Comparative studies on the expression patterns of three troponin T genes during mouse development

Abstract

In vertebrates, three troponin T (TnT) genes, cardiac TnT (cTnT), skeletal muscle fast-twitch TnT (fTnT), and slow-twitch TnT (sTnT), have evolved for the regulation of striated muscle contraction. To understand the mechanism for muscle fiber-specific expression of the TnT genes, we compared their expression patterns during mouse development. Our data revealed that the TnT expression in the developing embryo was not as restricted as that in the adult. In addition to a strong expression in the developing heart beginning at day 7.5 p.c (postcoitum), the cTnT transcript was detected at later stages in some skeletal muscles, where beginning at day 11.75 p.c. the fTnT and sTnT genes were also expressed. Only sTnT but not fTnT was found transiently in the developing heart. At day 13.5 p.c., expressions of all three genes were detected in the developing tongue and this co-expression continued to day 16.5 p.c. with the fTnT isoform being predominant. At this stage, overlapping and distinct expression patterns of both sTnT and fTnT genes were also evident in many developing skeletal muscles. These data suggest that different muscles during development undergo a complex change in TnT isoforms resulting in different contractile properties. Unexpectedly, the cTnT transcript was persistently found in the developing bladder, where presumably smooth muscle is present. In transgenic mice, expression of a LacZ gene driven by a rat cTnT promoter (−497 to +192 bp) was very similar to that of the endogenous cTnT gene, suggesting that this promoter contained regulatory elements sufficient for the control of tissue-specific cTnT expression during development. Anat Rec 263:72–84, 2001. © 2001 Wiley-Liss, Inc.

Troponin T (TnT) is the tropomyosin-binding subunit of the troponin complex, and plays a core function in the regulation of striated muscle contraction (Perry, 1998). In vertebrates, three fiber type-specific TnT isoforms have evolved, which are encoded by cardiac troponin T (cTnT) (Cooper and Ordahl, 1984; Jin et al., 1992), skeletal muscle fast-twitch troponin T (fTnT) (Wilkinson et al., 1984; Breitbart and Nadal-Ginard, 1986; Smillie et al., 1988; Bucher et al., 1999), and slow-twitch troponin T (sTnT) (Gahlmann et al., 1987; Huang et al., 1999; Yonemura et al., 2000) genes. Although these genes encode homologous proteins with conserved structures, the three TnT isoforms have distinct properties corresponding to functional differences in regulating muscle contraction (Schachat et al, 1987; Tobacman, 1988; Nassar et al., 1991; Malnic et al., 1998; Perry, 1998). In the adult mouse, cTnT is solely expressed in the heart, whereas fTnT is expressed in almost all skeletal muscles. Unlike large mammals, mouse sTnT expression is down regulated in most skeletal muscles during postnatal development. Therefore, in the adult mouse the sTnT isoform is only present in muscles enriched with slow fibers (Jin et al., 1998). Each of the three TnT genes can further generate a variety of transcripts by alternative splicing (see references in Schiaffino and Reggiani, 1996; Perry, 1998). Several lines of evidence suggest that different TnT isoforms have differential effects on the calcium sensitivity and the tension response curve of actomyosin ATPase (see references in Schiaffino and Reggiani, 1996; Perry, 1998; Solaro and Rarick; 1998; Moss, 1999). Thus, variation in TnT isoforms during muscle development and in different muscle types can modulate muscle calcium sensitivity and muscle performance (Thys et al., 1998; Marden et al., 1999).

The expression pattern of the three TnT genes during the early stages of striated muscle development has not been compared and described in detail. However, previous studies on TnT isoform distribution during late embryonic/fetal stages and the perinatal period of development suggest that expression of the cTnT, fTnT, and sTnT genes may not be as restricted as it is in the adult. For example, cTnT protein is detected in the developing skeletal muscles at late embryonic and fetal stages as well as in the heart; however, its expression in skeletal muscles is selectively repressed after late fetal stages (Toyota and Shimada, 1981; Cooper and Ordahl, 1984; Saggin et al., 1990; Sabry and Dhoot, 1991; Sutherland et al., 1991; Jin et al., 1992). It will be of interest to know whether the cTnT expression in the developing skeletal muscle overlaps the expression of fTnT and/or sTnT and whether the fTnT and/or sTnT genes are also transiently expressed in the embryonic heart. Furthermore, questions such as how early these fiber-specific TnT genes are expressed and how long the co-expression of 2 or 3 TnT genes in a given muscle endures remain to be addressed.

