Depletion of Zebrafish Titin Reduces Cardiac Contractility by Disrupting the Assembly of Z-Discs and A-Bands

Abstract

The genetic study of titin has been notoriously difficult because of its size and complicated alternative splicing routes. Here, we have used zebrafish as an animal model to investigate the functions of individual titin isoforms. We identified 2 titin orthologs in zebrafish, ttna and ttnb, and annotated the full-length genomic sequences for both genes. We found that ttna, but not ttnb, is required for sarcomere assembly in the heart as well as the subsequent establishment of cardiac contractility. In fact, ttna is the earliest sarcomeric mRNA that is expressed in the heart, which makes it an early molecular marker for cardiomyocyte differentiation. Surprisingly, ttna is required for later steps of sarcomere assembly, including the assembly of Z-discs and A-bands, but not for early steps such as the assembly of Z-bodies and nonstriated myosin filaments. Reduction of individual titin isoforms in vivo using morpholino-modified antisense oligonucleotides indicated that (1) both N2B exon—containing and N2A exon—containing isoforms of ttna are required for sarcomere assembly in the heart; (2) N2A exon—containing isoforms of both ttna and ttnb are required for sarcomere assembly in the somites; and (3) the N2B exon—containing isoforms of ttnb are expressed later than other titin isoforms and are probably involved in modulating their expression; however, these isoforms of ttnb are not required for sarcomere assembly. Collectively, our results reveal distinct functions of different titin isoforms and suggest that various phenotypes in “titinopathies” may be attributable to the disruption of different titin isoforms.

The sarcomere is the basic contractile unit of striated muscle and is composed of 3 filament systems: myosinbased thick filaments (A-band), actin-based thin filaments (I-band), and titin filaments. Titin is the largest known protein in humans and probably all vertebrates,¹ having a molecular mass of 3 to 4 megadaltons (MDa). Furthermore, a single titin molecule spans half of the sarcomere, a distance of approximately 1 μm. Protein domains of titin correspond with substructures in the sarcomere in an N- to C-terminal fashion, such that the N terminus of titin is anchored in the Z-disc where actin filaments are anchored, whereas its C terminus extends into the M-band where myosin filaments are anchored.² Thus, titin is proposed to serve as the template for sarcomere assembly by serving as a structural scaffold,³ and the A-band region of titin is proposed to function as a molecular ruler to define the length of myosin filaments.⁴ In addition to its structural functions, titin is postulated to function as a signaling molecule. Individual domains of titin associate with an extensive network of protein complexes,^5,6 some of which are involved in force transmission and stretch sensing, whereas others are potentially involved in gene regulation.^7-14 With these various structural and signaling functions, it is of no surprise that mutations in titin have been linked to a variety of sarcomeric diseases,^15,16 including dilated cardiomyopathy,^17,18 hypertrophic cardiomyopathy,¹⁹ and tibial muscular dystrophy or limb girdle muscular dystrophy type 2J.²⁰

The diversity of “titinopathies” may result from distinct functions of different titin domains or the complicated alternative splicing events used to generate the various isoforms of this giant molecule. Hundreds of alternative splicing products can be generated from a single titin gene locus, which are broadly categorized into N2A- and N2B-based isoforms and the short Novex isoforms.^21,22 The N2B domain is encoded by a single 2.7-kb exon in the I-band region and is thought to be expressed specifically in the heart. The N2A domain is encoded by eight exons in the I-band region and is thought to be expressed in both heart and skeletal muscle.²¹ Despite extensive efforts,^23-25 we still have limited knowledge on the functions of the different titin domains and alternative splicing isoforms, partially because of the lack of an ideal model organism for studying the in vivo functions of this large protein. We have exploited zebrafish as an in vivo vertebrate model for studying titin and its role in sarcomere assembly and sarcomeric diseases. Some key attributes of this model system are that (1) zebrafish embryos are transparent and develop quickly ex utero; (2) the zebrafish genome can be easily manipulated using either transgenic or knockdown technologies²⁶; and (3) zebrafish embryos can survive without a heart for 5 days, because the body is small and oxygen can be delivered by simple diffusion,²⁷ a unique feature making zebrafish particularly advantageous for studying sarcomere assembly in the heart.

