DEVELOPMENTAL DYNAMICS 208:244–254 (1997) Expression of the Mouse Fibronectin Gene and Fibronectin-lacZ Transgenes During Somitogenesis ROBERT A. PERKINSON AND PAMELA A. NORTON* Department of Medicine, Jefferson Medical College of Thomas Jefferson University, Philadelphia, Pennsylvania 19107 ABSTRACT Fibronectins (FNs) are essential for the proper development of embryonic mesenchymal tissues. A lacZ reporter gene has been fused to 4.9 kbp of DNA from the rat FN gene 58 flanking region, and this construct has been microinjected into fertilized mouse embryos to investigate the cis elements needed for the temporal and spatial regulation of FN in vivo. Histochemical staining of embryos for b-galactosidase activity demonstrated that four independent lines shared a specific pattern of lacZ expression, reflecting the activity of the fibronectin sequences contained within the transgene. Specifically, somites stained positively for lacZ, but expression was spatially and temporally non-uniform, with higher levels in more caudal somites after a total of ca. 13 somite pairs had formed. This rostral-caudal gradient of lacZ expression in somites of embryos beyond this stage resembled the distribution of endogenous FN mRNA, as detected by whole mount in situ hybridization. The transgene was not expressed in the developing heart where endogenous FN mRNA was detected. Unexpectedly, highly localized staining was observed within the neural tube beginning at ca. E10–10.5, and two of the lines exhibited additional areas of staining due to the individual integration sites. Thus, the 4.9 kbp FN fragment appears to recapitulate closely the complex pattern of FN expression observed during somitogenesis. A smaller fragment of 0.9 kbp also directed lacZ expression in caudal somites at E9.5, suggesting that these sequences are sufficient to establish the spatio-temporal pattern. Dev. Dyn. 208: 244–254, 1997. r 1997 Wiley-Liss, Inc. Key words: fibronectin; somite; transgenic mice; beta-galactosidase; in situ hybridization INTRODUCTION Fibronectins (FNs) are a family of large adhesive glycoproteins that possess the ability to interact with cells via integrin cell surface receptors as well as with other extracellular matrix components such as collagen and heparin (Hynes, 1990; Mosher, 1989; Paolella et al., 1993). FNs are present in vertebrate embryos at many sites of cell migration, and they have been postulated to be important for these processes (Adams and Watt, r 1997 WILEY-LISS, INC. 1993; DeSimone, 1994; Dufour et al., 1988; Hynes, 1994). Direct evidence for functional roles of FNs has been provided by perturbation experiments and genetic manipulation. Anti-FN antibodies or peptides that interfere with FN-integrin interactions can disrupt gastrulation of amphibian and avian embryos, and impede migration of mouse primary mesodermal cells (Boucaut et al., 1984a,b; Harrisson et al., 1993; Johnson et al., 1992; Klinowska et al., 1994). In contrast, mouse embryos homozygous for a disruption of the FN gene appear to gastrulate normally, but fail to form somites and exhibit other mesenchymal defects as well as defects in heart development (George et al., 1993). The failure to form somites is interesting in light of previous work implicating FNs in this process. FN accumulates in the region of the segmental plate that represents the site of active somitogenesis, adjacent to the caudalmost somite, as well as within and surrounding the somites themselves (Duband et al., 1987; Lash et al., 1984; Ostrovsky et al., 1983; Sternberg and Kimber, 1986). Addition of FN to presomitic mesoderm has been observed to enhance somite formation in culture (Lash et al., 1984). FN has also been implicated in heart formation and in cardiac cushion cell migration (Linask and Lash, 1988), and thus cardiac defects in the homozygous null mice were not surprising. In addition, evidence suggests that FNs promote the migration of neural crest cells through and around the somites via interactions with integrin cell surface receptors (Boucaut et al., 1984b; Duband et al., 1991; Duband et al., 1986; Dufour et al., 1988; Krotoski et al., 1986; Rovasio et al., 1983). In the adult, FN is a constituent of many basement membranes and FN levels are increased rapidly and locally during wound healing and in response to other fibrotic stimuli, suggesting an important role in these processes. However, the early death of the FN-null mice precludes assessment of the need for FN in later organogenesis or in the adult. The foregoing observations regarding the distribution of FNs suggest that the single copy Fn1 gene is regulated in a complex manner. The DNA elements needed for cell type-specific regulation of FN have begun to be identified from transfection studies and in Contract Grant sponsor: National Institutes of Health; Contract Grant number: GM-46402. *Correspondence to: Pamela A. Norton, Department of Medicine, Thomas Jefferson University, 1020 Locust Street, Philadelphia, PA 19107. Received 19 April 1996; Accepted 11 October 1996 FIBRONECTIN AND SOMITOGENESIS vitro studies using extracts of cells and tissues (Dean, 1989; Dean et al., 1989; Miao et al., 1993; Muro et al., 1992; Nakajima et al., 1992; Perkinson et al., 1996; Singh and Kanungo, 1991; Sporn and Schwarzbauer, 1995). However, the use of transgenic mice has proven necessary to identify the full complement of sequences needed for correct tissue-specific regulation of many genes (Kitsis and Leinwand, 1992). For instance, detailed analysis of the globin gene cluster has revealed that proper erythroid-specific, copy number dependent expression requires the presence of locus control regions that are 30–40 kbp from the coding regions (Dillon and Grosveld, 1993). This indicates that a complete description of the regulation of FN expression during embryogenesis and pathologic processes requires in vivo study. The E. coli lacZ gene has been used extensively as a reporter gene to investigate the sequences required to confer regulated expression during embryonic development (Bonnerot and Nicolas, 1993). The lacZ gene product, b-galactosidase, is a cell autonomous marker that is readily detected and that is functionally neutral in many cell types (Beddington et al., 1989; Goring et al., 1987). For these reasons, we chose to use a lacZ reporter gene to investigate the cis elements needed for the temporal and spatial regulation of the