Drs. Ju and Chong contributed equally to this work.
Recapitulation of fast skeletal muscle development in zebrafish by transgenic expression of GFP under the mylz2 promoter
Version of Record online: 12 MAR 2003
Copyright © 2003 Wiley-Liss, Inc.
Volume 227, Issue 1, pages 14–26, May 2003
How to Cite
Ju, B., Chong, S. W., He, J., Wang, X., Xu, Y., Wan, H., Tong, Y., Yan, T., Korzh, V. and Gong, Z. (2003), Recapitulation of fast skeletal muscle development in zebrafish by transgenic expression of GFP under the mylz2 promoter. Dev. Dyn., 227: 14–26. doi: 10.1002/dvdy.10273
- Issue online: 8 APR 2003
- Version of Record online: 12 MAR 2003
- Manuscript Accepted: 20 DEC 2002
- Manuscript Received: 25 APR 2002
- National University of Singapore
- Agency for Science, Technology, and Research (A*STAR) of Singapore
- myosin light chain;
- enhanced green fluorescent protein (EGFP);
- spadetail (spt);
- chordino (din);
- cAMP-dependent protein kinase A (PKA)
A 1,934-bp muscle-specific promoter from the zebrafish mylz2 gene was isolated and characterized by transgenic analysis. By using a series of 5′ promoter deletions linked to the green fluorescent protein (gfp) reporter gene, transient transgenic analysis indicated that the strength of promoter activity appeared to correlate to the number of muscle cis-elements in the promoter and that a minimal −77-bp region was sufficient for a relatively strong promoter activity in muscle cells. Stable transgenic lines were obtained from several mylz2-gfp constructs. GFP expression in the 1,934-bp promoter transgenic lines mimicked well the expression pattern of endogenous mylz2 mRNA in both somitic muscle and nonsomitic muscles, including fin, eye, jaw, and gill muscles. An identical pattern of GFP expression, although at a much lower level, was observed from a transgenic line with a shorter 871-bp promoter. Our observation indicates that there is no distinct cis-element for activation of mylz2 in different skeletal muscles. Furthermore, RNA encoding a dominant negative form of cAMP-dependent protein kinase A was injected into mylz2-gfp transgenic embryos and GFP expression was significantly reduced due to an expanded slow muscle development at the expense of GFP-expressing fast muscle. The mylz2-gfp transgene was also transferred into two zebrafish mutants, spadetail and chordino, and several novel phenotypes in muscle development in these mutants were discovered. Developmental Dynamics 227:14–26, 2003. © 2003 Wiley-Liss, Inc.
The generation of transgenic organisms has been a powerful approach in studying the genetic basis of embryonic development in Caenorhabditis elegans, Drosophila, mice, and other model organisms. In particular, by using a living color reporter gene, gfp (green fluorescent protein), under a tissue-specific promoter, development of a particular tissue and cell lineage easily can be traced in living animals (Chalfie et al., 1994). This approach is particularly useful in zebrafish (Danio rerio), as it has several combined advantages such as short generation time, transparent embryos, and external development (for review, see Gong et al., 2001).
Although the value of transgenic technology has been recognized increasingly in fish developmental biology, the application of this technology in earlier years was not very successful, to some extent due to the use of heterologous promoters. In recent years, several gfp transgenic zebrafish lines under tissue-specific, zebrafish-origin promoters have been generated and GFP expression has been observed in living embryos following the expression patterns of the endogenous genes from which the promoters derive. These transgenic zebrafish lines include gata1-gfp for erythroid-specific expression (Long et al., 1997); actin-gfp for muscle and ubiquitous expression (Higashijima et al., 1997); islet1-gfp, HuC-gfp, and α1-tubulin-gfp for neuron-specific expression (Higashijima et al., 2000; Park et al., 2000; Goldman et al., 2001); insulin-gfp and pdx1-gfp for endocrine pancreatic expression (Huang et al., 2001); robopsin-gfp for retinal expression (Kennedy et al., 2001); twhh-gfp for GFP expression in the notochord, floor plate, and several other tissues (Du and Dienhart, 2001); and krt8-gfp for epithelial expression in skin and digestive tract (Gong et al., 2002). In general, GFP expression in all of these transgenic lines faithfully mimics the endogenous gene expression; thus, GFP expression in live embryos and fish provides a powerful means to recapitulate the gene expression program.
Previously, we have isolated and characterized a zebrafish MLC2f cDNA clone encoding a myosin light chain 2 polypeptide (Xu et al., 1999). Now the cDNA clone was renamed mylz2 for myosin light polypeptide 2, by the Zebrafish Nomenclature Committee (Xu et al., 2000). The expression pattern of mylz2 has been characterized in detail in embryos as well as in the adult. Its expression is initiated in developing somites around 16 hours postfertilization (hpf), and the expression is limited to the fast skeletal muscles. At later stages, mylz2 mRNA is also expressed in fin buds and head muscles, including eye, jaw, and gill muscles. No mylz2 expression is detected in the slow muscle that is located in the superficial layer of somites. Therefore, mylz2 mRNA is a specific molecular marker to monitor fast skeletal muscle development (Xu et al., 1999, 2000).
In the present study, a 1,934-bp mylz2 promoter was isolated and used to create several gfp reporter gene constructs. The promoter activity was characterized by both transient and stable transgenic assays. In stable transgenic lines, GFP expression faithfully reflected the endogenous expression of mylz2. In addition, we also demonstrated the feasibility to use the mylz2-gfp transgenic line to characterize molecular events in muscle development by disruption of a signaling pathway and by transferring the mylz2-gfp transgene into zebrafish mutants with defects in muscle development.
