Slow Skeletal Troponin I Gene Transfer, Expression, and Myofilament Incorporation Enhances Adult Cardiac Myocyte Contractile Function

Proc. Natl. Acad. Sci. USA Vol. 94, pp. 5444–5449, May 1997 Physiology

Slow skeletal troponin I gene transfer, expression, and myofilament incorporation enhances adult cardiac myocyte contractile function

MARGARET V. WESTFALL*, ELIZABETH M. RUST, AND JOSEPH M. METZGER

Department of Physiology, School of Medicine, University of Michigan, Ann Arbor, MI, 48109-0622

Communicated by James A. Spudich, Stanford University, Stanford, CA, March 7, 1997 (received for review November 5, 1996)

ABSTRACT The functional significance of the develop- changes remains unclear, because developmental transitions in mental transition from slow skeletal troponin I (ssTnI) to other contractile protein isoforms occur over the same time cardiac TnI (cTnI) isoform expression in cardiac myocytes interval as the TnI isoform transition. remains unclear. We show here the effects of adenovirus- TnI isoforms are also postulated to influence cardiac myo- mediated ssTnI gene transfer on myofilament structure and filament pH sensitivity (4, 5). Contractile function decreases function in adult cardiac myocytes in primary culture. Gene markedly during acute myocardial ischemia (4), and acidosis transfer resulted in the rapid, uniform, and nearly complete plays a significant role in this decreased function by reducing replacement of endogenous cTnI with the ssTnI isoform with myofilament Ca2ϩ sensitivity (4–7). Solution studies indicate no detected changes in sarcomeric ultrastructure, or in the that TnI may play a role in this phenomenon because acidic isoforms and stoichiometry of other myofilament proteins pH-induced decreases in Ca2ϩ binding to and͞or subsequent compared with control myocytes over 7 days in primary conformational changes within troponin C (TnC) are in- culture. In functional studies on permeabilized single cardiac creased in magnitude in the presence of TnI (8). Studies using myocytes, the threshold for Ca2؉-activated contraction was permeabilized muscle preparations also have implicated TnI in significantly lowered in adult cardiac myocytes expressing this pH effect. For example, the acidosis-induced shift in ssTnI relative to control values. The tension–Ca2؉ relation- myofilament Ca2ϩ sensitivity is greater in cTnI-expressing ship was unchanged from controls in primary cultures of adult myocardium than in slow skeletal muscle fibers (9, 10) or cardiac myocytes treated with adenovirus containing the adult fetal͞neonatal myocardium (3, 5), which express ssTnI (1, 2). cardiac troponin T (TnT) or cTnI cDNAs. These results Together, these results suggest that TnI influences the myo- -indicate that changes in Ca2؉ activation of tension in ssTnI- filament Ca2ϩ sensitivity response to acidosis in an isoform expressing cardiac myocytes were isoform-specific, and not dependent manner. However, direct confirmation that TnI due to nonspecific functional changes resulting from overex- isoforms influence the acidosis-mediated shift in Ca2ϩ sensi- pression of a myofilament protein. Further, Ca2؉-activated tivity is difficult to assess in cardiac myocytes because other tension development was enhanced in cardiac myocytes ex- contractile protein isoforms also differ across muscle lineages pressing ssTnI compared with control values under conditions and during myocardial development (11). There are no pre- mimicking the acidosis found during myocardial ischemia. vious reports documenting the effect of ssTnI alone on Ca2ϩ- These results show that ssTnI enhances contractile sensitivity activated tension in adult cardiac myocytes. to Ca2؉ activation under physiological and acidic pH condi- Thus, the major goal of the present study is to define the tions in adult rat cardiac myocytes, and demonstrate the functional role of ectopically expressed ssTnI in adult cardiac utility of adenovirus vectors for rapid and efficient genetic myocytes as a means of further understanding the function of modification of the cardiac myofilament for structure͞ TnI isoforms within the contractile apparatus at physiological function studies in cardiac myocytes. and acidic pH. Recombinant adenovirus is used to deliver a ssTnI expression cassette into adult cardiac myocytes in pri- The thin filament protein troponin I (TnI) plays an essential mary culture. Western blot and immunohistochemistry anal- role in the regulation of striated muscle contraction. One yses of cultured myocytes show rapid and efficient expression approach to understanding the role of TnI within the contrac- and incorporation of ssTnI into the myofilament. Further- tile apparatus is to study functional differences among TnI more, ectopic ssTnI expression occurs without any detectable isoforms. TnI isoforms are expressed in a muscle lineage- effects on the isoform composition or stoichiometry of other specific and developmentally regulated pattern (1, 2). In contractile assembly proteins, or on the highly ordered struc- cardiac muscle, there are two developmentally regulated TnI ture of the contractile lattice. Importantly, functional mea- isoforms. The slow skeletal TnI (ssTnI) isoform is expressed in surements of Ca2ϩ-activated tension under physiological- fetal myocardium and the cardiac TnI (cTnI) isoform is activating conditions directly demonstrate, for the first time to expressed exclusively in adult myocardium (1, 2). During our knowledge, that TnI isoforms influence adult cardiac cardiac development, there are also marked alterations in myocyte contractile function. Specifically, we observe a lower cardiac contractile function, including an increase in the threshold for Ca2ϩ activation in adult cardiac myocytes ex- steepness and a rightward shift in the tension-Ca2ϩ relation- pressing ssTnI, a phenotype similar to that of fetal͞neonatal ship (1, 3). Thus, there is a correlation between TnI isoform myocardium (1, 3). Functional studies further show that pH expression and changes in myofilament function. However, the sensitivity is reduced in the adult cardiac myocytes expressing precise role of TnI isoforms in mediating these functional

Abbreviations: CMV, cytomegalovirus; cTnI, cardiac troponin I; LC1, The publication costs of this article were defrayed in part by page charge myosin light chain 1; LC2, myosin light chain 2; mu, map unit; nH, Hill 2ϩ payment. This article must therefore be hereby marked ‘‘advertisement’’ in coefficient; pCa, Ϫlog[Ca ]; pCa50, pCa required to produce 50% of accordance with 18 U.S.C. §1734 solely to indicate this fact. maximal tension; ssTnI, slow skeletal troponin I; TnC, troponin C; TnI, troponin I; TnT, troponin T; Tm, tropomyosin. Copyright ᭧ 1997 by THE NATIONAL ACADEMY OF SCIENCES OF THE USA *To whom reprint requests should be addressed at: Department of 0027-8424͞97͞945444-6$2.00͞0 Physiology, University of Michigan, 1301 East Catherine St., 7730 PNAS is available online at http:͞͞www.pnas.org. Medical Science II, Ann Arbor, MI 48109-0622.