A detailed understanding of the expression of the TnT genes during development may further provide a foundation for studying their transcriptional regulation. In this study, the expression of cTnT, fTnT, and sTnT genes during mouse development was systematically investigated for the first time by in situ hybridization assays. Although expressions of the three TnT genes overlapped in embryonic and fetal stages, distinct patterns were observed. The expression of cTnT gene was first detected in the precardiac mesoderm at day 7.5 postcoitum (p.c.). The fTnT and sTnT messages could not be detected until day 11.75 p.c. in the developing somites, where the cTnT gene was also expressed. At later stages, in addition to its strong expression in the heart, the cTnT gene was expressed in the muscles in the body and tail as were the fTnT and sTnT genes. The sTnT but not the fTnT transcripts were also found in the developing ventricle at day 13.5 p.c, but by day 16.5 p.c. cardiac expression of sTnT was no longer detected. We previously identified a proximal promoter (-497 to +192 bp) of the rat cTnT gene, which was capable of driving cardiac-specific expression of a LacZ reporter gene in transgenic mice (Wang et al., 2000). Here we compared the pattern of LacZ expression during development in the cTnT-LacZ transgenic mice with the endogenous expression pattern of the cTnT gene and found that the transgene expression driven by the cTnT proximal promoter had a pattern similar to that of the endogenous cTnT gene.

MATERIALS AND METHODS

TnT Isoform-Specific cDNA Probes

Full-length cDNA (clone 6-4) encoding an embryonic mouse cTnT isoform was cloned from a λZAPII (Stratagene) phage library using specific antibody screening (Jin et al., 1996). A full-length cDNA (clone A3) encoding a mouse embryonic fast skeletal muscle TnT isoform was isolated by screening a λZAPII phage library using an adult mouse fast TnT cDNA probe previously cloned by reverse transcription-coupled polymerase chain reaction (Wang and Jin, 1997). A full-length cDNA encoding slow skeletal muscle TnT (high molecular weight isoform 2) was isolated from the neonatal mouse skeletal muscle cDNA library by antibody screening (Jin et al., 1998). All three cDNA clones were autoexcised from the λZAPII phage as recombinant pBlueScript SK (-) plasmids. The TnT isoform cDNA inserts were fully sequenced. A 303-bp cTnT-specific cDNA fragment from nucleotide (nt) 802 downstream of the ATG initiation codon to the end of 3′-noncoding region was isolated by PstI and XhoI digestion of the clone 6-4 (Fig. 1A). A 337-bp fTnT-specific cDNA fragment containing 82 bp of the 5′-noncoding region and 255 bp of the coding region was isolated by EcoRI and SacI digestion of the clone A3 (FA1e17 isoform) cDNA (Fig. 1A). A 120-bp sTnT-specific fragment containing 40 bp of the 5′-noncoding region and 80 bp of the coding region was isolated by EcoRI and AvaI digestion of the high molecular weight isoform 2 cDNA (Fig. 1A). These cDNA fragments were then subcloned into the pBlueScript SK vector and amplified for use as specific probes.

Figure 1. Dot blot analysis with the cTnT-, fTnT-, and sTnT-specific probes. A: Diagram of the mouse cTnT, fTnT, and sTnT cDNAs. The gene-specific sequences were identified by sequence comparison and indicated by bold black boxes on the cDNA diagram. B: Dot blot analysis with the cTnT-, fTnT-, and sTnT-specific probes. Plasmid containing full-length cDNA of cTnT, fTnT, or sTnT was loaded on a nitrocellulose membrane at an amount from 1 to 1/64 ng and hybridized with the cTnT-, fTnT-, and sTnT-specific probes. When a small amount of plasmid (≤ ¼ng) was applied, no cross-hybridization was observed among cTnT, fTnT, and sTnT.

Dot Blot and Northern Blot Analysis

The plasmids containing cTnT, fTnT, and sTnT cDNAs were 4-fold serially diluted and dot-blotted on a nitrocellulose membrane. The cTnT-, fTnT-, or sTnT- specific fragments were labeled with α-³²P-dATP (NEN, Boston, MA) using a random priming DNA labeling kit (Boehringer Mannheim, Indianapolis, IN). Hybridization was performed at 50°C for 2 hr in a QuikHyb (Stratagene, La Jolla, CA) hybridization solution. The membrane was then washed twice in 2 × SSC, 0.05% SDS at room temperature for 15 min each, followed by 30 min at 50°C in 0.1 × SSC, 0.1% SDS. The membrane was then exposed to X-ray film at −70°C with an intensifying screen. For Northern blot analysis, an adult multiple tissue blot purchased from Clontech (Palo Alto, CA) was examined with radio-labeled cTnT- or fTnT-specific probes. The hybridization and wash were carried out at 60°C. The same membrane was reprobed with a β-actin cDNA probe to serve as a loading control.

