CanB results in significant cardiomyopathy, includ-ing impaired cardiomyocyte growth, impaired contractility,and lethality after birth (Schaeffer et al. 2009; Maillet et al.2010).

Calcineurin signaling in mammals involves calcineurin-dependent dephosphorylation of NFAT transcription factors(Molkentin et al. 1998; Okamura et al. 2000). In contrast,Drosophila do not have calcineurin-regulated isoforms ofNFAT and therefore use NFAT-independent pathways (Keyseret al. 2007). Myocyte enhancer factor (Mef2) is a well-knownNFAT-independent pathway that has been implicated incalcineurin-mediated cardiac and skeletal muscle hypertro-phy (Wilkins and Molkentin 2002; Sakuma and Yamaguchi2010). Cardiac calcineurin expression has been shown toactivate Mef2 reporter activity (Passier et al. 2000). The ex-pression of a dominant-negative Mef2 inhibited CanAact-induced cardiac enlargement and overexpressing Mef2caused cardiac chamber dilation (Van Oort et al. 2006).In skeletal muscle, calcineurin activates Mef2 with exercise(Wu et al. 2001; Sakuma et al. 2008). Mechanistically, cal-cineurin was found to co-immunoprecipitate with Mef2 andinduce activation of Mef2 through dephosphorylation (Wuet al. 2001).

Previously, two independent screens were conducted toidentify modifiers of calcineurin phenotypes in tissues otherthan the heart (Sullivan and Rubin 2002; Gajewski et al.2003). Sullivan and Rubin (2002) performed a dominantmodifier screen in the Drosophila eye and found five sup-pressor and four enhancer loci. Two modifier genes, CanB2and sprouty, were identified. However, modifier genes withinthe seven other broadly mapped loci remained uncharacter-ized. Gajewski et al. (2003) found seven different deletionintervals that suppressed the lethal phenotype of constitu-tively active calcineurin driven by the general mesodermaldriver 24B. CanB2 was determined to be a modifier andpreliminary experiments suggested that Mef2 might be themodifier for another interval. Importantly, only one intervaloverlapped between these two studies on chromosome 3L,cytolocation 66F.

Many pathways are conserved in mammalian and Dro-sophila cardiac development (Bodmer and Venkatesh 1998;Cripps and Olson 2002; Zaffran and Frasch 2002; Zaffranet al. 2002). In fact, strategies based on fly genetics havebeen used to identify genes that cause or modify cardiomy-opathies (Bier and Bodmer 2004; Wessells and Bodmer2004; Wolf et al. 2006; Yu et al. 2010). The Drosophila heartcan be efficiently monitored in real time in intact awakeDrosophila, using optical coherence tomography (OCT) (Wolf

et al. 2006). Importantly, Drosophila is well adapted for ge-netic studies: it has a relatively short generation time, well-developed genetic manipulation methods, and well-developedgenetic resources, including mutation stocks and vectors formaking transgenics (Kohler 1994; St. Pierre et al. 2014).Therefore, we conducted studies using fly genetics to iden-tify novel modifiers of cardiac calcineurin.

We generated Drosophila that expressed CanAact in theheart under the control of the cardiomyocyte-specific driver,tinC (tinC-CanAact), an approach that is analogous to that inprevious studies of transgenic mice (Molkentin et al. 1998).tinC-CanAact flies had enlarged cardiac chamber dimensionsand reduced cardiac contractility. In addition, expressingCanAact in the heart with tinC-Gal4 . UAS-CanAact inducedan increase in cardiac wall thickness. Based on the priorgenetic screens (Sullivan and Rubin 2002; Gajewski et al.2003), we then designed a deficiency screen targeting theoverlapping region on chromosome 3L. Here we show thatdeficiencies in a genetic region encoding galactokinase(Galk) rescued CanAact-induced cardiac enlargement anddecreased function. To test the hypothesis that Galk isa novel modifier of calcineurin in the heart, two indepen-dent transposable element insertions in Galk and cardiac-specific expression of RNAi directed against Galkwere studied.Genetic disruption of Galk rescued the cardiac enlarge-ment phenotype in tinC-CanAact flies and a transposableelement insertion in Galk rescued tinC-Gal4 . UAS-CanAact-induced increase in cardiac wall thickness. In addition,genetic disruption of Galk also suppressed an abnormalwing vein phenotype induced by the wing driver e16E-Gal4 . UAS-YCanAact. These findings suggest that galacto-kinase modifies calcineurin-mediated cardiomyopathy inadult Drosophila.

Materials and Methods

Drosophila stocks

The following Drosophila stocks were obtained from theBloomington Drosophila Stock Center: w1118 (Flybase ID:FBst0003605), Df(3L)ED4413, Df(3L)ED4414, Df(3L)ED4415,Df(3L)ED4416, Df(3L)ED4421, Df(3L)BSC130, Df(3L)BSC170,Df(3L)BSC390, Df(2R)X1,Mef2X1/CyOAdhnB, styΔ5/TM3,P{35UZ}2,P{EP}CanB2EP774, PBac{PB}Galkc03848, Mi{ET1}GalkMB10638,P{hsILMiT}2.4, P{en2.4-Gal4}e16E (e16E-Gal4), P{GAL4-dpp.blk1}40C.6 (dpp-Gal4), P{Act5C-GAL4}25FO1 (Act5C-Gal4),and P{GAL4-Mef2.R}3 (Mef2-Gal4). The double balancerline WR135 was kindly provided by Robin Wharton (OhioState University, Columbus, OH). The P{tinC-Gal4} line waskindly provided by Manfred Frasch (Mount Sinai MedicalSchool, New York, NY) (Yin and Frasch 1998). The P{tinC-GFP}(tinC-GFP) line was generated as previously described byinserting the 304-bp tinC genetic sequence into the pGreen-H-Pelican vector (Yin and Frasch 1998; Barolo et al. 2000; Yuet al. 2010) and the construct was injected at the DukeUniversity Model Systems Genomics Facility. The tinC-YCanAact,

592 T. E. Lee et al.










http://flybase.org/reports/FBal0018186.html

http://flybase.org/reports/FBab0035662.html








http://flybase.org/reports/FBti0007694.html





tinC-FCanAact, and UAS-YCanAact Drosophila lines were gener-ated at the Duke University Model Systems Genomics Facilityby injecting the corresponding constructs into Drosophilaembryos.

Calcineurin constructs

Constitutively active calcineurin was amplified from thecalcineurin gene Pp2B-14D as previously described, usingthe primers ATG TCT TCG AAT AAC CAG AGC AGC AG(forward) and TCA GTT GCG TAT CAC CTC CTT GCG CA(reverse) (Sullivan and Rubin 2002). Restriction enzymesites and Flag-tagged CanAact (FCanAact) were inserted byextending the N terminus with primers carrying the corre-sponding sequences. CanAact was amplified in topo vector(Invitrogen, Carlsbad, CA) and inserted at the 39 end toyellow fluorescent protein (YFP) in the pEYFP-C1 vector withappropriate restriction enzyme sites to form YFP-taggedCanAact (YCanAact). Full-length FCanAact or YCanAact PCRproducts were subsequently cloned into the pCaSpeR5Drosophila expression vector, inserting a tinC-hsp70 pro-moter at the 59 end as previously described (Yin et al.1997; Yu et al. 2010) (Figure 1A). The catalytic activity ofN-terminal YFP-tagged CanAact has been confirmed in stud-ies by a number of investigators, using well-characterizedNFAT reporter assays (Tokoyoda et al. 2000; Burkard et al.2005), NFAT phosphorylation assays (Tokoyoda et al. 2000),and NF-kB reporter assays (Kang et al. 2007). The UAS-YCanAact construct was generated similarly by insertingYFP-tagged CanAact into the pUAST vector following theUAS sequence.

OCT to measure cardiac function in adult Drosophila

End-diastolic dimension and end-systolic dimension in adultDrosophila were measured using OCT (Bioptigen, Durham,NC) as previously described (Wolf et al. 2006). Briefly, 7–10days posteclosion, female Drosophila were placed in GelWaxmedium and allowed to awaken. M-mode images throughthe conical chamber were collected for immobilized awakeDrosophila. End-diastolic and end-systolic dimensions (EDDand ESD, respectively) were measured in ImageJ, calibratedto a 125-mm-thick glass slide. Fractional shortening (FS)is calculated as (end-diastolic dimension 2 end-systolicdimension/end-diastolic dimension) 3 100% and used asa measure of cardiac contractility. All end-systolic dimen-sions are supplied in Supporting Information, Table S1,File S1.

Several types of controls were used and compared for ourstudy: w1118 (Flybase ID: FBst0003605; EDD = 54.10 62.95 mm; FS = 98.79 6 1.21%), w1118 used to create trans-genics from the Duke University Model Systems GeneticsFacility (Flybase ID: FBst0006326; EDD = 41.53 6 5.08mm; FS = 99.76 6 1.21%), tinC-Gal4 (EDD = 51.59 64.14 mm; FS = 96.70 6 1.53%), tinC-Gal4 heterozygouswith w1118 (EDD = 48.34 6 2.34 mm; FS = 98.92 60.74%), tinC-GFP (EDD = 52.88 6 5.85 mm; FS = 98.24 61.76%), and tinC-GFP heterozygous with w1118 (EDD =

56.34 6 4.32 mm; FS = 98.78 6 0.58%). We performeda one-way ANOVA comparing OCT measurements betweenall control groups and show that none were significantlydifferent from each other (Figure S1A).

Histology

Histology was performed according to standard paraffinembedding and hematoxylin and eosin staining proceduresfor the fly (Ashburner 1989). Briefly, 3–5 days posteclosion,Drosophila were washed with 70% EtOH before fixing in10% buffered formalin at 4� overnight. This fixes the heartsat their most relaxed state, end-diastole. To assess this, OCTimages were measured and compared between alive w1118

end-diastolic dimension (46.37 6 4.91 mm, N = 9) and thesize of the heart after EtOH fixation (50.19 6 6.09 mm,N = 9), P = NS, Student’s t-test. The next day, Drosophilawere dehydrated starting with a PBS wash and subsequentalcohol gradient into xylene. Drosophila were then incu-bated overnight at 60� under vacuum in liquid paraffin wax.Samples were then positioned appropriately in molds andallowed to harden. Consecutive sections (8 mm) were cutusing a microtome and adhered to poly-L-lysine-coatedglass slides. The slides were subsequently rehydrated intowater and stained with hematoxylin and eosin. The slideswere brought back through the alcohol gradient and xyleneand mounted in Cytoseal XYL mounting medium (ThermoFisher Scientific). Images were subsequently quantified inImageJ, calibrated with a hemocytometer grid measuring50 mm. Locating the position along the cardiac tube formeasurement was determined as previously described byexamining the portion of the sections where an en facesection of the cardiac tube was visible and measuring threesections (24 mm) posterior to this section (Yu et al. 2010).Wall thickness was calculated as the average of dorsal,ventral, left, and right walls, excluding the dorsal longitu-dinal muscle underlying the ventral side of the conicalchamber.