5444 Downloaded by guest on September 29, 2021 Physiology: Westfall et al. Proc. Natl. Acad. Sci. USA 94 (1997) 5445

ssTnI, indicating that TnI isoform expression significantly mg͞ml; Sigma) for 15 min. After increasing the [Ca2ϩ]inthe influences the myofilament response to acidosis. digestion solution (1 mM), and continuing the perfusion for 15 min, isolated ventricles were minced into pieces and gently METHODS shaken in digestion solution with occasional trituration using silanized pasteur pipets. Undigested ventricular tissue was Generation of Adenoviral Vectors. To construct recombi- removed using a 230-␮m mesh sieve. The cell suspension was nant adenovirus vectors, the plasmids pAdCMVssTnI and centrifuged, and the [Ca2ϩ] was increased to 1.75 mM after pJM17 (12) were cotransfected by calcium phosphate into a resuspending pelleted cells in KHB-A containing 2% BSA. HEK 293 cell line (Fig. 1A). The shuttle plasmid pAdCM- After centrifuging cells again, they were resuspended in culture VssTnI contained adenovirus serotype 5 sequences 0–1 map media containing serum [DMEM containing 50 units͞ml units (mu) and 9–16 mu flanking an expression cassette penicillin plus 50 ␮g͞ml streptomycin (P͞S), and 5% fetal containing the cytomegalovirus (CMV) promoter, the coding bovine serum]. Rod-shaped ventricular myocytes were plated sequence for the full-length rat ssTnI cDNA and its 3Ј un- onto laminin-coated coverslips in DMEMϩP͞Sϩfetal bovine translated region (13), and the simian virus 40 polyadenylyla- serum for 2 hr and then infected with recombinant adenovirus tion signal. The pJM17 plasmid was a 0–100 mu derivative of (Ϸ170 plaque-forming units͞cell) in serum-free media adenovirus serotype 5 containing a partial deletion in the E3 (DMEMϩP͞S). Serum-free medium (2 ml) was added 1 hr region and a 4.3-kb pBRX insert at 3.7 mu. The insert allowed later and changed every 2–3 days thereafter. replication of the plasmid in bacteria but made the viral Analysis of Protein Composition by Gel Electrophoresis and genome too large to be packaged into the virus capsid (12). Immunoblotting. Myofilament proteins were analyzed by col- After homologous recombination, the expression cassette in lecting 10 to 20 ventricular myocytes on a glass micropipet tip pAdCMVssTnI replaced the pBRX insert and the E1 region of and then transferring the myocytes to microcentrifuge tubes the genome (1–9 mu) in pJM17, thus making the recombinant containing 10 ␮l of sample buffer for separation by gel adenovirus, AdCMVssTnI, capable of being packaged but electrophoresis as described previously (17). For Western blot replication defective. The HEK 293 cell line is adenovirus analysis, glass micropipets were used to scrape cultured ven- serotype 5-transformed and expresses the E1 region (1–11.3 tricular myocytes from coverslips into 10 ␮l of sample buffer mu) in trans (12), which allows for replication of the recom- 1 to 7 days after plating. Samples were separated by gel binant adenovirus. A similar strategy was used to generate electrophoresis, and then transblotted onto poly(vinylidene AdCMVcTnI and AdCMVaTnT, which contain the adult rat difluoride) membrane for 2,000 V⅐h with immunodetection cTnI and cardiac troponin T (TnT) cDNAs described previ- carried out as described by Westfall et al. (18). Gels and blots ously by Murphy et al. (13) and Jin and Lin (14), respectively. were scanned with an Arcus II laser densitometer (AGFA- To harvest high titer recombinant virus, infected HEK 293 Gevaert NV;) and analyzed with Molecular Analyst software cells were collected and lysed by three cycles of freezing and (Bio-Rad). TnI, TnT, and tropomyosin (Tm) isoform compo- thawing. After removing cellular debris, the supernatant of the sition were determined using the primary TI-4 mAb (1:250; ref. viral lysate was stored at Ϫ20ЊC. Southern blot analysis was 2), JLT-12 mAb (1:200; Sigma), and CH-1 mAb (1:100; Sigma), used to identify recombinant adenovirus as described by respectively. Sambrook et al. (15), and the blot identifying ssTnI is shown Indirect Immunohistochemistry in Single Cardiac Myo- in Fig. 1B. cytes. A dual-mAb protocol (18) was used to determine the Primary Cultures of Rat Ventricular Cardiac Myocytes. extent of thin filament remodeling resulting from ectopic ssTnI Ventricular myocytes were isolated using a protocol adapted expression within single cardiac myocytes in primary culture. from Haworth et al. (16). Briefly, hearts from heparinized, The two primary mAbs used were cardiac-specific TI-1 mAb female Sprague–Dawley rats (200 g) were mounted on a and TI-4, a mAb recognizing all striated isoforms of TnI (2). modified Langendorff apparatus, perfused for 5 min with Anti-cTnI antibody (TI-1; 1:1,000) binding was detected with Krebs–Henseleit buffer (pH 7.40; 118 mM NaCl͞4.8 mM