In Situ Hybridization

Whole-mount in situ hybridization was performed as described previously (Wang et al., 1996). Linearized plasmid containing the cTnT-specific cDNA was used as template for making digoxigenin labeled sense or antisense riboprobe by in vitro transcription (Wang et al., 1999). The background was determined by in situ hybridization on comparable embryos with sense control probes. After whole-mount in situ hybridization, selected embryos were embedded in Paraplast Plus (Oxford, St Louis, MO) and sectioned at 9 μm as described previously (Wang et al., 1996).

For in situ hybridization on tissue sections with radioactive riboprobes, embryos were fixed overnight at 4°C in 4% paraformaldehyde in PBS, dehydrated, cleared in xylene, and embedded in Paraplast Plus. Seven-micron sections were cut and mounted onto Superfrost plus slides (Fisher, Pittsburgh, PA) with DEPC-treated water, dried overnight at 40°C, and stored at room temperature. Before use, the slides were baked at 60°C overnight and then processed for in situ hybridization as described (Sassoon and Rosenthal, 1993). Slides were hybridized at 50°C, washed in 5 × SSC at 50°C, and then washed in 2 × SSC, 50% formamide at 60°C. Autoradiography was done with Kodak NTB-2 emulsion. The in situ hybridization with sense control probes was performed on adjacent sections to determine the background signal. The images were collected by a digital microscope camera directly attached to a microscope and processed using Adobe Photoshop®.

Transgenic Mice

cTnT-LacZ transgenic mice containing the LacZ reporter gene driven by the cTnT promoter (Wang et al., 1994) were generated as described (Sigmund, 1993). These procedures were carried out in the transgenic core facility at the University of Iowa. Three independent transgenic lines were analyzed and gave similar results. For each line, heterozygous mice were crossed with nontransgenic C57BL6/J × SJL F1 mice and embryos from day 7 to 19 p.c. as well as neonates (day 1 and 14) and adults were analyzed.

Whole-Mount X-Gal Staining

Embryos at day 7.5 to 19 p.c. were dissected from the uterus. For embryos at day 7.5 p.c., they were left in the yolk sac, and for those older than day 12.5 p.c., the thoracic wall was opened, exposing the heart to allow maximal penetration of fixatives and reagents. For neonates, head and limbs were removed and the thoracic wall was opened. Embryos and neonates were fixed in freshly prepared fixative solution (PBS containing 0.2% glutaraldehyde, 0.02% sodium deoxycholate and 0.01% NP-40) on ice for 5 min (day 7.5 to 9.5 p.c.) to 30 min (neonatal and adult hearts). Following fixation, tissues were permeabilized in PBS containing 0.02% sodium deoxycholate and 0.01% NP-40 for at least 3–4 hr at room temperature. X-gal staining was performed in PBS solution containing 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, 2 mM MgCl₂ and 1 mg/ml X-gal at 30°C for 30 min to overnight. After whole-mount staining, selected embryos were embedded in Paraplast Plus and sectioned at 7 μm as described previously (Wang et al., 1996).

RESULTS

Specificities of cTnT, fTnT, and sTnT-Specific Probes

Since the cTnT, fTnT, and sTnT mRNAs share high sequence homology, cTnT-, fTnT-, and sTnT-specific cRNA probes are required for distinguishing the expression patterns of cTnT, fTnT, and sTnT by in situ hybridization. Through sequence comparison, cTnT-, fTnT- and sTnT-specific cDNA fragments were identified (Fig. 1A), and their specificity was tested by dot blot analysis (Fig. 1B). The labeled cTnT-, fTnT-, and sTnT-specific probes specifically recognized their own cDNAs and when appropriate amounts of cDNA (≤ 0.25 ng, 0.4 fmol) was applied, no cross-hybridization was observed. Since in the adult mouse cTnT and fTnT were solely expressed in cardiac and skeletal muscles, respectively, the specificity of the cTnT- and fTnT-specific probes was further confirmed by Northern blot analysis using an adult mouse multiple tissue blot (Clontech), which contains poly (A)⁺ RNAs isolated from heart, brain, spleen, lung, liver, skeletal muscle, kidney, and testis. The cTnT-specific probe specifically hybridized with the cTnT mRNA from the heart (Fig. 2A) while the fTnT-specific probe specifically hybridized with the fTnT mRNA from the skeletal muscles (Fig. 2B). There was no cross-hybridization between cTnT and fTnT probes, even in the tissues highly enriched in these messages. sTnT is not expressed in most of the adult mouse muscles (Jin et al., 1998) and was, therefore, not tested by Northern analysis on the commercial multiple tissue blot.