Minos excision

The Minos insertion in Galk, Mi{ET1}GalkMB10638, was ex-cised precisely according to standard transposable ele-ment excision procedures (Arca et al. 1997; Metaxakis et al.2005). Briefly, the Mi{ET1}GalkMB10638 males were crossedto virgin females containing the Minos transposase P{hsIL-MiT}2.4. After 3 days, adult flies were removed, and theembryos were heat-shocked in a 37� water bath for 1 hrfor 4 consecutive days. Male progeny were selected for thepresence of the Minos insertion and the Minos transposaseaccording to eye color (the Minos element expresses GFP,and Minos transposase expresses a red eye color from mini-white in the insertion) and crossed to the double balancer flystock WR135 (Sp/CyO;TM2/MKRS). Male progeny were se-lected for the absence of GFP and mini-white, indicatinga successful excision, and crossed again to WR135 virginfemales. Progeny harboring excisions were crossed to eachother to make a homozygous stock. These stocks were assayed

Galk Modifies CanA Cardiomyopathy 593


http://www.genetics.org/lookup/suppl/doi:10.1534/genetics.114.166777/-/DC1/genetics.114.166777-1.pdf














for the presence of a precise excision, using primers sequenc-ing through the affected genomic region.

qRT-PCR

Drosophila were collected 3–5 days after eclosion and fivewhole flies were collected or 10–15 fly hearts were dissectedfor each group. RNA was extracted using RNA-Bee RNA iso-lation reagent (Amsbio) according to the manufacturer’sprotocol. SuperScript II (Invitrogen) reverse transcriptionaccording to the manufacturer’s protocol was followed byreal-time PCR analysis, using a predesigned taqman probefor Galk (probe Dm01801608_g1; Applied Biosystems, FosterCity, CA).

Tissue-specific phenotypes

For wing vein phenotypes, Drosophila from the respectivecrosses were collected at 3–5 days after eclosion. The wingswere detached using forceps and the ventral surface wasplaced face down on GelWax plates. The wings were exam-ined for abnormalities under a dissection microscope at 403or imaged under a Leica M165FC fluorescence stereo micro-scope equipped with a Leica DFC310FX camera at 503. Arange of wing vein abnormalities were observed, whichwere divided into normal, abnormalities of the posteriorcrossvein (PCV), or abnormalities of both the PCV and thelongitudinal vein 5 (L5). The number of wings under eachcategory of wing vein phenotype was counted for eachgroup and the percentage of total wings counted was calcu-lated. Statistical significance to detect for a rescue of theabnormal wing vein phenotype (pooling the two differenttypes of wing vein abnormalities) with the Df(3L)ED4416deficiency or Mi{ET1}GalkMB10638 insertion was determinedusing Fisher’s exact test. Tissue-specific expression of CanAact

with dpp-Gal4, Act5C-Gal4, and Mef2-Gal4 was analyzed sim-ilarly by counting the number of progeny with the respectivephenotypes according to the crosses and genotypes describedin Figure 6.

Statistical analysis

All data were analyzed in GraphPad Prism 6 software. Student’st-test was performed for single comparisons and one-wayANOVA and post hoc analysis with Bonferroni correctionwere conducted for group comparisons. Fisher’s exact testwas employed to determine the rescue of the proportion offlies displaying wing or lethality phenotypes.

Results

Expression of CanAact in the adult fly heart causedcardiac enlargement

To identify genetic modifiers of cardiac calcineurin, we gen-erated two types of sensitized Drosophila lines expressingCanAact in the heart: (1) CanAact under the direct controlof the tinC promoter, a 304-bp intronic region within thetinman gene that controls cardiac-specific gene expressionwith a minimal hsp70 promoter (tinC-CanAact) (Yin et al.

1997; Lo and Frasch 2001; Wolf et al. 2006), tinC-CanAact

transgenic lines were generated under the direct control ofthe cardiac-specific tinC genomic fragment in the context ofan hsp70 minimal promoter to harbor either a flag- or YFP-tagged CanAact, FCanAact or YCanAact respectively, Figure1A; and (2) CanAact under control of the UAS promoter(UAS-CanAact) where cardiac-specific CanAact was expressedin Drosophila, using the Gal4/UAS system (tinC-Gal4. UAS-CanAact) (Brand and Perrimon 1993).

We performed measurements of cardiac chamber dimen-sions during complete relaxation (end-diastole) and full con-traction (end-systole), using OCT (Wolf et al. 2006); fractionalshortening was calculated as described inMaterials and Meth-ods and used as a measurement of cardiac contractility. Inagreement with previous studies, w1118 control Drosophiladisplayed close to 100% fractional shortening (Wolf et al.2006) (Figure 1B and Figure S1A). Drosophila expressingeither FCanAact or YCanAact displayed enlarged end-diastolicdimension and reduced fractional shortening (Figure 1B andFigure S1B). To account for possible positional effects of thetransgene, end-diastolic dimensions in several different lineswere tested. tinC-YCanAact or tinC-FCanAact insertions on thefirst, second, and third chromosomes all induced enlargedend-diastolic dimensions, suggesting that the transgene iscausing cardiac enlargement and the phenotype is not dueto disruption of an endogenous gene from transgenic in-sertion (Figure S1C). Cardiac enlargement persisted intinC-CanAact flies compared to controls over 5 weeks ofage (Figure S1D).

Histological analyses confirmed that the tinC-CanAact flyhearts were significantly larger than controls (Figure 1C).The heart wall was significantly thicker in tinC-Gal4 . UAS-CanAact but not in tinC-CanAact flies (Figure S1, E and F).This distinction may be due to augmented CanAact expres-sion with tinC-Gal4 . UAS-CanAact. Immunofluorescencestaining and confocal microscopy confirmed the cardiac-specific expression of tinC-CanAact (Figure S1G). To determinewhether the enlargement in heart size was due to increasein number of cells or increase in cell size, we examined thenumber of nuclei in tinC-CanAact and tinC-GFP control hearts.Confocal microscopy of tinC-CanAact hearts expressing cardiac-specific nuclear localized RFP showed the same number ofcardiomyocytes compared to controls, indicating that thephenotype was not the result of abnormalities in cardiomyo-cyte number (Figure S1H).

A region on chromosome 3L harbored novel modifiersof tinC-CanAact-mediated cardiac enlargement

To test the feasibility of finding a modifier using tinC-CanAact

as the sensitized line, we examined the effect of geneticdeficiency of the known downstream regulator of calcineurin,Mef2 (Van Oort et al. 2006). A deficiency in the known calci-neurin modifier Mef2 rescued tinC-CanAact-induced cardiacenlargement (Figure S2).

Previous screens identified that a region located at cytolo-cation 66F rescued 24B-CanAact-induced lethality (Gajewski

















et al. 2003) and corresponded to a suppressor and enhancerregion for GMR-CanAact-induced rough eye (Sullivan andRubin 2002). Therefore, we reasoned that a deficiencyscreen covering this region would likely identify a modifierof CanAact-induced cardiomyopathy. Genomic deficiency linessurrounding this region were initially examined for the abilityto suppress tinC-CanAact-mediated cardiac enlargement (Fig-ure 2A). Two lines, Df(3L)ED4416 and Df(3L)BSC130, sup-pressed tinC-CanAact-induced cardiac enlargement (Figure 2,B and C, and Figure S3A). Excluding the regions in which thedeficiencies did not rescue the phenotype, we narrowed theregion down to 80 kb that is predicted to encode 13 genes(Figure 2A). Of note, Df(3L)ED4421 spans the suppressorregion, but exerts a cardiac enlargement phenotype on itsown, indicating that genes within this large deficiency inducecardiac dilation independent of CanAact (Figure S3A). For thisreason, the Df(3L)ED4421 deficiency was not consideredwhen determining the suppressor region.

We validated the suppressor region corresponding toDf(3L)ED4416 and Df(3L)BSC130 by evaluating the cardiacphenotype in tinC-Gal4 . UAS-CanAact Drosophila in thecontext of these deficiencies. tinC-Gal4 . UAS-CanAact fliesdisplayed cardiac enlargement similar to that of tinC-CanAact

flies (Figure 3). Genetic disruption of the known calcineurinmodifiers Mef2, sty, or CanB2 rescued the fractional short-ening of tinC-Gal4 . UAS-CanAact flies (Figure 3 and FigureS3B), demonstrating the fidelity of using a sensitized Dro-sophila line to find modifiers of calcineurin. Cardiac enlarge-ment (end-diastolic dimension) of tinC-Gal4 . UAS-CanAact

hearts was rescued by Df(3L)ED4416 and Df(3L)BSC130(Figure 3). These results suggest that Df(3L)ED4416 and

Df(3L)BSC130 harbor modifiers of CanAact-induced cardiacenlargement.

Identification of galactokinase as a novel modifier ofcalcineurin-induced cardiac enlargement

The refined suppressor region contained a total of 13 genes(Figure 2A). Two genes, galactokinase (Galk) and arginine kinase(Argk), have high expression in the adult fly heart (flyatlas.org)(Chintapalli et al. 2007; Robinson et al. 2013). To determinewhetherGalk is a modifier of the tinC-CanAact cardiac phenotype,we tested two Drosophila lines that had transposable elementinsertions in Galk and decreased Galk transcript expression:PBac{PB}Galkc03848 and Mi{ET1}GalkMB10638. In the contextof tinC-CanAact, PBac{PB}Galkc03848 and Mi{ET1}GalkMB10638

rescued the tinC-CanAact-induced decrease in cardiac contrac-tility, and a precise excision of the Mi{ET1}GalkMB10638 inser-tion, Mi{ET1}Galkrev, reverted this rescue (Figure 4, A and B,and Figure S3C). Additionally, cardiac-specific expression ofRNAi directed against Galk also rescued tinC-CanAact-inducedcardiac enlargement (Figure 4, C and D). Galk expressionlevels were reduced in the transposable element insertionand Galk RNAi-expressing flies (Figure 4, E–G). To testwhether the tinC-Gal4 . UAS-YCanAact-induced hypertrophywould be rescued, we examined whether Mi{ET1}GalkMB10638

would prevent the CanAact-induced increase in cardiac wallthickness. Flies expressing tinC-Gal4 . UAS-YCanAact show res-cue of the increase in wall thickness (Figure 4H). Additionally,tinC-CanAact flies had reduced life span compared to controls,which was rescued in the context of Mi{ET1}GalkMB10638

(Figure S4). These results support the hypothesis that Galkis a modifier of calcineurin in the heart.