goat anti-mouse IgG mAb conjugated to Texas Red (1:100), KCl͞25 mM Hepes͞1.25 mM K2HPO4͞1.25 mM MgSO4͞11 whereas TI-4 mAb (1:500) binding was detected with fluores- mM glucose) containing 1 mM CaCl2 (KHB-A), followed by cein isothiocyanate-conjugated secondary antibody (1:200). KHB lacking added Ca2ϩ (KHB-B) for 5 min, and then with Immunofluorescence was analyzed on a Leitz Aristoplan 60 ml of recirculating KHB-B perfusate containing collage- microscope, and representative cells were photographed using nase (0.5 mg͞ml; type II, Worthington) and hyaluronidase (0.2 a Bio-Rad MRC 600 confocal microscope. -Characterization of Ca2؉-Sensitive Tension in Single Car diac Myocytes. Complete details of the experimental chamber and attachment procedure for mounting single cardiac myocytes and soleus fibers and for permeabilization and measure- ment of Ca2ϩ-activated tension have been reported elsewhere (9, 17). Statistics. An analysis of variance was used to test for significant differences between groups, with a post hoc Stu- dent–Newman–Keuls multiple comparison test to determine significance.

RESULTS FIG. 1. Generation and structural characterization of the recom- Before studying the effect of TnI isoforms on contractile binant adenovirus vector AdCMVssTnI. (A) Scheme for creating the function, it was first important to demonstrate that adenovirus- recombinant adenovirus AdCMVssTnI with plasmid pJM17 and plas- mediated gene transfer leads to incorporation of an exogenous mid pAdCMVssTnI, which contains the rat ssTnI expression cassette. myofilament protein into the contractile apparatus, which has (B) Recombinant AdCMVssTnI was identified by Southern blot analysis of viral lysates. Lane 1, undigested recombinant AdCMVssTnI previously been difficult to establish (19). To this end we used DNA; lane 2, BglII digest, and lane 3, BamHI digest. The BglII excised two immunological assays to assess transgene expression and the expected 1,400-bp fragment, containing the expression cassette incorporation into the myofilaments of adult cardiac myocytes. and the left end of the adenovirus genome, whereas digestion with First, Western blots were used to directly demonstrate ssTnI BamHI excised the full-length ssTnI cDNA of predicted size (800 bp). expression and incorporation into the contractile apparatus of Downloaded by guest on September 29, 2021 5446 Physiology: Westfall et al. Proc. Natl. Acad. Sci. USA 94 (1997)

AdCMVssTnI-treated adult cardiac myocytes (Fig. 2). In this assay, ssTnI content increased from 30% of total TnI content at 3 days to 76% at 4 days and 97% at 5 days, with high levels of expression maintained for at least 7 days post-AdCMVssTnI infection. This time course closely follows the 3.2-day half-life of TnI in cardiac muscle (20). Thus, from day 4 onward ssTnI was the dominant TnI isoform in the AdCMVssTnI-treated adult cardiac myocytes. In addition, ssTnI expression on Western blots was not different for permeabilized myocytes and membrane-intact samples, which was evidence that ssTnI was incorporated into the thin filament regulatory complex and was not simply accumulating in the cytoplasmic space. Western blots also demonstrated that the cTnI isoform was exclusively expressed throughout the culture period in controls (Fig. 2) and in cardiac myocytes treated with nonrecombinant adenovirus (results not shown). Densitometric analysis of the blots revealed that the stoichiometry of TnI expression was FIG. 3. TnI remodeling in adult single cardiac myocytes. Repre- unchanged from controls throughout the culture period (Fig. sentative confocal images of indirect immunostaining of TnI in control 2). Thus, culture conditions and͞or nonrecombinant adeno- (A and C) and AdCMVssTnI-infected (B and D) ventricular myocytes virus infection alone did not alter TnI isoform expression cultured for 4 (A and B)and7(Cand D) days. TI-4 mAb labeling and and͞or remodeling in adult cardiac myocytes. cardiac specific TnI labeling with TI-1 mAb are shown in the left and In the second assay, dual immunostaining with the cTnI- right portions of A-D, respectively. (Bar ϭ 17 ␮m.) specific TI-1 mAb and TI-4 mAb, which recognizes all striated isoforms of TnI, allowed cTnI to be distinguished from other mAb-associated immunofluorescence was much greater than TnI isoforms within a single ventricular myocyte. Adult cardiac the fluorescence observed after incubation with TI-1 mAb myocytes expressing cTnI stained positive with TI-1 and TI-4 (18). Taken together, these results indicate that TI-1 mAb mAbs (Fig. 3 A and C), whereas slow-twitch soleus muscle fluorescence is specifically associated with cTnI expression in fibers expressing ssTnI stained positive with TI-4 mAb, but not the myofilament. with TI-1 mAb (18). In fetal mouse ventricular myocytes, The TI-1 and TI-4 mAb