FN gene in transgenic mice. Previously, we established that a fragment of 4.9 kbp from the 58 end of the rat FN gene directs the expression of the lacZ reporter in rodent fibroblasts [pFNZ4.9 (Perkinson et al., 1996)]; we and others have also demonstrated that shorter FN fragments are also functional in such cells (Miao et al., 1993; Nakajima et al., 1992; Sporn and Schwarzbauer, 1995). In the present report, we describe the phenotype of mice that contain FN-lacZ transgenes pFNZ4.9 and pFNZ0.9 during the interval within which somitogenesis occurs. Histochemical staining for the lacZ gene product, b-galactosidase, reveals that the FN fragment directs the expression of lacZ in developing somites in a fashion that closely parallels endogenous FN mRNA, as detected by in situ hybridization. However, the transgene fails to recapitulate the complete repertoire of expression of the endogenous gene, as detailed below. RESULTS Whole Mount In Situ Hybridization Determination of FN mRNA Distribution We wished to employ lacZ as a reporter gene to identify sequences from the fibronectin gene that confer cell and tissue specific gene expression. However, interpretation of patterns of transgene expression requires comparison with expression of the endogenous Fn1 gene. The distribution of FN in the extracellular matrix may not coincide precisely with the location of cells that express the gene, due to differences in deposition or turnover. Thus, we evaluated FN mRNA levels by whole mount in situ hybridization. Embryos were isolated and fixed at E8.0–10.5, and hybridized to digoxigenin-labelled RNA probes either complementary (anti- 245 sense) or identical (sense control) to mouse FN mRNA. At ca. E8.0, the presomitic mesoderm hybridized intensely with the antisense probe, with lower levels of staining in somites and in the heart (Fig. 1a). All somites were labelled, and the somewhat more strongly stained caudal somites are indicated; the more rostral 2–3 somites are partially obscured by amnion remnants. At E9.5, specific labelling was detected in the caudal somites and presomitic mesoderm, with weak staining in the heart (Fig. 1b). At E10.5, significant background was present in these large specimens, but strong staining was observed with the antisense probe in the pre-somitic mesoderm and the caudal-most somites (Fig. 1c). These regions were not labelled with the sense control (Fig. 1d). Thus, it appears that FN mRNA levels are high prior to overt somite formation then decline as somites mature. Structure of the FN-lacZ Reporter Gene and Identification of Transgenic Mice We initially chose to test a fairly large segment of 58 flanking DNA from the FN gene for its ability to be expressed in vivo. A ca. 4.9 kbp fragment of the rat FN gene was isolated from a lambdaphage clone (Patel et al., 1987). This fragment includes ca. 4.8 kbp of 58 flanking sequence, the transcription start site and 136 nucleotides of 58 untranslated sequence. The sequence of the 58 flanking region is incomplete, but the proximal half has been determined in three overlapping segments (Nakajima et al., 1992; Patel et al., 1987; Sporn and Schwarzbauer, 1995). In general, the three reports are in agreement in regions of overlap. However, a short region of sequence divergence occurs between 21908 and 21897, with a PstI site present in only one of the two sequences, as shown in Figure 2b. The 4.9 kbp fragment that we isolated also lacks the PstI site at 21908 relative to the start of transcription. To determine which sequence variant corresponds to the endogenous FN gene, we performed PCR analysis. The forward primer corresponds to sequences that lie near a PstI site at ca. 24.8 kbp relative to the transcription start site (unpublished data); the reverse primer lies within the most 58 region of sequence concordance (see Fig. 2a). Amplification of rat genomic DNA resulted in a 2.5 kbp fragment that was identical in size to that derived from the plasmid template (Fig. 2c, compare lanes 2 and 3). In addition, the products amplified from both genomic DNA and plasmid were cleaved by BstXI to yield fragments of 2.0 and 0.5 kbp (data not shown; refer to map on Fig. 2a). Thus, the 4.9 kbp fragment retains the proper sequence organization of the rat FN gene, indicating that the sequence of Sporn and Schwarzbauer, extending from 22547 to 21080 and lacking the PstI site, is the correct one. Because the 4.9 kbp PstI fragment was derived from the same lambdaphage as the subclones that these workers sequenced, it very likely contains all the sequence features that they describe. 246 PERKINSON AND NORTON Fig. 1. Distribution of endogenous FN mRNA. FN mRNA distribution was examined by whole mount in situ hybridization using digoxigenin labelled RNA probes transcribed from mouse FN cDNA; a–c, antisense probe, d, sense probe. a: Lateral view of E8.5 embryo. Note the strong staining in the posterior region (P) with lower levels of staining in somites (arrowheads) and heart (H). b: Lateral view of E9.5 embryo. Hybridization is strongest in presomitic mesoderm (P) but is also apparent in caudal somites. c: Lateral view of E10.5 embryo; the head has been removed. While background is higher at this stage, staining is evident in the presomitic mesoderm of the tail (P) and in caudal somites. d: Dorsal view of E10.5 embryo; the head has been removed. Note lack of staining of presomitic mesoderm (P) and adjacent somites. Plasmid pFNZ4.9 contains a lacZ reporter gene flanked on the 58 end by the 4.9 kbp fragment and on the 38 end by 0.8 kbp sequence from the 38 end of the rat FN gene [Fig. 2a (Perkinson et al., 1996)]. Thus, FN sequences provide signals for the start of transcription and for polyadenylation, but the entire FN coding region has been replaced by lacZ. This plasmid was digested with NotI to remove vector sequences and the gel-purified insert DNA was injected into fertilized one-cell mouse embryos, which were then implanted into foster mothers. Screening of 41 pups by PCR analysis of tail DNA identified four mice that had acquired FNZ4.9 DNA (Table 1). All four founders were bred to C3H mice, and F1 progeny were assayed as pups (DNA) or as embryos (lacZ expression). This initial analysis demonstrated that the transgene was transmitted in all four lines and that it was expressed in the mice. In three lines, the frequency of positives was significantly less than 50%, suggesting