Developmentally Regulated, Alternative RNA Splicing-generated Pectoral Muscle-specific Troponin T Isoforms and Role of the NH2-terminal Hypervariable Region in the Tolerance to Acidosis*

The structure-function relationship of the alternative RNA splicing-generated NH2-terminal variable region of troponin T (TnT) is essential for understanding the physiological significance of developmental or muscle-specific TnT isoforms. Representing the hypervariable nature of the NH2-terminal region, a repeating transition metal-binding sequence (H(E/A)EAH)4–7 (Tx) has been found in chicken fast skeletal muscle TnT. In the present study, the developmentally regulated pectoral muscle-specific expression of this novel TnT isoform has been characterized. It was found that the variable amino terminus determined the isoelectric points of the TnT isoforms expressed, and the adult muscle-specific inclusion of the Tx sequence resulted in pectoralis TnTs, which were significantly more acidic in their NH2-terminal segment versus gastrocnemius TnTs. Experiments testing the effect of pH on TnT interaction with troponin I and tropomyosin indicated that although the interaction of acidic TnT isoforms with troponin I was less sensitive to the decrease of pH than the basic TnTs, the binding affinity of acidic TnT isoforms with tropomyosin was minimally affected by the decreased pH in contrast to basic TnT isoforms. Given that the majority of adult skeletal muscles express basic fast TnT isoforms, the switching between acidic and basic TnT isoforms may play a role in the functional adaptation of muscle to acidosis.

The structure-function relationship of the alternative RNA splicing-generated NH 2 -terminal variable region of troponin T (TnT) is essential for understanding the physiological significance of developmental or musclespecific TnT isoforms. Representing the hypervariable nature of the NH 2 -terminal region, a repeating transition metal-binding sequence (H(E/A)EAH) 4 -7 (Tx) has been found in chicken fast skeletal muscle TnT. In the present study, the developmentally regulated pectoral muscle-specific expression of this novel TnT isoform has been characterized. It was found that the variable amino terminus determined the isoelectric points of the TnT isoforms expressed, and the adult muscle-specific inclusion of the Tx sequence resulted in pectoralis TnTs, which were significantly more acidic in their NH 2 -terminal segment versus gastrocnemius TnTs. Experiments testing the effect of pH on TnT interaction with troponin I and tropomyosin indicated that although the interaction of acidic TnT isoforms with troponin I was less sensitive to the decrease of pH than the basic TnTs, the binding affinity of acidic TnT isoforms with tropomyosin was minimally affected by the decreased pH in contrast to basic TnT isoforms. Given that the majority of adult skeletal muscles express basic fast TnT isoforms, the switching between acidic and basic TnT isoforms may play a role in the functional adaptation of muscle to acidosis.
Troponin T (TnT) 1 is the tropomyosin (Tm)-binding subunit of the troponin (Tn) complex and participates in the Ca 2ϩ -dependent regulation of vertebrate striated muscle contraction (1). In the thin filament-based regulatory system, TnT interacts with troponin I (TnI) and troponin C (TnC), and Tm and may be considered as an organizer molecule (2,3). Cardiac, slow skeletal and fast skeletal muscle TnTs are encoded by different genes, and tissue-specific or developmentally regulated isoforms are expressed from each gene through alternative RNA splicing (4 -8). The alternative splicing of fast skeletal muscle TnT isoforms involves a pair of mutually exclusive COOH-terminal exons (16/17) and seven or possibly more exons (4 -8, fetal, w, X, and y) encoding a variable NH 2 -terminal region (5, 9 -12). As a result of complex splicing of the NH 2terminal variable region, a large number of fast skeletal muscle TnT isoforms are expressed in contrast to cardiac and slow skeletal muscle TnTs. Studies of TnT structure-function relationships have revealed that the COOH-terminal T2 fragment of TnT binds to the central region of Tm, whereas the central region of the TnT polypeptide has been shown to interact with the head-to-tail overlapping region of Tm (13)(14)(15)(16). The COOHterminal T2 region of TnT also anchors the TnI-TnC binary complex formation in a direct, Ca 2ϩ -independent manner (17,18). Resonance energy transfer experiments showed that TnT, but not TnC, caused an elongation of the TnI molecule (19). Therefore, TnT's core position within the thin filament regulatory system places it in an influential position in the regulation of contraction. In turn, TnT isoforms with structural variations may have a functional importance.