immunostaining profiles of control which primarily express ssTnI, the area and intensity of TI-4 rat cardiac myocytes in primary culture were positive and remained invariant throughout the entire 7-day culture period (n Ϸ 500 myocytes per day for days 1 and 4–7). In addition, confocal image analysis of staining with both antibodies showed the characteristic, periodic immunostaining pattern of striated muscle (Fig. 3 A and C). These results indicate that the cTnI isoform was present within the cardiac contractile assembly throughout the culture period. In marked contrast, the immunostaining profile of AdCMVssTnI-treated cardiac myocytes progressively changed after gene delivery (Fig. 3 B and D). Cardiac myocytes staining positive for cTnI with TI-1 mAb decreased from 96.4% at 4 days (n ϭ 1,084), to 87.6% at 5 days (n ϭ 1,110), 39.6% at 6 days (n ϭ 897), and 9.9% at 7 days (n ϭ 1140) after ssTnI gene transfer. Importantly, the TI-4 mAb immunostaining pattern in these cardiac myocytes did not change during this time and was indistinguishable from controls (Fig. 3). This is evidence that the endogenous cTnI isoform was effectively replaced by the exogenous ssTnI gene product. A diffuse or irregular immunostaining pattern would be expected if ssTnI was nonspecifically distributed throughout the cell, a result that was not obtained. To determine whether adenovirus alone influenced TnI isoform expression, we also examined the immunostaining pattern in myocytes treated with nonrecombinant adenovirus serotype 5 virus (n ϭ 138) FIG. 2. Western blot analysis of regulatory protein isoforms in and found the pattern of staining with both anti-TnI antibodies control and AdCMVssTnI-treated adult cardiac myocytes. (A) TnI was unchanged from controls during the entire culture period. isoform composition. Labels on top indicate days in primary culture We also tested for possible alterations in the isoform for control myocytes while labels on bottom indicate days in culture for composition and stoichiometry of other key myofilament AdCMVssTnI-infected myocytes. A soleus slow skeletal fiber is shown as a positive control for detection of ssTnI. (B) TnT and Tm isoform proteins. Sarcomere organization, myofilament protein ex- composition. The day in culture (days) and experimental group pression, and stoichiometry were unchanged in control ven- (control ϭϪ; AdCMVssTnI ϭϩ) are labeled. These results (TnI, tricular myocytes for at least 7 days in primary culture (un- TnT, and Tm) are obtained from membrane-intact samples but also published data). These characteristics were also stable in represent the pattern observed in permeabilized preparations (results ventricular myocytes treated with adenovirus carrying a re- not shown). A silver (Ag)-stained portion of the gel was used to porter gene (unpublished data). In the present study, immu- normalize data for variations in protein loaded per lane. The noblot analysis showed that the isoform expression and stoi- cTnIϩssTnI͞Ag stain ratio in AdCMVssTnI-treated myocytes over 7 chiometry of the adult isoforms of TnT and Tm were identical days (0.94 Ϯ 0.25, n ϭ 4) is not significantly different (P Ͼ 0.05) from in control and AdCMVssTnI-treated cardiac myocytes (Fig. 2). control myocyte values (1.56 Ϯ 0.28, n ϭ 4). The TnT͞Ag stain and Tm͞Ag stain ratios in AdCMVssTnI-infected myocytes over 7 days Further, SDS͞PAGE analysis showed normal expression of also are unchanged from controls (TnT͞Ag stain: control ϭ 1.27 Ϯ myosin heavy chain and normal isoforms and conserved 0.11, AdCMVssTnI ϭ 1.08 Ϯ 0.07, n ϭ 4; Tm͞Ag stain: control ϭ stoichiometry of TnC, and myosin light chains 1 (LC1) and 2 0.54 Ϯ 0.09, AdCMVssTnI ϭ 0.55 Ϯ 0.17; n ϭ 4). (LC2) in AdCMVssTnI-treated cardiac myocytes (Fig. 4 Right). Downloaded by guest on September 29, 2021 Physiology: Westfall et al. Proc. Natl. Acad. Sci. USA 94 (1997) 5447

FIG.4. (Right) Representative SDS͞PAGE gel analysis of myofilament isoform composition and stoichiometry of AdCMVssTnI- treated (lane 1) and control (lane 2) ventricular myocytes. Gels indicate that isoform expression of myosin heavy chain, TnC, LC1, and LC2 are not altered by ssTnI gene transfer. Laser-based densitometry was performed, and the integrated peak for each protein was used to evaluate myofilament stoichiometry. Ratios for TnC͞ (TnCϩLC1ϩLC2) were 0.06 and 0.08 in control and AdCMVssTnI- treated myocytes, respectively. Ratios for LC1 (LC1͞(LC1ϩLC2) and LC2 (LC2͞(LC1ϩLC2) were 0.56 and 0.45 in control myocytes and, 0.58 and 0.42 in AdCMVssTnI-treated myocytes, respectively. (Left) Transmission electron micrograph of a cardiac myocyte 3 days after adenovirus-mediated ssTnI gene transfer. The average sarcomere length of this cardiac myocyte is 1.7 ␮m. (Inset) Electron micrograph of a freshly isolated control cardiac