the founders were mosaics. The similar frequency of positives detected by the two methods indicates that the lacZ assay reliably identified transgenic individuals. The estimated copy number of the transgene is also indicated. In subsequent generations, all four lines transmitted the transgene at the expected Medelian frequency (data Fig. 2. Structure of the FN-lacZ transgene. a: Diagram of FNZ4.9 DNA microinjected into embryos. The start of transcription is indicated by the arrow. Open bar, FN 58 flanking sequences; stippled box, FN 58 untranslated sequences; line, lacZ; shaded box, FN 38 sequences, with the site of polyadenylation indicated. The entire transgene is approximately 8.5 kbp in length. b: Comparison of the rat FN sequence reported by Sporn and Schwarzbauer (top) and Nakajima et al. (bottom) at positions ca. 21,900 kbp relative to the transcription start site. The former sequence continues upstream, whereas the latter truncates at the PstI site (underlined). The sequences are identical beginning at the italicized residues and continuing 38. c: PCR of FN from rat genomic DNA and pFNZ4.9. Lane 1, no template DNA; lane 2, rat genomic DNA; lane 3, plasmid pFNZ4.9 DNA. In lanes 1 and 2, 20 µl of each PCR reaction using the eLongase enzyme mixture was loaded; in lane 3, 5 µl of a reaction using Taq DNA polymerase was loaded. At left, positions of molecular size markers are indicated (kbp). 247 FIBRONECTIN AND SOMITOGENESIS TABLE 1. Summary of FNZ4.9 Transgene Transmission to F1 Generation Animals Est. copy number DNA detectiona lacZ detectionb Transgene positive/Total Male 4 2 6/22 13/31 Male 23 4–5 4/29 2/17 Male 24 6–7 6/35 1/8 Female 30 2–3 15/34 NT 19/53 6/46 7/43 15/34 aDNA detection, transgene detected by PCR of genomic DNA from tails. detection, transgene detected by staining embryos with X-gal. NT, not tested. blacZ not shown). Analyses of lacZ expression in the F1 mice and subsequent generations are detailed below. Features of lacZ Reporter Gene Expression Common to All Lines Starting at E8.0, embryos were isolated, fixed and stained with X-gal to detect b-galactosidase expression. Several individuals from lines 24 and 30 are shown in Figure 3; these two lines were indistinguishable, and the staining features described were also observed in embryos derived from lines 23 and 4. By E8.0, FN-null mutant embryos have a deficiency of mesoderm and fail to initiate somitogenesis (George et al., 1993), indicating a requirement for FN at or prior to this stage. At E8.0, we observed lacZ expression in the head-fold mesenchyme and in the first somite pairs (Fig. 3a), consistent with the reported normal distribution of FN at this stage (George et al., 1993). A similar situation was observed at ca. E8.5–9.0 (Fig. 3b), with staining in all somite pairs (ca. 8–12) but with diminished staining in the head region. In slightly more advanced embryos (ca. E9.5, 13 or more somite pairs), lacZ expression was most evident in newly formed somites, but was diminished in the older, rostral somites (Fig. 3c). Thus, formation of $ca. 13 somite pairs and the nearcompletion of turning coincided with a transition in the pattern of FNZ4.9 transgene expression. It was noted that X-gal staining was present within the rostral portion of the presomitic mesoderm, where somitogenesis is occurring (Fig. 3c and d). Dissection of the caudal somites revealed that the transgene appeared to be expressed in cells dispersed throughout the somite (date not shown). Note that no X-gal staining was observed in the developing heart (Fig. 3b and c). By day 10.5, newly forming somites continued to be positive for transgene expression (Fig. 3e). In addition, a blue stripe was evident in the lateral view, beginning in the cranial region and extending to the region of the hind limb bud. Somewhat more dorsal views demonstrated two internal bilateral stripes, most readily visible through the roof of the fourth ventricle at the back of the head (Fig. 3f ). Transverse sections of stained embryos revealed that these bilateral tracts were within the neural tube (see below). Non-transgenic littermates were completely negative for X-gal staining. The pattern of staining in the neural tube and in the most posterior somites persisted through days E11.5–12.5 (data not shown). Thus, the FN sequences present in the FNZ4.9 transgene direct lacZ expression in the somites, but not in the heart; the expression in the central nervous system is discussed below. Line Specific Features of Reporter Gene Expression The staining pattern described above is shared by all four lines. Lines 30 and 24 exhibited only this pattern of staining, but lines 4 and 23 exhibited additional regions of distinct, line specific staining. Characterization of these lines has been less extensive, and was restricted largely to stages E9.5–10.5. In addition to the common staining pattern, embryos of line 23 exhibited variable X-gal staining near the crown of the head, in the facial region and in the fore and hind limbs (not shown). E10.5 embryos of line 4 exhibited intense staining in the developing forelimb which was strongest in the anterior and dorsal regions of the limb (Fig. 4). No staining of the forelimbs at this stage has been observed by in situ hybridization with the antisense FN probe (data not shown). The hind limb bud was stained much less intensely, even when a similar size and developmental stage was reached (data not shown). Additional staining was observed in the area where the cerebellum will form, and in more rostral somites; neither is typical of endogenous FN mRNA (Fig. 1c and data not shown). The staining patterns are summarized in Table 2. We attribute the staining patterns unique to lines 4 or 23 to the different integration sites of the transgenes in these lines, as they are not consistent with the pattern of expression of the endogenous gene. Ectopic Expression of the FN-lacZ Transgene in the Neural Tube An E10.5 embryo of line 4 similar to those shown in Figure 4 that had been stained with X-gal was postfixed and sectioned. Transverse sections across the region just rostral to the forelimb bud were reacted with anti-FN antibody and horseradish peroxidase-conjugated secondary antibody. FN staining was distributed widely, including within the heart and surrounding the neural tube. However, there was no obvious colocalization of the brown peroxidase reaction product with the blue