Previously, we identified titin as the gene for pickwick, a group of zebrafish mutants identified through mutagenesis screening.²⁸ We have established the pik^m171 allele as an embryonic model for human dilated cardiomyopathy. Here, we identified and annotated the full-length sequences of zebrafish titin orthologs, possibly the largest genes in this organism. Interestingly, 2 titin orthologs, ttna and ttnb, are present in this teleost fish model, and different expression patterns and functions exist for the different titin isoforms during embryogenesis. We show that ttna, but not ttnb, is the earliest sarcomere gene expressed in the heart, the depletion of which reduces cardiac contractility by disrupting sarcomere assembly. Finally, our genetic data suggest that titin is not required for the assembly of Z-bodies or nonstriated myosin filaments, but rather for the later steps of sarcomere assembly involving the formation of Z-discs and A-bands.

Materials and Methods

Cloning and Annotation of Titin

The zebrafish genome database zV4 was searched using human titin peptide sequences. We identified 4 contigs (dZ71F1, AL772356, AL732421, and AL714003) that encode peptide sequences highly orthologous to human titin. The latter 3 contigs have been assigned to chromosome 9 and the DNA sequences from the 4 contigs overlap with each other. We confirmed the location of these contigs to the same chromosomal locus by radiation hybrid mapping (data not shown). Four pairs of PCR primers were designed, targeting the intron regions of the 4 contigs. They were sent to Dr Zhou Yi at Children’s Hospital (Boston, Mass) to search for chromosomal location using the T51 RH panel. The sequences of the primers are available on request.

The zebrafish sequences were manually assembled into a single contig using DSgene. Exon—intron boundaries were determined using GENSCAN. The predicted sequences were confirmed by aligning them with the corresponding human titin sequence using ClustalW. The identity of the domains that are encoded by each exon was manually determined using the annotation of human titin for comparison.²² The complete sequences have been deposited into GenBank (accession no. DQ649453).

Injection of Morpholinos

Morpholino antisense oligonucleotides that targeted splicing donor sites were purchased from Gene Tools LLC and prepared and injected as previously described.²⁶ The sequences were as follows:

• ttna_N2B MO (targets e45): 5′-CAAGAGGTTGGAATTATGGGAATAT

• ttna_N2A MO (targets e79): 5′-GTGGAAGACCGGTAAGATTACATCT

• ttnb_N2B MO (targets e37): 5′-ACATTTTCTGTAAGAATAAAAGGTG

• ttnb_N2A MO (targets e47): 5′-ACCAAAGTCACAATCAAAGGTAATT

Antibody Staining and Histology

Whole-mount immunofluorescence experiments were performed as previously described.²⁹ The following antibodies were used at the indicated dilutions: anti—sarcomeric α-actinin (clone EA53; Sigma) at 1:1000 and F59³⁰ at 1:10. The heart was dissected using forceps. Images were acquired using an AxioplanII Zeiss microscope. Embryos were embedded in JB4 (Polysciences) and 5-μm sections of the embedded specimens were collected and subjected to hematoxylin and eosin staining. Images were acquired using an AxioplanII Zeiss microscope equipped with a Nikon 8700 digital camera.

Results

Complete Gene Sequence of Two Titin Orthologs in Zebrafish

Human titin spans 283 kb of genomic region, comprising 363 exons and having a coding capacity of 38,138 amino acids.²² We searched the zebrafish genome database and identified 4 contigs that encode peptide sequences highly orthologous to human titin. Subsequently, we determined that 2 titin genes, ttna and ttnb, are present in zebrafish and are located in tandem in a head-to-tail fashion on chromosome 9 (Figure 1A). Zebrafish ttna spans a 176-kb region of the genome, consists of 235 exons, and is capable of encoding 35 173 amino acids, whereas zebrafish ttnb spans 137 kb, consists of 201 exons, and is capable of encoding 29 094 amino acids. Next, we determined the identity of each exon by manually aligning the predicted peptide sequences encoded by the 2 zebrafish titin genes with human titin (Figure 1B). Similar to human titin, zebrafish ttna and ttnb mainly consist of repeats of immunoglobulin C2 (Ig) domains and fibronectin type III (FN3) domains. In the Z-disc, A-band, and M-line regions, the exon—intron structures, the number of exons, and even the length of each exon are quite similar among ttna and ttnb of zebrafish and human titin (see the table in File S1 of the online data supplement, available at http://circres.ahajournals.org). The sequences of ttna and ttnb diverge considerably from that of human titin in the I-band region. This is in agreement with previous reports that, in mammals, alternative splicing of titin frequently occurs in this region.¹⁷ ttnb does not encode the 6-kb Novex III exon, which suggests that ttnb does not have the capacity to encode the short Novex isoforms of titin.²² By RT-PCR analysis, we show that all 3 major exon skipping events in human titin are conserved in zebrafish (see supplemental Figure I).