myocyte with an average sarcomere length of 1.8 ␮m. Examination of other electron micrographs up FIG. 5. Enhanced tension generation in AdCMVssTnI-treated to 5 days postinfection with AdCMVssTnI showed no differences in single, adult cardiac myocytes in culture. (A) Original fast time-base sarcomeric ultrastructure compared with control myocytes in primary recordings of Ca2ϩ-activated isometric tension development in a culture. Ventricular myocytes in primary culture were fixed, embed- control (a) and an AdCMVssTnI-treated (b) single cardiac myocyte. ded, and mounted as described previously (3). In a, the pCa of the activating solution are 4.0, 6.0, 5.8, 5.3, 4.5, 4.0, 5.0, 4.6, 4.4, and 4.0, respectively. (B) pCas are 4.0, 7.0, 6.6, 6.3, 6.0, These results are consistent with targeted remodeling of the 4.0, 5.7, 5.2, 4.8, and 4.0, respectively. Records i-v in a and b were contractile apparatus in the absence of other changes in obtained at pH 7.00, and records vi-x are obtained at pH 6.20. Active myofilament composition. Finally, transmission electron mi- tension was obtained by subtracting resting tension in relaxing crographs (Fig. 4 Left) demonstrated normal sarcomeric struc- solution 9.0 (b, xi) from total tension (see arrow in a, i) at each pCa. ture in AdCMVssTnI-treated myocytes, indicating that gene On average, tension was determined at eight different pCas at pH transfer did not alter the highly ordered myofibrillar lattice. 7.00 and 6.20. Maximum Ca2ϩ-activated tension is 29 kN͞m2 in Taken together, these results provide strong evidence that the control and 30 kN͞m2 in the AdCMVssTnI-treated myocytes shown. 2 2 cardiac myocyte contractile apparatus is preserved after ade- Vertical calibration bar is 27 kN͞m for a and 19.7 kN͞m for b.A summary of the tension-pCa relationships are shown for control (B) novirus-mediated transfer of the ssTnI cDNA, in agreement and AdCMVssTnI-treated (C) myocytes at pH 7.0 (F) and pH 6.20 with earlier transgenic studies on other myofilament genes (21, (E). Active tension (P) at each submaximal pCa is expressed as a 22). This point is important because myofibrillar disarray is fraction of the maximum active, isometric tension at pCa 4.0 (Po)in associated with overexpression of a normal or mutant myofil- each myocyte. The shape and position of the tension–pCa relation- ament protein in the indirect flight muscle of Drosophila ship in controls cultured for 6 days are unchanged from values melanogaster (23). Based on our results and those from trans- obtained from acutely isolated myocytes. Thus, control data shown genic experiments (21, 22), it appears that cardiac myocytes in B are pooled from control myocytes cultured for 0–6 days (n ϭ exquisitely regulate the amount͞content of each contractile 11–12 observations͞point). Results for AdCMVssTnI-treated single cardiac myocytes are from 6-day primary cultures (n ϭ 11–14 protein within the cell. observations per point). (D) Summary of pCa50 in control and The demonstration of TnI remodeling independent of other AdCMVssTnI-treated cardiac myocytes and in single soleus skeletal alterations in the contractile machinery allowed us to examine muscle fibers at pH 7.00 (empty bars) and pH 6.20 (filled bars). (E) 2ϩ the effects of TnI isoform composition on Ca -activated Summary of ⌬pCa50 (pCa50 pH 7.00 Ϫ pCa50 at pH 6.20) in con- tension in adult cardiac myocytes. Functional studies focused trol (n ϭ 12) and AdCMVssTnI-treated (n ϭ 11), and AdCM- on cardiac myocytes at day 6 postinfection because immuno- VaTnT-treated (n ϭ 4; day 6) cardiac myocytes and in single soleus fluorescence and Western blot analysis demonstrated marked skeletal muscle fibers (n ϭ 5). Values in B-E are expressed as incorporation of ssTnI at this time point. Representative mean Ϯ SEM. The pCa50 and nH (see Results) values are derived isometric tension recordings over a range of Ca2ϩ concentra- from the tension-pCa curves as described in Methods, and n indicates the number of observations per point for the tension-pCa relation- tions are shown in Fig. 5A for control and AdCMVssTnI- ship in B and C. Cross indicates significantly different from control treated ventricular myocytes. Normalized maximum isometric at pH 7.00 (D; P Ͻ 0.01). Asterisk indicates significantly different tension was not different in AdCMVssTnI-treated cardiac (P Ͻ 0.001) from control (E) or different from control at pH 6.20 myocytes compared with control values (Fig. 5A) or to values (D). Downloaded by guest on September 29, 2021 5448 Physiology: Westfall et al. Proc. Natl. Acad. Sci. USA 94 (1997)