X-gal staining within the ventro-lateral 248 PERKINSON AND NORTON TABLE 2. Summary of X-gal Staining Patterns of FNZ4.9 Transgenic Mice Line 4 23 24 30 Heart 2 2 2 2 Somites 1 1 1 1 Neural tube 1 1 1 1 Other forelimb, head limbs, head 2 2 within the neural tube (Fig. 1a and not shown). Thus, the combined data indicate that FNZ4.9 transgene accurately reflects expression of the endogenous gene in the somites, but not in other tissues. A 0.9 kb Fragment Is Sufficient to Confer Somite-Specific Expression Fig. 3. Expression of the FNZ4.9 transgene in E8.0–10.5 embryos of lines 24 and 30. Embryos from the two lines were stained with X-gal at various developmental stages; representative individuals are shown. a: E8.0 embryo with 2–3 somites. Note staining in head folds (arrows). b: Lateral view of an E9.0 embryo with ca. 11 somites. The embryo is incompletely turned, bringing both columns of somites into view. Note similar staining of somites at all levels, and lack of staining in the heart (labelled with adjacent H). c: Lateral view of an E9.25 embryo with ca. 15 somites. Staining in rostral somites has diminished, and there is lack of staining in the heart (H). d: Higher magnification view of caudal end of an E9.25 embryo. Note that staining is present in the presomitic mesoderm, caudal to the last visible somite, the boundaries of which are defined by the arrows. e: Lateral view of an E10.5 embryo. Staining is observed in newly formed caudal somites, and as well as within the neural tube (see text). f: A more dorsal view of an E10.5 embryo, revealing the bilateral neural tube staining. neural tube (Fig. 5). Thus, FN does not appear to accumulate at a specific site within the neural tube. In addition, high levels of FN mRNA were not detected Two transgenics were obtained with pFNZ0.9, which contains only 880 bp of FN promoter sequence upstream of the transcription start site. Lines were established from each founder (designated lines 255 and 272), and F1 and F2 progeny analyzed for lacZ expression at E9–11. Some individuals from both lines exhibited X-gal staining in caudal somites and presomitic mesoderm; Figure 6a shows an individual from line 255 with staining similar to that seen for FNZ4.9 mice at a comparable stage (compare to Fig. 3c). A total of 12 E9.0–9.5 individuals were positive for somitic and presomitic mesoderm staining out of 15 total positive for X-gal staining. In contrast, the somite staining pattern was less well preserved in later embryos; only 8/19 total E10-12 X-gal positives exhibited any staining in caudal somites, and staining tended to be weak relative to that seen with FNZ4.9 embryos. However, a different pattern was seen in some individuals beginning at E10, with staining of lateral structures that may represent the dorsal root ganglia. The intensity of staining is highly variable, from almost none to very strong (Fig. 6b and c); such variation was observed in both lines and extreme examples were found within single litters. Staining of head structure was also observed, but no staining was detected in the heart, and the ectopic staining within the neural tube was diminished relative to the FNZ4.9 mice. These staining features were never observed in studies of the FNZ4.9 mice. Single individuals were found with both somitic and lateral staining patterns, suggesting that they are not mutually exclusive. Note that the in situ hybridization data did not reveal any evidence for lateral expression of FN (compare Fig. 1b and c with Fig. 6b and c). Thus, the 0.9 kbp fragment was sufficient to establish transgene expression in caudal somites and the presomitic mesoderm in embryos of E9–9.5, but that this was not sufficient to reliably direct expression to nascent somites at later times. DISCUSSION Four independent lines of mice were obtained that contain the FNZ4.9 transgene. Studies of embryos FIBRONECTIN AND SOMITOGENESIS 249 Fig. 4. Expression of the FNZ4.9 transgene in E10.5 embryos from line 4. Embryos were collected at E10.5 and stained with X-gal. Lateral and dorsal views of line 4 embryos. Note the staining in caudal somites and in the neural tube. Strongly stained forelimbs are marked by arrows. revealed that the FN sequences present in the transgene are sufficient to establish and maintain an appropriate rostral-caudal gradient of expression in the somitic mesoderm. However, the transgene is not expressed in the developing heart although endogenous FN mRNA is present [Fig. 1 and (ffrench-Constant and Hynes, 1988)]. A truncated version of this transgene also directed reporter gene expression to the somites, but the instability of the phenotype over time suggests that some regulatory elements have been removed. it is interesting that FN expression is high during the compaction process. The lacZ-positive cells appeared to be distributed throughout the somite. This distribution is not in precise agreement with reports of the FN protein distribution in the chick and mouse, where FN has been reported to surround the somites, with less material within the somite (Duband et al., 1987; Duband et al., 1986; Krotoski et al., 1986; Ostrovsky et al., 1983; Stepp et al., 1994; Sternberg and Kimber, 1986). However, b-galactosidase is a cell-autonomous marker; the cells that synthesize the protein retain it. In contrast, FN is an extracellular protein, and its distribution may not reflect its pattern of synthesis, due to differential deposition or turnover in the extracellular matrix. Alternatively, there may be reduced immunoreactivity of the FN in certain regions, possibly due to masking by other molecules. Most of the studies regarding the localization of FN protein (see above) and mRNA (ffrench-Constant and Hynes, 1989; ffrench-Constant and Hynes, 1988) in somites have been done in the chick. Although the mouse and chick are likely to be very similar, our study has revealed that FN gene expression undergoes a distinct transition during somitogenesis. Beyond the ca. 12 somite stage, the caudal-most somites express much higher levels of FN mRNA and b-galactosidase compared to the more rostral structures. We might thus expect that FN distribution in the matrix will vary with developmental stage and axial level. In agreement with this idea, another group has described a shift in the localization of FN in older, more rostral somites in the mouse (Sternberg and Kimber, 1986). It is also of interest that