Figure 1 tinC-CanAact flies display a cardiac enlargementphenotype. (A) Domain structure of calcineurin, tinC-YFP-tagged CanAact (tinC-YCanAact), and tinC-Flag-taggedCanAact (tinC-FCanAact). CanAact, constitutively active cal-cineurin; B, CanB binding domain; M, calmodulin bindingdomain; I, autoinhibitory domain. tinC is composed ofa 304-bp genomic DNA element driving transgene expres-sion in all cells constituting the heart tube as previouslydescribed (Yin et al. 1997; Lo and Frasch 2001; Wolf et al.2006). (B) Representative OCT m-mode images and summarydata for end-diastolic dimension and fractional shorteningof w1118 control (N = 16) and heterozygous tinC-YCanAact

from cross with w1118 (N = 19) flies. CanAact flies showenlarged end-diastolic dimensions and reduced fractionalshortening. (C) Representative transverse paraffin sec-tions of homozygous tinC-GFP control (N = 11) and tinC-YCanAact (N = 9) fly hearts (blue arrows indicate the flyheart). Summary data for lumen area and perimeter areshown in micrometers. The tinC-YCanAact fly displayeda significantly enlarged cardiac lumen area and perimeter.(Student’s t-test, *P , 0.05, **P , 0.01, ***P , 0.001compared to either w1118 or tinC-GFP control. Data repre-sent mean 6 SEM.)

































We then tested whether Mi{ET1}GalkMB10638 would res-cue a non-calcineurin-mediated cardiomyopathy of the tro-ponin I mutant hdp2 that shows a flight muscle abnormalityand cardiac dilation (Beall and Fyrberg 1991; Wolf et al.2006). Mi{ET1}GalkMB10638 did not rescue the cardiac dila-tion phenotype of heterozygous hdp2 flies (Figure S5), indi-cating that Galk modification is specific to CanAact-inducedcardiomyopathy.

We also tested Argk as a possible modifier for calcineurin.A Drosophila line with an insertion in Argk, PBac{WH}Argkf05525,had normal cardiac function and did not rescue tinC-CanAact-induced cardiac enlargement (Figure S6, A–C).

In addition, the dorsocross genes (Doc1, Doc2, and Doc3) arehomologous to mammalian T-box genes and known to regu-late embryonic cardiac development (Reim and Frasch 2005).

Accordingly, we examined the effects of these candidate geneson the tinC-CanAact cardiac enlargement phenotype, usingOCT. Doc1, Doc2, and Doc3 RNAi lines were obtained andcrossed into tinC-Gal4, tinC-CanAact flies to test for rescue ofthe CanAact phenotype. Genetic disruption of the Doc geneswith RNAi did not rescue tinC-CanAact-induced cardiac en-largement (Figure S6D). Interestingly, several cardiac-specificRNAi lines directed against the Doc genes caused an enlargedcardiac phenotype on their own (Figure S6D).

Tissue-specific suppression of CanAact-inducedphenotypes in noncardiac tissues with geneticdeficiency of Galk

Next, we used a Gal4/UAS system for driving CanAact withnoncardiac tissue drivers. The expression of CanAact under

Figure 2 Molecularly defined deficiencies Df(3L)ED4416and Df(3L)BSC130 rescued calcineurin-mediated abnor-malities in adult flies. (A) Genetic map of deficiencystocks tested (adapted from Gbrowse, http://flybase.org/cgi-bin/gbrowse/dmel). Dashed lines indicate thesuppressor region. Genes within the suppressor regionare shown in the magnified view below. (Green bars,rescuing deficiencies; red bars, nonrescuing deficiencies;black bar, deficiency that causes cardiomyopathy on itsown.) (B and C) Summary data for (B) end-diastolic di-mension and (C) fractional shortening of w1118 control(first column) tinC-YCanAact alone and in the context ofmolecularly defined genomic deficiencies. All deficien-cies were tested in the heterozygous states. (N = 17–51 for each group.) Two deficiency lines rescued thetinC-YCanAact phenotype, narrowing down the originalsuppressor region to the region in between the dottedlines in A. Note that Df(3L)ED4421 covers the deficiencyregion but was dilated on its own without CanAact ex-pression (Figure S3) and was not considered in definingthe deficiency region. (One-way ANOVA with Bonferronicorrection; *P , 0.05, **P , 0.0001; data representmean 6 SEM.)
















http://flybase.org/cgi-bin/gbrowse/dmel

http://flybase.org/cgi-bin/gbrowse/dmel


control of the engrailed wing driver e16E-Gal4 caused anabnormality in the PCVs and the longitudinal wing vein L5(Figure 5). Importantly, Df(3L)ED4416 significantly rescuedthe e16E-Gal4 . UAS-YCanAact-induced abnormal wingvein phenotype. The percentage of flies that had normalwing veins significantly increased from 11 to 44%, and thepercentage of flies with abnormal wing vein phenotypesdecreased from 89 to 56% with a genetic deficiency en-compassing Galk (Figure 5). To further investigate the in-volvement of Galk in wing vein abnormality, the effect ofdisruption of Galk with a transposable element insertionwas examined in e16E-Gal4 . UAS-YCanAact wings. Disrup-tion of Galk with Mi{ET1}GalkMB10638 in e16E-Gal4 . UAS-YCanAact flies also rescued the wing vein phenotypes, albeitat a lesser rate compared to the deficiency, with the per-centage of normal wing vein flies increasing from 5 to 31%and the percentage of abnormal winged flies decreasingfrom 95 to 69% in the context of Mi{ET1}GalkMB10638 (Fig-ure 5, bottom right).

Expression of CanAact under control of the ubiquitousdriver, Act5C-Gal4, caused larval lethality (Figure 6A), whileexpression under the ectodermal driver, dpp-Gal4, caused acomplete inability of the wing to expand (Figure 6B), andexpression under the muscle driver, Mef2-Gal4, caused lethal-ity at a pupal stage (Figure 6C). In contrast to the observedrescue of the abnormal wing vein phenotype, Df(3L)ED4416was not sufficient to rescue the wing or lethality phenotypesdriven by Act5C-Gal4, dpp-Gal4, orMef2-Gal4, suggesting that

deficiency of Galk suppresses CanAact-induced phenotypes ina tissue-specific manner. We also examined the ability ofMi{ET1}GalkMB10638 to suppress theMef2-Gal4. UAS-YCanAact

lethality phenotype and show that Mi{ET1}GalkMB10638 wasnot able to rescue the Mef2-Gal4 . UAS-YCanAact-inducedlethality (Figure 6C), further indicating that Galk may notmodify Mef2-Gal4 . UAS-YCanAact-induced lethality.

Discussion

Calcineurin is a known mediator of cardiac hypertrophyin mammalian hearts and understanding the signals thatregulate calcineurin has the potential to alter cardiacpathology. Here we show that (1) CanAact in the Drosophilaheart induced cardiac enlargement and reduced cardiaccontractility; (2) Galk disruption suppressed the CanAact-mediated cardiac enlargement, increase in wall thickness,and posterior wing vein phenotypes; and (3) Galk regulationof CanAact-induced phenotypes was tissue specific. Usingthe resources that are available in fly genetics, the abilityto phenotype in vivo cardiac chamber sizes, and the uniquefact that flies lack calcineurin-regulated NFAT, we identifieda potential new regulator of calcineurin in the Drosophilaheart. We also observed that CanAact induced sustainedcardiac enlargement in Drosophila during aging and re-duced life span. Furthermore, known regulators of cal-cineurin signaling, including Mef2, rescued the observedcardiac abnormalities, supporting our hypothesis that these

Figure 3 tinC-Gal4 . UAS-CanAact displayed cardiac en-largement that was rescued by tinC-CanAact-rescuing defi-ciencies and diminished contractility phenotype that wasrescued by known calcineurin modifiers. (A and B) Sum-mary data for (A) end-diastolic dimension and (B) fractionalshortening of w1118 controls (N = 9); tinC-Gal4 . UAS-YCanAact (N = 32) alone; heterozygous tinC-Gal4 (N = 7);and tinC-Gal4 . UAS-YCanAact in the presence of Mef2(N = 11), sty (N = 12), CanB2 (N = 23), Df(3L)ED4416 (N = 18),and Df(3L)BSC130 (N = 13). All transgenes and mutationswere heterozygous. Known modifiers of calcineurin sig-naling, Mef2, sty, and CanB2, rescued tinC-Gal4 . UAS-YCanAact cardiac contractility (sty rescued fractional shorteningbut not end-diastolic dimension). In addition, the tinC-YCanAact-rescuing deficiencies Df(3L)ED4416 and Df(3L)BSC130 also rescued tinC-Gal4 . UAS-YCanAact cardiacenlargement [Df(3L)BSC130 rescued end-diastolic dimen-sion but not fractional shortening]. (*P , 0.05, **P ,0.01, ***P , 0.0001; one-way ANOVA with Bonferronicorrection; data represent mean 6 SEM.)















approaches can identify modifiers of calcineurin-mediatedcardiac abnormalities.

Two prior screens of the fly scored changes in eye mor-phology or lethality caused by activated calcineurin andidentified several genomic regions that harbor potential mod-

ifiers. Four enhancer regions and five suppressor regions wereidentified in a mutagenesis screen of the fly eye (Sullivan andRubin 2002), whereas seven suppressor regions were discov-ered in a deficiency screen for lethality (Gajewski et al. 2003).These studies identified cytological regions in the fly genome

Figure 4 Genetic disruption of Galk rescued tinC-CanAact-induced reduction in cardiac contractility and tinC-Gal4 .UAS-CanAact- induced hypertrophy. (A and B) Summary datafor (A) end-diastolic dimension and (B) fractional shortening offlies expressing w1118 control, tinC-YCanAact, the transpos-able elements PBac{PB}Galkc03848, Mi{ET1}GalkMB10638, orprecise excision of Mi{ET1}GalkMB10638 (Mi{ET1}Galkrev), inthe context of tinC-YCanAact. The two transposable elementsin Galk rescued tinC-YCanAact -mediated cardiac contractility(fractional shortening), while a precise excision revertedthe rescue. [(A) *P , 0.05, **P , 0.01; (B) *P , 0.05,**P , 0.001, ***P , 0.0001 compared to tinC-YCanAact

or Mi{ET1}GalkMB10638 as indicated with an overbar; N =14–27 for each group; one-way ANOVA with Bonferronicorrection for multiple comparisons, all data representmean 6 SEM.] (C and D) Summary data for (C) end-diastolic dimension and (D) fractional shortening of w1118

control, CanAact only, UAS-Galk no driver (heterozygouswith w1118), tinC-Gal4 driver only (heterozygous with w1118),Galk RNAi only flies, and flies expressing tinC-YCanAact inthe context of cardiac Galk RNAi. RNAi to Galk significantlyrescued tinC-YCanAact-mediated cardiac enlargement. (*P ,0.01, **P, 0.0001 compared to tinC-YCanAact; N = 7–20 foreach group; one-way ANOVA with Bonferroni correction formultiple comparisons, all data represent mean 6 SEM.) (E)qRT-PCR for Galk expression of w1118 control, homozygousPBac{PB}Galkc03848, or homozygous Mi{ET1}GalkMB10638.Galk expression is downregulated by transposable elementinsertions. (*P, 0.001 compared to w1118 control; N = 3 ineach group; one-way ANOVA with Bonferroni correctionfor multiple comparisons, all data represent mean 6 SEM.)(F) qRT-PCR of Galk expression in w1118 control flies, fliesheterozygous for tinC-YCanAact alone, and flies in the con-text of heterozygous Mi{ET1}GalkMB10638 or a precise ex-cision of Mi{ET1}GalkMB10638, Mi{ET1}Galkrev. (*P , 0.05,**P , 0.001 compared to heterozygous tinC-YCanAact/Mi{ET1}GalkMB10638; N = 8 in each group; one-wayANOVA with Bonferroni correction for multiple comparisons,all data represent mean 6 SEM.) (G) Galk expression fortinC-Gal4 heterozygous with w1118 (driver only) and het-erozygous UAS-Galk RNAi knockdown with a tinC-Gal4driver (Galk RNAi). Galk expression is downregulated bytransposable element insertions. (N = 3 for each group;Student’s t-test *P , 0.001, Galk RNAi vs. tinC-Gal4 driveronly.) (H) Representative hematoxylin- and eosin-stainedhistological sections and quantification of wall thicknessfor heterozygous tinC-Gal4 (from cross with w1118, N = 8),tinC-Gal4 . UAS-YCanAact (N = 13), or tinC-Gal4 .UAS-YCanAact in the context of Mi{ET1}GalkMB10638