A regulated large to small, acidic to basic cardiac TnT isoform switch occurs during both avian and mammalian heart development (4,20,21). Gene cloning and cDNA sequencing have revealed that alternative splicing of an exon encoding an acidic segment in the NH 2 -terminal variable region is responsible for this cardiac TnT isoform switch (7,22). The primary structure of mouse fast skeletal muscle TnT isoforms has also shown a large to small, acidic to basic isoform switch during the postnatal maturation to adulthood as a result of alternative splicing of multiple exons encoding the NH 2 -terminal variable region (12). Although the developmental TnT isoform switch is conserved in mammalian and avian cardiac and skeletal muscles, the functional purpose of the acidic and basic TnT isoform is unknown. It has been demonstrated that the contractile properties of skinned trabeculae from embryonic or neonatal animals are less sensitive to acidosis versus the adult (23,24). Nonetheless, the functional differences of TnT isoforms remain unclear because the fragment corresponding to the variable NH 2 -terminal region of TnT has not been directly associated with a defined function (13). Therefore, investigation of intact TnT isoforms has provided insights into their functional significance. Differences in the Ca 2ϩ sensitivity of actomyosin ATPase have been observed using reconstituted systems containing two bovine cardiac TnT isoforms differing in the NH 2terminal hypervariable region (25). Further, a link between cardiac TnT point mutations and human familial cardiomyopathies has been found (26,27), suggesting that subtle changes in TnT primary structure may cause abnormalities in muscle growth and function. In addition, the in vitro motility assay has demonstrated a possible functional consequence of missense mutations in the NH 2 -terminal region of TnT (28). Therefore, increasing evidence points toward the role of TnT isoforms in the altered contractile properties of muscles under different physiological and pathological conditions. Our previous studies had identified a transition metal-binding (Cu 2ϩ Ͼ Ni 2ϩ Ͼ Zn 2ϩ ϭ Co 2ϩ ) element (His-(Glu/Ala)-Glu-Ala-His) 4 , designated as Tx) in the NH 2 -terminal variable region of chicken fast skeletal muscle TnT (29). Our studies demonstrated that the threedimensional structure of the NH 2 terminus of this TnT isoform can be altered by its interaction with the metal ions, leading to remote secondary conformational changes that affect TnT's binding to Tm (30) and TnI (49). This interaction between the NH 2 -terminal region and other domains of TnT implies that different isoforms of TnT generated through alternative splicing of the NH 2 -terminal variable region may confer fine changes in overall conformation of the protein and influence TnT's associations in the thin filament. To further test this hypothesis, the primary structure, developmentally regulated alternative splicing, and tissue distribution of the chicken Txpositive TnT isoforms were examined in the present study. These TnT isoforms are unique in their NH 2 -terminal primary structure and bear more negative charge (i.e., more acidic isoelectric points, pI) compared with other fast skeletal muscle TnTs in adult animals. The functional significance of TnT isoform expression was further studied by determining the effect of pH on the binding affinity of acidic and basic TnT isoforms to TnI and Tm. The results demonstrated that whereas a decrease in pH affected TnI and Tm binding of basic TnT isoforms, the acidic pH had significantly less effect on acidic TnT function, suggesting a novel role of TnT isoform expression in the functional adaptation of muscle fibers.

RT-PCR Cloning and Sequence Characterization of Chicken TnT
Isoform cDNAs-To characterize the various fast skeletal muscle TnT isoforms expressed in the pectoralis and gastrocnemius muscles, RT-PCR was done on total RNA isolated by the Trizol reagent method (Life Technologies, Inc.) from the pectoralis or gastrocnemius muscles of an adult White Leghorn chicken. Two synthetic oligonucleotide primers, e2F and e18R ( Fig. 1), were designed according to the published sequence (9) around the conserved translation initiation and termination codons to amplify full-length coding sequence for the fast TnT cDNA. RT-PCR was done as described (12), and TnT cDNAs were cloned into the EcoRV site of pBluescript SK(ϩ) vector (Stratagene). To determine which of the mutually exclusively spliced exons (16 or 17) was expressed in the TnT isoforms, the cloned cDNAs were mapped by PstI restriction enzyme cleavage at a recognition site present in exon 16 but not exon 17 of the chicken fast skeletal muscle TnT gene. To determine the exon organization in the 5Ј variable region of the TnT isoform cDNAs, DNA sequencing was done by the dideoxy chain termination method (Sequenase 2.0, Amersham Pharmacia Biotech). The cDNA sequences were analyzed by the DNAstar computer program (Lasergene) to obtain the primary structure and predicted physical properties of the TnT isoforms.
Expression of Cloned Chicken Breast Muscle TnT cDNA in Escherichia coli-A cDNA encoding the major chicken breast muscle TnT isoform was cloned in-frame into the pAED4 prokaryotic expression vector (31). To demonstrate the authenticity of the cloned cDNA, E. coli BL21(DE3)pLysS was transformed with the expression plasmid, and the culture was induced with 0.2 mM isopropyl-1-thio-␤-D-galactopyranoside as described (30). Protein extracts of the induced bacterial cells were prepared for SDS-polyacrylamide gel electrophoresis (PAGE). The bacterial lysate, along with a chicken breast muscle total protein sample, was resolved by 14% SDS-PAGE with an acrylamide:bisacrylamide ratio of 180:1 and transferred to nitrocellulose membrane (0.45-m pore size) for Western blot analysis. A rabbit anti-TnT antiserum RATnT and a Tx-element specific monoclonal antibody (mAb) 6B8 (49) were used to identify the cloned and native chicken TnT isoforms. The Western blot transfer, first and second antibody reactions, and washing conditions were previously described (30).