Figure 2. Multiple tissue Northern blot analysis using the cTnT- and fTnT-specific probes. A: Multiple tissue Northern blot analysis using the cTnT-specific probe. An adult multiple tissue blot (Clontech) containing poly (A)⁺ RNAs isolated from various mouse tissues was hybridized with the cTnT-specific probe. The cTnT-specific probe only recognized the cTnT mRNA in the heart. B: Multiple tissue Northern blot analysis using the fTnT-specific probe. The fTnT-specific probe only recognized the fTnT mRNA in the skeletal muscle. In both experiments, hybridization using β-actin probe serves as an RNA loading control.

Expression of the cTnT, sTnT, and fTnT Genes During Development

Although no expression could be detected at day 7.0 p.c. with the cTnT probe, at day 7.5 p.c., cTnT expression was clearly visible in the lateral plate mesoderm in the region of the cardiogenic plate (Fig. 3A and B). By day 8.0 p.c. the cardiac precursor cells have fused to form a linear heart tube and, at this stage of development, cTnT expression was specifically seen in the heart tube as revealed by whole mount in situ hybridization (Fig. 3C). From cross-sections it could be seen that the expression was limited to the myocardial cells of the forming heart (Fig. 3D). At day 10.5 p.c., cTnT was strongly expressed in the heart and weakly in a few of the most posterior somites by section in situ hybridization (data not shown). However no signal was seen at this stage with the fTnT and sTnT probes (data not shown). In a separate experiment using a more sensitive whole-mount in situ hybridization, cTnT messages were detected in all the somites of day 10.5 p.c. and day 12.5 p.c. embryos (data not shown). As will be described below, the transgene expression driven by rat cTnT promoter also can be detected in all the somites of day 10.5 p. c. embryo (see Fig. 8).

Figure 3. In situ hybridization on a section of a mouse embryo at day 7.5 p.c. using the ³⁵S-labeled cTnT-specific probe (A,B). The embryo was left in the yolk sac and cross-sectioned from rostral to caudal. A: Bright-field view showing the embryo structures. B: Corresponding dark-field view showing the cTnT mRNA signals. At this stage, the cTnT gene is specifically expressed in the lateral plate mesoderm (cardiogenic plate). Bar = 115 μm. NF, neural fold; LM, lateral plate mesoderm. Whole mount in situ hybridization and a section of a mouse embryo at day 8 p.c. using the cTnT-specific probe (C,D). C: The positive signal is solely detected in the linear heart tube. D: Cross-section of the embryo shown in C. The positive signal is in the myocardial cells of the heart tube. A black line in C indicates the orientation of the section. Bar = 115 μm. H, heart tube; NT, neural tube; FG, foregut; Mc, myocardium.

Using ³⁵S-labelled probes on day 11.75 p.c. sections, cTnT expression was again very strong in the heart and the tip of the tail, and at a lower level in the posterior somites (Fig. 4A and B). In addition, strong signal was seen in the mesenchyme around the cardinal veins and the sinus venosus and a weak signal in the somites (Fig. 4C and D). The faint signals in other tissues were the same as the background signals obtained by in situ hybridization on adjacent sections with sense control probe (data not shown). With the sTnT probe there was weak but significant signal in the center of the mandible, the ventral portion of the heart, and the tip of the tail (Fig. 4E and F). As compared to the signal in the sense control section (data not shown), the fTnT probe showed only a very faint signal in the tip of the tail, suggesting that the fTnT gene may be expressed later than the sTnT gene (Fig. 4G and H).