Isolation and Sequence Characterization of a 1,934-bp Promoter Region of mylz2
To investigate the gene expression program of mylz2 in muscle cells, a 1,934-bp mylz2 promoter region was isolated and characterized in the present study. The complete sequence of the promoter region is shown in Figure 1. Sequence analysis revealed six putative E-boxes (CANNTG) for binding of bHLH transcription factors (MyoD and Myogenin). There are also five potential MEF2 binding sites that match at least 9 of the 10-nucleotide consensus, (C/T)TA(T/A)4TA(A/G) (Schwarz et al., 1993; Olson et al., 1995). Both bHLH transcription factors and MEF2 factors are important in regulation of muscle-specific gene expression (Olson et al., 1995).
Transient Transgenic Analysis of the mylz2 Promoter
To investigate the tissue specificity of the isolated mylz2 promoter, it was cloned into a gfp reporter gene vector, pEGFP-1. The resulting construct, pMYLZ2-1934, was injected into zebrafish embryos at one- to two-cell stage, and muscle-specific expression was observed (Fig. 2C). The 1,934-bp sequence appeared to be a very strong promoter. When the DNA construct was injected at our regular concentration, 500 μg/ml, 100% GFP-expressing embryos displayed strong GFP fluorescence in essentially all muscle fibers in the trunk and it seemed that there was no mosaic distribution of the injected DNA. The earliest expression of transgenic GFP under the 1,934-bp promoter was observed around 19 hpf in a small percentage of injected embryos. GFP appeared first in a few bundles of muscle fibers in the first few anterior somites (data not shown), just as the first appearance of the endogenous mylz2 mRNA as detected by whole-mount in situ hybridization (Xu et al., 2000). By 28 hpf, the majority of the injected embryos showed obvious GFP expression and the expression appeared to increase in the first few days after injection. By 48 hpf (hatching), GFP expression was spread to the entire trunk region. By 72 hpf and onward, expression of GFP was also detected in the head muscles, including eye, jaw, and gill muscles (data not shown). Ectopic expression of GFP was observed in approximately 10% of injected embryos, mainly in the heart, lens, and skin cells (Table 1).
|Constructs||Injected||Survivalb||Expressionc||Tissue distribution of GFPd||Profiles of GFP expression in skeletal musclee|
|MYLZ2-1934 (500 μg/ml)||36||21 (58%)||16 (76%)||16 (100)||2 (13)||3 (19)||4 (25)||0 (0)||0 (0)||16 (100)||0 (0)||0 (0)||0 (0)|
|MYLZ2-1934 (100 μg/ml)||402||229 (57%)||141 (62%)||141 (100)||13 (9)||6 (4)||21 (15)||0 (0)||0 (0)||32 (23)||53 (37)||56 (40)||0 (0)|
|MYLZ2-1338||294||142 (48%)||78 (55%)||78 (100)||4 (5)||2 (3)||7 (9)||0 (0)||0 (0)||11 (14)||18 (23)||49 (63)||0 (0)|
|MYLZ2-871||222||92 (41%)||52 (57%)||52 (100)||2 (4)||0 (0)||3 (6)||3 (6)||0 (0)||1 (2)||16 (30)||35 (68)||0 (0)|
|MYLZ2-482||229||165 (72%)||104 (63%)||104 (100)||5 (5)||0 (0)||5 (5)||3 (3)||0 (0)||0 (0)||14 (13)||90 (87)||0 (0)|
|MYLZ2-283||134||80 (60%)||55 (69%)||55 (100)||9 (16)||0 (0)||2 (4)||2 (4)||2 (4)||0 (0)||8 (15)||47 (85)||0 (0)|
|MYLZ2-77||236||108 (46%)||47 (44%)||47 (100)||15 (32)||0 (0)||1 (2)||3 (6)||5 (11)||0 (0)||5 (11)||42 (89)||0 (0)|
|MYLZ2-3||246||176 (72%)||30 (17%)||14 (47)||6 (20)||0 (0)||0 (0)||0 (0)||17 (57)||0 (0)||0 (0)||14 (47)||17 (53)|
To further analyze the 1,934-bp promoter region, a series of 5′ deletion constructs was made and tested by transient transgenic assay. To get a quantitative view of the relative strength of the promoter constructs, we classified GFP-expressing embryos into four categories based on the expression pattern in the trunk muscle: high, medium, weak, and nil. The high category included embryos displaying GFP expression in nearly all muscle fibers throughout the trunk; medium, embryos showing GFP expression in patch usually in >10 muscle fibers in the mid-trunk region; low, embryos having fewer than 10 dispersed GFP-positive fibers; and nil, embryos without GFP expression in the skeletal muscle but displaying ectopic GFP expression. The non–GFP-expressing embryos were not scored, because the lack of GFP expression in some of the embryos may be due to unsuccessful microinjection. The examples of high, medium, and weak expression in embryos of 2–3 days old are shown in Figure 2C. In general, the intensity of GFP expression also correlated to the number of expressing muscle fibers or the high, medium, and weak categories.