(N = 6). tinC-Gal4 . UAS-YCanAact caused a cardiac hyper-trophy phenotype that was rescued byMi{ET1}GalkMB10638.Blue arrowheads point to the heart. *P , 0.05, one-wayANOVA with Bonferroni correction, all data representmean 6 SEM. (All transgenes were heterozygous unlessotherwise noted.)


and, based on these important findings, we focused ourattention on a region that was common to both studies butlacked identification of the candidate modifier of calcineurin.Using molecularly defined genomic deficiencies, transpos-able element insertions, precise excisions, and transgenicRNAi, we identified that galactokinase was a candidatemodifier of calcineurin-mediated cardiac abnormalities inthe fly. Galactokinase may function either as an enhancerof calcineurin signaling or in a pathway downstream ofcalcineurin activation. Interestingly, previous studies havenot implicated an interaction between calcineurin andgalactokinase.

Our results demonstrate that deficiency of Mef2 sup-pressed cardiac-CanAact-induced cardiac enlargement. SinceDrosophila do not express calcineurin-regulated NFAT, thisimplies that Mef2 functions independent of NFAT in oursystem. In agreement with our results, studies in skeletalmuscle and C2C12 myogenic cells have shown that calci-neurin binds to Mef2, leading to hypophosphorylation andenhanced transcription activity (Wu et al. 2000, 2001).However, studies in Jurkat T-lymphocytes suggest thatcalcineurin-regulated Mef2 activation requires recruitment ofNFATc2 (Blaeser et al. 2000; Youn et al. 2000). In the mouseheart, the major pathway appears to be NFAT since express-ing dominant negative Mef2 rescued only cardiac dilationand not hypertrophy (Van Oort et al. 2006). In the samemouse study, overexpressing Mef2 induced a cardiac dila-tion phenotype without hypertrophy, suggesting that Mef2mainly contributes to the dilation phenotype induced bycalcineurin. Interestingly, flies expressing tinC-Gal4 . UAS-

Mef2 had significantly increased end-diastolic dimensions(Figure S7), similar to the tinC-CanAact phenotype, althoughthe flies overexpressing Mef2 did not display reduced frac-tional shortening, suggesting that Mef2 is necessary but notsufficient for calcineurin-induced reduction in contractility.These results suggest that an NFAT-independent pathwaythrough Mef2 in Drosophila may regulate cardiac enlarge-ment and that additional factors outside of Mef2 may beinvolved in controlling cardiac function.

Several potential genes were excluded as modifiers fromour original candidate region. The Doc genes are known toregulate Drosophila cardiac development (Reim and Frasch2005), and we showed that knocking down Doc expressionwith RNAi caused cardiac enlargement and did not rescuetinC-CanAact. This implies that Doc genes are important forcardiac development but we do not have evidence for themregulating calcineurin. Argk has high expression in the adultDrosophila heart. Argk is an enzyme that transfers the phos-phate on ATP to arginine, creating an energy-rich buffer formaintaining ATP concentration (Newsholme et al. 1978).Although it is conceivable that Argk functions to providesufficient energy for the fly myocardium, our results showedthat disrupting Argk with a transposable element did notproduce a phenotype and did not rescue calcineurin-inducedcardiac enlargement. These results suggest that Argk doesnot regulate calcineurin-induced cardiac enlargement. How-ever, we note that the gene expression was decreased byonly 30% in the heterozygous PBac{PB}Argkf05255 line used(Figure S6C) and it is possible that the lack of rescue wasdue to incomplete knockdown of gene expression.

Figure 5 The deficiency Df(3L)ED4416 and Minos inser-tion in Galk Mi{ET1}GalkMB10638 rescued e16E-Gal4 .UAS-YCanAact -induced wing vein loss. Wing vein pheno-types of heterozygous e16E-Gal4 driver only control ore16E-Gal4 . UAS-YCanAact flies: normal, abnormality ofthe posterior crossvein (PCV), or abnormality of both PCVand longitudinal vein 5 (L5). Progeny were counted fromthe cross e16E-Gal4 3 UAS-YCanAact/CyO;Df(3L)ED4416/MKRS or e16E-Gal43UAS-YCanAact/CyO;Mi{ET1}GalkMB10638/MKRS. The graph represents percentage of all heterozygouse16E-Gal4 . UAS-YCanAact flies counted with normal orabnormal wings in the context of heterozygous MKRS bal-ancer (no deficiency), Df(3L)ED4416, orMi{ET1}GalkMB10638.Df(3L)ED4416 partially rescues the e16E-Gal4. UAS-YCanAact-induced wing vein loss phenotype [no deficiency, N = 46;Df(3L)ED4416, N = 54]. Mi{ET1}GalkMB10638 also partiallyrescued the e16E-Gal4 . UAS-YCanAact -induced wing veinloss phenotype (no Minos, N = 129; Mi{ET1}GalkMB10638,N = 242). (Arrowheads point to the shortened abnormalPCV or L5, respectively. Significant rescue of the wing veinabnormality with Df(3L)ED4416 and Mi{ET1}GalkMB10638

was determined by Fisher’s exact test, *P , 0.0001.)





A deficiency encompassing Galk suppressed the CanAact-induced phenotypes in heart and posterior wing. However,this deficiency did not suppress phenotypes driven by ubiq-uitous Act5C-Gal4, ectodermal dpp-Gal4, or mesodermalMef2-Gal4 drivers. These observations could be explainedby the tissue-specific context in which calcineurin was ex-pressed. The expression of each driver occurs in differenttissues. Although e16E-Gal4 and dpp-Gal4 expression ofCanAact induced wing phenotypes, e16E-Gal4 drives expres-sion in the posterior wing while dpp-Gal4 drives expressionanterior to the anterior–posterior boundary during imaginaldisc development at the third instar larva stage (Brower1986; de Celis 1997). Many signaling factors are differ-entially expressed during wing development in these twoseparate compartments to guide correct patterning. For ex-ample, engrailed and hedgehog are expressed posteriorly,guiding formation of the posterior wing veins (Brower 1986;Lee et al. 1992), while dpp and the EGFR inhibitor knot areexpressed specifically in the anterior wing imaginal disc (deCelis 1997; Mohler et al. 2000). It is possible that Galkmodification of calcineurin signaling is regulated by thesedifferentially expressed factors. In addition, Act5C-Gal4 andMef2-Gal4 driving CanAact induced lethality that was notsuppressed by deficiency of Galk. These drivers induce tran-script expression in multiple tissues during early embryonicdevelopment (Burn et al. 1989; Bour et al. 1995). One ex-planation for the findings is that the signals at an early stageof development involve pathways that are not modulated byGalk.

We examined the survival of tinC-CanAact Drosophila andin the context of the insertion Mi{ET1}GalkMB10638. Of note,in mice, cardiac-specific expression of CanAact also inducedpremature sudden death (Molkentin et al. 1998). Flies hetero-zygously expressing tinC-CanAact had a significantly reducedlife span compared to the control group. This reduction wassuppressed in the context of one copy of Mi{ET1}GalkMB10638

(Figure S4), further suggesting that deficiency in Galk sup-pressesmultiple cardiac calcineurin-induced phenotypes. How-ever, a caveat to the interpretation of this experiment is thepossibility of genetic background confounding the beneficialeffects seen by disruption of Galk.

Galactokinase belongs to the GHMP ATP-dependentkinase family (named after the four representative kinasesin this family: galactokinase, homoserine kinase, mevalo-nate kinase, and phosphomevalonate kinase) (Holden et al.2004). In the fly, galactokinase phosphorylates galactoseand N-acetyl-galactosamine (GalNAc), allowing further uti-lization in either metabolism (energy production) or glyco-sylation (protein modification) (Holden et al. 2004). Thesepathways potentially lead to cardiac enlargement: eithergalactokinase promotes a higher level of phosphorylatedgalactose, galactose-1-p, leading to a diseased state, or down-stream reactions involving UDP-galactose incorporation intoglycosylated proteins promote cardiac enlargement, or bothmechanisms may be required. A previous screen examin-ing Drosophila cardiac development discovered that muta-tions in HMG-CoA reductase (the rate-limiting enzyme inthe mevalonate pathway) induce a cardiac phenotype by

Figure 6 The deficiency Df(3L)ED4416 did not rescue Act5C-Gal4. UAS-YCanAact-induced lethality, dpp-Gal4. UAS-YCanAact -induced unexpandedwing, orMef2-Gal4. UAS-YCanAact-induced pupal lethality. (A) Progeny number from the cross Act5C-Gal4/CyO3 UAS-YCanAact/CyO;Df(3L)ED4416/TM2. Driving UAS-YCanAact with an actin (Act5C-Gal4) driver resulted in a lethal phenotype, which was not rescued by Df(3L)ED4416. (B) Percentage oftotal abnormal-winged progeny from the cross dpp-Gal4/TM6B3 UAS-YCanAact/CyO;Df(3L)ED4416/TM2. Expressing CanAact with a dpp driver resultedin a shriveled abnormal wing phenotype. This was not rescued by Df(3L)ED4416. (C) Total progeny from the cross Mef2-Gal4 3 UAS-YCanAact/CyO;Df(3L)ED4416/MKRS or Mef2-Gal4 3 UAS-YCanAact/CyO;Mi{ET1}GalkMB10638/MKRS. Expressing CanAact with a Mef2- driver resulted in a pupal lethalphenotype that was not rescued by Df(3L)ED4416 or Mi{ET1}GalkMB10638. Data were analyzed using Fisher’s exact test.








geranylgeranylation of the G-protein Gg1, suggesting a path-way by which post-translational modifications can alter thefly heart (Yi et al. 2006).