Figure 4. In situ hybridization on sagittal and cross-sections of day 11.75 p.c. embryos using the cTnT-, sTnT-, or fTnT-specific probe. A,C,E,G: Bright-field views showing the embryo structures. B,D: Dark-field view of A and C showing the cTnT mRNA signal. Strong signal was detected in the heart, tip of the tail, somites, and in the mesenchyme around the cardinal vein and sinus venosus. F: Dark-field view of E showing the sTnT mRNA signal. Very weak signal was detected in the mandible, heart, and tip of tail. H: Dark-field view of G showing the fTnT mRNA signal. Very faint signal was found in the tip of the tail. The other signals seen in B,D,F,H were considered as background. The background signal was determined by in situ hybridization on adjacent section with each of TnT sense control probes (data not shown). Bar = 0.5 mm. Hd, head; M, mandible; H, heart; S, somites; Cv, cardinal veins; Sv, sinus venosus.

By embryonic day 13.5 there was strong cTnT expression in the heart and tongue, and weak but significant expression in the crus of the diaphragm, tail muscles, muscles around the vertebrae, and urogenital sinus (Fig. 5A and B). Faint cTnT signal was visible in the developing muscles of the body and limbs (Fig. 5C and D). Strong sTnT expression could be seen in the developing muscles of the body and limbs, tail muscles, muscles around the vertebrae, crus of the diaphragm, and tongue (Fig. 5E-H). Also noteworthy was the clear expression of sTnT in the ventral portion of the ventricle, particularly in the ventricular groove (Fig. 5F and H). It was at this stage of development that fTnT expression was first strongly seen in the developing muscles of the limb, body, tail, tongue, and crus of the diaphragm (Fig. 5I–L). At this stage, the fTnT gene had no expression in the heart in contrast to the sTnT gene.

Figure 5. In situ hybridization on sagittal and cross-sections of day 13.5 p.c. embryos using the cTnT-, sTnT-, or fTnT-specific probes. A,C,E,G,I,K: Bright-field views showing the embryo structures. B,D: Dark-field view of A and C showing the cTnT mRNA signal. F,H: Dark-field view of E and G showing the sTnT mRNA signal. J,L: Dark-field view of I and K showing the fTnT mRNA signal. Bar = 1 mm; D, diaphragm; T, tongue; H, heart; V, vertebrae; Us, urogenital sinus; M, muscle in body.

At the latest embryonic stage (day 16.5 p.c.), we observed that cTnT expression was again very strong in the heart and bladder, with somewhat weaker expression in the tongue and lower jaw (Figs. 6, 7A,B) and even fainter expression in the muscles of the body and limbs (Fig. 7A and B). The sTnT probe gave some signal in the tongue and lower jaw, with slightly stronger signal in the muscles around the ribs (Fig. 6C and D) and very weak signal in the muscles of the body and limbs (Fig. 7C and D). No sTnT expression was seen in the heart at this stage. fTnT expression was very high in the tongue and lower jaw and in the muscles around the vertebrae, ribs, and pubic bone (Fig. 6E and F). Some signal was also seen in the body of the diaphragm (Fig. 6E and F). Very strong signal was seen with the fTnT probe in the muscles of the body and limbs (Fig. 7E and F). At this stage sTnT expression was weak as compared to cTnT and fTnT expression and, in contrast to the cTnT genes, neither the sTnT nor the fTnT gene was expressed in the bladder (Fig. 6). The expression of cTnT in the bladder but not other smooth muscles during embryonic development might suggest that smooth muscle in the bladder has a distinct structure and function. The significance of TnT expression in smooth muscle development remains to be investigated.

Figure 6. In situ hybridization on sagittal sections of a day 16.5 p.c. embryo using the cTnT-, sTnT-, or fTnT-specific probe. A,C,E: Bright-field views showing the embryo structures. B: Dark-field view of A showing the cTnT mRNA signal. D: Dark-field view of C showing the sTnT mRNA signal. F: Dark-field view of E showing the fTnT mRNA signal. Bar = 1 mm; T, tongue; J, lower jaw; R, ribs; H, heart; V, vertebrae; D, diaphragm; B, bladder.

Figure 7. In situ hybridization on cross-section of a day 16.5 embryo using the cTnT- sTnT-, or fTnT-specific probe. A,C,E: Bright-field views showing the embryo structures. B: Dark-field view of A showing the cTnT mRNA signal. D: Dark-field view of C showing the sTnT mRNA signal. F: Dark-field view of E showing the fTnT mRNA signal. Bar = 0.5 mm; H, heart; M, muscle in body; L, muscle in limb.

