| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Two murine and human homologs of mab-21, a cell fate determination gene involved in Caenorhabditis elegans neural development
Introduction
Results
Cloning and sequence analysis of mab-21 homologs in human and mouse
Protein sequence analysis
Mab21l1 and Mab21l2 map to mouse chromosome 3
MAB21L2 maps to human chromosome 4q31
Mab21l2 is expressed in the CNS and neural crest in midgestation embryogenesis
Mab21l1 and Mab21l2 remain expressed in the adult cerebellum and eye
Subcellular localization
Discussion
Materials And Methods
General procedures
Tissue preparation and RNA extraction
RNA fingerprinting
Sequence analysis
Genetic mapping
FISH mapping
In situ hybridization of mouse tissue sections
Whole-mount in situ hybridization
Cellular immunofluorescence
Acknowledgments
References
Two murine and human homologs of mab-21, a cell fate determination gene involved in Caenorhabditis elegans neural development
Received June 23, 1999; Revised and Accepted September 30, 1999
We report the cloning and genetic characterization of one human and two murine homologs of the mab-21 cell fate specification gene. mab-21 participates in the formation of sensory organs in the male nematode tail, and is essential for other developmental functions elsewhere in the Caenorhabditis elegans embryo. The expanding mab-21 gene family, which is strikingly conserved in evolution, includes two putative Drosophila members. The two mammalian genes, encoding 41 kDa nuclear basic proteins, are expressed in partially overlapping territories in the embryonic brain, eye and limbs, as well as in neural crest derivatives. Recent genetic data implicating mab-21 as a downstream target of TGF-[beta] signaling, together with the distribution of mab-21 transcripts in the mouse embryo, propose these novel genes as relevant factors in various aspects of vertebrate neural development.
INTRODUCTION
The nematode Caenorhabditis elegans represents an ideal model for the study of genetic interactions occurring during embryonic development. In that system, the analysis of genetic mutants has led to the isolation of factors playing key roles in the formation of specific body structures, and the analysis of epistatic interactions has clarified their relationships with other genes and hierarchical status within specific morphogenetic pathways. The identification of the male abnormal 21 (mab-21) mutation (1) and the analysis of the corresponding gene are examples of the power of the nematode as a biological system in molecular embryology. By genetic analysis of spontaneous loss-of-function mutants, this gene has been implicated in cell fate determination events crucial to the correct formation of peripheral sense organs, named sensory rays, within the nematode male tail. Although each ray develops from an identical ray cell sublineage, each ray forms at a distinct site, has a distinct morphology and mediates specific behavioral responses during mating. Additionally, each ray must exhibit a specific cell recognition code, allowing it to assemble into separate organs. Mutations in several genes cause alteration in this combinatorial code, leading to anomalous fusions between different rays. HOM-C/Hox genes play a primary role in these processes, and mab-21 mutations act as dominant modifiers of their effects (2). In particular, mab-21 interacts genetically with an Abd-b homolog, egl-5 (2), and possibly with a Pax6 homolog, namely mab-18 (3).
Homozygous recessive mutations of mab-21 were associated with short-range homeotic transformations, resulting in three cells descended from the sensory ray 6 precursor cell adopting fates of anterior homologs (ray 4), and a fourth, lineally unrelated hypodermal cell transforming into a neuroblast. mab-21 mutations have additional pleiotropic effects affecting body shape, fecundity and motor coordination, suggesting that mab-21 plays developmental roles outside the tail region of male nematodes. The alleles leading to those changes were hypomorphic ones, whilst null alleles resulted in early embryonic lethality.
Recently published work has provided a link between mab-21 and the TGF-[beta] pathway initiated by the cet-1 ligand, and transduced by a receptor heterotetramer containing SMA-6 and DAF-4, and by activator SMAD proteins SMA-2, SMA-3 and SMA-4 (4). cet-1 is virtually identical to another TGF-[beta] ligand gene cloned independently and dubbed dbl-1 (5). The two ligands are homologous to TGF-[beta] proteins of the decapentaplegic (DPP)/bone morphogenetic protein (BMP) class (for review, see ref. 6). Epistatic analysis indicated that the above pathway antagonizes mab-21 function, with mab-21 acting downstream of it, whereas expression data suggested that TGF-[beta] interacts with mab-21 at the post-transcriptional level (4).