Affinity Chromatography Purification and Analysis of Metal-binding
TnT Isoforms-Chicken and turkey breast muscle TnTs were purified using an approach that takes advantage of the Zn 2ϩ binding property of Tx-positive TnTs. Breast muscles were diced into small pieces and homogenized in 10 volumes (v/w) of 1 M NaCl, 50 mM sodium phosphate buffer, pH 7.0, and centrifuged at 12,000 ϫ g for 20 min at 4°C. To 150 ml of the supernatant, solid urea was added to 6 M, clarified by centrifugation and loaded onto a 10-ml chelating fast flow Sepharose (Pharmacia) Zn 2ϩ affinity column charged as described previously (29) and equilibrated in the sample buffer (6 M urea, 1 M NaCl, 50 mM sodium phosphate buffer, pH 7.0). The column was washed with the sample buffer until the A 280 nm of the elute was Ͻ0.2. The metal-binding chicken or turkey pectoralis TnTs were eluted with a step imidazole gradient of 10 -100 mM imidazole in the sample buffer. The purified TnTs identified by SDS-PAGE at the 30 -50 mM imidazole elutes were collected, dialyzed against 0.5% formic acid, and lyophilized. To demonstrate the presence of the unique metal-binding Tx element in specific TnT isoforms, analytical Zn 2ϩ affinity column chromatography was done. Purified chicken and turkey pectoralis and rabbit skeletal muscle TnTs (32), as well as cloned chicken fast skeletal muscle TnT1, -2, -3, and -4 (9, 33), were mixed (2 mg each) in a buffer containing 6 M urea, 1 M NaCl, 20 mM sodium phosphate buffer, pH 7.0, for a final sample volume of 1.5 ml. The mixture was loaded onto a 1-ml chelating fast flow Sepharose Zn 2ϩ affinity column. Following collection of the flowthrough, the column was washed with 5 volumes of the above buffer including 2 mM imidazole. The bound proteins were eluted by a linear imidazole gradient of 2-80 mM (25 ml total) in the same buffer and analyzed by 14% SDS-PAGE with an acrylamide to bisacrylamide ratio of 37.5:1 to determine the Zn 2ϩ affinity profile of the various TnT isoforms.
Western Blot Analysis of TnT, TnI, and Tm Isoform Expression in Chicken Skeletal Muscles-Samples of adult chicken skeletal muscles were obtained from a White Leghorn chicken following identification of the muscles according to Ref. 34. The excised muscle samples were homogenized immediately in SDS-PAGE buffer containing 1% SDS to ensure the integrity of the proteins. The protein extracts were resolved by 14% SDS-PAGE with an acrylamide:bisacrylamide ratio of 180:1 to examine the expression of TnT, TnI, and Tm isoforms by Western blots. To determine the developmental expression pattern of the TnT isoforms, chicken muscle samples were obtained from embryonic through adult stages. From embryonic day 10 and 14 White Leghorn chicken embryos, the pectoralis and gastrocnemius muscles were excised under a dissecting microscope. Postnatal pectoralis muscle samples were obtained from White Leghorn chicks at 0, 7, 14, and 28 days following hatching. The polyclonal RATnT, the Tx-specific mAb 6B8, an anti-Tm mAb CH1 (Ref. 35; a gift from Dr. Jim Lin, University of Iowa), and a specific anti-TnI mAb TnI-1 2 were used as the first antibodies in the Western blotting as described above.
Two-dimensional Gel Electrophoresis Analysis of Acidic and Basic TnT Isoforms-Purified chicken pectoralis TnT as well as cloned chicken fast skeletal muscle TnT isoforms TnT2, TnT3, and TnT4 were analyzed by two-dimensional gel electrophoresis with non-equilibrium pH gradient gel electrophoresis (NEPHGE; Ref. 36) as the first dimension and SDS-PAGE the second dimension. Using Mini Protean II 2-D mini-gel apparatuses (Bio-Rad) as described previously (31), first dimension NEPHGE with tube gels containing pH 3.5-10 ampholine (Pharmacia) were electrophoresed at 350 V for 3 h and 750 V for 10 min. The gel pieces were then equilibrated in SDS-PAGE sample buffer for 10 min and transferred onto the flat top of a 14% SDS-PAGE gel (acrylamide:bisacrylamide ϭ 37.5:1) for the second dimension separation. The gel was stained with Coomassie Blue R250 to reveal the resolved acidic and basic chicken skeletal muscle TnT isoforms.