Expression Pattern of a Reporter Gene Driven by the Rat cTnT Promoter Is Similar to the Endogenous cTnT Gene During Mouse Development

We previously have shown that the proximal −497 bp promoter of the rat cTnT gene could drive cardiac-specific expression of an LacZ reporter gene in adult mice (Wang et al., 2000). The expression pattern of the cTnT-LacZ transgene was observed in developing mice of 3 independent founders. The LacZ reporter gene driven by the −497-bp promoter was first detected in the lateral plate mesoderm at day 7.5 p.c., and as can be seen from the cross-sections, the expression was specific to the mesoderm cells of the cardiogenic plates (Fig. 8A). At day 8.0 p.c., LacZ expression was restricted to the myocardial cells in the linear heart tube (Fig. 8B). As development proceeded, LacZ expression was distributed uniformly throughout both atria and ventricles of developing mouse embryos (Fig. 8C and D). From day 10.5 p.c., LacZ expression was also detected in developing somites (Fig. 8C), agreeing with the whole-mount in situ data. These results suggest that β-gal staining and whole-mount in situ are more sensitive assays than in situ done on sections. At later stages (day 13.5p.c.), LacZ was expressed in the developing skeletal muscles at a level much lower than that in the developing heart (Fig. 8D). In transgenic mice at day 1 and day 14, LacZ expression was specifically seen in the heart, and pulmonary vessels (Fig. 8E and F). On the other hand, the LacZ expression in skeletal muscle was undetectable. We have not yet confirmed cTnT endogenous gene expression in the pulmonary vessels by in situ hybridization. In the adult, LacZ was expressed only in the heart but not other tissues including intestine, lung, liver, kidney, spleen, skeletal muscle, stomach, and brain (data not shown).

Figure 8. β-gal staining of cTnT-LacZ transgenic embryos. A: Cross-section of a day 7.5 p.c. transgenic embryo after β-gal staining. Bar = 45 μm. The positive cells are the lateral plate mesoderm cells in the pair of cardiogenic plates. NF, neural fold; LM, lateral plate mesoderm. B: Cross-section of a day 8 p.c. transgenic embryo after β-gal staining. Bar = 115 μm; NT, neural tube; FG, foregut. The positive cells are in the myocardium (Mc). C: β-gal staining of a day 10.5 p.c. transgenic embryo. At this stage, the β-gal staining was strongly detected in the developing heart, and also in the developing somites. Bar = 1 mm; H, heart; S, somites. D: Ventral view of a day 13.5 p.c. transgenic embryo after β-gal staining. The β-gal staining was strongly detected in the developing heart, and also in the developing skeletal muscles at a low level. Bar = 1 mm; H, heart; M, skeletal muscle. E: The β-gal staining of the heart of a day 1 LacZ newborn. Bar = 465 μm. F: The β-gal staining of the pulmonary vessels of a day 1 LacZ newborn. Bar = 585 μm.

DISCUSSION

In the present study, the expression patterns of cTnT, fTnT, and sTnT genes during mouse development at the transcriptional level were described for the first time using in situ hybridization. The probes used here were specific for their respective genes as revealed by both dot blot and Northern blot analyses (Figs. 1,2). Furthermore, each probe was designed to recognize all alternatively spliced isoforms derived from the particular TnT gene. Therefore, the expression patterns determined by these probes should be specific and representative. However, it should be noted that the presence of detected messages in tissues might not warrant the expression of the gene of interest at the translational level. Since cTnT, sTnT, and fTnT proteins are highly conserved, antibodies used to examine the expression patterns of TnT genes often cross-react each other (Saggin et al., 1990; Sabry and Dhoot, 1991). Although gene-specific monoclonal antibodies to the cTnT protein have been reported (Saggin et al., 1990; Malouf et al., 1992), their specificity has not been extensively tested on individual skeletal muscle fibers during early stages of mouse development. Thus, the determination of TnT protein expression patterns during mouse development will be dependent on the availability of gene-specific antibodies.