A human homolog of mab-21 exists (7), and has been partially characterized, especially with regard to its putative role in human neurologic and psychiatric diseases. The human MAB21L1 gene was cloned from a retinal cDNA library in a search for transcripts containing CAG triplet repeats (7). In MAB21L1, a highly polymorphic CAG trinucleotide repeat was found in the 5[prime] untranslated region (5[prime]-UTR), displaying allele variation in normal individuals. Potter (8) analyzed 928 chromosomes from controls and patients with a variety of neurologic diseases. The analysis of family pedigrees established that short repeats were transmitted in a Mendelian fashion, whereas longer repeats (n [ge] 45) displayed meiotic instability. An additional study (9) confirmed that longer repeats are meiotically unstable and tend to expand, but failed to associate the expanded allele with any phenotypic abnormalities. These authors hypothesized that the CAG repeat in MAB21L1 may behave as a premutation, and that longer expansions may be associated with a clinical phenotype (9).
In a search for novel regulatory genes involved in embryonic brain development, our group has isolated the first murine homolog of mab-21. The gene, initially labeled Mab21, has been renamed Mab21l1, in keeping with the approved human nomenclature. In a recent expression study (10), we described the distribution of the corresponding transcript during embryogenesis. Mab21l1 is expressed starting early in development at several neural and extraneural sites. More recently, we have cloned a second member of this family (Mab21l2) in the murine and human genomes.
The present study describes the expanding mab-21 gene family. In particular, we present the nucleotide and peptide sequences of murine Mab21l1 and Mab21l2, and of human MAB21L2, and describe the phylogenetic relationships among the new and known family members, as well as their map localizations.Furthermore, we describe the comparative distribution profiles of the two murine transcripts, and analyze the subcellular distribution of the corresponding proteins.
RESULTS
Cloning and sequence analysis of mab-21 homologs in human and mouse
In recent years, to isolate genes expressed within restricted domains along the antero-posterior axis of the embryonic CNS, we devised a PCR-based differential screening protocol (11) subsequently utilized by our group and by others (12-16) in various developmental and differentiation studies. In that study, we compared RNAs extracted from various segments along the AP axis of the mouse brain at embryonic day 13 (E13). Among others, we isolated and cloned one band (band 270) corresponding to a transcript mostly restricted to the basal forebrain, midbrain and hindbrain. Sequencing and database analysis identified clone 270 as the mouse homolog of C.elegans mab-21 (1; GenBank accession no. U19861). An additional human homolog, cloned independently (7), was found in the database, containing a triplet repeat in its 5[prime]-UTR (CAGR1, subsequently renamed MAB21L1) (GenBank accession no. U38810). This gene was mapped by others to chromosome 13q13 (7). We used clone 270 to screen an E11.5 whole embryo cDNA library (Stratagene), and isolated three phage clones spanning a 1.8 kb contig, containing a 1.08 kb coding sequence preceded by in-frame stop codons. Whilst we provisionally dubbed the mouse gene Mab21 (10), the same gene was eventually renamed Mab21l1 by the Mouse Nomenclature Committee based on homology to the human 13q13 gene (7).
Later, through the analysis of the mouse EST database, we gathered evidence of a second mouse gene, that we cloned from the same E11.5 whole embryo cDNA library, using an EST sequence (GenBank accession no. W64940) as a probe, and named Mab21l2 (GenBank accession no. AF149877). Moreover, in the human EST database, we found one cDNA sequence (GenBank accession no. AA679158) homologous to the 5[prime] region of Mab21l2. By PCR-screening with two primers designed from this sequence, we isolated four P1 artificial chromosome (PAC) clones (860J22, 869K18, 869L18 and 1173I17) containing the human ortholog of mouse Mab21l2, provisionally named MAB21L2. We directly sequenced one of these genomic PAC clones (860J22) with walking primers derived from the human EST clone (GenBank accession no. AA679158). As a result, we assembled a 2050 bp genomic contig spanning the entire coding region (1.08 kb) and 865 bp upstream of it, containing several in-frame stop codons. Differently from MAB21L1, the 5[prime]-UTR of MAB21L2 does not contain a CAG triplet repeat.