Solid Phase Protein Binding Assays-To monitor the effect of pH on the interaction of acidic and basic TnT isoforms with TnI and Tm, an enzyme-linked immunosorbant assay-based protein binding assay was developed. Purified rabbit ␣-Tm (37) or chicken fast skeletal muscle TnI (38) was diluted to 5 g/ml in 100 mM KCl, 3 mM MgCl 2 , 20 mM Tris-HCl, pH 7.5 (buffer A), and aliquoted (100 l/well) into triplicate wells of a 96-well microtiter plate (Falcon 3915) and allowed to coat overnight at 4°C. The coated wells were washed in buffer A containing 0.05% Tween 20 (buffer T) and blocked for 1 h at room temperature with buffer T containing 1% bovine serum albumin (150 l/well). The plates were then washed three times with buffer T and incubated for 2 h at room temperature with serial dilutions (2 to 0.0019 M) of chicken pectoralis muscle TnT or rabbit skeletal muscle TnT diluted in buffers containing 100 mM KCl, 3 mM MgCl 2 , 0.1% bovine serum albumin, 0.05% Tween 20, and 20 mM PIPES at a final pH of 7.0, 6.6, or 6.2. Washed three times as above, the plates were incubated for 1 h at room temperature with a predetermined dilution of the RATnT antiserum in buffer T containing 0.1% bovine serum albumin. Following washing, the bound first antibody was detected by incubation at room temperature for 45 min with horseradish peroxidase-labeled anti-rabbit IgG second antibody (Sigma). The bound peroxidase activity was detected by H 2 O 2 -2,2Ј-azinobis-(3-ethylbenzthiazolinesulfonic acid) substrate in 0.1 M citrate buffer, pH 4.0, and monitored for absorbance at 405 nm at a series of time points. Data from the linear portion of the time course were plotted by Igor Pro software (Wavemetrics, Inc.), and differences between the TnT concentrations required for 50% of maximum binding were assessed statistically by the Student's t test.

Cloning of a Large Number of cDNAs Encoding Chicken Fast Skeletal Muscle TnT Isoforms and Primary Structure of the Major TnT Isoform in Chicken Breast
Muscle-Twenty original chicken fast skeletal muscle TnT cDNAs have been cloned by RT-PCR from both the adult pectoralis and gastrocnemius muscles. The exon organization of the 5Ј hypervariable region and splicing of the 3Ј mutually exclusive exons 16 and 17 have been determined. The high expression level of Tx-TnTs in the pectoralis was evident as 85% of the cloned transcripts contained exon X. The frequency of the fast skeletal muscle TnT isoform cDNAs expressed is summarized in Table I. In the pectoralis muscle, isoform TnT8e16 expressed the Tx element coding for 7 of the histidine repeats (H(E/A)EAH) (Fig. 1) and accounted for 70% (14/20) of the transcripts cloned from the pectoralis RNA. The TnT8e16 cDNA was completely sequenced, and its structural organization is shown in Fig. 1. To verify that the TnT8e16 cDNA represents the major TnT isoform expressed in the chicken breast muscle, the cDNA was expressed in E. coli, and the cloned TnT8e16 isoform migrates in SDS-PAGE identically with the native major chicken breast muscle TnT, as recognized by the RATnT antiserum under stringent Western blotting conditions. The E. coli expressed TnT8e16 was also specifically recognized in immunoblots by the Tx element-specific mAb 6B8, confirming that like the native major chicken breast muscle TnT isoform, the cloned TnT isoform is Tx positive.

Expression Frequencies and Splicing Relationships of the Alternative Exons in the Chicken Fast Skeletal Muscle TnT
Gene-The sequences of the 5Ј variable regions of the large sample of cloned chicken fast skeletal muscle TnT cDNAs revealed the relationships between the splicing pathways of the multiple alternative exons. In TnT cDNAs cloned from the adult pectoralis, a high preference for exon 16 (95%) and inclusion of exon X encoding the Tx element (85%) was found (Table  I). On the other hand, of TnT transcripts isolated from gastrocnemius total RNA, 65% included exon 16, whereas only 10% showed inclusion of exon X. There was no definitive relationship between the Tx exon expression and the COOH-terminal alternative splicing pattern. On the other hand, when the TnT cDNAs cloned from the two muscles are compared with each other, several interesting patterns can be summarized. The expression of exon 16, which was heavily favored in pectoralis TnTs, was less dominant in the gastrocnemius. Similar to the mammalian fast TnT gene expression (12), exon 5 of the chicken fast TnT gene is included in all of the 40 adult isoforms, whereas exon 8 is the least favored (2/40). Further, of the 20 transcripts examined from the pectoralis, none showed expression of exon 4, whereas 50% of the cDNAs cloned from gastrocnemius RNA included exon 4. Exon y, which was identified in embryonic chicken fast TnT cDNAs (11), was not found in the total of 40 adult cDNAs sequenced. The expression of exon w (Ref. 11, Table I) was found only in association with exon 5. Therefore, although no exclusive relationship exists between the expression of the alternative exons with respect to each other, tissue-specific control of the splicing pattern seems to be tightly regulated and worth further investigation. To date, no sequence information is available regarding the genomic organization of the chicken fast skeletal muscle TnT gene. Given the information in Table I and in the literature (9, 11), a possible organization of the 5Ј alternative exons of the chicken fast skeletal muscle TnT gene are shown in Fig. 2. According to the splicing boundaries for the exon(s) encoding the Tx histidine repeat region, splicing of the Tx cassette through alternative acceptor/donor sites, such as those indicated by the arrowheads in Fig. 2, may underlie the variation of the Tx inclusion pattern.