In additional experiments, we show that a transposableelement insertion in galectin (a galactoside-binding lectin)suppressed the tinC-CanAact cardiac phenotype (Figure S8, Aand B). Mammalian Galectin3 has been shown to bind toGalk-regulated N-acetyllactosamine side chains on EGFR,preventing endocytosis and enhancing signaling of isolatedmouse mammary cells (Partridge et al. 2004). Driving acti-vated EGFR in the Drosophila heart has been shown to in-duce a cardiac hypertrophy phenotype (Yu et al. 2013) andsprouty, a regulator of EGFR signaling, has been shown tomodify calcineurin-induced rough eye (Sullivan and Rubin2002). Although speculation, one potential mechanism bywhich galactokinase functions to modulate calcineurin-induced cardiac enlargement is by influencing cotransla-tional glycosylation of EGFR or another yet unidentifiedcell surface protein that is bound by galectin (Figure S8C).

A yeast homolog of Galk, Gal3p, has been found to act asa transcriptional activator by interacting with Gal80p (Zenkeet al. 1996); transcriptional regulation is activated with thebinding of galactose and ATP to Gal3p. This suggests thepossibility that Galk may act as a modifier of transcriptionfor known pathways including Mef2. Whether Galk func-tions as part of the transcriptional machinery remains to bedetermined.

In conclusion, we have developed a model for screeningfor novel modifiers of constitutively active calcineurin in theDrosophila heart and identified galactokinase as a novelmodifier of constitutively active calcineurin-induced cardio-myopathy, shortened life span, and wing vein abnormality inadult Drosophila. These findings have set up a system fordelineating the pathways involved in CanAact-induced car-diomyopathy in Drosophila.

Acknowledgments

This work was supported by grants HL083065 (to H.A.R.)and HL116581 (to M.J.W.) from the National Institutesof Health and by American Heart Association predoctoralaward 11PRE7100004 (to T.E.L.).

Literature Cited

Arca, B., S. Zabalou, T. G. Loukeris, and C. Savakis, 1997 Mo-bilization of a Minos transposon in Drosophila melanogasterchromosomes and chromatid repair by heteroduplex formation.Genetics 145: 267–279.

Ashburner, M., 1989 Drosophila. Cold Spring Harbor LaboratoryPress, Cold Spring Harbor, NY.

Barolo, S., L. A. Carver, and J. W. Posakony, 2000 GFP and beta-galactosidase transformation vectors for promoter/enhanceranalysis in Drosophila. Biotechniques 29: 726, 728, 730, 732.

Beall, C. J., and E. Fyrberg, 1991 Muscle abnormalities inDrosophila melanogaster heldup mutants are caused by miss-

ing or aberrant troponin-I isoforms. J. Cell Biol. 114: 941–951.

Bier, E., and R. Bodmer, 2004 Drosophila, an emerging model forcardiac disease. Gene 342: 1–11.

Blaeser, F., N. Ho, R. Prywes, and T. A. Chatila, 2000 Ca(2+)-dependent gene expression mediated by MEF2 transcriptionfactors. J. Biol. Chem. 275: 197–209.

Bodmer, R., and T. V. Venkatesh, 1998 Heart development inDrosophila and vertebrates: conservation of molecular mecha-nisms. Dev. Genet. 22: 181–186.

Bour, B. A., M. A. O’Brien, W. L. Lockwood, E. S. Goldstein, R.Bodmer et al., 1995 Drosophila MEF2, a transcription factorthat is essential for myogenesis. Genes Dev. 9: 730–741.

Brand, A. H., and N. Perrimon, 1993 Targeted gene expression asa means of altering cell fates and generating dominant pheno-types. Development 118: 401–415.

Brower, D. L., 1986 Engrailed gene expression in Drosophilaimaginal discs. EMBO J. 5: 2649–2656.

Burkard, N., J. Becher, C. Heindl, L. Neyses, K. Schuh et al.,2005 Targeted proteolysis sustains calcineurin activation. Cir-culation 111: 1045–1053.

Burn, T. C., J. O. Vigoreaux, and S. L. Tobin, 1989 Alternative 5Cactin transcripts are localized in different patterns during Dro-sophila embryogenesis. Dev. Biol. 131: 345–355.

Chin, E. R., E. N. Olson, J. A. Richardson, Q. Yang, C. Humphrieset al., 1998 A calcineurin-dependent transcriptional pathwaycontrols skeletal muscle fiber type. Genes Dev. 12: 2499–2509.

Chintapalli, V. R., J. Wang, and J. A. Dow, 2007 Using FlyAtlas toidentify better Drosophila melanogaster models of human dis-ease. Nat. Genet. 39: 715–720.

Cripps, R. M., and E. N. Olson, 2002 Control of cardiac develop-ment by an evolutionarily conserved transcriptional network.Dev. Biol. 246: 14–28.

de Celis, J. F., 1997 Expression and function of decapentaplegicand thick veins during the differentiation of the veins in theDrosophila wing. Development 124: 1007–1018.

Drazner, M. H., J. E. Rame, E. K. Marino, J. S. Gottdiener, D. W.Kitzman et al., 2004 Increased left ventricular mass is a riskfactor for the development of a depressed left ventricular ejec-tion fraction within five years: the Cardiovascular Health Study.J. Am. Coll. Cardiol. 43: 2207–2215.

Gajewski, K., J. Wang, J. D. Molkentin, E. H. Chen, E. N. Olsonet al., 2003 Requirement of the calcineurin subunit genecanB2 for indirect flight muscle formation in Drosophila. Proc.Natl. Acad. Sci. USA 100: 1040–1045.

George, Jr., A. L., 2013 Molecular and genetic basis of suddencardiac death. J. Clin. Invest. 123: 75–83.

Grossman, W., and W. J. Paulus, 2013 Myocardial stress and hy-pertrophy: a complex interface between biophysics and cardiacremodeling. J. Clin. Invest. 123: 3701–3703.

Holden, H. M., J. B. Thoden, D. J. Timson, and R. J. Reece,2004 Galactokinase: structure, function and role in type IIgalactosemia. Cell. Mol. Life Sci. 61: 2471–2484.

Kang, Y. J., B. Kusler, M. Otsuka, M. Hughes, N. Suzuki et al.,2007 Calcineurin negatively regulates TLR-mediated activa-tion pathways. J. Immunol. 179: 4598–4607.

Keyser, P., K. Borge-Renberg, and D. Hultmark, 2007 The Dro-sophila NFAT homolog is involved in salt stress tolerance. InsectBiochem. Mol. Biol. 37: 356–362.

Klee, C. B., T. H. Crouch, and M. H. Krinks, 1979 Calcineurin:a calcium- and calmodulin-binding protein of the nervous sys-tem. Proc. Natl. Acad. Sci. USA 76: 6270–6273.

Klee, C. B., G. F. Draetta, and M. J. Hubbard, 1988 Calcineurin.Adv. Enzymol. Relat. Areas Mol. Biol. 61: 149–200.

Klee, C. B., H. Ren, and X. Wang, 1998 Regulation of the calmod-ulin-stimulated protein phosphatase, calcineurin. J. Biol. Chem.273: 13367–13370.





Kohler, R. E., 1994 Lords of the Fly: Drosophila Genetics and theExperimental Life. University of Chicago Press, Chicago.

Lee, J. J., D. P. von Kessler, S. Parks, and P. A. Beachy,1992 Secretion and localized transcription suggest a role inpositional signaling for products of the segmentation genehedgehog. Cell 71: 33–50.

Levy, D., R. J. Garrison, D. D. Savage, W. B. Kannel, and W. P.Castelli, 1990 Prognostic implications of echocardiographi-cally determined left ventricular mass in the Framingham HeartStudy. N. Engl. J. Med. 322: 1561–1566.

Lo, P. C., and M. Frasch, 2001 A role for the COUP-TF-relatedgene seven-up in the diversification of cardioblast identities inthe dorsal vessel of Drosophila. Mech. Dev. 104: 49–60.

Maillet, M., J. Davis, M. Auger-Messier, A. York, H. Osinska et al.,2010 Heart-specific deletion of CnB1 reveals multiple mecha-nisms whereby calcineurin regulates cardiac growth and func-tion. J. Biol. Chem. 285: 6716–6724.

Messerli, F. H., and R. Ketelhut, 1991 Left ventricular hypertrophy:an independent risk factor. J. Cardiovasc. Pharmacol. 17(Suppl.4): S59–S66; discussion S66–S57.

Metaxakis, A., S. Oehler, A. Klinakis, and C. Savakis, 2005 Minosas a genetic and genomic tool in Drosophila melanogaster. Ge-netics 171: 571–581.

Mohler, J., M. Seecoomar, S. Agarwal, E. Bier, and J. Hsai,2000 Activation of knot (kn) specifies the 3–4 intervein regionin the Drosophila wing. Development 127: 55–63.

Molkentin, J. D., J. R. Lu, C. L. Antos, B. Markham, J. Richardsonet al., 1998 A calcineurin-dependent transcriptional pathwayfor cardiac hypertrophy. Cell 93: 215–228.

Musaro, A., K. J. McCullagh, F. J. Naya, E. N. Olson, and N. Rosenthal,1999 IGF-1 induces skeletal myocyte hypertrophy through cal-cineurin in association with GATA-2 and NF-ATc1. Nature 400:581–585.

Newsholme, E. A., I. Beis, A. R. Leech, and V. A. Zammit,1978 The role of creatine kinase and arginine kinase in mus-cle. Biochem. J. 172: 533–537.

Okamura, H., J. Aramburu, C. Garcia-Rodriguez, J. P. Viola, A.Raghavan et al., 2000 Concerted dephosphorylation of thetranscription factor NFAT1 induces a conformational switch thatregulates transcriptional activity. Mol. Cell 6: 539–550.

Parsons, S. A., D. P. Millay, B. J. Wilkins, O. F. Bueno, G. L. Tsikaet al., 2004 Genetic loss of calcineurin blocks mechanical over-load-induced skeletal muscle fiber type switching but not hyper-trophy. J. Biol. Chem. 279: 26192–26200.

Partridge, E. A., C. Le Roy, G. M. Di Guglielmo, J. Pawling, P.Cheung et al., 2004 Regulation of cytokine receptors by GolgiN-glycan processing and endocytosis. Science 306: 120–124.

Passier, R., H. Zeng, N. Frey, F. J. Naya, R. L. Nicol et al.,2000 CaM kinase signaling induces cardiac hypertrophy andactivates the MEF2 transcription factor in vivo. J. Clin. Invest.105: 1395–1406.