In this study, we have not characterized the promoter regions of vertebrate mab-21 genes. The only genomic sequence analyzed by our group (human MAB21L2) does not exhibit significant features of any recognizable promoter elements.
Both human genes (MAB21L1 and MAB21L2) as well as their murine counterparts present putative coding regions of 1077 nucleotides preceded by in-frame stop codons. The putative initiation codon is located in an acceptable context for initiation of transcription (17), with a dATP at position -3, preceded by several dCTPs, and a purine (A) at position +4. We transcribed and translated in vitro both murine cDNAs and obtained single protein products for each cDNA, whose sizes were compatible with the sizes of the deduced peptide sequences (41 kDa). Likewise, in western analysis performed in adult neural tissues, a single 41 kDa band is detected by an anti-peptide antibody raised against the 14 amino acid C-terminal peptide of MAB21L1. This antibody crossreacts with in vitro translated MAB21L2 (data not shown).
Given the possibility that additional mab-21 family members exist in the murine and human genomes, we searched human and murine EST databases using the BLASTN and TBLASTN programs. This search failed to identify any additional mab-21 homologs in the corresponding genomes. Albeit suggestive, these results cannot be taken as formal, conclusive evidence that only two mammalian mab-21 paralogs exist.
In summary, the expanding mab-21 familyincludes at least two mouse genes and two human genes (Fig. 1). The phylogenetic tree constructed with the four mammalian members of the family and with C.elegans mab-21, as well as the pairwise comparisons shown in Table 1 reveal that Mab21l1 and Mab21l2, probably ensuing from an ancestral duplication, are roughly equidistant from their C.elegans homolog.
Figure 1. Sequence comparison of MAB-21 proteins from different species. (Top) Murine MAB21 proteins and their alignment to the human and C.elegans proteins. Identical residues are boxed. Numbered boxes represent putative phosphorylation sites identified through the PROSITE database analysis programs: 1, putative tyrosine phosphorylation site; 2 and 3, putative serine phosphorylation sites. See text for details. (Bottom) Phylogenetic distances among members of the MAB21 family. The alignment and evolutionary tree were obtained using the GCG programs Pileup and Distances.
Table 1. Similarities among MAB-21 proteins identified in different species
C, C.elegans; m, mouse; h, human. Each box contains percent degree of similarity (upper) and percent degree of identity (lower, italics). Pairwise comparisons were done using the GCG Bestfit computer program.
Finally, by analyzing the Drosophila genome EST database, we identified five partial cDNA sequences assembling into two distinct clusters. Cluster 1 contains three cDNAs (GenBank accession nos AI134107, AA941125 and AI531549) while cluster 2 contains two cDNAs (GenBank accession nos AI533533 and AI532715). These contigs identify two new fruitfly members of the mab-21 family. Their conceptual translations yield two distinct peptides (93 and 41 residues, respectively) clearly homologous to the N-terminus of known MAB-21 proteins. Albeit incomplete, these cDNA sequences suggest that two, possibly redundant, mab-21-like paralogs are present in the Drosophila genome, providing a tentative explanation of why this developmental gene was not identified in the course of various mutagenesis screens conducted in the fly (18).
Protein sequence analysis
Both human and mouse genes encode 359 amino acid proteins, with predicted molecular weights of 41 kDa. The conceptually translated proteins are mostly hydrophilic and have predicted isoelectric points of 9.45. A search for similarities to known functional domains found no hits, except for several putative serine phosphorylation sites. Two of these are found perfectly conserved in the C.elegans protein (Fig. 1). Residues 106-111 represent a consensus sequence for cAMP-cGMP-dependent protein kinase (box 2 in Fig. 1). A protein kinase C target site is possibly represented by residues 124-126 (box 3 in Fig. 1). Furthermore, we identified a putative tyrosine phosphorylation site, again conserved from man to nematode, that fits the consensus binding sequence for the SH2 domain of SEM-5 (residues 79-81) (19,20) (box 1 in Fig. 1).
Mab21l1 and Mab21l2 map to mouse chromosome 3
By haplotype and linkage analysis of the BSS backcross DNA panel generated and maintained at the Jackson Laboratory (Bar Harbor, ME) (21), we mapped the two mouse genes to murine chromosome 3 (Fig. 2).Although syntenic, the two genes are not clustered, being separated by an 11 cM genetic interval, with Mab21l1 centromeric to Mab21l2.