Zn 2ϩ Binding Profile of Tx-containing TnTs-Results of the analytical Zn 2ϩ affinity chromatography (Fig. 3) showed that the four Tx-negative TnTs (rabbit TnT and cloned chicken  To highlight possible splicing boundaries in the 5Ј region of the chicken fast skeletal muscle TnTs, three Tx-positive TnT isoforms are compared with the full sequence known to date. TnT8e16 is the major adult pectoralis TnT isoform (Fig. 1), whereas TnT1 (9) and NP1 (11) are two embryonic and neonatal pectoralis muscle-specific TnT isoforms, respectively. By amino acid and nucleotide sequence alignment among the three Tx-positive TnT isoforms, potential alternative splicing sites are indicated by arrowheads.
FIG. 3. Zn 2ϩ affinity binding profile for various skeletal muscle TnTs. Turkey breast, chicken breast, rabbit skeletal, as well as cloned chicken fast skeletal muscle TnT isoforms (TnT1-TnT4), were analyzed for their Zn 2ϩ binding properties. A mixture of the purified TnTs was loaded onto a Zn 2ϩ affinity column, and the flow-through and imidazole gradient eluting fractions were analyzed by SDS-PAGE. The rabbit skeletal muscle TnT, as well as the Tx-negative chicken TnT 2-4, did not bind to the column, whereas the Tx-positive turkey and chicken breast muscle TnTs and TnT1 specifically bound to the column and were eluted only by competition with the imidazole gradient. Protein bands apparent underneath the full-length turkey and chicken TnTs in the elution profile are proteolytic digests that retain the Tx element. Individual samples of the purified proteins used in the assay are shown at right. breast muscle TnTs as well as a difference in their Zn 2ϩbinding avidities. The higher binding avidity of the turkey breast muscle TnTs to the Zn 2ϩ column may indicate that the turkey Tx-positive TnT isoforms contain more of the histidine repeats as compared with the chicken breast muscle TnT isoforms.
Expression of TnT, TnI, and Tm Isoforms in Adult Chicken Skeletal Muscles-Troponin T, TnI, and Tm isoform distributions of representative adult chicken skeletal muscles were determined by Western blots using specific antibodies, and the expression profiles for the various muscles are presented in Fig. 4. Tropomyosin isoforms were indicated as ␣and ␤-Tm (39) and TnI as fast and slow isoforms (40). Characterization of TnT isoforms is more complex due to the large number of variants generated by alternative RNA splicing (Table I). Nonetheless, the 6B8 antibody distinguished muscles expressing the high M r , Tx-positive TnT isoforms from muscles expressing the lower M r , Tx-negative adult fast TnT isoforms. Although the majority of the muscles showed a large diversity in TnT isoforms expressed, muscles expressing the high M r Tx-positive TnTs were restricted to the pectoral limb. Muscles from the head and neck region, as well as muscles from the pelvic limb, did not show detectable levels of Tx-TnT expression in the Western blots. In contrast, Tm expression was not as diverse. Of the muscles tested, only the pectoralis superficialis (breast major) showed exclusive ␣-Tm expression, whereas all other muscles tested contained mixed ␣and ␤-Tm isoforms. The distribution of TnI isoforms showed that all of the pectoral limb muscles tested expressed exclusively fast TnI, whereas many pelvic limb muscles express slow TnI in addition to the fast isoform. Altogether, the results suggest that the alternative RNA splicing-regulated Tx-TnT expression is separate from the transcriptional regulation of TnI and Tm isoforms.
Expression of Tx-positive TnT Isoforms in the Developing Chicken Pectoralis-Examination of TnT isoform expression in the developing chicken pectoralis demonstrated a shift in TnT isoform expression from a mix of high and low M r TnTs to a dominant high M r TnT in the adult (Fig. 5). Identification by mAb 6B8 revealed that the high M r isoforms are Tx positive and are not detected in the embryonic chicken pectoralis. Onset of Tx-TnT expression occurs after 14 days in ovo development as comparable levels of multiple Tx-positive TnT isoforms are detected. With development to the adult, the Tx-TnTs become dominant (between 7 and 14 days after hatching), and a downshift in the M r of Tx-positive TnT isoform expression is seen, ending in the high level expression of the one major pectoralis muscle-specific isoform TnT8e16 (Fig. 5 and Table I). The FIG. 4. Expression patterns of TnT, TnI, and Tm isoforms in adult chicken muscles. The expression of TnT, TnI, and Tm isoforms in multiple adult chicken muscles was characterized by Western blots. A, the representative patterns of total TnT, Tx-TnT, TnI, and Tm are presented. Three TnT isoform expression patterns were identified by the RATnT polyclonal antibody, indicated as a mixture of high and low M r TnTs (Type I), only high M r TnTs (Type II), or only low M r TnTs (Type III). The high M r TnTs represent Tx-positive TnTs, as identified by mAb 6B8 staining on a type I muscle. The expression of TnI is identified by the TnI-1 mAb as fast (F) and slow (S) isoforms and Tm by the CH1 mAb as the ␣and ␤-isoforms. B, the TnT, TnI, and Tm isoform expression profiles for various muscles. Muscles were grouped as belonging to the head/neck region, the pectoral limb, or the pelvic limb according to (34). Ϫ/ϩ, negative or positive in Tx-TnT expression; n.d., not determined.  (Table I). Throughout the development of the gastrocnemius, the mAb 6B8 did not detect any significant level of Tx-TnT expression, in agreement with the cDNA sequence data of the gastrocnemius muscle fast TnTs (Table I). Consistently, the RATnT antiserum detected only low M r TnT expression throughout development of the gastrocnemius. The results conclude that the expression of Tx-positive TnT isoforms is specifically activated while the pectoral muscles mature. Tropomyosin and TnI expressions were also examined in the developmental muscle samples. Fast TnI was seen throughout development for both pectoralis and gastrocnemius muscles. Although both muscles expressed a mix of ␣and ␤-Tm at the neonatal stage, ␤-Tm was down-regulated in the pectoralis (data not shown), resulting in only ␣-Tm expression in the adult pectoralis. In contrast, the adult gastrocnemius showed mixed ␣and ␤-Tm expression (Table I). Therefore, separate developmental regulations are observed for the Tx-TnT, TnI, and Tm isoform expressions.