obtained in acutely isolated myocytes (24). The similar max- DISCUSSION imum tension values for control and adenovirus-treated myo- We report here that ssTnI gene transfer results in two signif- cytes provides functional evidence that myofilament content icant alterations in the Ca2ϩ-activated mechanical function of and sarcomere architecture are maintained after transfer of adult cardiac myocytes. First, the threshold for Ca2ϩ-activated the ssTnI cDNA into cardiac myocytes. tension and molecular cooperativity are lowered in these The tension recordings also were used to derive the tension- 2ϩ myocytes. Tension development is evident over the pCa range Ϫlog[Ca ] (pCa) relationship, and a summary of this rela- of 7.0–6.5 in ssTnI-expressing cardiac myocytes, whereas over tionship in control and AdCMVssTnI-treated adult single this same Ca2ϩ range the cTnI-expressing adult cardiac myo- cardiac myocytes is shown in Fig. 5 B and C. A significant cyte is fully relaxed. Second, the acidic pH-induced desensi- change at physiological pH (7.00) was observed in nH (Fig. 5 tization of Ca2ϩ-activated tension is markedly reduced in B and C), which is an index of myofilament cooperativity (ref. ssTnI-expressing cardiac myocytes and is similar to that ob- 17 and references therein). The nH was significantly lower in tained in slow soleus fibers that also express ssTnI. These AdCMVssTnI-treated (nH ϭ 1.6 Ϯ 0.2; P Ͻ 0.05) cardiac effects appear to be specific, as ssTnI gene transfer results in myocytes compared with controls (nH ϭ 3.0 Ϯ 0.2) and was rapid and efficient expression and incorporation of ssTnI into similar to the nH in soleus fibers (6, 9), and in fetal (3) and the cardiac myofilament without detectable effects on the neonatal myocardium (1), which also express ssTnI (1, 2). stoichiometry or isoforms of other key contractile͞regulatory Further analysis of the tension-pCa relationship demonstrated proteins, or on sarcomere ultrastructure. Finally, normalized that the pCa required to produce 50% of maximal tension maximum Ca2ϩ-activated tension is not different between the 2ϩ (pCa50), a measure of myofilament Ca sensitivity, was ssTnI- and cTnI-expressing adult cardiac myocytes, indicating increased in AdCMVssTnI-treated cardiac myocytes com- that peak mechanical performance is not altered by ssTnI gene pared with untreated controls (Fig. 5D). The pCa50 and nH in transfer. myocytes treated with nonrecombinant adenovirus (results not The overall effect of the ssTnI-induced changes in cardiac shown) were not significantly different from untreated control myocyte contractile function at physiological pH is to lower the 2ϩ values. In a separate set of control experiments, cTnI gene threshold for Ca -activated tension without substantial 2ϩ transfer had no effect on the tension-pCa relationship after 7 changes in tension development at high Ca concentrations. A reduction in the threshold for Ca2ϩ activation may result days (control: pCa50 ϭ 5.71 Ϯ 0.04, nH ϭ 2.5 Ϯ 0.4, n ϭ 6; AdCMVcTnI: pCa ϭ 5.77 Ϯ 0.02, n ϭ 2.6 Ϯ 0.4, n ϭ 5, P Ͼ from altered interactions within a thin filament functional 50 H group, with a functional group defined as seven actins, one Tm, 0.05). Thus, the increased pCa observed in AdCMVssTnI- 50 and one Tn (25). One possible interaction that could affect the infected cardiac myocytes relative to control values indicates Ca2ϩ activation threshold is the interaction between TnC and that TnI isoform expression influences myofilament sensitivity 2ϩ TnI. A stronger interaction between TnC and ssTnI compared to Ca . However, because the pCa50 in AdCMVssTnI-treated with cTnI would lower the threshold for Ca2ϩ-activated tension myocytes was less than in soleus muscles (P Ͻ 0.05; Fig. 5D), development in the adult cardiac myocytes expressing ssTnI. isoforms of other contractile proteins likely contribute to the The inhibitory carboxy terminal region of TnI is one possible 2ϩ ͞ myofilament Ca sensitivity differences observed between domain that could be involved in strengthening the ssTnI-TnC cardiac and slow skeletal muscles (3, 9). Functionally, the most interaction relative to cTnI-TnC. Solution studies indicate this important conclusion to draw from the increased Ca2ϩ sensi- region of TnI is necessary for the full inhibitory actions of TnI tivity at pH 7.00 is that the threshold for Ca2ϩ activation is (26, 27) and interacts with the amino terminus of TnC that significantly reduced in AdCMVssTnI-treated cardiac myo- contains the regulatory Ca2ϩ binding site (27). Thus, isoform cytes compared with controls. differences within this domain could directly affect Ca2ϩ- We also examined the influence of TnI isoform composition mediated activation of the thin filament. In addition, other on the Ca2ϩ-activated tension response to acidosis. Our results contractile protein isoforms, such as TnT, appear to influence 2ϩ show that the pCa50 in cardiac myocytes expressing ssTnI was submaximal Ca activation (28, 29). Thus, interactions be- markedly increased at pH 6.20, relative to