For analysis of the deletion constructs, we injected less DNA (100 μg/ml) to lower the GFP expression level to reduce the expression saturation effect from the high strength of the mylz2 promoter. Due to the decrease of the amount of injected DNA, the percentage of high expressing embryos injected with pMYLZ2-1934 was reduced from 100% to 23% (Table 1). As summarized in Table 1 and Figure 2, it seems that there is a general correlation between the decrease of GFP expression and the 5′ deletions, indicating that the level of GFP expression may depend on the number of muscle cis-elements, such as the E-box and MEF2 binding sites shown in Figure 1, although the possibility of presence of other important cis-elements cannot be ruled out from this study. The deletions up to −283 bp and −77 bp still produced GFP in trunk muscles in 100% of expressing embryos. This correlates well with the presence of a proximal MEF2 element, which is the only perfect MEF2 element in the 1,934-bp region (Fig. 1). With further deletion to −3 bp, where the TATA box has been removed, only 17% of injected embryos displayed weak GFP signal compared with at least 44% of embryos expressing GFP injected with other constructs. Moreover, most of the GFP signal in pMYLZ2-3 injected embryos was ectopic, such as in the notochord and skin. Thus, the expression under the −3 bp promoter may result from a weak cryptic promoter.
Stable mylz2-gfp Transgenic Lines
To generate stable transgenic lines for further analysis of the expression program of mylz2, embryos injected with pMYLZ2-1934 and showing GFP expression were retained and raised to adulthood. GFP expression was maintained in several adult founders, in which a patched pattern of GFP expression was observed in skeletal muscles, as shown in Figure 3A. The GFP expression in some individuals is so strong that it was readily observable even under normal daylight. Among 10 founder fish that were screened by crossing with wild-type fish, two of them produced GFP-expressing offspring. The gfp transmission rate to F1 was approximately 10% for the two transgenic founders. This finding was consistent with the earlier finding that the transgene in founder transgenic fish was mosaic in a germline (reviewed by Gong and Hew, 1995). When GFP-positive F1 was bred with a wild-type fish, we always observed a 50% transmission ratio in F2 as predicted based on Mendelian genetics. Because Mendelian ratio was maintained in all subsequent generations (now up to F6 generation), the gfp transgene likely stably integrated into the host chromosome at a single locus. Hence, the two mylz2-1934 transgenic lines were named Line A and Line B. The genomic Southern blot hybridization indicated that there were multiple copies (>10) of mylz2-gfp transgene in Line B, whereas only a single copy of mylz2-gfp most likely integrated in Line A (data not shown).
In both transgenic lines, GFP expression in F1 offspring is highly specific to skeletal muscles and no ectopic expression was noted (Fig. 3). In Line A, GFP expression was first observed in trunk muscles around 20 hpf. The expression level was very low at the beginning (see a 24 hpf embryo in Fig. 3C) but gradually increased as embryonic development proceeded (see a 48 hpf embryo in Fig. 3D). For comparison, mylz2 mRNA expression, as detected by whole-mount in situ hybridization, in a 20 hpf embryo is also displayed (Fig. 3B). By approximately 58 hpf, GFP expression could be detected in early developing head muscles (Fig. 3E). This pattern of GFP expression is consistent with the endogenous mylz2 mRNA expression as detected by whole-mount in situ hybridization (Fig. 3B; Xu et al., 2000), although there are approximately 4 hr of temporal delay between the expression of mylz2 mRNA and transgenic GFP. This delay is likely due to the translation of mRNA into protein and accumulation of GFP to the detectable level. In Line B, a similar expression pattern of GFP was detected except for a relatively weak expression of GFP in the trunk skeletal muscle in the first few days of development (Fig. 3E). In both lines the level of GFP expression increased during development and, by 1 month of age, GFP expression easily could be observed even under normal daylight (Fig. 3F). The visible green color was maintained throughout the adulthood. Although GFP synthesized in the muscle appears to be doubled in homozygous transgenic zebrafish (data not shown), there is no obvious difference in green fluorescence between homozyotes and heterozygotes.
Efforts were also made to generate stable transgenic lines with the deletion constructs, pMYLZ2-871, pMYLZ2-482, and pMYLZ2-283. A stable line (Line C) with the 871-bp promoter was obtained after screening of 25 founder fish. GFP expression in F1 was very weak, starting at approximately 36 hpf in the trunk muscles (Fig. 3G). Unlike the two transgenic lines with the 1,934-bp promoter, the level of GFP expression in Line C increased little during subsequent development. GFP expression in head muscle was also faint after 72 hpf. In addition, transgenic embryos were also identified by polymerase chain reaction (PCR) from 2 of 22 founders injected with the 482-bp promoter, but no GFP expression was observed from these embryos up to 1 week after fertilization, indicating either that the gfp transgene was not activated or that the GFP expression was too low to detect. No germ-line transmission was obtained from pMYLZ2-283 injected founders (n = 32).
Development of Head Skeletal Muscles as Revealed by GFP Expression
As reported previously (Xu et al., 1999, 2000), mylz2 mRNA is also expressed in various head skeletal muscles starting from ∼53 hpf. The distribution of mylz2 mRNA in a 72 hpf embryo, as revealed by whole-mount in situ hybridization, is shown in Figure 4A,B. The nomenclature of the head muscles is based on Schilling and Kimmel (1997). Similarly, in mylz2-gfp transgenic embryos at the same stage, the same set of head muscles positive for mylz2 mRNA were also positive for GFP (Fig. 4C,D).
Because a stable gfp transgenic line under a tissue-specific promoter is useful to study the program of gene expression and to trace tissue development (for review, see Gong et al., 2001), the present gfp transgenic lines should be useful to study development of skeletal muscle in vivo. By using the two 1,934-bp promoter transgenic lines, we were able to trace the head muscle development in live embryos and to determine the temporal sequence of development of the cranial muscles (Fig. 4). As reported previously by Schilling and Kimmel (1997), by using several molecular and histochemical markers, the first pair of muscle developed in the head region is adductor mandibulae (am; Fig. 4E), followed by medial rectus (mr; Fig. 4F), and others (Fig. 4G–J). The temporal appearance of GFP-positive muscles was generally consistent with the report of Schilling and Kimmel (1997). In terms of the order of appearance of GFP expression, the eye muscles generally develop first, followed by jaw muscles and gill muscles. Like endogenous mylz2 mRNA, transgenic GFP expression was also observed in fin buds (Fig. 4I,J). Thus, these transgenic lines provide an excellent model to observe the dynamic development of skeletal muscles in a live fish.