Physical Characteristics of the Cloned Fast Skeletal Muscle TnT Isoforms-A comparison of the protein primary structures deduced from the cDNA sequences showed a significant distinction between the fast skeletal muscle TnT isoforms expressed in the pectoralis versus the gastrocnemius. Troponin T isoform switching through mammalian fast skeletal muscle development has shown an acidic to basic charge transition from the embryo to the adult (12). Although no embryonic TnT transcripts were examined by cDNA sequencing in this study, the large to small M r switch of both Tx-positive and Tx-negative TnT isoforms in the pectoralis during development (Fig. 5) indicates a similar developmental regulation in avian muscle. As shown in Table I, the fast TnT isoforms expressed in the chicken adult gastrocnemius were predominantly basic, with pI values ranging from 8.34 (TnT11e16) to 9.29 (TnT9e16), similar to the adult mouse fast skeletal muscle TnT isoforms (12). Isoform TnT8e16 (pI 6.91) and TnT10e17 (pI 7.00) represent the acidic isoforms but were only 15% (3/20) of the population. In contrast to the gastrocnemius fast TnT isoforms, the specific inclusion of exon X in the majority of chicken pectoralis muscle TnTs results in a significant decrease in the pI. With the exception of three Tx-negative isoforms (TnT4e16, 4e17, and 5e16), the majority of the breast muscle TnTs included the Tx element and had more acidic pI, ranging from 6.91 (TnT7e16 and 8e16) to 7.19 (TnT6e16). To evaluate the overall effect of Tx expression in the pectoral and pelvic limb muscles, the weighted average pI for fast TnT isoforms expressed in the pectoralis muscle was 7.26 versus 8.47 in the gastrocnemius, reflecting a significant difference (p Ͻ 0.01) in their NH 2terminal charge.
To verify that the calculated pI of the TnT isoforms is truly representative of the physical properties of the proteins, four purified acidic and basic TnT isoforms were analyzed experimentally by two-dimensional gel electrophoresis (Fig. 6). The chicken pectoralis muscle TnTs (mainly TnT8e16, Table I), as well as the embryonic isoform TnT3 (Ref. 9; calculated pI ϭ 6.49), migrated less from the cathode during the NEPHGE analysis, consistent with their predicted, higher NH 2 -terminal negative charge, i.e., more acidic isoelectric points. In contrast, TnT2 and TnT4 (Table I) moved faster toward the anode, consistent with their basic pI. Therefore, the data confirm that a significant overall charge difference exists between TnT isoform proteins varying in the NH 2 -terminal variable region. Given that the adult cardiac and fast skeletal muscle TnT isoforms are basic compared with their embryonic counterparts (12,20), the acidic TnT expression in the adult chicken pectoralis is unique and provides a useful system to examine the structural and functional significance of the alternatively spliced TnT isoforms.
The Effect of pH on the Interaction of Acidic and Basic TnT Isoforms with TnI-The effect of physiological pH changes on the interaction of acidic (chicken pectoralis muscle) and basic (rabbit skeletal muscle) TnT isoforms to TnI is demonstrated by the enzyme-linked immunosorbant assay-mediated solid phase protein binding assays (Fig. 7). The curves show that at pH 7.0, 0.021 M chicken breast muscle TnT and 0.020 M FIG. 6. Charge profile of acidic and basic TnT isoforms. Purified chicken pectoralis TnT (CSTnT) as well as the cloned chicken fast skeletal muscle isoforms TnT2, -3, and -4 were analyzed by SDS-PAGE (left lane) and two-dimensional gel electrophoresis in which NEPHGE with pH 3.5-10 ampholine was used as the first dimension prior to SDS-PAGE. The anode (Ϫ) and cathode (ϩ) sides for the NEPHGE are indicated. The acidic pectoralis TnT and TnT3 showed significantly higher negative charges than the basic TnT2 and -4, migrating in agreement with their predicted pIs.

FIG. 7. Effect of decreases in pH on the interaction of acidic and basic TnTs to TnI.