differences have been observed in the expression of cadherin-11 and N-cadherin in newly formed vs. older somites (Kimura et al., 1995). Thus, it FNs Gene Expression During Somitogenesis An early defect manifested by FN-null mice is the absence of somites, along with a general deficiency of mesoderm (George et al., 1993). Interestingly, we observed that FN mRNA levels were elevated in the presomitic mesoderm, consistent with a role for FN in somitogenesis. Somites are formed from the paraxial mesoderm, beginning between E8.0 and E8.5, with a new pair forming at approximately hourly intervals (Tam, 1981). It has been proposed that FN is involved during early somitogenesis, at the compaction stage that precedes epithelialization (Lash et al., 1984). Expression of the lacZ reporter and FN mRNA was evident in the pre-somitic mesoderm (Figs. 2 and 4), as is FN protein (Duband et al., 1987; Lash et al., 1984; Ostrovsky et al., 1983). The domain of FN mRNA expression extends more caudally than that of lacZ, which may reflect mRNA accumulation prior to the synthesis of significant levels of protein. Alternatively, expression of the transgene may be slightly delayed with respect to the endogenous gene; such a phenomenon has been reported for the MyoD gene (Asakura et al., 1995). We have shown previously that the activity of the FN promoter is increased with increased cell density in mouse fibroblasts (Perkinson et al., 1996). Thus, 250 PERKINSON AND NORTON show that the FN gene is induced in newly formed somites, and that lacZ expression appears somewhat elevated in the anterior portion of each somite (Fig. 3d). These findings are consistent with a role for FNs in crest cell migration, although FN protein has been reported to be evenly distributed throughout the somite (Krotoski et al., 1986). However, our data are consistent with formation of a transient gradient. FN Gene Expression in Other Regions Fig. 5. Expression of the endogenous FN gene. Immunohistochemical localization of FN. An E10.5 embryo from line 4 was stained with X-gal, postfixed, then embedded in paraffin and sectioned. This transverse section was incubated with polyclonal anti-FN antibody, and detected via horseradish peroxidase conjugated secondary antibody. Note reactivity surrounding the heart (H, bottom) and the neural tube (N, top), and the lack of co-localization within the neural tube between FN (brown) and b-galactosidase (blue). may be important to evaluate several developmental stages of both chick and mouse embryos to detect any transitions that occur in the expression of various adhesion molecules. Neural crest cells migrate from the dorsal surface of the neural tube in an axial wave that follows somitogenesis. Immunolocalization studies suggested that FN plays a role in neural crest cell migration through and around the somites. Later studies demonstrated that FN promotes neural crest cell migration, and that specific antibodies and peptides disrupt FN-cell interactions (Boucaut et al., 1984a,b; Duband et al., 1991; Duband et al., 1986; Dufour et al., 1988; Krotoski et al., 1986; Rovasio et al., 1983). If FN is involved in crest cell migration, it must be present prior to crest cell migration, and should be concentrated in the anterior portion of the somite (Bronner-Fraser et al., 1991). Our data The FNZ4.9 transgene contains ca. 4.9 kbp of DNA 58 to the transcription start site, 0.1 kbp of 58 untranslated sequence and less than 1.0 kb of 38 sequences; these are sufficient to confer appropriate FN gene expression in newly forming somites. Regulatory elements that have been identified as important for FN promoter function include an enhancer element that lies nearly 2.0 kbp from the start site (Sporn and Schwarzbauer, 1995), as well as more proximal elements that are important for promoter activity in cultured cells (Bowlus et al., 1991; Dean et al., 1989; Miao et al., 1993; Muro et al., 1992; Nakajima et al., 1992). All these elements are present in the FNZ4.9 transgene; the absence of lacZ expression in the early embryonic heart indicates that the construct probably lacks one or more as yet unidentified tissue-specific sequence element. The posited additional regulatory element(s) could lie further upstream of the transcription start site than the nearly 5.0 kbp included in FNZ4.9. Alternatively, it could lie within the transcribed region. A DNAseI hypersensitive site within the first intron is present in at least one adult tissue where FN is expressed (liver) and absent from one where the gene is largely silent (cerebrum) (Singh and Kanungo, 1991). Analyses of the FNZ0.9 transgene indicates that ,1.0 kbp is sufficient to direct reporter gene expression in the somites, but that maintenance of gene expression is more variable. It is possible that the variable phenotype of FNZ0.9 embryos at later stages is a consequence of removing the fibroblast enhancer (Sporn and Schwarzbauer, 1995). Additional constructs will be needed to test the role of these individual elements. Staining in the neural tube was unexpected; typically, FN protein has not been detected within the neural tube, and we did not detect significant levels of mRNA. However, FN and its receptor were detected in one study (Duband et al., 1986), and an in situ hybridization study in the chick noted labelling in precisely this area when an embryo of a similar developmental stage was examined (ffrench-Constant and Hynes, 1989). Thus, it is possible that the endogenous FN gene normally is expressed in a restricted fashion in the neural tube, but that expression of the FN-lacZ transgene is enhanced, possibly due to the absence of a silencer element. One trivial explanation, that heterologous sequences were fused during cloning to proximal FN promoter elements, is discounted by the PCR data shown in Figure 1c. The identity of the labelled cells has not been established, but based on their ventral FIBRONECTIN AND SOMITOGENESIS 251 Fig. 6. Expression of the FNZ0.9 transgene in E9.5–10.5 embryos. Embryos from lines 255 and 272 were stained with X-gal at various developmental stages; representative individuals are shown. a: Lateral view of an E9.5 embryo with ca. 20 somites. Staining in presomitic mesoderm (P) and caudal somites is evident and there is lack of staining in the heart (H). b: Frontal-lateral view of a ca. E10.0 embryo. Staining is observed internally and in the cranial region, and