control values (Fig. tween TnI and other contractile proteins also may be involved 5D; filled bars). Further, the magnitude of the pCa50 difference in the TnI isoform-dependent change in the threshold for between myocytes expressing ssTnI and control myocytes was Ca2ϩ-activated tension. greater at pH 6.20 (Fig. 5D, filled bars) than the magnitude The reduced threshold for Ca2ϩ activation of tension also difference observed at pH 7.00 (Fig. 5D, empty bars). To more may help to explain the decrease in cooperativity observed in clearly show the prominent effect of ssTnI expression on the cardiac myocytes expressing ssTnI. Cycling crossbridges are 2ϩ present in these cardiac myocytes over a range of Ca2ϩ pH sensitivity of Ca -activated tension, the ⌬pCa50 was calculated (Fig. 5E) and found to be significantly reduced in concentrations, which does not activate the cTnI-expressing 2ϩ the ssTnI-expressing myocytes, such that it was comparable to cardiac myocytes. Both Ca and crossbridge binding to actin the ⌬pCa in slow soleus fibers (Fig. 5E). These results directly cause cooperative activation of the thin filament, in part by 50 influencing the activation of neighboring functional groups show that TnI isoforms play a significant role in determining (30). Brandt and colleagues (31) have suggested that cooper- the sensitivity of the cardiac contractile assembly to acidosis. ative activation of the thin filament decreases when both of the Acidosis had no apparent effect on n in control or AdCM- H activating ligands, Ca2ϩ and attached crossbridges, are present. VssTnI-treated cardiac myocytes (pH 6.20: control nH ϭ 3.6 Ϯ Thus, the presence of cycling crossbridges at low Ca2ϩ in 0.3, AdCMVssTnI nH ϭ 2.0 Ϯ 0.2; see above for pH 7.00 data). cardiac myocytes expressing ssTnI may be one possible expla- In other control experiments, we studied the effect of acidic nation for the observed decrease in cooperativity. pH on contractile function using cardiac myocytes infected Our results provide direct evidence to support the hypoth- with AdCMVaTnT, an adenoviral vector containing an adult esis that ssTnI isoform expression also plays an important role rat cardiac TnT cDNA (14). TnT was chosen because it is in the reduced pH sensitivity observed in fetal͞neonatal another key regulatory protein in the contractile apparatus and cardiac (3, 5) and slow skeletal (9, 10) muscle compared with has a turnover rate comparable to TnI (20). The ⌬pCa50 in adult cardiac myocytes. The acidosis-induced decrease in myocytes treated with this adenoviral vector was not different myofilament Ca2ϩ sensitivity is considerably lessened in car- from controls (Fig. 5E), indicating that overexpression of an diac myocytes expressing ssTnI rather than cTnI and is not endogenous cardiac isoform does not, in itself, alter myofila- different from the shift observed in slow skeletal muscle. The ment Ca2ϩ sensitivity. molecular mechanism by which TnI isoforms confer varied pH Downloaded by guest on September 29, 2021 Physiology: Westfall et al. Proc. Natl. Acad. Sci. USA 94 (1997) 5449

sensitivity of contraction in cardiac muscle is not known. One 1. Reiser, P. J., Westfall, M. V., Schiaffino, S. & Solaro, R. J. (1994) possibility is that the interactions made between TnC and Am. J. Physiol. 267, H1589–H1596. ssTnI are less affected by acidic pH compared with those 2. Saggin, L., Gorza, L., Ausoni, S. & Schiaffino, S. (1989) J. Biol. between TnC and cTnI. Chem. 264, 16299–16302. An important implication of our findings is that the in- 3. Metzger, J. M., Lin, W.-I. & Samuelson, L. C. (1994) J. Cell Biol. 2ϩ 126, 701–711. creases in the threshold for Ca activation, cooperativity, and 4. Lee, J. A. & Allen, D. G (1991) J. Clin. Invest. 88, 361–367. sensitivity to pH found during normal cardiac development are 5. Solaro, R. J., Lee, J. A., Kentish, J. C. & Allen, D. G. (1988) Circ. due, at least in part, to the developmental transition in TnI Res. 63, 779–787. isoforms. These structural and functional changes within the 6. Donaldson, S. K. B. & Hermansen, L. (1978) Pflu¨gers Arch. 376, myofilament are also likely to be important for the function of 55–65. the developing heart expressing ssTnI. The reduced threshold 7. Fabiato, A. & Fabiato, F. (1978) J. Physiol. (London) 276, for myofilament Ca2ϩ activation may enhance myofilament 233–255. force development during cardiac maturation, when intracel- 8. El-Saleh, S. C. & Solaro, R. J. (1988) J. Biol. Chem. 263, 3274– lular Ca2ϩ handling systems are still maturing (32). In addition, 3278. relaxation may be slowed slightly in myocardium expressing 9. Metzger, J. M. (1996) J. Physiol. (London) 492, 163–172. 10. Wattanapermpool, J., Reiser, P. J. & Solaro, R. J. (1995) Am. J. ssTnI compared with relaxation rates observed in the presence Physiol. 