The effect of pH on acidic and basic TnT isoform interaction with TnI was tested by solid phase binding assays. Both rabbit skeletal muscle TnT (A) and chicken pectoralis TnT (B) show decreases in TnI binding affinity at acidic pH levels, whereas the acidic chicken pectoralis TnT is less affected than the basic rabbit TnTs (Table II). rabbit skeletal muscle TnT were required for 50% of maximal binding to TnI. Decreases in rabbit fast skeletal muscle and chicken pectoralis TnT binding to TnI at acidic pH were observed as the concentration of TnT required for 50% of maximal binding increased to 0.150 and 0.168 M at pH 6.6 and 6.2, respectively, for the basic rabbit skeletal muscle TnTs and 0.051 and 0.045 M at pH 6.6 and 6.2, respectively, for acidic chicken breast muscle TnT (Table II). Although there was no significant difference in the concentrations of acidic and basic TnTs required for half-maximal binding to TnI at the normal intracellular pH (7.0; p Ͼ 0.8), the concentrations required for half-maximal binding increased significantly at pH 6.6 and 6.2 versus pH 7.0 for both classes of TnT isoforms (p Ͻ 0.05 for chicken and p Ͻ 0.01 for rabbit TnTs, Table II). In addition, there was a significant difference between the response of acidic and basic TnT isoform binding to TnI at lower pH levels (p Ͻ 0.01). Interestingly, there was no significant difference in TnI binding at pH 6.2 versus pH 6.6 for either TnT class (p Ͼ 0.35 for chicken, p Ͼ 0.30 for rabbit TnTs), indicating that once the interaction is moved away from normal physiological conditions, the effects seem not to be graded depending on the pH. Therefore, the interaction between TnTs and TnI does not follow a simple dose response to the decreasing pH from 7 to 6.6 or 6.2. On the other hand, the similar values for half-maximal binding at pH 6.6 and 6.2 indicates that the effect of pH on the TnT-TnI interaction is a specific response to the physiological pH variation, inducing a lower affinity binding state.
Significant Difference of the Effect of pH on the Interaction of Acidic and Basic TnT Isoforms with ␣-Tm-In the enzymelinked immunosorbant assay protein binding experiments, the concentration of rabbit skeletal muscle TnTs required for halfmaximal binding to ␣-Tm were 0.095 and 0.142 M at pH 6.6 and 6.2, respectively, which were significantly different (p Ͻ 0.05) versus 0.027 M at pH 7.0 (Fig. 8A). In contrast, acidic chicken pectoralis muscle TnTs bound to ␣-Tm with similar concentrations for half-maximal binding at pH 7.0, 6.6, and 6.2 (Fig. 8B). The concentrations for half-maximal binding were 0.023, 0.031, and 0.033 M at pH 7.0, 6.6, and 6.2, respectively, without significant differences (p Ͼ 0.20 at pH 6.2 versus 6.6 and p Ͼ 0.10 at pH 6.6 versus 7.0). Therefore, the interaction of acidic TnT isoforms with ␣-Tm shows significant tolerance to acidosis in contrast to the normal basic adult fast skeletal muscle TnT isoforms. Similar to the TnT-TnI interaction, there was no significant difference between the concentrations required for half-maximal binding to ␣-Tm at pH 6.6 versus 6.2 for either TnT class (p Ͼ 0.20 for acidic, p Ͼ 0.10 for basic TnTs), further supporting the specific switch of binding states of TnT due to the environmental pH.

A Cluster of Seven Transition Metal Binding Sites in the NH 2 -terminal Variable Region of Chicken Fast Skeletal Muscle
TnT-In this study, we have identified three Tx-positive chicken fast TnT isoforms (TnT6, -7, and -8) containing seven H(E/A)EAH metal-binding sites in the NH 2 -terminal variable region. The Tx element is a unique feature of avian fast skeletal muscle TnTs in the two orders Galliformes and Craciformes (29) and is therefore not an essential element required for the function of the protein. The presence of this large unique segment implies that the sequences in the NH 2 terminus of TnTs have an inherent flexibility, insofar as the insertion of nonessential sequences seems to not destroy TnT's function. This also indicates that the alternatively spliced NH 2 -terminal variable region of TnT is not essential for TnT's basic function, agreeing with previous reports that an engineered TnT truncated at its NH 2 terminus remains functional (41). However, when residing in an intact TnT molecule, the physical properties of the NH 2 terminus of different TnT isoforms may play a role in modulating TnT's function. In this respect, the unique metal-binding Tx-TnT isoforms provide a novel system for TnT structure-function studies as the metal-binding Tx element may be used as a manipulatable site for TnT structure-function studies (30,49).