very weak staining is seen laterally in the vicinity of the more rostral somites. c: Lateral view of an E10.5 embryo. Lateral staining is very strong in this individual. location, they might represent a motor neuron subset. This raises the possibility that the expression of the FNZ4.9 transgene is under the influence of one or more members of the LIM homeobox genes, which appear to define subgroups of motor neurons in a combinatorial fashion (Tsuchida et al., 1994). Further studies are necessary to establish the identity of the transgenepositive cells as well as the sequences responsible for this highly restricted pattern of expression. defined as E0.5.) Embryos from timed matings of C3H/HeJ mice (Taconic Farms) were dissected in phophate buffered saline (PBS), 2.0 mM EGTA. Extraembryonic membranes were removed, and older embryos were decapitated. Embryos were fixed in fresh 4% paraformaldehyde, 2.0 mM EGTA in PTW (PBS, 0.1% Tween 20) at 4°C for 5–18 hr, and stored in methanol at 220°C for up to 1 month. Rehydrated embryos were treated with 10 µg/ml Proteinase K (Gibco/BRL) in PTW for 10–20 min, depending on the age of the embryo, then post-fixed for 20 min in 4% paraformaldehyde, 0.1% glutaraldehyde in PTW. Embryos were washed with PTW and equilibrated in hybridization mix (50% formamide, 1.3 3 SSC, 5 mM EDTA, 50 µg/ml yeast RNA, 0.2% Tween-20, 0.5% CHAPS, 100 µg/ml heparin), then prehybridized in 1.0 ml of hybridization mix for 2 hr at 65°C. Embryos were incubated overnight at 65°C in 1.0 ml of prewarmed hybridization mix containing 0.1–0.25 µg/ml digoxygenin-labelled RNA probe. Following hybridization, embryos were washed twice at 65°C for 30 min with 1.5 ml hybridization mix, then equilibrated with MABT (100 mM sodium maleate, pH 7.5, 150 mM NaCl, 0.1% Tween). Embryos were gently rocked at room temperature for 1 hr in 1.5 ml MABT, 2% Boehringer Blocking Reagent (BBR), then for 1.5 hr with 1.5 ml MABT, 2% BBR, 20% heat-treated sheep serum. Antibody binding was overnight at 4°C in 1.0 ml MABT, 2% BBR, 20% serum and 1/2,000th dilution of alkaline phosphataseanti-DIG antibody (Boehringer). Embryos were washed three times with MABT, twice with NTMT (0.1 M NaCl, 0.1 M Tris-Cl, pH 9.5, 0.05 M MgCl2, 0.1% Tween-20), and transferred to 1.5 ml NTMT/NBT/BCIP (NTMT, 0.225 mg p-nitro blue tetrazolium chloride, 0.115 mg 5-bromo-4-chloro-3-indolyl phosphate). To stop color EXPERIMENTAL PROCEDURES Digoxygenin-Labelled RNA Probe Preparation A 480 bp EcoRI-HindIII mouse fibronectin cDNA fragment was subcloned into pBS2 (Stratagene); this fragment includes 27 bp of the alternative EIIIB exon (Górski et al., in press). Labelled RNA probes were synthesized by in vitro transcription of the EcoRI or HindIII linearized DNA with T3 or T7 RNA polymerase, respectively. A 20 µl in vitro transcription reaction contained 1 µg DNA, 2 µl of NTP label mix [10 mM ATP, 10 mM CTP, 10 mM GTP, 6.5 mM UTP, 3.5 mM digoxygenin-labelled UTP (Boehringer Mannheim)], 1 µl of RNase Block (40 units, Stratagene) and 1 µl of RNA polymerase (50 units, Stratagene). Reactions were incubated at 37°C for 2 hr then stopped with 2 µl of 0.2 M EDTA, pH 8.0. The RNA was precipitated with 2.5 µl of 4M LiCl and 75 µl of 100% ethanol at 220°C overnight. The RNA was collected by centrifugation, washed once with 70% ethanol, dried, and stored in 100 µl H2O and 1 µl of RNase inhibitor. Whole Mount In Situ Hybridization With Non-radioactive Detection Whole mount in-situ hybridizations were performed on E9.5–E10.5 mouse embryos essentially as described (Henrique et al., 1995). (The date of the vaginal plug is 252 PERKINSON AND NORTON development, embryos were washed with PTW, refixed, dehydrated and stored in 100% methanol. Construction of the Transgene and Verification of Its Structure The plasmids pFNZ4.9 and pFNZ0.9 contain a lacZ reporter flanked by sequences from the rat FN gene, and the construction of these plasmids and their characterization in cultured cells has been reported elsewhere (Perkinson et al., 1996). Briefly, a 4.9 kilobase pair (kbp) PstI fragment was excised from lrFN-9 (Patel et al., 1987) and inserted upstream of lacZ; this fragment includes the transcription initiation site and 136 nucleotides of 58 untranslated sequence. A SacII-HindIII fragment containing most of the 38 untranslated sequence and 204 nucleotides beyond the site of polyA addition was excised from lrFN-1 (Patel et al., 1987) and inserted downstream of lacZ. Thus, lacZ transcripts from pFNZ4.9 should contain 136 nucleotides of 58 and 513 nucleotides of 38 untranslated sequence derived from rat FN (the complete regions are 207 and 691 nucleotides, respectively). The lacZ reporter was designed so as to be translated efficiently in eukaryotic cells, and the b-galactosidase fusion protein that results is typically nuclear or peri-nuclear in localization (McInnis et al., 1995). Similarly, pFNZ0.9 contained a StuI-PstI fragment of 880 bp inserted upstream of the lacZ reporter. To confirm the structure of the 58 flanking region of the rat FN gene, primers were designed based on the sequence of the furthest upstream portion of the 4.9 kbp fragment for the forward primer (unpublished data) and on published sequence for the reverse primer (Nakajima et al., 1992; Sporn and Schwarzbauer, 1995). Forward (58-GATTCTGTGCATGGGAGC-38) and reverse (58-CCACAAATATGGCAGATC-38) primers were used in PCR reactions with Taq DNA polymerase (Stratagene) or eLongase Polymerase mixture (Gibco/ BRL), following manufacturer’s recommended conditions. Templates were either 0.2 µg rat liver genomic DNA (gift of D. DeSimone) or 25 pg of plasmid DNA; amplification was for 35 cycles (94°C, 30 sec; 52°C, 30 sec; and 68 or 72°C, 3 min). Aliquots were analyzed by agarose gel electrophoresis and visualized by photographing the ethidium bromide stained gel. The photograph was scanned and the TIFF image was reversed with NIH Image. Production and Breeding of Trangenic Mice Prior to microinjection, the FN-lacZ fragment was separated from vector sequences by digestion with NotI (pFNZ4.9) or NotI plus BglII (pFNZ0.9) and gel purification. Microinjection of DNA into one-cell mouse embryos and embryo transfer was performed at the Jefferson Cancer Institute Transgenic Mouse Facility. Embryos were derived by mating C57BL/6J 3 C3H/ HeJ F1 hybrids. In subsequent generations, transgenic individuals were bred to C3H males and females. Embryos