Inhibition of GFP Expression by Overexpression of Dominant Negative Form of Protein Kinase A
It has been demonstrated that both Sonic hedgehog (Shh) and cAMP-dependent protein kinase A (PKA) affect development of fast and slow muscles (Blagden et al., 1997; Du et al., 1997; Barresi et al., 2000). Shh inhibits the PKA activity and promotes slow muscle differentiation. Overexpression of Shh or inhibition of PKA has the same phenotype, i.e., conversion of the fast muscle precursors to the slow muscle. Therefore, injection of a dominant negative PKA RNA would stimulate development of slow muscle and inhibit development of fast muscle. If our transgenic model faithfully recapitulates fast muscle development, we expect to see the reduction of GFP expression by injection of dominant negative PKA RNA.
Thus, RNA encoding the dominant-negative PKA regulatory subunit PKI (Hammerschmidt et al., 1996a) was injected into heterozygotic mylz2-gfp embryos, obtained from a cross of a homozygous mylz2-gfp parent (Line A) with a wild-type fish. In the uninjected control group, all 41 embryos expressed GFP (100%) in the fast muscle (Fig. 5A,C–E). Among the 54 injected embryos, 13 (24.1%) showed mild reduction of GFP expression and 41 embryos (75.9%) displayed strong reduction of GFP expression (Fig. 5B,F–H). Due to the mosaic distribution of injected DNA (Westerfield et al., 1992), reduction of GFP expression was usually observed only on one side (Fig. 5B). Several embryos that had reduced GFP expression were then sectioned and double-stained with F59 antibody to display slow muscle and anti-GFP antibody to display fast muscles. By this stage (30 hpf), fast muscles started to acquire a mild expression of F59 antigen and thus a weak stain of fast muscle is also observed (Devoto et al., 1996). However, the slow muscle fibers express this marker more intensely (Fig. 5D,G). Consistent with the previous work on wild-type embryos (Devoto et al., 1996), the mylz2-gfp control embryos also show the slow muscle marker in the superficial layer of somitic tissue (Fig. 5D,E). However, in PKI RNA-injected embryos, the slow muscle in the GFP reduced side was extended into the fast muscle area (Fig. 5G,H). GFP expression was found only in a few remaining fast muscle cells that were negative for the F59 marker on that side, indicating a correlation between reduction of the GFP expression and F59 staining of slow muscle. Thus, these experiments indicate that the mylz2-gfp transgenic fish can be used for analysis of muscle development in the zebrafish.
Characterization of mylz2-gfp Transgenic Mutants
An important application of gfp transgenic lines is genetic analysis of development of specific tissues by introduction of the gfp transgene into mutants through conventional breeding. The simplicity of monitoring the GFP fluorescence should facilitate the characterization of mutant phenotypes in vivo. To demonstrate the usefulness of the mylz2-gfp transgenic line in characterization of mutants that affects skeletal muscle development, we transferred the mylz2-gfp transgene by genetic crossing into two mutants, spadetail (spt) and chordino (din).
Spt encodes a tbx6-related gene, and its mutation causes severe defects in development of the anterior skeletal mesoderm, whereas the posterior region is less affected (Kimmel et al., 1989; Hopwood and Gurdon, 1990; Griffin et al., 1998). Our analysis of the mylz2-gfp transgenic spt-/- embryos revealed several defects of skeletal muscles in the anterior trunk (Fig. 6A–C,E,G). First, there are gaps between anterior somites (40 of 44, 90.9%). Frequently, the two anterior-most somites were dissociated from the posterior ones, and this finding was observed from the early stages of somitogenesis before the development of head muscles (Fig. 6A–C). Despite the severe effect on the initial positioning of the myoblasts, it seems that differentiation of individual myofibrils is not affected, as judged by intensity of GFP expression and morphology of myofibrils. This observation suggests that muscle differentiation in these somites does not require the contact with the block of posterior somites and that differentiation of the posterior somite is also independent of the previous one. Second, the GFP-positive fast muscles were distributed asymmetrically in spt-/- embryos (38 of 44, 86.6%) (Fig. 6D,E). An expansion of blocks of GFP-positive cells on one side correlated with a decrease or disappearance of muscle tissue on the other side (Fig. 6E). Third, we also detected abnormal distribution of muscles in the anterior abdominal region (41 of 44, 93.2%; Fig. 6G), compared with that in the wild-type siblings (Fig. 6F).
Chordino is the zebrafish version of Chordin, and its mutation (din) affects the bone morphogenic protein signaling pathway, leading to defects of the head and trunk linked to deficiency of the notochord (Hammerschmidt et al., 1996b; Gonzalez et al., 2000). Analysis of mylz2-gfp progeny on din-/- background revealed disorganization of several cranial muscles in approximately two-thirds of analyzed GFP expressing din-/- embryos (16 of 25, 64%). For example, the angle between the two intermandibularis posterior (imp) muscles is much reduced compared with that in wild-type embryos (Fig. 7A,B). In more severe cases, the gill muscles, particularly the dorsal pharyngeal wall (dpw) and sternohyoideus (sh), are also affected (Fig. 7C). In the tail, formation of the horizontal myoseptum is affected in essentially all examined din-/- embryos. In control embryos, GFP expression was absent from the horizontal myoseptum, where muscle pioneer cells (slow muscle) were located (Fig. 7D), whereas in the din-/- embryos the myoseptum is disorganized, the shape of somites was also abnormal, and GFP expression is more heterogeneous. These findings are probably due to deficiency of the notochord, from which Shh secretion is required for development of early slow muscles (Barresi et al., 2001). Consistent with this deficiency, we also found that the din-/- embryos have a much reduced or completely diminished notochord in the tail region, compared with the wild-type embryos (Fig. 7F,G), as reported previously by Hammerschmidt et al., (1996b).