Developmentally Regulated Pectoral Muscle-specific Expression of Tx-TnT Isoforms-An interesting finding in this study is the restricted expression of Tx-TnT to the muscles of the pectoral limb of the adult chicken, which are related to the nonessential, short lived flight of this species. In contrast, the muscles of the head, neck, diaphragm, and pelvic limb, which are used essentially by the chicken, did not show significant expression of Tx-TnTs (Fig. 4). The developmentally up-regulated expression of high M r Tx-TnTs in the pectoralis is unique since TnT isoform changes during development usually involve  8. Effect of decreases in pH on the interaction of acidic and basic TnTs to ␣-Tm. Solid phase protein binding assays at different pH levels showed that the binding of basic rabbit skeletal muscle TnT to Tm was lowered under decreased pH (A), similar to that seen for TnT-TnI interactions (Fig. 7). The binding of acidic chicken pectoralis TnT to Tm, however, was not significantly affected by the acidosis (B). a high M r to low M r , acidic to basic switch in the isoforms expressed (4,12,20,21). The inclusion of the Tx segment in TnT results in a significant change in the size and charge of the NH 2 -terminal domain (Fig. 6), which may modulate the overall conformation of TnT and the thin filament regulation of contraction. The regulation of the splicing pathways that control the expression of the Tx element remains to be investigated. NH 2 -terminal Alternative Splicing Generates Acidic and Basic TnT Isoforms-The splicing pattern of fast skeletal muscle TnTs is more diverse than both cardiac and slow skeletal muscle TnTs due to the multiple alternatively spliced exons in the NH 2 -terminal region as well as the additional COOH-terminal variable region. Summarized from cloning and sequencing data, the large number of mouse fast skeletal muscle TnT isoforms fall into two non-overlapping acidic and basic groups, which are determined by their NH 2 -terminal splicing patterns (12). Our analysis of chicken fast TnT primary structure ( Table  I) further demonstrates that the NH 2 -terminal variation determines acidic and basic TnT isoform expression in avian muscles as well. Unlike most skeletal muscle wherein only basic TnTs are expressed in the adult, the developmentally regulated, tissue-specific high level expression of exon X, which encodes a major acidic stretch (Fig. 1), produces an acidic group of TnT isoforms in the adult chicken pectoralis (Table I). Since a homologous segment to exon X is not found in any other TnT genes from the vertebrate phyla, it may be a result of an isolated evolutionary event. The splicing regulation of this exon may have been suppressed in the chicken muscles that are subjected to greater functional demands (Fig. 4). Loss of this splicing regulation in the adult pectoral muscles may result in the high level inclusion of this exon and, perhaps, an unintentional trend to express TnT isoforms with acidic pIs. It is interesting to note that exons 4 and 8, both encoding highly acidic segments (EEYEEE and EEAPEE, respectively), are spliced out from the Tx-positive TnT isoforms expressed in the adult muscles, possibly to prevent the expression of extremely acidic TnT isoforms in combination with the Tx segment ( Table  I). The mechanism may be an adaptation in response to the introduction of exon X into the genome.
Role of the NH 2 -terminal Charge in TnT Function-The role of the conserved acidic to basic TnT isoform switch during development and the physiological function of the NH 2 -terminal variable region are essential questions to the structurefunction relationship of TnT. Similar to embryonic fast skeletal muscle TnT isoforms, adult cardiac and slow skeletal muscle TnTs are more acidic in their pIs, and therefore the biological significance of basic TnT isoform expression in the mature fast skeletal muscles is of interest. Our data demonstrate that acidic and basic TnT isoforms contributed to an adaptation to environmental pH as shown by their particular binding to TnI and Tm, the two major interacting partners of TnT in the thin filament assembly (Figs. 7 and 8). Although the acidic TnT showed higher tolerance to acidosis, the effect of environmental pH changes on interactions of the thin filament is not a universal change in binding affinities. It is interesting that for both acidic and basic TnTs, binding to TnI was more sensitive to acidosis versus binding to ␣-Tm. This may reflect the two-site binding model for TnT-Tm (13) versus one site TnT-TnI binding (42). Nonetheless, the saturable binding profiles of both acidic and basic TnTs to ␣-Tm (Fig. 8) clearly demonstrate that the effects of acidosis are not solely a reflection of the number(s) of available binding sites. Therefore, the observed changes in cooperativity and calcium sensitivity of muscle during acidosis (23, 43, 44) may be better understood by mapping the effects of decreased pH on the interactions between thin filament proteins. The observation that contractile function of neonatal hearts is less sensitive to the effects of acidosis when compared with the adult heart (23, 45) may be explained by different protein isoform components in the Ca 2ϩ -activation system. The neonatal heart is known to express embryonic cardiac TnT isoforms with more acidic NH 2 termini as compared with the adult cardiac TnT, which may contribute to this adaptation. In diabetic rats, force developed by skinned trabeculae was not as sensitive to acidosis as compared with control rats (46). As a correlation, evidence is available that a switch of cardiac TnT isoform expression occurs in diabetic rats and failing hearts from human patients (47,48). The abnormal inclusion of an acidic or basic exon to a TnT molecule to change the charge profile of the NH 2 -terminal variable region may have significant consequences for the functional performance of the fiber, due to alterations in the interaction of TnT with other thin filament regulatory proteins. Therefore, in pathological situations, expression of TnT isoforms with altered NH 2 -terminal charge may further compromise the contractile function.