Transient Expression of cTnT in Developing Skeletal Muscles

The cTnT message was detected as early as day 7.5 p.c. in the lateral plate mesoderm. As the heart tube formed and developed into a four-chamber heart, cTnT was strongly expressed in the myocardial cells. Although the first somite could be morphologically identified in day 8.0 p.c. (Tam et al., 2000), TnT genes were not expressed until day 10.5 p.c., suggesting that the full differentiation of muscle had not yet taken place in developing somites. The first TnT transcripts detected in developing somites was cTnT. This somite expression of cTnT gene was further confirmed by whole-mount in situ hybridization on day 10.5 p.c. embryos (data not shown) and by the LacZ expression in day 10.5 p.c. embryos of transgenic mice carrying the cTnT-LacZ transgene (Fig. 8C). At day 11.75 p.c. a faint signal of both sTnT and fTnT mRNAs could be detected in the developing somites. Overall, as development progresses, the level of cTnT expression in developing mouse skeletal muscles maintains a minimal but significant amount. This is in contrast to the previous finding in chick that cTnT is the predominant form expressed in developing skeletal muscles (Toyota and Shimida, 1981; Cooper and Ordahl, 1984; Swiderski and Solursh, 1990; Wong and Ordahl, 1996). The transient expression of the cTnT gene in skeletal muscles was no longer detectable in neonates as revealed by the transgene analyses on days 1 and 14 transgenic mice. Similar low-level and transient expression of the cTnT gene in developing skeletal muscles was also found in the rat by Western blot analyses (Saggin et al., 1990; Sabry and Dhoot, 1991) and by Northern blot analysis (data not shown). Using a monoclonal antibody CT3 against bovine cardiac TnT, somite expression of cTnT was also found in developing Xenopus embryos at stage 24–26 (Kolker et al., 2000).

The significance of transient expression of the cTnT isoform in developing skeletal muscles remains to be determined. Previously, we have shown that force-expressed rat cTnT can associate with skeletal muscle tropomyosin and assemble onto microfilaments in cultured skeletal myogenic cells (Warren and Lin, 1993), suggesting that the expressed cTnT can be functional in the skeletal muscle environment. Therefore, expression of cTnT protein in the developing skeletal muscle may be an important supplement for contraction until sufficient skeletal muscle isoforms are obtained for a particular muscle fiber. Indeed, when fTnT and sTnT are strongly expressed at day 16.5 p.c., the cTnT expression in the skeletal muscle is significantly decreased and the cTnT is not a major isoform in these tissues. Although the question of why the cTnT isoform is expressed in developing skeletal muscles remains to be investigated, this phenomenon appears to be conserved among rat (Saggin et al., 1990), mouse, rabbit (Briggs et al., 1987), chick (Toyota and Shimida 1981; Cooper and Ordahl, 1984; Swiderski and Solursh 1994), and possibly xenopus (Kolker et al., 2000). Developing skeletal muscles may need contraction for muscle growth and for the morphogenesis of skeletal tissues such as tendons, cartilage, and bones. In development, cTnT and/or sTnT isoforms may be more appropriate for the relatively low load-bearing embryonic muscles, than the fTnT isoform.

All Three TnT Genes Are Expressed in the Developing Tongue

At day 13.5 p.c., all three TnT messages were strongly expressed in the tongue. Interestingly, sTnT and fTnT mRNAs seemed to be restricted to the anterior and dorsal, respectively, portions of the tongue, whereas the region of cTnT transcripts appeared to be a combination of sTnT- and fTnT-positive regions. When sections of a day 13.5 p.c. embryo were immunohistochemically stained with a monoclonal antibody against cTnT, a similar cTnT expression pattern was detected (data not shown), suggesting that the detected cTnT messages appeared to be translated into protein. As development proceeds, fTnT becomes the predominant form in the tongue. At day 16.5 p.c., the overlapping expression of all three TnT messages in the tongue is much more apparent, with the cTnT isoform enriched in the periphery of the tongue. The expression pattern in the tongue for fTnT at this stage appears to be a combination of both cTnT and sTnT expression patterns. Furthermore, the expression level of fTnT exhibits a progressive decrease from the base to tip of the tongue, suggesting that the development of the striated muscles in the tongue may follow a base to tip progression. This overlapping expression pattern of the three genes seems to be continuous with that in the lower jaw. This result suggests that both jaw muscles (originating from somites 4 and 6) and lingual muscles (originating from occipital somites 1 through 7) may have a common origin (Carlson, 1988). Coincidentally, the pattern of cTnT expression in the tongue is very similar to that reported previously for a homeobox-containing transcription factor, NKx2.5 (Lints et al., 1993), which is essential for normal cardiac morphogenesis and myogenesis (Lyons et al., 1995). Thus, Nkx2.5 may be one of the transcription factors regulating the expression of the cTnT gene. Further sequence analysis on the rat and mouse cTnT promoter reveals the presence of a consensus Nkx2.5 binding sequence (Jin et al., 1995; Wang et al., 2000). This may partly explain the observed expression of the cTnT gene in the tongue. The significance of all three TnT isoforms in the developing tongue remains to be determined. It may provide a functional role for the developing tongue, particularly considering that the tongue is a partly free and highly mobile organ and that there is no bone in the tongue to allow lingual muscles to attach. Although the major isoform detected in the adult rabbit tongue is fTnT, a significant amount of the cTnT protein is also found to be associated with tongue myofibrils (Briggs et al., 1987). Thus, it is likely that at least both fTnT and cTnT may still be co-expressed in the adult mouse tongue.