Figure 2. Mab21l2 and Mab21l1 mapping in mouse. (A) Haplotype and linkage analysis of Mab21l1 and Mab21l2 and flanking loci on mouse chromosome 3 through the analysis of the BSS backcross (Jackson Laboratory). Empty squares, Mus spretus allele; solid squares, C57BL/6J allele; shaded squares, genotype not determined. Numbers to the right, between rows, indicate recombination fractions ± SEM, and LOD scores. Columns represent different haplotypes observed on chromosome 3. Numbers below columns define the numbers of individuals sharing each haplotype. (B) Position of Mab21l1 and Mab21l2 on chromosome 3 with respect to nearby markers independently mapped by others on the BSS backcross. Numbers on the left represent approximate genetic distances (cM) from the most centromeric chromosome 3 marker in this cross. Chromosomal locations of the human homologs, MAB21L1 and MAB21L2 are also indicated (13q13 and 4931, respectively).
Mapping of Mab21l1 and Mab21l2 was made possible by the existence of TaqI and MspI polymorphisms, respectively, between the parental DNAs (C57BL/6JEi and SPRET/Ei). Such polymorphisms allowed BS heterozygotes to be distinguished from SS homozygotes among the N2 progeny of the cross. With the probes used in these experiments, the polymorphism consisted of a single SPRET/Ei-specific TaqI or MspI band, and of two C57BL/6JEi-specific TaqI or MspI fragments. The human homolog of Mab21l1 has been mapped to 13q13 (7) (confirmed in the present study, see below). Our data provide the first evidence of synteny between human 13q and murine chromosome 3.
MAB21L2 maps to human chromosome 4q31
By PCR-screening of the CEPH YAC library and by FISH, we obtained the physical map of MAB21L2 to human chromosome 4q31, a chromosomal region of known synteny to mouse chromosome 3. We isolated four human 4q31 YACs (931B5, 921C3, 769H11 and 970F2) belonging to the WC4.6 contig positive for MAB21L2. Since YACs 710G2 and 799F1, belonging to the same contig, are negative for the gene, we conclude that the physical domain containing MAB21L2 is bracketed by STS markers D4S1606 and RPS3A1.
Because of the apparent discrepancy between the mapping position of murine and human Mab21l1, while reproducing the mapping assignment in mouse, we decided to confirm the mapping of human MAB21L1, previously assigned by others to 13q13 through linkage in CEPH families and radiation hybrid mapping (7). Several human 13q13 YACs (748EG, 881F2, 927A3, 955F1, 802G8 and 903C7) belonging to the WC13.2 contig were positive for MAB21L1, while YACs 772B4 and 931A6 were negative. Therefore, the physical domain containing MAB21L1 is delimited by STS markers WI5283 and IB237, from the WC13.2 contig. The map assignments for both human genes were confirmed by cytogenetic means: fluorescence in situ hybridization (FISH) done using the 860J22 PAC clone (MAB21L2-positive) and the 802G8 and 748E4 YACs (MAB21L1-positive) as DNA probes, confirmed their localization to 4q31 and 13q13, respectively (Fig. 3).
Figure 3. Cytogenetic localization of MAB21L1 and MAB21L2 in the human genome. FISH with YAC clone no. 802G8 for MAB21L1 and PAC clone no. 860J22 for MAB21L2. Arrows indicate the chromosomal positions of the two genes: 13q13 and 4q31, for MAB21L1 and MAB21L2, respectively.
Mab21l2 is expressed in the CNS and neural crest in midgestation embryogenesis
An initial analysis of Mab21l1 expression has been described previously (10). In the present paper, we characterize the distribution of the Mab21l2 transcript in depth. Three previously published images (Figs 4A and C, and Fig. 5A) of Mab21l1 transcript distribution in embryogenesis are re-used here, with permission, for cross-reference (10). We assayed the expression of Mab21l2 in mouse development by various means. By whole-mount in situ hybridization (Fig. 4) of E9.5 embryos, the Mab21l2 transcript (A[prime]) was clearly visible in the dorsal midbrain, even in the posterior portion of the mesencephalon, where Mab21l1 is expressed at lower levels (A). Mab21l2 signal was also detected in the retinal primordium of the optic cup. Neither Mab21l1 nor Mab21l2 signal could be detected in the alar telencephalon, at this or later stages.