for lacZ analyses were derived from timed matings or from mice sacrificed when obviously pregnant. In the latter case, all the members of the litter were examined for length, number of somites (in the case of younger embryos) or appearance of limbs (older embryos), which were used as criteria to establish the approximate gestational age of the group (Kaufman, 1992). Detection of lacZ Transgene Pups were weaned at 3 to 4 weeks of age and tail biopsies were performed essentially as described (Hogan et al., 1994) and resuspended in 100 µl TE (10 mM Tris, pH 8.0, 0.1 mM EDTA). PCR was carried out in a volume of 100 µl using 1.0 µl genomic DNA, 200 µM each dNTP, 10 µl 10 3 reaction buffer and 2.5 U Taq DNA Polymerase (Perkin-Elmer/Cetus or Stratagene) and 50 pmol of each lacZ-specific primer (forward, 58-CGTAATAGCGAAGAGGCCCG-38; reverse, 58-GCCCGTTGCACCACAGATGA-38). Amplification was for 30 cycles (1 min, 94°C; 1 min, 62°C; and 1 min, 72°C). Twenty microliter aliquots were analyzed by agarose gel electrophoresis and visualized by staining with ethidium bromide. A 363 bp product was obtained with individuals whose genome contained a copy of the transgene. Transgene copy number was determined by dot blot hybridization analysis of DNA isolated from F1 progeny mice, using a random primer labelled fragment of lacZ as a probe; standard protocols were followed (Sambrook et al., 1989). Blots were exposed to a Molecular Dynamics PhophorImager screen for quantitation. Histochemical Assay for b-Galactosidase Non-transgenic C3H females were mated to transgenic males; the date of plug is considered E0.5 (embryonic day 0.5). Females were sacrificed at E8–12 and embryos were dissected out in PBS. Treatment of embryos was similar to that described (Bonnerot and Nicolas, 1993; Hogan et al., 1994). Embryos were fixed in 10 ml of cold freshly prepared 4% paraformaldehyde in PBS containing 0.02% NP40, 0.01% sodium deoxycholate, 2 mM MgCl2, and 5 mM EGTA (PBS rinse mix), according to age (from 5 min for E8 to 20 min for E12.5). After fixation, embryos were washed three times with 10 ml of PBS rinse mix at room temperature. Staining was in PBS containing 50 mM K3Fe(CN) 6, 50 mM K4Fe(CN) 6, 0.02% NP40, 0.01% sodium deoxycholate, 2 mM MgCl2, 1 mM EGTA and 1.0 mg/ml X-gal (5-bromo4-chloro-3-indolyl-b-D-galactoside; stock, 40 mg/ml in dimethylformamide) at 37°C from 2 to 48 hr. Following color development, embryos were post-fixed in 4% paraformaldehyde in PBS. In some cases, embryos were dehydrated in ethanol, and cleared in methyl salicylate. Immunohistochemistry Embryos that had been postfixed following X-gal staining were dehydrated, embedded in paraffin and 7.0 µm serial sections were cut and dried. Prior to immunodetection, sections were deparaffinized, and rehydrated into PBS. After blocking with 5% dry milk FIBRONECTIN AND SOMITOGENESIS powder in Tris-buffered saline, sections were incubated first with a polyclonal antibody directed against mouse FN (Telios). Following three washes in PBS, horseradish peroxidase-conjugated goat anti-rabbit IgG (BioRad) was added. After three washes, substrate (0.025% diaminobenzidine, 0.03% H2O2 ) was applied, and development was monitored visually. ACKNOWLEDGMENTS We thank A. Gehris for help with generating sections, K. Cheah for providing detailed protocols for lacZ detection, and R. McInnis for technical assistance during the early stages of this project. We are grateful to D. Henrique, D. Ish-Horowitz, S. Sporn and J. Schwarzbauer for communicating data prior to publication and to R. Hynes for the lambdaphage clones. Many thanks go to our colleagues V. Bennett, A. Gehris, and G. Grunwald for helpful comments on manuscript. REFERENCES Adams, J.C., and Watt, F.M. (1993) Regulation of development and differentiation by the extracellular matrix. Development 117:1183– 1198. Asakura, A., Lyons, G.E., and Tapscott, S.J. (1995) The regulation of MyoD gene expression: Conserved elements mediate expression in embryonic axial muscle. Dev. Biol. 171:386–398. Beddington, R.S.P., Morgenstern, J., Land, H., and Hogan, A. (1989) An in situ transgenic enzyme marker for the midgestation mouse embryo and the visualization of inner cell mass clones during early organogenesis. Development 106:37–46. Bonnerot, C., and Nicolas, J.-F. (1993) Application of lacZ gene fusions to postimplantation development. In ‘‘Guide to Techniques in Mouse Development,’’ Wassarman, P.M., and DePamphilis, M.L. (eds). San Diego: Academic, pp. 451–469. Boucaut, J.C., Darribere, T., Boulebache, H., and Thiery, J.P. (1984a) Prevention of gastrulation but not neurulation by antibodies to fibronectin in amphibian embryos. Nature 307:364–367. Boucaut, J.C., Darribère, T., Poole, T.J., Aoyama, H., Yamada, K.M., and Thiery, J.P. (1984b) Biologically active synthetic peptides as probes of embryonic development: A competitive peptide inhibitor of fibronectin function inhibits gastrulation in amphibian embryos and neural crest cell migration in avian embryos. J. Cell Biol. 99:1822– 1830. Bowlus, C.L., McQuillan, J.J., and Dean, D.C. (1991) Characterization of three different elements in the 58-flanking region of the fibronectin gene which mediate a transcriptional response to cAMP. J. Biol. Chem. 266:1122–1127. Bronner-Fraser, M., Stern, C.D., and Fraser, S. (1991) Analysis of neural crest cell lineage and migration. J. Craniofac. Genet. Dev. Biol. 11:214–222. Dean, D.C. (1989) Expression of the fibronectin gene. Am. J. Respir. Cell Mol. Biol. 1:5–10. Dean, D.C., Blakeley, M.S., Newby, R.F., Ghazal, P., Hennighausen, L., and Bourgeois, S. (1989) Forskolin inducibility and tissue-specific expression of the fibronectin promoter. Mol. Cell. Biol. 9:1498–1506. DeSimone, D.W. (1994) Adhesion and matrix in vertebrate development. Curr. Opin. Cell Biol. 6:747–751. Dillon, N., and Grosveld, F. (1993) Transcriptional regulation of multigene loci: Multilevel control. Trends Genet. 9:134–137. Duband, J.-L., Rocher, S., Chen, W.-T., Yamada, K.M., and Thiery, J.P. (1986) Cell adhesion and migration in the early vertebrate embryo: Location and possible role of the putative fibronectin receptor complex. J. Cell Biol. 102:160–178. Duband, J.-L., Dufour, S., Hatta, K., Takeichi, M., Edelman, G.M., and Thiery, J.P. (1987) Adhesion molecules during somitogenesis in the avian embryo. J. Cell Biol. 104:1361–1374. Duband, J.