268, C323–C330. of cTnI. Finally, one of the most important physiological 11. Nadal-Ginard, B. & Mahdavi, V. (1989) J. Clin. Invest. 84, functions of ssTnI in the maturing heart may be to preserve 1693–1700. contractile function during periods of acidosis (5). The de- 12. McGrory, W. J., Bautista, D. S. & Graham, F. L. (1988) Virology creased developed tension and reduced rates of tension de- 163, 614–617. velopment and relaxation induced by acidosis in the intact 13. Murphy, A. M., Jones, L., II, Sims, H. F. & Strauss, A. W. (1991) heart are less severe in fetal and neonatal hearts compared Biochem. 30, 707–712. with adults (5, 33), and this adaptation may be important for 14. Jin, J.-P. & Lin, J. J.-C. (1989) J. Biol. Chem. 264, 14471–14477. preserving cardiac function during transient periods of acido- 15. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular sis, such as during the birth process (34). Cloning: A Laboratory Manual (Cold Spring Harbor Lab. Press, Plainview, NY), 2nd. Ed., pp. 9.31–9.57. In addition to demonstrating the functional role of ssTnI in 16. Haworth, R. A., Hunter, D. A. & Berkoff, H. A. (1980) J. Mol. cardiac myocytes, our results show that adenovirus-mediated Cell. Cardiol. 12, 715–723. gene transfer can be used to rapidly and specifically remodel 17. Metzger, J. M. (1995) Biophys. J. 68, 1430–1442. both myofilament structure and function in the highly differ- 18. Westfall, M. V., Samuelson, L. C. & Metzger, J. M. (1996) Dev. entiated adult cardiac myocyte. Future studies can now be Dyn. 206, 24–38. designed using adenovirus-mediated gene transfer to investi- 19. Marian, A. J., Yu, Q.-T., Mann, D. L., Graham, F. L. & Roberts, gate the functional domain(s) within TnI, which affect myo- R. (1995) Circ. Res. 77, 98–106. filament Ca2ϩ sensitivity, cooperativity, and pH sensitivity in 20. Martin, A. F. (1981) J. Biol. Chem. 256, 964–968. adult cardiac myocytes. In addition, this approach may now be 21. Metzger, J. M., Parmacek, M., Barr, E., Pasyk, K., Lin, W., useful as a general model for studying the function(s) of other Cochrane, K., Field, L. & Leiden, J. (1993) Proc. Natl. Acad. Sci. USA 90, 9036–9040. key myofilament proteins. Recently, Marian and colleagues 22. Muthuchamy, M., Grupp, I. L., Grupp, G., O’Toole, B. A., Kier, (19) delivered a mutant myosin heavy chain into adult cardiac A. B., Neumann, J. & Wieczorek, D. F. (1995) J. Biol. Chem. 270, myocytes, although they were unable to structurally or func- 30593–30603. tionally distinguish expression of the mutant myosin from 23. Beall, C. J. & Fyrberg, E. (1991) J. Cell Biol. 114, 941–951. endogenous myosin expression. Thus, the success of this 24. Sweitzer, N. K. & Moss, R. L. (1990) J. Gen. Physiol. 96, 1221– approach critically depends on the protein isoform studied and 1245. its turnover rate within the contractile assembly. In this regard, 25. Bremel, R. D. & Weber, A. (1972) Nat. New Biol. 238, 97–101. structure͞function studies may be more complicated when 26. Sheng, Z., Pan, B.-S., Miller, T. E. & Potter, J. D. (1992) J. Biol. studying myosin, myosin light chain, or actin, myofilament Chem. 267, 25407–25413. proteins that all have longer endogenous turnover rates than 27. Farah, C. S., Miyamoto, C. A., Ramos, C. H. I., da Silva, A. C. R., Qyaggio, R. B., Fujimori, K., Smillie, L. B. & Reinach, F. C. TnI (20). Additional studies will be needed to determine (1994) J. Biol. Chem. 269, 5230–5240. whether adenovirus-mediated gene delivery can be used to 28. Tobacman, L. S. & Lee, R. (1987) J. Biol. Chem. 262, 4059–4064. deliver other contractile protein isoforms and͞or mutants as a 29. Schachat, F. H., Diamond, M. S. & Brandt, P. W. (1987) J. Mol. means of elucidating the functional role of these proteins. Biol. 198, 551–554. 30. Hill, T. L., Eisenberg, E. & Greene, L. (1980) Proc. Natl. Acad. We thank Mr. Faris Albayya and Dr. Krystyna Pasyk for expert Sci. USA 77, 3186–3190. technical assistance, Drs. Anne Murphy for cTnI and ssTnI cDNAs, 31. Brandt, P. W., Reuben, J. P. & Grundfest, H. (1972) J. Gen. Jim Lin for adult cardiac TnT cDNA, Christina Addison and Frank Physiol. 59, 305–317. Graham for shuttle plasmids, Stefano Schiaffino for TnI antibodies, 32. Fabiato, A. & Fabiato, F. (1978) Ann. N.Y. Acad. Sci. 307, Craig Logsdon for shared equipment, and Linda Samuelson and Philip 491–522. Wahr for helpful comments on earlier versions of this manuscript. This 33. Jarmakani, J. M., Nakazawa, M., Nagatomo, T. & Langer, G. A. work was supported by grants from the National Institutes of Health (1978) Am. J. Physiol. 235, H469–H474. and American Heart Association to J.M.M. J.M.M. is an Established 34. Glass, H. G., Snyder, F. F. & Webster, E. (1944) Am. J. Physiol. Investigator of the American Heart Association. 140, 609–615. Downloaded by guest on September 29, 2021