Reim, I., and M. Frasch, 2005 The Dorsocross T-box genes are keycomponents of the regulatory network controlling early cardio-genesis in Drosophila. Development 132: 4911–4925.

Robinson, S. W., P. Herzyk, J. A. Dow, and D. P. Leader,2013 FlyAtlas: database of gene expression in the tissues ofDrosophila melanogaster. Nucleic Acids Res. 41: D744–D750.

Sakuma, K., and A. Yamaguchi, 2010 The functional role of cal-cineurin in hypertrophy, regeneration, and disorders of skeletalmuscle. J. Biomed. Biotechnol. 2010: 721219.

Sakuma, K., M. Akiho, H. Nakashima, R. Nakao, M. Hirata et al.,2008 Cyclosporin A modulates cellular localization of MEF2Cprotein and blocks fiber hypertrophy in the overloaded soleusmuscle of mice. Acta Neuropathol. 115: 663–674.

Schaeffer, P. J., J. Desantiago, J. Yang, T. P. Flagg, A. Kovacs et al.,2009 Impaired contractile function and calcium handling in

hearts of cardiac-specific calcineurin b1-deficient mice. Am. J.Physiol. Heart Circ. Physiol. 297: H1263–H1273.

Semsarian, C., M. J. Wu, Y. K. Ju, T. Marciniec, T. Yeoh et al.,1999 Skeletal muscle hypertrophy is mediated by a Ca2+-dependent calcineurin signalling pathway. Nature 400: 576–581.

Shibasaki, F., U. Hallin, and H. Uchino, 2002 Calcineurin as a mul-tifunctional regulator. J. Biochem. 131: 1–15.

St. Pierre, S. E., L. Ponting, R. Stefancsik, and P. McQuilton, andFlyBase Consortium, 2014 FlyBase 102–advanced approachesto interrogating FlyBase. Nucleic Acids Res. 42: D780–D788.

Sullivan, K. M., and G. M. Rubin, 2002 The Ca(2+)-calmodulin-activated protein phosphatase calcineurin negatively regulatesEGF receptor signaling in Drosophila development. Genetics161: 183–193.

Sussman, M. A., H. W. Lim, N. Gude, T. Taigen, E. N. Olson et al.,1998 Prevention of cardiac hypertrophy in mice by calcineurininhibition. Science 281: 1690–1693.

Taigen, T., L. J. De Windt, H. W. Lim, and J. D. Molkentin,2000 Targeted inhibition of calcineurin prevents agonist-inducedcardiomyocyte hypertrophy. Proc. Natl. Acad. Sci. USA 97: 1196–1201.

Tokoyoda, K., Y. Takemoto, T. Nakayama, T. Arai, and M. Kubo,2000 Synergism between the calmodulin-binding and autoin-hibitory domains on calcineurin is essential for the inductionof their phosphatase activity. J. Biol. Chem. 275: 11728–11734.

van Berlo, J. H., M. Maillet, and J. D. Molkentin, 2013 Signalingeffectors underlying pathologic growth and remodeling of theheart. J. Clin. Invest. 123: 37–45.

van Oort, R. J., E. van Rooij, M. Bourajjaj, J. Schimmel, M. A.Jansen et al., 2006 MEF2 activates a genetic program promot-ing chamber dilation and contractile dysfunction in calcineurin-induced heart failure. Circulation 114: 298–308.

Watanabe, Y., B. A. Perrino, B. H. Chang, and T. R. Soderling,1995 Identification in the calcineurin A subunit of the domainthat binds the regulatory B subunit. J. Biol. Chem. 270: 456–460.

Wessells, R. J., and R. Bodmer, 2004 Screening assays for heartfunction mutants in Drosophila. Biotechniques 37: 58–60, 62,64 passim.

Wilkins, B. J., and J. D. Molkentin, 2002 Calcineurin and cardiachypertrophy: Where have we been? Where are we going? J.Physiol. 541: 1–8.

Wilkins, B. J., Y. S. Dai, O. F. Bueno, S. A. Parsons, J. Xu et al.,2004 Calcineurin/NFAT coupling participates in pathologi-cal, but not physiological, cardiac hypertrophy. Circ. Res. 94:110–118.

Wolf, M. J., H. Amrein, J. A. Izatt, M. A. Choma, M. C. Reedy et al.,2006 Drosophila as a model for the identification of genescausing adult human heart disease. Proc. Natl. Acad. Sci. USA103: 1394–1399.

Wu, H., F. J. Naya, T. A. McKinsey, B. Mercer, J. M. Shelton et al.,2000 MEF2 responds to multiple calcium-regulated signals inthe control of skeletal muscle fiber type. EMBO J. 19: 1963–1973.

Wu, H., B. Rothermel, S. Kanatous, P. Rosenberg, F. J. Naya et al.,2001 Activation of MEF2 by muscle activity is mediatedthrough a calcineurin-dependent pathway. EMBO J. 20: 6414–6423.

Yi, P., Z. Han, X. Li, and E. N. Olson, 2006 The mevalonate path-way controls heart formation in Drosophila by isoprenylation ofGgamma1. Science 313: 1301–1303.

Yin, Z., and M. Frasch, 1998 Regulation and function of tinmanduring dorsal mesoderm induction and heart specification inDrosophila. Dev. Genet. 22: 187–200.


Yin, Z., X. L. Xu, and M. Frasch, 1997 Regulation of the twisttarget gene tinman by modular cis-regulatory elements duringearly mesoderm development. Development 124: 4971–4982.

Youn, H. D., T. A. Chatila, and J. O. Liu, 2000 Integration ofcalcineurin and MEF2 signals by the coactivator p300 duringT-cell apoptosis. EMBO J. 19: 4323–4331.

Yu, L., T. Lee, N. Lin, and M. J. Wolf, 2010 Affecting Rhomboid-3function causes a dilated heart in adult Drosophila. PLoS Genet.6: e1000969.

Yu, L., J. Daniels, A. E. Glaser, and M. J. Wolf, 2013 Raf-mediatedcardiac hypertrophy in adult Drosophila. Dis. Model. Mech. 6:964–976.

Zaffran, S., and M. Frasch, 2002 Early signals in cardiac develop-ment. Circ. Res. 91: 457–469.

Zaffran, S., X. Xu, P. C. Lo, H. H. Lee, and M. Frasch,2002 Cardiogenesis in the Drosophila model: control mechanismsduring early induction and diversification of cardiac progenitors.Cold Spring Harb. Symp. Quant. Biol. 67: 1–12.

Zenke, F. T., R. Engles, V. Vollenbroich, J. Meyer, C. P. Hollenberget al., 1996 Activation of Gal4p by galactose-dependent inter-action of galactokinase and Gal80p. Science 272: 1662–1665.

Communicating editor: N. Perrimon


GENETICSSupporting Information


Galactokinase Is a Novel Modifier ofCalcineurin-Induced Cardiomyopathy in Drosophila

Teresa E. Lee, Lin Yu, Matthew J. Wolf, and Howard A. Rockman

Copyright © 2014 by the Genetics Society of AmericaDOI: 10.1534/genetics.114.166777

T. E. Lee et al. 2SI

Figure S1 Drosophila expressing constitutively active calcineurin (CanAact) in the heart displayed cardiac abnormalities. A) Quantification of end-‐diastolic dimensions and fractional shortenings from m-‐mode images of several different control flies: w1118 (FBst0003605 used in all of our experiments), w1118 (FBst0006326 used for injection by the Duke Model Systems Genomics Facility, all other data of w1118 refer to FBst0003605), tinC-‐Gal4 heterozygous with w1118 FBst0003605 (tinC-‐Gal4/+), homozygous tinC-‐Gal4 (tinC-‐Gal4), tinC-‐GFP heterozygous with w1118 FBst0003605 (tinC-‐GFP/+), and tinC-‐GFP homozygous (tinC-‐GFP). The end-‐diastolic dimensions and fractional shortenings were not significantly different between groups (One-‐way ANOVA with Bonferroni correction). B) Quantification of end-‐diastolic dimension and fractional shortening from OCT m-‐mode images for w1118 control and heterozygous tinC-‐Flag-‐CanAact from cross with w1118 (tinC-‐FCanAact/+) Drosophila. tinC-‐FCanAact/+ flies also displayed a cardiac enlargement phenotype compared to control (student’s t test). C) Quantification of end-‐diastolic dimensions for different transgenic lines expressing tinC-‐YFP-‐CanAact and tinC-‐Flag-‐CanAact. Several different transgenic lines including insertions on the first, second, and third chromosomes all induce cardiac enlargement. Lines 32-‐1 (Chr3) for tinC-‐YCanAact and 25-‐4 (Chr1) for tinC-‐FCanAact were used in this study. (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. One-‐way ANOVA with Bonferroni correction. Numbers denote different transgenic lines derived after initial plasmid injection of

0

20

40

60

80

0

20

40

60

80

100

w1118

tinC-FCanAact /+

En

d-d

iasto

lic

Dim

en

sio

n (

mic

ron

s)

Fra

ctional

Short

enin

g (

%)

En

d-d

iasto

lic

Dim

en

sio

n (

mic

ron

s)

Fra

ctional

Short

enin

g (

%)

1 2 3 4 5

weeks

0

5

10

15

tinC-GFP

tinC-YCanAact

Th

ickn

ess (

mic

ron

s)

Th

ickn

ess (

mic

ron

s)

N=11 9

N=7 4 4 4

*

*

A

D

E

tinC-Gal4/+ (Driver alone)

tinC-Gal4>UAS-YCanAact

tinC-Gal4 - + - +

UAS-YCanAact - - + +

Os Os Os

nuclear RFP

F

Os Os Os

A1

A1

A2 A3

A2 A3 tinC-YCanAact

tinC-GFP

tinC-YFP-CanAact

tinC-GFP

G H

0

5

10

w11

18

25-4

(Chr1

)

14-5

(Chr2

)

14-4

(Chr2

,3)

33-3

(Chr3

)

0

50

100

**

*** ****

*** *** *** *

En

d-d

iasto

lic

Dim

en

sio

n (

mic

ron

s)

En

d-d

iasto

lic

Dim

en

sio

n (

mic

ron

s)

B

0

20

40

60

80E

nd

-dia

sto

lic

dim

en

sio

n (

mic

ro

ns

)

0

20

40

60

80

100

Fra

cti

on

al

Sh

orte

ing

(%

)

w11

18 (3

605)

w1118 (

6326)

tinC-G

al4/+

tinC-G

al4

tinC-G

FP/+

tinC-G

FP

C

N= 11 10 14 11 11 8

w11

18

32-1

(Chr3

)

13-1

(Chr2

)

16-1

(Chr3

)

0

50

100

tinC-YCanAact

tinC-FCanAact

N= 9 4 9 3

N= 9 9 7 5 5

N.S.

N.S.