mylz2-gfp Transgenic Lines as a Model for Development of Fast Skeletal Muscle
Previously, we have demonstrated that mylz2 mRNA is specifically expressed in fast skeletal muscle and, thus, is an excellent marker for development of fast skeletal muscle (Xu et al., 1999, 2000). During embryogenesis, mylz2 mRNA was first detected in the first few anterior somites around 16 hpf or at the 14-somite stage. Among 10 muscle-specific genes that we previously compared, mylz2 belongs to the intermediate class in terms of the timing of ontogenic activation during somitogenesis (Xu et al., 2000). The expression is limited to the fast muscle precursors and is detected only in the deep muscle cells in the somite. The slow muscle in zebrafish, derived from adaxial cells that migrate to the surface of somites, can be defined by an anti-slow myosin antibody F59 (Devoto et al., 1996) or by an expressed sequence tag (EST) clone encoding slow myosin binding protein C (Xu et al., 2000). It has been confirmed that mylz2 mRNA is not expressed in these cells (Xu et al., 2000). At later stages, mylz2 mRNA is also expressed in fin buds and head muscles including eye, jaw, and gill muscles.
In the present study using the 1,934-bp mylz2 promoter, we have generated stable gfp transgenic lines. GFP expression in these transgenic lines mimics very closely the endogenous mylz2 mRNA expression. Transgenic GFP expression was first detected in somites around 20 hpf, approximately 4 hr later than the timing of detection of endogenous mylz2 mRNA. This finding is likely due to the duration of translation and accumulation of GFP to the level sufficient for visualization. At later stages, GFP was also detected in fin buds and head muscles in a pattern identical to that of endogenous mylz2 mRNA expression. Thus, our mylz2-gfp transgenic lines faithfully recapitulate the development of fast skeletal muscle.
To validate the transgenic model, we have attempted to manipulate the Shh signaling pathway, which has been demonstrated to be involved in skeletal muscle development in zebrafish (Blagden et al., 1997; Du et al., 1997; Barresi et al., 2000). Overexpression of Shh can transform the entire myotome into the slow muscle at the expense of the fast muscle. The signaling downstream of Shh is regulated by PKA, which is repressed by Shh. Manipulation of PKA activity also affects development of the slow and fast muscles. For example, an inhibition of PKA activity by overexpression of RNA for a dominant negative form of PKA, PKI, results in activation of Hh signaling leading to an increase of slow muscle, whereas elevated expression of PKA blocks the development of slow muscle (Hammerschmidt et al., 1996a; Du et al., 1997; Barresi et al., 2000). Consistent with these reports, GFP expression was significantly reduced when PKI RNA was injected into transgenic embryos. The reduction of GFP expression paralleled an increase of the slow muscle. These observations provide one more illustration on an embryonic plasticity of the myocytes that, during early stages of development, can be reprogrammed to different types of differentiated cells, depending on modulation of developmental pathways. In addition, our studies demonstrate the usefulness of the mylz2-gfp transgenic line for future studies of factors affecting fast muscle development. By using the same approach, other unknown factors can be rapidly screened for their role in formation of the musculature.
Analysis of GFP Expression in Mutants
A large number of zebrafish mutants that affect various developmental pathways have been generated successfully (Driever et al., 1996; Haffter et al., 1996). The characterization of various aspects of mutant phenotypes and molecular mechanisms involved represents a formidable challenge. Especially as the number of useful antibodies is restricted, most analyses are performed at the transcription level and within the first few days of development. Current methods of analysis of fixed material also suffer several major setbacks. These include scarcity of information about the molecular events at the protein level, incompatibility of these methods with continuous monitoring of developmental events in vivo, and inaccessibility of certain areas such as the trunk of late embryos/larva to the molecular probes, including antisense RNA or antibody. This is mainly due to a rapid accumulation of extracellular matrix that prevents the penetration of macromolecular probes. To demonstrate how to overcome these limitations, we transferred the mylz2-gfp reporter gene into two mutants and demonstrated that this strategy yielded a new and valuable research tool for analysis of mutant phenotype. In particular, we were able to reveal a role of spadetail in organization of somites into a continuous row as mutation of this gene leads to an asymmetric distribution of somatic muscles along the midline. We also detected disorganization of cranial muscles and some changes in the tail region in din-/- embryos. Thus, our observations in GFP-expressing embryos under a specific gene promoter improve the level of understanding of mutant phenotypes and opens a new avenue for further analysis of these relatively well-characterized mutants. This approach may be even more efficient when applied to analysis of mutants with subtle changes in organization of tissue or to other mutants that were not characterized previously. In addition, mylz2-gfp fish may be used as a starting material for the mutant screen for genes involved in formation and organization of the skeletal musculature.
Promoter Analysis of mylz2
It is well known that both bHLH and MEF2 families of transcriptions factors are important in transcriptional activation of muscle-specific genes such as mylz2 (Crow and Stockdale, 1986; Schwarz et al., 1993) As shown in Figures 1 and 2, several putative cis-elements for binding of the two families of transcription factors are noted in the 1,934-bp mylz2 promoter region. Transient expression assay using promoter deletion constructs indicated that the level of transgene expression correlated well with the length of the promoter. Thus, it is likely that the strength of mylz2 expression is determined by the number of available cis-elements in the promoter region. The same trend was also observed from our stable transgenic lines with different lengths of mylz2 promoter: the longer the promoter, the higher the level of GFP expression.