Figure 4. Comparison of Mab21l1 and Mab21l2 expression in the mouse embryo. Whole-mounts, whole-mount in situ hybridization of mouse embryos at 9.5 (A and A[prime]), 10 (B and B[prime]), and 10.5 days post-coitum (C and C[prime]), with the Mab21l1 (A-C) and Mab21l2 (A[prime]-C[prime]) probes. bra, branchial arches; hl, hindlimb bud; fl, forelimb bud; in, intestine; mb, midbrain; or, optic recess. The arrowhead indicates retinal expression domain; the arrow points to the posterior mesencephalon, which expresses Mab21l2 and not Mab21l1. (A) and (C) are reprinted from Mechanisms of Development, 79, Mariani, M. et al., Mab21, the mouse homolog of a C.elegans homeotic regulator, participates in cerebellar, midbrain and eye development, pp. 131-135, © 1998, with permission from Elsevier Science (10).
Figure 5. Comparison of Mab21l1 and Mab21l2 expression in the mouse embryo. Tissue sections, mRNA in situ hybridization of embryonic tissue sections at days 10.5 (A and A[prime]), 12.5 (B, B[prime], D, D[prime], F and F[prime]) and 14.5 days post-coitum (C, C[prime], E and E[prime]) with Mab21l1 and Mab21l2 probes. All sections are sagittal, rostral is on the left, except (E and E[prime]), coronal sections of the right eye, and (F and F[prime]), coronal sections of the caudal spinal cord. Arrows, posterior portion of the mesencephalon; arrowheads, alar mesencephalon; bra, branchial arches; cb, cerebellum; fl, forelimb bud; gr, genital ridge; in, intestine; l, lens; max, maxilla; mb, midbrain; r, retina; rh, rhomboencephalon; sc, spinal cord; t, telencephalon. (A) is reprinted from Mechanisms of Development, 79, Mariani, M. et al., Mab21, the mouse homolog of a C.elegans homeotic regulator, participates in cerebellar, midbrain and eye development, pp. 131-135, © 1998, with permission from Elsevier Science (10).
At E10 (Fig. 4B and B[prime]), the two genes are expressed to different temporo-nasal extents in the embryonic retina and, correspondingly, to different anteroposterior extents in the mesencephalon. Mab21l1 (Fig. 4B) covers the temporal aspect of the retina and the anterior two thirds of the alar midbrain, whereas Mab21l2 (B[prime]) is expressed in the temporal and nasal aspect of the retina, and throughout the length of the dorsal midbrain.
At E10.5 (Fig. 4C and C[prime]) Mab21l1 and Mab21l2 transcripts could be revealed in the same territories expressing the two genes earlier on. In the eye, Mab21l1 expression keeps marking the temporal aspect of the retina, while a new expression domain appears in the lens. Extraneural sites of Mab21l2 expression include the branchial region, limb buds and intestine. In the limb buds, Mab21l2 (Fig. 4C[prime])marks the apical region that does not express Mab21l1 (C).
To compare Mab21l2 to Mab21l1 transcript distribution in detail, we performed radioactive in situ hybridizations of mouse tissue sections at developmental stages ranging from E10.5 to E14.5 (Fig. 5). Strong signal was seen at all sites observed by whole-mount hybridization. High expression levels could be observed in the alar mesencephalon (arrowheads in Fig. 5A-C and A[prime]-C[prime]) and in the rhombencephalic basal plate. At the lumbar level, Mab21l2, as well as Mab21l1, is expressed in the spinal cord. Outside the brain, the optic cup and presumptive lens also express Mab21l1 and Mab21l2 at high levels (Fig. 5E and E[prime]). Peripherally, signal was observed for both genes in the olfactory epithelium (data not shown). Other expression sites included the four branchial arches (that contain neural crest derivatives) and the limbs.