50 microns

N=8 8

**

0

25

50

75

100

0

20

40

60

80

100

1 2 3 4 5

weeks

w1118FCanAact


Drosophila embryos.) D) Quantification of end-‐diastolic dimension and fractional shortening from OCT images of the heart taken 1-‐5 weeks after eclosion. tinC-‐FCanAact-‐induced cardiac enlargement persisted with age. (N=7-‐17 for each group. P<0.0001, tinC-‐FCanAact compared to w1118 control, two-‐way ANOVA. No significant interaction was noted between groups.) E) Measurements for cardiac wall thickness from hematoxylin and eosin sections of tinC-‐GFP and tinC-‐YCanAact flies from Figure 1C. Wall thickness was not significantly different in tinC-‐YCanAact flies compared to control. F) Histological sections of tinC-‐Gal4 (Driver alone) and tinC-‐Gal4>UAS-‐YCanAact flies with quantified results measuring heart wall thickness. tinC-‐Gal4>UAS-‐YCanAact resulted in a hypertrophied phenotype with increase in wall thickness (*P<0.05 compared to control, one-‐way ANOVA with Bonferroni correction, all data represent mean ±SEM). G) GFP staining of tinC-‐GFP and tinC-‐YCanAact Drosophila hearts in the first abdominal segment (A1) showing cardiac enlargement in tinC-‐YCanAact flies using confocal microscopy. H) Fly hearts expressing tinC-‐nuclear RFP in the presence of either tinC-‐GFP or tinC-‐YCanAact were imaged with confocal microscopy. Ostia cells (Os) along with the cuticle pattern mark the boundaries to abdominal compartments (A1-‐A3). The number of cells in each abdominal segment per heart was not altered in tinC-‐YCanAact flies.


Figure S2 tinC-‐CanAact-‐mediated cardiac enlargement was rescued by Mef2 deficiency. A-‐D) Representative OCT m-‐mode images of w1118 (control, first columns in E and F), heterozygous tinC-‐FCanAact, heterozygous Mef2 deficiency stock, and heterozygous Mef2 deficiency rescue of tinC-‐FCanAact, showing that deficiency of the known calcineurin suppressor gene Mef2 also suppressed tinC-‐FCanAact-‐induced cardiac enlargement. (Mef2X1= Df(2R)X1,Mef2X1/CyOAdhB) E-‐F) Summary data for end-‐diastolic dimension and fractional shortening from m-‐mode OCT images. Cardiac-‐specific CanAact produced a cardiac enlargement phenotype which was rescued by a deficiency in the known calcineurin modifier Mef2. (*p<0.05, **p<0.01, ***P<0.001, one-‐way ANOVA with Bonferroni correction. All data represent mean ±SEM.)

E

F

125

mic

rons

End

-dia

stol

ic

Dim

ensi

on (m

icro

ns)

Frac

tiona

l S

horte

ning

(%)

- + - + - - + +

1 sec

w1118 (control)

tinC-FCanAact

Mef2X1

tinC-FCanAact/+

Mef2X1/+

tinC-FCanAact/+;+/Mef2X1

******A

B

C

D

**

020406080

100

020406080

100

N= 15 21 16 25

**** *


Figure S3 End-‐diastolic dimension of the deficiency or transposable element insertion stocks alone. A) End-‐diastolic dimensions and fractional shortenings for w1118 control, tinC-‐YCanAact, or Drosophila lines utilized in the deficiency screen without YCanAact. One deficiency, Df(3L)ED4421, exerts cardiac enlargement on its own. B) End-‐diastolic dimensions and fractional shortenings of OCT m-‐mode images for control w1118, tinC-‐Gal4>UAS-‐YCanAact, or Drosophila lines containing genetic disruptions in known suppressors of calcineurin: Mef2, sty, and CanB. Disruption of calcineurin suppressor genes alone did not result in a cardiac phenotype. C) End-‐diastolic dimensions and fractional shortenings for w1118 control, tinC-‐YCanAact, transposable element insertions in Galk, or the corresponding precise excision of Mi{ET1}GalkMB10638, Mi{ET1}Galkrev. The transposable elements themselves did not result in a cardiac enlargement phenotype. (All transgenes were heterozygous. *P<0.05, **P<0.0001 compared to w1118 control (first column), one-‐way ANOVA and post-‐hoc analysis with Bonferroni correction, all data represent mean ±SEM.)

Df(3L)E

D4416

Df(3L)B

SC130

Df(3L)E

D4421

Df(3L)E

D4414

Df(3L)E

D4415

Df(3L)E

D4413

Df(3L)B

SC170

Df(3L)B

SC390

tinC-YCanAact - + - - - - - - - - Deficiency - -

*Fr

actio

nal

Shor

teni

ng (%

)En

d-di

asto

lic

Dim

ensi

on (m

icro

ns) **

**

A

sty65 /+

Df(2R)X

1,Mef2

X1 /+

P{EP}C

anB2E

P774 /+

w11

18

sty65 /+

Df(2R)X

1,Mef2

X1 /+

P{EP}C

anB2E

P774 /+

w11

180

20

40

60

80

100

20

40

60

80

100

tinC-G

al4>U

AS-Can

Aac

t

tinC-G

al4>U

AS-Can

Aac

t

*

*

Frac

tiona

l Sh

orte

ning

(%)

End-

dias

tolic

D

imen

sion

(mic

rons

)

B

C

0

20

40

60

0

20

40

60

80

100

tinC-Y

CanAac

t /+

PBac{P

B}Galk

c038

48 /+

Mi{ET1}G

alkMB10

638 /+

Mi{ET1}G

alkrev /+

0

***

w11

18

tinC-Y

CanAac

t /+

PBac{P

B}Galk

c038

48 /+

Mi{ET1}G

alkMB10

638 /+

Mi{ET1}G

alkrev /+

w11

18

Frac

tiona

l Sh

orte

ning

(%)

End-

dias

tolic

D

imen

sion

(mic

rons

)

N= 9 32 4 14 11

N= 14 20 23 7 7

0

20

40

60

80

0

20

40

60

80

100

N= 35 31 13 15 10 23 44 31 37 16


Figure S4 Genetic disruption of Galk rescued tinC-‐YCanAact-‐induced decrease in life span. Survival curve of flies expressing tinC-‐YCanAact in the context of Mi{ET1}GalkMB10638. tinC-‐YCanAact flies had a decreased life span and the Mi{ET1}GalkMB10638 insertion in Galk improved survival of tinC-‐YCanAact flies. P<0.0001 for male and female tinC-‐YCanAact/+ compared to Mi{ET1}GalkMB10638/tinC-‐YCanAact or +/Mi{ET1}GalkMB10638, Mantel-‐Cox log rank test, N=60 in each group. All transgenes were heterozygous.

Perc

ent s

urvi

val (

%) 100

50

00 20 40 60 80

Days

Males

Perc

ent s

urvi

val (

%)

100

50

00 20 40 60 80

Days

FemalestinC-YCanAact/Mi{ET1}GalkMB10638

+/Mi{ET1}GalkMB10638

tinC-YCanAact/+


Figure S5 Disruption of Galk did not rescue TnI-‐mediated cardiac dilation. Quantification from OCT m-‐mode images of w1118 control (N=9), heterozygous TnI mutant (hdp2) with w1118 (N=16), heterozygous Mi{ET1}GalkMB10638 with w1118 (N=13), and heterozygous hdp2 and Mi{ET1}GalkMB10638 (N=22). Heterozygous hdp2 produced a relatively subtle cardiac dilation phenotype and is not rescued in the context of Mi{ET1}GalkMB10638. (Mi{ET1}Galk=Mi{ET1}GalkMB10638; Mi{ET1}Galk/+ = Mi{ET1}GalkMB10638 heterozygous with w1118; hdp2= TnI mutant; hdp2/+ = hdp2 heterozygous with w1118. *P<0.05, **P<0.01, one-‐way ANOVA with Bonferroni correction, all data represent mean ±SEM.)

0

20

40

60

80

w1118

hdp2/+

Mi{ET1}Galk/+

hdp2/+;+/Mi{ET1}Galk

0

5

10

15

0

20

40

60

80

100E

nd-d

iast

olic

Dim

ensi

on (m

icro

ns)

w1118

hdp2/+

Mi{ET1}Galk/+

hdp2/+;+/Mi{ET1}Galk

End

-sys

tolic

Dim

ensi

on (m

icro

ns)

Frac

tiona

l Sh

orte

ning

(%)

** **

* ** *

* ** *


Figure S6 Genetic disruption of Argk, Doc1, Doc2, and Doc3 did not rescue the tinC-‐CanAact phenotype. A-‐B) Summary data of end-‐diastolic dimensions and fractional shortenings for a transposable element in Argk. A transposable element in Argk did not suppress tinC-‐YCanAact-‐induced cardiac enlargement. C) qRT-‐PCR of flies heterozygous for a transposable element in Argk (PBac{WH}Argkf05255). Argk expression is decreased in PBac{WH}Argkf05255 flies. N=3 for each group. D) Summary data of end-‐diastolic dimensions and fractional shortenings for fly lines expressing the corresponding RNAi to dorsocross (Doc) genes. Knocking down Doc genes did not rescue tinC-‐FCanAact-‐induced cardiac enlargement, and caused cardiac enlargement with insertions in Doc2 and Doc3. Doc1v16746 RNAi= P{GD4334}v16746; Doc1JF02222 RNAi= P{TRiP.JF02222}; Doc2v37634 RNAi= P{GD4335}v37634; Doc3JF02223 RNAi= P{TRiP.JF02223}attP2. All transgenes were heterozygous. (*p<0.05, **P<0.01, ***P<0.001, ****P<0.0001, student t test for single comparisons and one-‐way ANOVA with Bonferroni correction for multiple comparisons. Data represent mean ±SEM.)

tinC-Gal4 - - + + + + + + + +

Doc1v16747 RNAi - - + + - - - - - - Doc1JF02222 RNAi - - - - + + - - - - Doc2v37634 RNAi - - - - - - + + - -

Doc3JF02223 RNAi - - - - - - - - + +

tinC-FCanAact - + - + - + - + - +

End

-dia

stol

ic

Dim

ensi

on (m

icro

ns)

Frac

tiona

lS

horte

ning

(%)

0

20

40

60

80

100

N= 12 65 13 15 22 17 35 17 37 24

***

*

Argk expression (Fold /control)

C

tinC-YCanAact - + - + - + - + PBac{PB}Argkf05255

- - + + - - + +

A B

020406080100

Frac

tiona

l S

horte

ning

(%)

End

-dia

stol

ic

Dim

ensi

on (m

icro

ns)

00.40.20.6

10.81.2

w1118 Argkf05255/+

D

N= 16 19 11 10

**

*

0

20

40

60

80

100

0

50

100

** ******** ****

***

****


Figure S7 Drosophila overexpressing Mef2 in the heart has an enlarged end-‐diastolic dimension. Quantification of end-‐diastolic dimension and fractional shortening from OCT m-‐mode images for tinC-‐Gal4 (driver only) heterozygous from cross with w1118 (N=11), UAS-‐Mef2 (no driver) heterozygous from cross with w1118 (N=4), and flies expressing tinC-‐Gal4>UAS-‐Mef2 (N=11). The tinC-‐Gal4>UAS-‐Mef2 fly heart had a significantly larger end-‐diastolic dimension when compared to driver only and no driver controls. In contrast, fractional shortening was not changed with cardiac Mef2 expression. (All transgenes were heterozygous. tinC-‐Gal4/+ = tinC-‐Gal4 heterozygous from cross with w1118; UAS-‐Mef2/+ = UAS-‐Mef2 heterozygous from cross with w1118. *p<0.01, **P<0.001 compared as indicated by the upper bar, one-‐way ANOVA with Bonferroni correction, all data represent mean ±SEM.)