Figure 1.

Zebrafish has 2 titin orthologs. A, Genomic structure of ttna and ttnb on chromosome 9. The 2 zebrafish titin orthologs are located in tandem array. B, Schematics of the exon—intron structures of ttna and ttnb. The detailed exon—intron information is summarized in the supplemental table (File S1). The genomic sequences have been deposited into GenBank. The accession no. is DQ649453. PEVK indicates Pro-Glu-Val-Lys; UTR, untranslated region.

Titin Isoforms Are Differentially Expressed During Embryogenesis

We next examined the expression patterns of N2A- and N2B-based isoforms of ttna and ttnb during zebrafish embryogenesis. Riboprobes targeting the N2B or the N2A region were synthesized, as indicated by green boxes in supplemental Figure IB. The signals detected following whole-mount in situ hybridization using these riboprobes should reflect the mRNA expression pattern of a group of N2B or N2A exon—containing isoforms of ttna or ttnb, which were named ttna_N2B, ttna_N2A, ttnb_N2B, and ttnb_N2A, respectively.

In somites, the expression of both ttna and ttnb can be detected at the 1 somite (S) stage in the adaxial mesoderm cells, corresponding with the expression of MyoD, a transcription factor that demarcates commitment to the myoblast cell fate (Figure 2G, 2M, 2Y, and 2e). The expression is earlier than all known sarcomere genes, including desmin, tpma, tnnc, acta1, and myhz1, whose expression in the somites start at the 6 to 15 S stages.³¹ The expression of ttna and ttnb in the somites becomes segmented at the 10 S stage, although to a lesser extent than MyoD, continues to get stronger until the 16 S stage and reaches a peak at 24 hours postfertilization (hpf). Starting at 24 hpf, the expression patterns of N2B-based isoforms of both ttna and ttnb differ from those of N2A-based isoforms. The expression of ttnb_N2B in the somites becomes weaker in comparison to the other isoforms (Figure 2c), whereas the expression of ttna_N2B isoforms totally switches off in the somites (Figure 2K). At the cellular level, the mRNA signals of either ttna_N2A or ttnb_N2A are located on the border of myoseptum (Figure 2Q’ and 2i’), whereas ttnb_N2B is expressed in the middle part of each somite segment, similar to MyoD and other sarcomere genes such as mylz2, desmin, and myhz1 (Figure 2c’ and 2o” and elsewhere³¹). The mRNA levels of N2A-based isoforms decrease between 48 and 72 hpf, when expression in the appendage muscles becomes apparent (data not shown).

Figure 2.

mRNA expression of titin isoforms in the somites. Shown are images acquired following whole-mount in situ hybridization using riboprobes represented by the green boxes in supplemental Figure I. Dorsal views are shown of embryos at 1, 5, 10, and 16 S stages, whereas lateral views are shown of embryos at 24 and 48 hpf. Q’, c’, i’, and o’ are highmagnification images of the somite staining in 24-hpf embryos, whereas Q”, c”, i”, and o” are images of cross sections of the same embryo. As a control, we generated 2 riboprobes that target exons in the M-line region, which are predicted to be constitutive exons that exist in both N2A- and N2B-based isoforms (named ttna_M and ttnb_M). Indeed, the expression pattern of ttn_M riboprobes is a combination of the ttn_N2A and ttn_N2B riboprobes, as shown in A through F and S through X.

The expression of ttna in the cardiac progenitors can be detected at the 5 S stage in the anterior lateral plate mesoderm (ALPM), corresponding with the expression of nkx2.5, the earliest cardiogenic transcription factor (Figure 3). The onset of ttna expression in the heart is earlier than that of other sarcomere genes such as vmhc and cmlc2, whose expression in the ALPM begins around the 10 to 16 S stages (Figure 3U through 3d and elsewhere³²). Both N2A- and N2B-based isoforms of ttna are expressed in the heart as early as the 5 S stage (Figure 3A,F), whereas expression of the ttnb_N2B isoforms in the heart starts at 30 hpf, much later than ttna, and the expression level is much weaker (Figure 3O). Expression of ttnb_N2A isoforms was not detected in the heart.