Similar expression of all threeTnT genes can be detected in the muscles at the tip of the tail and the crus of the developing diaphragm at day 13.5 p.c. (Fig. 5). The crus may represent a specialized musculature, serving as the tendinous attachments of the diaphragm to the lumbar vertebrae forming the sites for the aortic opening. At this stage, however, the body of the diaphragm itself did not show expression of these TnT genes. Later, both sTnT and fTnT messages but not cTnT message were found in the body of diaphragm (Fig. 6). This is consistent with the previous report that the sTnT but not the cTnT isoform is expressed in the diaphragm of day 17 p.c. rat embryo (Sabry and Dhoot, 1991).

Expression of cTnT in the Developing Bladder

The cTnT gene, but not the fTnT or sTnT gene, was expressed in the developing urogenital sinus and bladder as detected by in situ hybridization. The function of the bladder is to store urine at a low pressure and then to periodically expel the urine through coordinated and sustained contractions. The musculature in the bladder wall is smooth muscle. However, during development, a transient co-expression of skeletal muscle and smooth muscle actin has been previously reported in human fetal bladder (Shapiro, 1999). At 16 weeks of gestation, the bladder no longer expresses skeletal muscle actin and contains only smooth muscle actin, indicative of smooth muscle differentiation. The author thus suggests that a developmentally programmed transdifferentiation may occur in the developing bladder muscle (Shapiro, 1999). The cTnT messages detected in the bladder of a mouse embryo at day 16.5 p.c. may represent a stage of this transdifferentiation. In a preliminary Western blot analysis with a monoclonal antibody to cTnT, CT3, we can detect cTnT protein in a total extract of day 16.5 p.c. embryonic bladder but not in the adult bladder (data not shown). This result suggests that functional striated muscle may exist in day 16.5 p.c. embryonic bladder. It will be of interest to determine whether other striated proteins are also present in the developing bladder.

It is known that human fetal bladders are functional (processing 5 ml urine per hour) at 20 weeks of gestation (Nguyen and Kogan, 1999), when the bladder contains no skeletal muscle actin (Shapiro, 1999). To our knowledge, there was no previous report that mouse embryonic bladders at day 16.5 p.c. are functional or that bladder smooth muscle at this stage is fully differentiated. If this is the case, persistence of a striated muscle protein, cTnT, in the bladder of a mouse embryo at a late stage of development may point to a specialized function for bladder smooth muscle. It is worthy to note that the adult body wall muscle of the ascidian Halocynthia roretzi is a morphologically nonstriated smooth muscle and contains a troponin complex consisting of three subunits T, I, and C (Endo et al., 1996). Experiments are currently designed to distinguish whether the presence of cTnT protein in the developing bladder represents a transdifferentiation or a co-expression.

There Is a Low-Level Expression of sTnT and No Expression of fTnT in Developing Heart

At day 13.5 p.c., sTnT but not fTnT mRNA was present in the heart. Compared to the cTnT signal, the sTnT signal in the heart was much weaker and only in the ventral portion of the ventricles. This weak sTnT expression in the heart was no longer detectable in the day 16.5 p.c. embryo. Our finding on the temporal pattern of sTnT expression in the developing mouse heart seems to be inconsistent with a previous report showing the presence of small amounts of embryonic sTnT isoforms in the postnatal cardiac muscle of the rat (Sabry and Dhoot, 1991). Since these embryonic sTnT isoforms detected by the monoclonal antibody CDC4 have not been further characterized in terms of their primary sequences and gene families, this discrepancy may be simply explained by non-specific reacting bands recognized by the antibody used (Sabry and Dhoot, 1991).

Acknowledgements

The authors thank Douglas L. Schulte and Jenny L.-C. Lin for technical assistance and Dr. Karla J. Daniels for critical discussion of the manuscript. This work was supported by an American Heart Association grant and NIH grants HL62178 (SCOR in congenital heart disease) and HD18577 to J.J.-C.L. and by a grant from the Medical Research Council of Canada to J.-P.J.

• 1

  Q.W. and R.S.R. contributed equally to this work. Qin Wang's present address is Department of Pharmacology, Vanderbilt University Medical Center, Nashville, TN 37232.