The main differences between the two expression profiles consisted of a strong Mab21l2 signal in the posterior third of the tectum where the Mab21l1 transcript is less abundant (arrows in Fig. 5A[prime]-D[prime]), and in the presumptive gut wall (Fig. 5A[prime] and B[prime]) where Mab21l1 is not expressed. The intestinal wall contains post-synaptic parasympathetic ganglia of neural crest origin. On the other hand, Mab21l2 expression could not be detected in the genital ridge, where Mab21l1 mRNA is present (Fig. 5A).
Mab21l1 and Mab21l2 remain expressed in the adult cerebellum and eye
The Mab21l1 and Mab21l2 probes used for northern analysis were obtained from the 5[prime]-UTR regions of both cDNAs in order to maximize specificity. Furthermore, washes were conducted at high stringency conditions (final wash: 15 min at 0.1× SSC, 0.1% SDS, 65°C). In such experimental conditions, no cross-hybridization was observed between Mab21l1 and Mab21l2. Northern analysis performed on adult mouse tissues (Fig. 6) revealed that none of 13 extraneural adult tissues examined (thymus, lung, heart, liver, spleen, intestine, kidney, adrenals, ovary, uterus, testis, prostate and muscle) expresses detectable levels of either Mab21l1 or Mab21l2 (data not shown). However, the two genes are expressed in the adult cerebellum and eye, with lower levels in the adult forebrain. Based on the amounts of actin transcript, used for normalization, it seems clear that both genes are expressed at the highest levels in the adult eye. Comparatively, Mab21l1 is expressed more abundantly in the cerebellum than Mab21l2.
Figure 6. Mab21l1 and Mab21l2 remain expressed in the adult CNS and retina. Northern blot analysis of Mab21l1 and Mab21l2 transcript levels (A) conducted on total RNAs extracted from adult mouse forebrain, cerebellum and eyes. Actin mRNA levels (B) are shown for normalization. See text for discussion of banding patterns.
In the three adult tissues that express the two genes, complex banding patterns could be observed. The Mab21l1-specific probe detects a doublet in RNA extracted from cerebellum and eye, consisting of a 3.0 kb band and a 2.85 kb band. Only the upper band of the doublet is found in forebrain RNA. The Mab21l2-specific probe detects a complex pattern in northern analysis, consisting of at least three bands in eye and cerebellum RNA (3.25, 2.7 and 2.2 kb), and of one band in forebrain RNA (2.2 kb). The top band (3.25 kb) is more prominent in RNA extracted from the eye.
Such banding patterns might reflect tissue-specific alternative splicing. Although one cannot rule out cross-hybridization with other, closely related, as yet undiscovered transcripts, this hypothesis seems less likely considering the high stringency conditions in which the experiment was carried out. Further studies will be required in order to clarify this point.
Subcellular localization
We conducted indirect cellular immunofluorescence studies to determine the subcellular localization of mammalian MAB-21 proteins. We transfected NIH 3T3 embryonic fibroblasts with expression constructs encoding MAB21L1 and MAB21L2 proteins tagged at their N-termini by the addition of a myc epitope (Fig. 7). After 48 h in culture, we fixed the cells and analyzed them by cellular immunofluorescence using a monoclonal antibody raised against the myc epitope. The experiment revealed that the epitope-tagged proteins localize clearly in the nucleus although, at higher concentrations, significant signal for both proteins can be observed in the cytoplasm.
Figure 7. MAB-21 proteins localize mainly in the nucleus. Cellular immunofluorescence analysis of NIH 3T3 cells transfected with Mab21l1 (A) and (B) and Mab21l2 (C) and (D) cDNA constructs encoding N-terminally myc-tagged mouse MAB-21 proteins. The transfected cell (B) displays dual localization to the nucleus and cytoplasm. Similar distributions were observed in cells expressing high levels of the MAB21L2 protein. Fixed cells were immunostained with a monoclonal anti c-myc antibody. Cells shown in the figure are representative of >100 observations.
DISCUSSION
The present paper introduces and describes three new mammalian (one human and two murine) members of the expanding gene family homologous to mab-21, a nematode gene involved in cell fate determination events required for the formation of sensory structures in the male nematode tail. Mapping experiments indicate that Mab21l1 and Mab21l2 are not clustered either in the human or in the mouse genome, and that the ancestral duplication originating the second paralog led to the generation of fully independent transcribed and cis-acting regulatory sequences.