In our transient transgenic assay, the −77-bp promoter region seemed to maintain a relatively strong muscle-specific activity. In the −77-bp region, there is a perfect MEF-2 binding site and this site probably critical to produce a relatively strong muscle expression. Consistent with the observation, this region also has a strong promoter activity in adult skeletal muscle as assayed by direct injection of naked DNA (Xu et al., 1999). Mutation of the MEF2 site resulted in a marked reduction of promoter activity (Xu et al., 1999). Our observation is also in agreement with that of Arnold et al. (1988), who previously reported that 64 bp of 5′ flanking region from the chicken cardiac MLC2A gene is sufficient to allow muscle-specific transcription.
As mylz2 mRNA was detected in many regions in developing embryos, including somites, fin buds, and head region, it is interesting to understand whether there are different cis-elements for its expression in different regions. To explore the question, we generated many promoter deletions and carried out both transient and stable transgenic expression. By transient expression, deletion of the promoter region up to −482 bp showed a clear GFP expression in head muscles (data not shown). We did not observe GFP expression in head muscle after injection of the −283-bp and −77-bp constructs. However, this finding may be due to a general low level of GFP expression with the two constructs rather than to the lack of cis-element for head muscle expression. Consistent with this, we noted that head muscle expression appeared to be observed only in embryos showing strong GFP expression. It is likely that many injected embryos failed to express GFP in the head muscles because of a decrease of foreign DNA by late stages of embryogenesis when head muscles started to develop. Therefore, despite different origins of somitic and head muscles (Schilling and Kimmel, 1997), it is likely that the same set of cis-elements is used in regulation mylz2 transcription in both somitic and head muscles. Consistent with this, the transgenic line with the short promoter (871-bp) showed an identical GFP expression pattern as the two transgenic lines with the 1,934-bp promoter, although the level of GFP expression is very low. Thus, it seems that there are no distinct cis-elements for activation of mylz2 gene in different skeletal muscles. The simplicity of the mylz2 promoter is in sharp contrast to several other characterized zebrafish promoters from the genes showing more complex expression patterns. For example, the promoter of GATA-2 is expressed in at least three distinct types of cells and at least three distinct sets of regulatory elements are present in a 7.5-kb promoter region for hematopoietic, enveloping layer, and neuronal expression (Meng et al., 1997). In the promoter of islet1, some cis-elements for certain neurons are located over 15 kb from the transcription start site (Higashijima et al., 2000).
It is interesting to note that GFP expression in the heart was observed in 15–25% of the GFP-expressing embryos injected with the 1,934-bp promoter construct (Table 1). This rate is significantly higher than the number of ectopic expression in other tissues particularly by consideration of the tiny size of the heart tissue. This finding may be because cardiac and skeletal muscles share many common regulatory elements and transcription factors; thus a cross-activation of the mylz2 promoter in the heart would be much easier than in other tissues when a high copy number of a foreign gene is available by transient transgenic expression. However, in all three stable transgenic lines, there is no GFP expression detected in the heart (data not shown), suggesting that the expression can be strictly controlled when only one or a few copies of transgene are present.
For transient transgenic analysis, it is frequently observed that the transgene is expressed in a mosaic pattern. This phenomenon has been attributed to the mosaic segregation of exogenously introduced DNA, as observed by injection of fluorescein-labeled DNA (Westerfield et al., 1992). In our study, when the 1,934-bp promoter construct was injected into the embryos at our regular, high DNA concentration, 500 μg/ml, GFP expression was observed in essentially 100% of muscle fibers in all injected embryos as if there was no mosaic DNA segregation. However, for the deletion constructs, we always observed the mosaic expression pattern of muscle expression. This may be explained by the strength of the promoter. When the promoter is strong, few copies will be sufficient for a visible GFP expression; when the promoter is weak, more copies will be required for a detectable expression. Therefore, the mosaic segregation of injected DNA may not be in an “all or none” manner and probably will be just a quantitatively uneven segregation.
Zebrafish and Embryo Production
Zebrafish were purchased from a local ornamental fish farm and maintained according to the Zebrafish Book (Westerfield, 1995). Embryos were produced from several pairs of male and female fish reared with the photoperiod of 10-hr dark and 14-hr light.
Promoter Isolation and Deletion
mylz2 promoter was isolated by an improved linker-mediated PCR method as described previously (Liao et al., 1997). The 1,934-bp promoter region used in the present study was amplified from EcoRI-digested zebrafish genomic DNA. The gene-specific primers and linker primers have been described previously by Xu et al. (1999). The isolated 1,934-bp promoter region was sequenced completely and ligated into the gfp reporter gene vector, pEGFP-1 (Clonetech), at the EcoRI and BamHI site, and the resulting construct was named pMYLZ2-1934. To analyze the promoter region, 5′ unidirectional deletion was carried out from pMYLZ2-1934 to delete the 5′ region progressively by using the Double-stranded Nested Deletion Kit according to the manufacturer's instruction manual (Pharmacia). Six deletion constructs were selected for microinjection and the start nucleotides of the deletion constructs were determined by DNA sequencing. The deletion constructs were named pMYLZ2-1338, pMYLZ2-871, pMYLZ2-482, pMYLZ2-283, pMYLZ2-77, and pMYLZ2-3, where the numbers indicate the position of the start nucleotide in each construct by referring the transcription start site as +1 (Fig. 1).