Figure 3.

mRNA expression of titin isoforms in the heart. Shown are dorsal views of embryos at the indicated stages following whole-mount in situ hybridization using riboprobes represented by the green boxes in supplemental Figure I. A no-tail riboprobe was included in all experiments to stain the notochord in the midline, which was used as a reference to indicate the location of the cardiac field. nkx2.5, cmcl2, and vmhc were included as controls. As indicated by nkx2.5 staining, the heart field in zebrafish appears as 2 strips in the anterior lateral plate mesoderm (ALPM) at the 5 S stage and later migrates to the midline and fuses to form the heart tube. The onset of cardiac expression for each riboprobe is indicated by arrowheads.

ttna but Not ttnb Is Required for Establishment of Cardiac Contractility

To examine the functions of each of the 4 groups of titin isoforms, we injected morpholino antisense oligos that target the splice donor sites represented by () in supplemental Figure IB. We performed real-time RT-PCR analysis to evaluate the knock down efficiency of the targeted splicing events. We found that reduced expression of individual titin isoforms was not compensated for by an upregulation in the expression of the other titin isoforms in the same tissue (see supplemental Figure II).

When expression of the ttna_N2B isoform was reduced to 9%, 99% (167/168) of morphants exhibited cardiac defects that resembled those of pickwick (pik) mutants. Pericardiac edema developed and the contractility of the heart was strongly reduced (for movies of the mutant and morphants, see supplemental Files S2 through S6). Similar cardiac defects were also observed in 84% (276/330) of the ttna_N2A isoform morphants when expression was reduced to 5%. We quantified cardiac contractility by measuring shortening fraction of the ventricular cardiomyocytes. As shown in Figure 4A, shortening fraction was reduced from 25% in wild-type embryos to 5% and 7% in ttna_N2A and ttna_N2B morphants, respectively. Reduction of the ttnb_N2B isoform to 3% resulted in mild cardiac edema in 94% (247/263) of embryos; however, cardiac contractility was only slightly affected (Figure 4A). As expected from its expression pattern (Figure 3P through 3T), expression of the ttnb_N2A isoform was not required for cardiac function. Only 5% (13/276) of morpholino-injected embryos developed cardiac edema, probably a consequence of experimental disturbance during the microinjection procedure. Collectively, these results demonstrate that ttna, but not ttnb, is required for the establishment of cardiac contractility during zebrafish embryogenesis. In contrast, heart rate was normal in all morphants, indicating that titin is not required for the establishment of cardiac rhythm (Figure 4B).

Figure 4.

ttna is required for the establishment of cardiac contractility. A, Shortening fraction of ventricular cardiomyocytes in titin morphants and mutant. B, Heart rates in titin morphants and mutant. Shown are the mean±SD. *P<0.01 when compared with wild type (WT).

ttna Is Required for the Maturation of Z-Discs and A-Bands but Not Early Stages of Sarcomere Assembly

As titin has been proposed to function as a template for sarcomere assembly, we next performed antibody staining to characterize the sarcomere assembly process in wild-type and titin morpholino-injected cardiomyocytes. Because the heart is located within a protective chamber whose surface is often obscured by the hatching glands, we dissected out the heart after antibody staining for imaging purposes. As shown in Figure 5A through 5C, a network of assembled sarcomeres can be detected in a wild-type heart at 48 hpf. At this stage, Z-bodies have assembled to form Z-discs with a width of approximately 3 nm, whereas thick filaments have incorporated to form A-bands (for electron micrographs, see Xu et al²⁸ and Wanga et al³³). The formation of Z-disc and A-band structures was disrupted in both ttna_N2A and ttna_N2B morphants, but remained normal in both ttnb_N2A and ttnb_N2B morphants (Figure 5D through 5O). However, the early steps of sarcomere assembly appeared to be undisturbed in both ttna_N2A and ttna_N2B morphants, as a nonstriated thick filament network was established, and a thin filament network assembled together with α-actinin, which appeared as a dotted pattern (Figure 5D through 5I).

Figure 5.

ttna is required for sarcomere assembly in the heart. Shown in A through O are images of live fish and fixed samples following antibody staining of morphants in which different titin isoforms were targeted by morpholinos. Shown in P through R are images of a live fish and fixed samples following antibody staining of pik^m171 homozygous mutant embryos. Z-discs were revealed by staining for α-actinin (B, E, H, K, N, and Q), whereas A-bands were revealed by staining using the F59 antibody, which recognizes myosin (C, F, I, L, O, and R). Insets are high-magnification images. Scale bar=10 μm. WT indicates wild type.