In addition to cloning two full coding mouse cDNAs and a complete coding region in the human genome, our group, by database analysis, has also identified two potential mab-21 homologs in the Drosophila genome, that are clearly distinct from each other. This preliminary finding suggests that the developmental function encoded by mab-21 may be overspecified in Drosophila, which in turn may explain why no mab-21 mutants have been identified in the numerous mutagenesis screens conducted in that organism.
In the mouse embryonic neuraxis, both mab-21 genes are expressed in the dorsal half of the spinal cord, in a domain reminiscent of Pax3 transcript distribution (22). In vertebrates, this region contains mostly prospective sensory neurons and other neural crest precursor cells. Caenorhabditis elegans mab-21 is required for the proper patterning of sensory rays in the male tail, a process requiring dynamic changes in cell adhesion, cell fate determination and migration properties. In vertebrates, neural crest cells go through analogous changes in shape, motility and adhesive features while initiating and carrying out their migration.
In the nematode, mab-21 engages in genetic interactions with the Abd-B homolog egl-5. Vertebrate Hox genes homologous to egl-5 include paralogs 9-13, specifying the posterior trunk and prospective lumbo-sacral segments of the axial skeleton, as well as the limbs (23,24). Besides being expressed in the spinal cord, Mab21l1 is expressed in two parallel proximodistal strips along the limb axis, whereas Mab21l2 is found in a broader domain including the apical region of the limb bud, that expresses several Hox genes (24).
Furthermore, anteriorly, mab-21 gene expression peaks at the level of the rhomboencephalon and midbrain. In the hindbrain, the mab-21 transcripts are distributed early (E12.5) in the rhombic lip, a region from which granule cell precursors migrate tangentially to form the external germinal layer of the developing cerebellum (25,26). In this mitotic territory, granule cells keep proliferating, differentiate further, and enter their radial, centripetal migration through the molecular layer to form the internal granular layer in the mature cerebellum, where expression of Mab21l1 and Mab21l2 is maintained through adulthood.
In the midbrain, mab-21 genes are transcribed throughout the thickness of the ventricular wall, and at higher levels in the mantle layer. Mab21l1 is expressed mostly in the anterior two thirds of the tectum, whereas Mab21l2 covers the entire length of the dorsal midbrain. The temporal half of the retinal ganglionic layer projects retinotectal fibers to the anterior half of the tectum, whereas the nasal side of the retina projects fibers to the posterior half of the tectum (for review, see ref. 27). Thus, the territories expressing Mab21l1 in the retina correspond to those expressing it in the tectum. Likewise, Mab21l2, being expressed further ventrally in the retina, covers a greater extent of the tectum along the anteroposterior axis. It is therefore tempting to speculate that mab-21 genes and their combinatorial code may play a role in the proper retinotectal migration of ganglion cell axons.
mab-21 genes are also expressed in the forebrain. In the diencephalon, the main site of expression observed in early embryogenesis corresponds to the optic recess/cup which gives rise to the prospective retina and lens.
In regard to expression sites observed outside the central nervous system, those in the branchial arches and dorsal root ganglia are consistent with a possible involvement of vertebrate mab-21 in the specification/differentiation of neural crest derivatives. The striking expression levels of Mab21l2 in the embryonic gut may reflect a role in the maturation of post-synaptic neurons in parasympathetic ganglia, that derive from the cephalic and caudal crest. Extraneural signal for Mab21l1 is also present in the genital ridge, a territory specified by gradients of homeobox gene expression (28).
Both mammalian mab-21 homologs described in this study encode 41 kDa proteins containing putative protein kinase sites, to suggest that MAB-21 proteins may be part of a signaling cascade, and their function may be modulated post-translationally, via serine/threonine or tyrosine phosphorylation. Recent data obtained in the nematode by epistatic analysis suggest that MAB-21 function is antagonized by TGF-[beta] signaling at the post-transcriptional level (4). Thus, the TGF-[beta] cascade may inactivate MAB-21 proteins by phosphorylating them, or by regulating molecular interactions of MAB-21 with other intracellular members of the pathway.
Our data suggest that mammalian MAB-21 proteins are localized predominantly, albeit not exclusively, in the nucleus. Indeed, the analysis of transfected cells expressing high levels of MAB-21 proteins clearly reveals that a quota of the protein localizes in the cytoplasm. This may be due either to saturation of the nuclear localization machinery because of the high intracellular protein levels, or to a regulated entry in the nucleus. The latter, in turn, may be linked to specific stages in the cell cycle, or modulated by signaling pathways, possibly through protein-protein interactions or post-translational modifications.