Microinjection and Observation of GFP Expression
The plasmid constructs were linearized with the restriction enzyme BglII and injected into embryos of the one- or two-cell stage at a concentration of 500 μg/ml, as previously described by Ju et al. (1999). For functional analysis of deletion constructs, the DNA concentration is 100 μg/ml. Expression of GFP was observed and photographed under a Zeiss Axiovert 25 fluorescence microscope or a confocal microscope Zeiss LSM 510 (Germany).
Screening of Germline Transgenic Zebrafish
To screen for germline transgenic zebrafish, usually a founder fish, developed from the GFP-expressing fry after microinjection, was crossed with a wild-type fish and embryos/larva up to 96 hpf were examined for GFP expression under a fluorescent microscope. For each cross, at least 100 embryos were examined. For screening of germline transmission of some deletion constructs, the PCR method was used. Genomic DNA was extracted from pooled 24 hpf embryos following a protocol described in the Zebrafish Book (Westerfield, 1995). Briefly, embryos were digested in the DNA extraction buffer (10 mM Tris, pH 8.2; 10 mM ethylenediaminetetraacetic acid; 200 mM NaCl; 0.5% sodium dodecyl sulfate; and 200 mg/ml proteinase K) for 3 hr at 50°C. The solution was extracted once with phenol/chloroform and precipitated with 100% ethanol. PCR was carried out by using a forward primer located in the mylz2 promoter region (5′-CTCACTTAAGGTTGGGTC) and a reverse primer located in the pEGFP-1 vector (5′-GCTGCCTCGTCTTGCAGTTC). Founder fish that produced embryos positive by PCR assay were then bred again to obtain embryos for GFP fluorescence observation.
Whole-Mount In Situ Hybridization
Whole-mount in situ hybridization using digoxigenin (DIG) -labeled riboprobe was carried out as previously described (Korzh et al., 1998). The mylz2 cDNA clone, E72, used in the present study, is one of our EST clones in pBluescript vector (Gong et al., 1997). The plasmid DNA was linearized with BamHI, followed by in vitro transcription reaction with T7 RNA polymerase to generate the antisense RNA probe. The embryos were fixed in 4% paraformaldehyde, hybridized with the DIG-labeled riboprobe in a hybridization buffer (50% formamide, 5× SSC [1× = 15 mM NaCl, 15 mM sodium citrate, pH 7.6], 50 mg/ml tRNA and 0.1% Tween 20) at 70°C, followed by incubation with anti-DIG antibody conjugated with alkaline phosphatase and by staining with the substrates nitroblue phosphate (NBT) and 5-bromo, 4-chloro, 3-indolil phosphate (BCIP) to produce purple and insoluable precipitates.
Overexpression of PKI and Immunostaining
Dominant negative protein kinase A (PKI) RNA was transcribed from plasmid, pPSP64T-PKI (Hammerschmidt et al., 1996a). After linearization with NotI, 5′ capped RNA was synthesized by using Sp6 mESSAGEmACHINE kit following manufacturer's instruction manual (Ambion, TX). The synthetic RNA was suspended in diethylpyrocarbonate-treated water and then mixed with injection buffer (0.15 M Tris-HCl, pH 7.5) containing phenol red to a final concentration of 100 μg/ml. Microinjection was performed in embryos at the one-cell stage, and approximately 4.4 nl of RNA solution was injected into each embryo.
For immunostaining, 30 hpf embryos were embedded in 1.5% agar-sucrose and sectioned at 12 μm by a cryostat. The cryosections were first incubated with blocking solution (5% bovine serum albumin [BSA], 1% dimethylsulfoxide (DMSO) in PBST [0.8% NaCl; 0.02% KCl; 0.0144% Na2HPO4, 0.024% KH2PO4, pH 7.4; and 0.1% Tween 20]) for 2 hr at room temperature. Mouse F59 monoclonal antibody against slow myosin (Crow and Stockdale, 1986) was used as primary antibody at a 1:10 dilution in PBST with 2% BSA and 1% DMSO. After overnight incubation at 4°C, the sections were washed with PBST several times. Secondary anti-mouse antibody labeled with fluorescein was applied at a 1:500 dilution for 1 hr at room temperature. The labeled embryos were observed under a scanning confocal microscope (Leica TCS4D).
Transfer of mylz2-gfp Transgene Into Zebrafish Mutants
To transfer mylz2-gfp transgene into a zebrafish mutant, usually a homozygous mylz2-gfp male (containing the 1,934-bp promoter) was crossed with several heterozygous mutant females. GFP-positive F1 (gfp/+, mutant/+, or gfp/+, +/+), which could be easily selected even under daylight due to the green color of their muscles, were crossed to obtain F2 homozygous mutants with GFP expression (gfp/+, mutant/mutant) that were screened for mutant phenotype and GFP fluorescence under the fluorescent dissecting microscope. Some embryos were further characterized by cryosection and by staining with 3.5 μM DAPI (4′,6-diamidine-2-phenylidole-dihydrochloride).
X.W., Y.X., and H.W. were supported by postgraduate scholarships from NUS. We thank Dr. M. Hammerschmidt for the gift of plasmid pSP64T-PKI and Dr. Frank Stockdale for F59 antibody.
- 1988. The promoter of the chicken cardiac myosin light chain 2 gene shows cell-specific expression in transfected primary cultures of chicken muscle. Nucleic Acids Res 16: 2411–2429. , , .
- 2000. The zebrafish slow-muscle-omitted gene product is required for Hedgehog signal transduction and the development of slow muscle identity. Development 127: 2189–2199. , , .