0

20

40

60

0

20

40

60

80

100* ** N.S.

Frac

tiona

l Sh

orte

ning

(%)

End

-dia

stol

ic D

imen

sion

(mic

rons

)

tinC-Gal4/+

UAS-Mef2/+

tinC-Gal4>UAS-Mef2

tinC-Gal4/+

UAS-Mef2/+

tinC-Gal4>UAS-Mef2


Figure S8 A transposable element in galectin suppressed tinC-‐CanAact-‐induced cardiac enlargement. A) End-‐diastolic dimensions and B) fractional shortenings of Drosophila harboring a transposable element insertion in galectin in the context of tinC-‐YCanAact. The transposable element P{wHy}galectinDG25505 suppressed the tinC-‐YFP-‐CanAact-‐mediated cardiac enlargement. (*p<0.05, **p<0.0001, one-‐way ANOVA with Bonferroni correction. Data represent mean ±SEM.) C) While speculative, a possible mechanism that links Galk and galectin is by their action on galactose/GalNAc. Galactokinase regulates the phosphorylation of galactose/GalNAc, allowing for conversion of phosphorylated galactose/GalNAc into UDP-‐galactose/GalNAc. UDP-‐galactoase/GalNAc is then transferred on to cell surface receptor proteins that are recognized and bound by galectin, possibly altering signal transduction downstream of the receptor proteins to modify downstream cardiac enlargement.

020406080

100 * ***

tinC-YCanAact - + - + - + - + P{wHy}galectinDG25505 - - + + - - + +

Frac

tiona

l S

horte

ning

(%)

End

-dia

stol

ic

Dim

ensi

on (m

icro

ns)

N= 39 23 17 19

A B

020406080

100** * *

C


Table S1 End-‐systolic dimensions for all figures.

N

Average (microns) SEM

P value (vs. w1118)

P value (vs. YCanA)

Figure 1 w1118 16 1.9 0.79

**

YCanA/+ 19 13.06 2.79 **

Figure 2 w1118 35 3.66 1.25 ****

YCanA/+ 31 24.3 2.39 ****

Df(3L)ED4416/YCanA 45 4.9 1.15 ****

Df(3L)BSC130/YCanA 29 3.29 1.22 ****

Df(3L)ED4421/YCanA 17 12.54 3.1 *

Df(3L)ED4414/YCanA 38 20.6 2.06 ****

Df(3L)ED4415/YCanA 51 16.03 1.51 **** *

Df(3L)ED4413/YCanA 37 15.08 1.87 *** *

Df(3L)BSC170/YCanA 23 15.57 3.22 **

Df(3L)BSC390/YCanA 19 23.65 3.175 ****

Figure 3 w1118 35 3.66 1.25

****

UAS-‐YCanA/+ (no driver) 31 24.3 2.39

****

tinC-‐Gal4/+ (driver only) 45 4.9 1.15

****

tinC-‐Gal4>UAS-‐YCanA 29 3.29 1.22 **** tinC-‐Gal4>UAS-‐YCanA_Mef2 17 12.54 3.1

****

tinC-‐Gal4>UAS-‐YCanA_Sty 39 21.03 2.06

**

tinC-‐Gal4>UAS-‐YCanA_CanB 51 16.03 1.51

****

tinC-‐Gal4>UAS-‐YCanA_9221 38 14.72 1.85

****

tinC-‐Gal4>UAS-‐YCanA_9295 23 15.57 3.22

Figure 4AB w1118 14 3.71 1.19

*

tinC-‐YCanA/+ 20 12.42 1.97 * tinC-‐YCanA/pBac{PB}Galk^c03848 25 5.90 1.33

*

tinC-‐YCanA/Mi{ET1}Galk^MB10638 21 2.00 0.71

***

tinC-‐YCanA/Mi{ET1}Galk^rev 14 20.28 3.29 **** *

Figure 4CD w1118 14 3.71 1.19

****

tinC-‐YCanA/+ 20 18.45 2.52 **** UAS-‐Galk RNAi/+ (no driver) 9 8.60 1.82

tinC-‐Gal4/+ (driver only) 7 5.47 1.46

*


tinC-‐Gal4>UAS-‐Galk RNAi (RNAi only) 7 5.79 3.49

*

tinC-‐YCanA_tinC-‐Gal4>UAS-‐Galk RNAi 17 7.72 2.36

** YCanA=YCanAact, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, one-‐way ANOVA with Bonferroni correction.


Supporting Methods

Drosophila stocks

The Drosophila stocks P{GD4334}v16746, P{TRiP.JF02222}, P{GD4335}v37634,

P{TRiP.JF02223}attP2, and P{wHy}galectinDG25505 were obtained from the Bloomington Stock

center. The Drosophila stock PBac{PB}Argkf05255 was obtained from the Exelixis Collection at

Harvard Medical School. w1118 (FBst0006326) was kindly provided by Dr. Eric Spana, Duke

University Model Systems Genomics Facility. All experiments in the main body and text refer to

the w1118 stock FBst0003605 that was routinely used in our lab unless otherwise stated. The

tinC-‐YCanAact lines 32-‐1 (on chromosome 3), 13-‐1 (on chromosome 2), and 16-‐1 (on

chromosome 3) and tinC-‐FCanAact lines 25-‐4 (on chromosome 1), 14-‐4 (double insertions on

chromosomes 2 and 3), and 33-‐3 (on chromosome 3) were generated as separate lines from

injecting w1118 (FBst0006326) fly embryos from the Duke University Model Systems Genomics

Facility. Drosophila lines 32-‐1 (on chromosome 3) for tinC-‐YCanAact and 25-‐4 (on chromosome

1) for tinC-‐FCanAact were used in the main text and figures for screening and rescue

experiments. The UAS-‐Mef2 line was kindly provided by Dr. Michael Taylor (BOUR et al. 1995;

GUNTHORPE et al. 1999). The TnI mutant hdp2 was a kind gift from Dr. Jim Vigoreaux and had

previously been shown to induce flight muscle (BEALL and FYRBERG 1991) and cardiac

enlargement (WOLF et al. 2006) phenotypes.

Confocal microscopy

Flies that harbored tinC-‐GFP or tinC-‐YCanAact were collected 1-‐2 days after eclosion, and

dissected as previously described (OCORR et al. 2007; YU et al. 2010) by removing the ventral

Wendy

Typewritten Text

File S1

Wendy

Typewritten Text

Wendy

Typewritten Text


side of the abdomen and exposing the heart tube. Immunohistochemistry was performed

according to standard procedures as previously described (ALAYARI et al. 2009; YU et al. 2010)

using 5% BSA in TBS as the blocking reagent with a rabbit αGFP antibody (Invitrogen, Inc.), and

Alexa Fluor488 goat αrabbit secondary antibody (Invitrogen, Inc.). After immuno-‐staining, the

Drosophila heart was visualized under a Zeiss LSM510 confocal microscope. Consecutive Z-‐

stacks were taken at 1µm intervals, and projected to form a 3D image in the LSM image

browser software.

Cardiac cell number

Fly hearts expressing tinC-‐nuclear RFP in the presence of either tinC-‐GFP or tinC-‐YCanAact

were imaged with confocal microscopy. Ostia cells (Os) along with the cuticle pattern mark the

boundaries to abdominal compartments (A1-‐A3). tinC-‐GFP or tinC-‐YCanAact flies were crossed

with tinC-‐Gal4, UAS-‐nuclear RFP flies and the offspring were collected 1-‐2 days after eclosion.

Hearts were exposed by dissection as previously described for confocal microscopy and imaged

without staining under the confocal microscope (OCORR et al. 2007; YU et al. 2010). Nuclei from

visualization of RFP were counted to represent the number of cardiomyocytes present in each

heart.

Life span

Drosophila were collected 1-‐2 days after eclosion. 10 males and 10 females were kept in

the same vial and food changed every 2-‐3 days. A total of 6 vials were monitored for every


group. The number of remaining flies was monitored daily. Data was analyzed using GraphPad

Prism software and statistical analysis was performed with a Mantel-‐Cox test for significance.


Supporting References

ALAYARI, N. N., G. VOGLER, O. TAGHLI-‐LAMALLEM, K. OCORR, R. BODMER et al., 2009 Fluorescent

labeling of Drosophila heart structures. J Vis Exp.

BEALL, C. J., and E. FYRBERG, 1991 Muscle abnormalities in Drosophila melanogaster heldup

mutants are caused by missing or aberrant troponin-‐I isoforms. J Cell Biol 114: 941-‐951.

BOUR, B. A., M. A. O'BRIEN, W. L. LOCKWOOD, E. S. GOLDSTEIN, R. BODMER et al., 1995 Drosophila

MEF2, a transcription factor that is essential for myogenesis. Genes Dev 9: 730-‐741.

GUNTHORPE, D., K. E. BEATTY and M. V. TAYLOR, 1999 Different levels, but not different isoforms, of

the Drosophila transcription factor DMEF2 affect distinct aspects of muscle

differentiation. Dev Biol 215: 130-‐145.

OCORR, K., N. L. REEVES, R. J. WESSELLS, M. FINK, H. S. CHEN et al., 2007 KCNQ potassium channel

mutations cause cardiac arrhythmias in Drosophila that mimic the effects of aging. Proc

Natl Acad Sci U S A 104: 3943-‐3948.

WOLF, M. J., H. AMREIN, J. A. IZATT, M. A. CHOMA, M. C. REEDY et al., 2006 Drosophila as a model for

the identification of genes causing adult human heart disease. Proc Natl Acad Sci U S A

103: 1394-‐1399.

YU, L., T. LEE, N. LIN and M. J. WOLF, 2010 Affecting Rhomboid-‐3 function causes a dilated heart in

adult Drosophila. PLoS Genet 6: e1000969.

Galactokinase Is a Novel Modi er of Calcineurin-Induced ... · wing vein phenotype induced by the...

Documents

Transcript of Galactokinase Is a Novel Modi er of Calcineurin-Induced ... · wing vein phenotype induced by the...