-L., Dufour, S., Yamada, S.S., Yamada, K.M., and Thiery, 253 J.P. (1991) Neural crest cell locomotion induced by antibodies to beta 1 integrins. A tool for studying the roles of substratum molecular avidity and density in migration. J. Cell Sci. 98:517–532. Dufour, S., Duband, J.-L., Kornblihtt, A.R., and Thiery, J.P. (1988) The role of fibronectins in embryonic cell migrations. Trends Genet. 4:198–203. ffrench-Constant, C., and Hynes, R.O. (1988) Patterns of fibronectin gene expression and splicing during cell migration in chicken embryos. Development 104:369–382. ffrench-Constant, C., and Hynes, R.O. (1989) Alternative splicing of fibronectin is temporally and spatially regulated in the chicken embryo. Development 106:375–388. George, E.L., Georges-Labouesse, E.N., Patel-King, R.S., Rayburn, H., and Hynes, R.O. (1993) Defects in mesoderm, neural tube and vascular development in mouse embryos lacking fibronectin. Development 119:1079–1091. Goring, D.R., Rossant, J., Clapoff, S., Breitman, M.L., and Tsui, L.C. (1987) In situ detection of b-galactosidase in lenses of transgenic mice with a g-crystallin/lacZ gene. Science 235:456–458. Górski, G.K., Aros, M.C., and Norton, P.A. (1997) Characterization of mouse fibronectin alternative mRNAs reveals an unusual isoform present transiently during liver development. Gene Expression, in press. Harrisson, F., Van Nassauw, L., Van Hoof, J., and Foidart, J.-M. (1993) Microinjection of anti-fibronectin antibodies in the chicken blastoderm: Inhibition of mesoblast cell migration but not of cell ingression at the primitive streak. Anat. Rec. 236:685–696. Henrique, D., Adam, J., Myat, A., Chitnis, A., Lewis, J., and IshHorowicz, D. (1995) Expression of a Delta homologue in prospective neurons in the chick [see comments]. Nature 375:787–790. Hogan, B., Beddington, R., Constantini, E., and Lacy, E. (1994) ‘‘Manipulating the Mouse Embryo: A Laboratory Manual,’’ 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Hynes, R.O. (1990): ‘‘Fibronectins.’’ New York: Springer Verlag. Hynes, R.O. (1994) Genetic analysis of cell-matrix interactions in development. Curr. Opin. Genet. Dev. 4:569–574. Johnson, K.E., Boucaut, J.C., and DeSimone, D.W. (1992) The role of the extracellular matrix in amphibian gastrulation. Curr. Top. Dev. Biol. 27:91–127. Kaufman, M.H. (1992) ‘‘The Atlas of Mouse Development.’’ London: Academic. Kimura, Y., Matsunami, H., Inoue, T., Shimamura, K., Uchida, N., Ueno, T., Miyazaki, T., and Takeichi, M. (1995) Cadherin-11 expressed in association with mesenchymal morphogenesis in the head, somite, and limb bud of early mouse embryos. Dev. Biol. 169:347–358. Kitsis, R.N., and Leinwand, L.A. (1992) Discordance between gene regulation in vitro and in vivo. Gene Exp. 2:313–318. Klinowska, T.C., Ireland, G.W., and Kimber, S.J. (1994) A new in vitro model of murine mesoderm migration: The role of fibronectin and laminin. Differentiation 57:7–19. Krotoski, D.M., Domingo, C., and Bronner-Fraser, M. (1986) Distribution of a putative cell surface receptor for fibronectin and laminin in the avian embryo. J. Cell Biol. 103:1061–1071. Lash, J.W., Seitz, A.W., Cheney, C.M., and Ostrovsky, D. (1984) On the role of fibronectin during the compaction stage of somitogenesis in the chick embryo. J. Exp. Zool. 232:197–206. Linask, K.K., and Lash, J.W. (1988) A role for fibronectin in the migration of avian precardiac cells. I. Dose-dependent effects of fibronectin antibody. Dev. Biol. 129:324–329. McInnis, R., Perkinson, R.A., Kuo, B.A., and Norton, P.A. (1995) Unexpected nuclear localization of a chimeric b-galactosidase lacZ reporter gene product in mammalian cells. Biochem. Mol. Biol. Int. 36:483–490. Miao, S., Suri, P.K., Shu-Ling, L., Abraham, A., Cook, N., Milos, P., and Zern, M.A. (1993) Role of the cyclic AMP response element in rat fibronectin gene expression. Hepatology 17:882–890. Mosher, D.F. (1989) ‘‘Fibronectin.’’ London: Academic. Muro, A.F., Bernath, V.A., and Kornblihtt, A.R. (1992) Interaction of the 2170 cyclic AMP response element with the adjacent CAATT box in the human fibronectin gene promoter. J. Biol. Chem. 267: 12767–12774. 254 PERKINSON AND NORTON Nakajima, T., Nakamura, T., Tsunoda, S., Nakada, S., and Oda, K. (1992) E1A-responsive elements for repression of rat fibronectin gene transcription. Mol. Cell. Biol. 12:2837–2846. Ostrovsky, D., Cheney, C.M., Seitz, A.W., and Lash, J.W. (1983) Fibronectin distribution during somitogenesis in the chick embryo. Cell Differ. 13:217–223. Paolella, G., Barone, M.V., and Baralle, F.E. (1993) Fibronectin. In: ‘‘Extracellular Matrix. Chemistry, Biology and Pathobiology with Emphasis on the Liver,’’ Zern, M.A., and Reid, L.M. (eds). New York: Marcel Dekker, pp. 3–24. Patel, R.S., Odermatt, E., Schwarzbauer, J.E., and Hynes, R.O. (1987) Organization of the fibronectin gene provides for exon shuffling during evolution. EMBO J. 6:2565–2572. Perkinson, R.A., Kuo, B.A., and Norton, P.A. (1996) Modulation of transcription of the rat fibronectin gene by cell density. J. Cell. Biochem. 63:74–85. Rovasio, R.A., Delouvee, A., Yamada, K.M., Timpl, R., and Thiery, J.P. (1983) Neural crest cell migration: Requirements for exogenous fibronectin and high cell density. J. Cell Biol. 96:462–473. Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) ‘‘Molecular Cloning: A Laboratory Manual,’’ 2nd ed. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory Press. Singh, S., and Kanungo, M.S. (1991) DNase I hypersensitive sites of the 58 region of the fibronectin gene of the liver of the rat. Biochem. Biophys. Res. Commun. 181:131–137. Sporn, S., and Schwarzbauer, J.E. (1995) Identification of an enhancer involved in tissue-specific regulation of the rat fibronectin gene. Nucleic Acids Res. 23:3335–3342. Stepp, M.A., Urry, L.A., and Hynes, R.O. (1994) Expression of alpha 4 integrin mRNA and protein and fibronectin in the early chicken embryo. Cell Adhes. Commun. 2:359–375. Sternberg, J., and Kimber, S.J. (1986) Distribution of fibronectin, laminin and entactin in the environment of migrating neural crest cells in early mouse embryos. J. Embryol. Exp. Morph. 91:267–282. Tam, P.P.L. (1981) The control of somitogenesis in mouse embryos. J. Embryol. Exp. Morph. 65(Suppl):103–128. Tsuchida, T., Ensini, M., Morton, S.B., Baldassare, M., Edlund, T., Jessell, T.M., and Pfaff, S.L. (1994) Topographic organization of embryonic motor neurons defined by expression of LIM homeobox genes. Cell 79:957–970.