- 2001. Distinct mechanisms regulate slow-muscle development. Curr Biol 11: 1432–1438. , , , .
- 1997. Notochord induction of zebrafish slow muscle mediated by Sonic hedgehog. Genes Dev 11: 2163–2175. , , , .
- 1994. Green fluorescent protein as a marker for gene expression. Science 263: 802–805. , , , , .
- 1986. Myosin expression and specialization among the earliest muscle fibers of the developing avian limb. Dev Biol 113: 238–254. , .
- 1996. Identification of separate slow and fast muscle precursor cells in vivo, prior to somite formation. Development 122: 3371–3380. , , , .
- 1996. A genetic screen for mutations affecting embryogenesis in zebrafish. Development 123: 37–46. , , , , , , , , , , , .
- 2001. Zebrafish tiggy-winkle hedgehog promoter directs notochord and floor plate green fluorescence protein expression in transgenic zebrafish embryos. Dev Dyn 222: 655–666. , .
- 1997. Positive and negative regulation of muscle cell identity by members of the hedgehog and TGF-beta gene families. J Cell Biol 139: 145–156. , , , .
- 2001. Transgenic zebrafish for studying nervous system development and regeneration. Transgenic Res 10: 21–33. , , , , .
- 1995. Transgenic fish in aquaculture and developmental biology. Curr Topic Dev Biol 30: 177–214. , .
- 1997. Rapid identification and isolation of zebrafish cDNA clones. Gene 201: 87–98. , , , , , .
- 2001. Green fluorescent protein (GFP) transgenic fish and their applications. Genetica 111: 213–225. , , .
- 2002. Green fluorescent protein (GFP) expression in germ-line transmitted transgenic zebrafish under a stratified epithelial promoter from keratin8 (krt8). Dev Dyn 223: 204–215. , , , , , , .
- 2000. Head and trunk in zebrafish arise via coinhibition of BMP signaling by bozozok and chordino. Genes Dev 14: 3087–3092. , , , , , , .
- 1998. Molecular identification of spadetail: regulation of zebrafish trunk and tail mesoderm formation by T-box genes. Development 125: 3379–3388. , , , .
- 1996. The identification of genes with unique and essential functions in the development of the zebrafish, Danio rerio. Development 123: 1–36. , , , , , , , , , , , , , , , , .
- 1996a. Protein kinase A is a common negative regulator of Hedgehog signaling in the vertebrate embryo. Genes Dev 10: 647–658. , , .
- 1996b. dino and Mercedes, two genes regulating dorsal development in the zebrafish embryo. Development 123: 95–102. , , , , , , , , , , , , , , .
- 1997. High-frequency generation of transgenic zebrafish which reliably express GFP in whole muscles or the whole body by using promoters of zebrafish origin. Dev Biol 192: 289–299. , , , , .
- 2000. Visualization of cranial motor neurons in live transgenic zebrafish expressing green fluorescent protein under the control of the islet-1 promoter/enhancer. J Neurosci 20: 206–218. , , .
- 1990. Activation of muscle genes without myogenesis by ectopic expression of MyoD in frog embryo cells. Nature 347: 197–200. , .
- 2001. Analysis of pancreatic development in living transgenic zebrafish embryos. Mol Cell Endocrinol 177: 117–124. , , , , .
- 1999. Faithful expression of green fluorescent protein (GFP) in transgenic zebrafish embryos under control of zebrafish gene promoters. Dev Genet 25: 158–167. , , , , , , , .
- 1989. A mutation that changes cell movement and cell fate in the zebrafish embryo. Nature 337: 358–362. , , , , .
- 2001. Isolation of a zebrafish rod opsin promoter to generate a transgenic zebrafish line expressing enhanced green fluorescent protein in rod photoreceptors. J Biol Chem 276: 14037–14043. , , , , .
- 1998. Expression of zebrafish bHLH genes ngn1 and nrd defines distinct stages of neural differentiation. Dev Dyn 213: 92–104. , , , , .
- 1997. An alternative linker-mediated polymerase chain reaction method using a dideoxynucleotide to reduce amplification background. Anal Biochem 253: 137–139. , , .
- 1997. GATA-1 expression pattern can be recapitulated in living transgenic zebrafish using GFP reporter gene. Development 124: 4105–4111. , , , , , .
- 1997. Promoter analysis in living zebrafish embryos identifies a cis-acting motif required for neuronal expression of GATA-2. Proc Natl Acad Sci U S A 94: 6267–6272. , , , , .
- 1995. Regulation of muscle differentiation by the MEF2 family of MADS box transcription factors. Dev Biol 172: 2–14. , , .
- 2000. Analysis of upstream elements in the HuC promoter leads to the establishment of transgenic zebrafish with fluorescent neurons. Dev Biol 227: 279–293. , , , , , , , , , , , , .
- 1997. Musculoskeletal patterning in the pharyngeal segments of the zebrafish embryo. Development 124: 2945–2960. , .
- 1993. In: KarinM, editor. Gene expression: general and cell-type-specific. Boston, MA: Birkauser. p 93–115. , , .
- 1995. The zebrafish book. Eugene, OR: University of Oregon Press. .
- 1992. Specific activation of mammalian Hox promoters in mosaic transgenic zebrafish. Genes Dev 6: 591–598. , , , , .
- 1999. Fast skeletal muscle-specific expression of a zebrafish myosin light chain 2 gene and characterization of its promoter by direct injection into skeletal muscle. DNA Cell Biol 18: 85–95. , , , , , , , .
- 2000. Asynchronous activation of 10 muscle-specific protein (MSP) genes during zebrafish somitogenesis. Dev Dyn 219: 201–215. , , , , .