It was surprising to find that ttna is not required for early steps of sarcomere assembly, because we observed that ttna is among the earliest sarcomere genes to be expressed in both heart and somites. Possibly, ttna levels were not sufficiently reduced using morpholino technology; therefore, we examined the sarcomere structures in the heart of homozygous pik^m171 embryos. Analysis of the complete genomic sequence for both ttna and ttnb indicated that a nonsense mutation is present in pik^m171 in the N2B exon of ttna, and this should lead to a truncated ttna_N2B protein in homozygous pik^m171 mutant embryos. As shown in supplemental Movie S2 and quantified in Figure 4A, heart contractility in homozygous pik^m171 mutant embryos is reduced to a degree less severe than that of ttna morphants. Further, antibody staining revealed an actin/myosin network similar to those in ttna morphants in the pik^m171 homozygous mutant hearts, despite the disrupted Z-discs and A-bands (Figure 5P through 5R). These data confirm that ttna is not required for the early assembly of Z-bodies and thick filaments in the heart or the establishment of cardiac contractility, but rather for later stages of sarcomere formation.

N2A-Based Isoforms of Both ttna and ttnb Are Required for Sarcomere Assembly in Skeletal Muscle

In zebrafish, the process of sarcomere assembly in the skeletal muscle starts first in the anterior somites and progresses to the posterior somites, with well-defined Z-discs and A-bands detectable by antibody staining at 48 hpf (Figure 6A through 6C). In a 30-hpf embryo, Z-bodies start to assemble with thick filaments at the posterior somites (Figure 6P), whereas Z-discs with discernable A-bands have already assembled in the anterior somites of the same embryo (Figure 6Q). The sarcomere is fully assembled in a 72-hpf embryo with perfectly aligned Z-disc, I-band, A-band, and M-band structures (Figure 6R). Additionally, the Z-discs are closely associated with the T-tubule, an invaginated plasma membrane system that is important for the regulation of contraction (Figure 6R and 6S and reported previously³⁴). Reduction of ttna_N2A isoforms to 5% resulted in paralysis phenotypes, as indicated by 90% (295/330) of the injected embryos lacking the ability to twitch at day 1 and having reduced swimming capacity at day 2. Antibody staining revealed disrupted sarcomere structures in the skeletal muscles of these morphants, and the shape of the muscle fiber was severely distorted (Figure 6G through 6I and 6M through 6O). Consistent with the antibody staining, electron micrographs revealed that free Z-bodies consisting of actin filaments and α-actinin, also named I-Z-I brushes, as well as large amounts of unincorporated thick filaments were present in the somites of these morphants (Figure 6T through 6W). Reduction of ttnb_N2A to 2% resulted in similar somite defects in 92% (254/276) of embryos. The sarcomere structure was not completely disrupted, as residual sarcomeres could still be detected by staining for α-actinin and myosin. Reduction of the N2B-based isoforms of ttna and ttnb did not result in paralysis phenotypes or detectable defects in the somites, even when the targeted splicing events were reduced to 9% and 3%, respectively (Figure 6D through 6F and 6J through 6L). In summary, the above data suggest that N2A- but not N2B-based isoforms are required for sarcomere assembly in the somites.

Figure 6.

N2A-based titin isoforms are required for sarcomere assembly in the somites. A through O, Histological and antibody staining of morphants in which different titin isoforms were targeted by morpholinos. Z-discs were revealed by staining for α-actinin (B, E, H, K, and N), whereas the A-bands were revealed by staining using the F59 antibody, which recognizes myosin in slow muscles (C, F, I, L, and O). Scale bar=10 μm. P through S, Process of sarcomere assembly in skeletal muscles of a wild-type zebrafish embryo (WT). Shown are electron micrographs of posterior somites of a 30-hpf embryo (P), anterior somites of the same embryo (Q), and somites in a 72-hpf embryo (R). S, Magnified view of a T-tubule. T through W, Sarcomere assembly is disrupted in titin mutants. Shown are electron micrographs of 30-hpf embryos of morphants with reduced ttna-N2A (T and U) or ttnb-N2A (V and W) isoforms. Arrows indicate Z-bodies (T and V) and Z-discs (P and Q). Brackets indicate representative thick filaments (U and W) and A-bands (P and Q). Arrowheads indicate the T-tubule structure (R and S).

Discussion

Zebrafish Is a Useful Model Organism to Study Functions of Titin

In this article, we report the identification and full-length genomic sequences of 2 titin orthologs in zebrafish, ttna and ttnb, their expression patterns, and their roles in sarcomere assembly in heart and skeletal muscle. Both ttna and ttnb are highly orthologous to human titin and encode peptide sequences corresponding to Z-disc, I-band, A-band, and M-band regions of the sarcomere.²² This is in contrast to titin orthologs in Drosophila and Caenorhabditis elegans, where the Z-disc and I-band regions are encoded by a titin-like gene and the A-band and M-band regions are encoded by different genes, such as strechin, twitchin, and projectin.^1,35 Thus, zebrafish may represent the simplest model organism that contains titin molecules with intact features of human titin.

Functions of Zebrafish Titin in Sarcomere Assembly

Although there are 2 zebrafish titin homologs, only ttna is required for the establishment of cardiac contractility during zebrafish embryogenesis. By investigating the process of de novo sarcomere assembly, we demonstrated that zebrafish titin is required for the lateral growth of Z-bodies to form the Z-discs and the registration of thick filaments to form the A-bands, but is not required for the assembly of Z-bodies or the myosin filaments. These nonstriated thick filaments actually formed a network that resembled the stress fiber—like structure (SFLS) that forms in the early stages of sarcomere assembly.³⁶ A similar conclusion was recently reached on the basis of studies of titin mutants lacking M-line domains using cultured mouse ES cells.³⁷ Because our ttna morphants/mutants lack both A-band and M-line regions of titin, our data further suggest that the A-band region of titin is not required for the assembly of nonstriated myosin filaments, despite the fact that this region contains 170 domains that can bind to myosin.² In addition, the rudimentary actin/myosin network was functional, as indicated by the residual cardiac contraction present in the morphants/mutants.

Our conclusion regarding the role of titin in myosin assembly differs, however, with that of 2 other studies in which cultured cells were used.^23,38 Lack of myosin filament assembly was observed in a myofibroblast cell line (BHK-21_Bi) expressing truncated titin proteins and in adult rat cardiomyocytes (ARC) where titin expression was reduced by injection of antisense oligonucleotides. The difference should not result from inefficiency of morpholino technology, because similar sarcomere defects have been observed in pik^m171 mutant embryos. It is also less likely that ttnb compensates for the functions of ttna, because ttnb is not required for sarcomere assembly in the heart during embryogenesis. Furthermore, the expression of ttnb did not change in ttna morphants, and double injection of morpholinos against both ttna_N2B and ttnb_N2B did not result in more severe cardiac defects (data not shown). We suggest that the observed differences are a result of the use of different model organisms or the use of an in vivo model organism versus cultured cells. Confirmation of this reasoning, however, awaits the generation of null titin mutants in another whole animal model organism.

Different Titin Isoforms Have Different Expression Patterns and Functions

Despite its limitations, morpholino technology provides a novel opportunity to reveal specific functions of individual titin isoforms in an in vivo model organism. Disruption of ttna_N2B isoforms, through the use of either morpholinos or pik mutants, affects only the heart, and not somites. Conversely, reduction of ttnb_N2A isoforms specifically affects somites, but not the heart, despite the fact that sarcomere assembly in somites starts around 16 S, which is earlier than sarcomere assembly in the heart, which starts at approximately 26 S. These data suggest that sarcomere assembly in the heart and somites are 2 independent developmental events.

In heart, both N2A- and N2B-based isoforms of ttna are required for sarcomere assembly. This is consistent with previous reports that N2BA-based isoforms, which contain both N2B- and N2A-based exons, are the predominant isoforms expressed during cardiogenesis in mammals.^39-41 In contrast, ttnb_N2B is not required for either cardiac contractility or sarcomere assembly in the heart, despite mild cardiac edema in the morphants. Furthermore, the ttnb_N2A isoforms were never detected in the embryonic heart. In skeletal muscle, reduction of N2A-based isoforms of either ttna or ttnb is sufficient for sarcomere disarray and functional paralysis, suggesting that their functions are not redundant. In contrast, the N2B-based isoforms do not seem to play an essential role in sarcomere assembly in the somites.

In summary, our data demonstrate that different isoforms of titin have distinct expression patterns and functions, consistent with findings in human patients indicating that mutations in different titin regions lead to different diseases. It is thus suggested that disruption of different titin isoforms could be among the molecular mechanisms that contribute to various phenotypes of titinopathies. With its demonstrated capability to directly link mutation of a particular titin isoform to both cardiac functional changes and structural changes in sarcomere assembly, we anticipate that further studies of titin in zebrafish will provide more insights into pathological pathways in human titinopathies, including hypertrophic cardiomyopathy, dilated cardiomyopathy, and muscular dystrophy.