Although mosaic analysis in C.elegans provides genetic evidence of a non-cell-autonomous activity of the mab-21 gene (1), we have found no evidence to support the notion that the corresponding proteins, devoid of a canonical leader peptide, are secreted by the cell. Thus, their likely role is in the context of pathways transducing signals from the membrane to the nucleus, or vice versa.
In summary, the expanding mab-21 family contains two murine and two human homologs, as well as likely homologs in Drosophila melanogaster. The gene family exhibits striking phylogenetic conservation from man to nematode. Genetic and functional analysis in vivo and in vitro is underway to determine the role of these factors, and their mode of interaction with regulatory pathways active in CNS development.
MATERIALS AND METHODS
General procedures
Standard molecular techniques, including nucleic acid purification, restriction analysis, gel electrophoresis, DNA ligation, cloning, subcloning, dideoxy-sequencing, probe radiolabeling, northern and Southern analysis, and library screening were performed according to established protocols (29) Automated sequencing with Dyedeoxy primers or Dyedeoxy terminators was performed on an ABI 373 machine. Hybridizations of northern and Southern blots (Pall) were performed at 65°C in 125 mM sodium phosphate (pH 7.2), 250 mM NaCl, 7% SDS, 10% PEG. Filters were washed at 65°C to final stringencies of 0.2× SSC for 10 min. Phage plaque hybridizations and subsequent washes were carried out under comparable stringency conditions.
Experimentation on animal tissue conformed to the stipulations of the Institutional Animal Care and Use Committee of the San Raffaele Scientific Institute, Milan, Italy.
Tissue preparation and RNA extraction
Preparations of embryonic CNS were done as follows: under a dissection microscope, brain tissue was separated from surrounding mesoderm and non-neural ectoderm. Embryonic hindbrain, midbrain and forebrain territories from two CD1 litters were separated and pooled. Fresh tissue preparations were lysed in guanidine isothiocyanate. RNA extraction was carried out on a cesium chloride gradient (29).
RNA fingerprinting
Clone 270 was derived through a modification (11) of the RNA fingerprinting protocol (30) from the RNAs extracted as detailed above. RNA fingerprinting was conducted as follows: a reverse transcription reaction was carried out using a (dT)16 primer on total RNAs extracted by the cesium chloride method (29) and digested with 4 IU DNase I/µg total RNA. Radioactive PCR reactions, in duplicate, were performed from 1 µl of each RT reaction in 50 µl final volume with an arbitrary 12mer (DR34 sequence: 5[prime]-GAC GAG GCT GGA) (final concentration 4 µM). PCR conditions were 3 min at 94°C, 2 min at 80°C at which Taq polymerase was added (hot start), followed by 35 cycles of 1 min at 94°C, 1 min at 50°C, 30 s at 72°C, with a final elongation step of 5 min at 72°C. [[alpha]-32P]dCTP (0.1 µl) was added to each reaction. Amplified products were separated on a 5% denaturing acrylamide gel and visualized by autoradiography. Differentially displayed bands were cut from the gel and electro-eluted. The bands were reamplified using the same 12mer primers and blunt-end cloned into pBluescript II SK+ (Stratagene, La Jolla, CA) as described (31).
Sequence analysis
Data bank searches (GenBank, GenEMBL, SwissProt and PIR) were run through the BLAST server (32). Additional sequence analysis and contig assembly was done using the MacVector program (Oxford Molecular, Oxford, UK) and the Sequencher program (Gene Code, Ann Arbor, MI), respectively.
Genetic mapping
Genetic mapping was performed on 96 DNAs, corresponding to the parentals and 94 N2 progeny of a (C57BL/6j×SPRET/Ei)F1×SPRET/Ei (BSS) backcross generated and distributed by the Jackson Laboratory (Bar Harbor, ME) (21). A TaqI RFLP was identified, as described above. Its segregation was followed, and linkage analysis performed, with the MapManager v.2.6 program (Jackson Laboratory)(33).
FISH mapping
Chromosome preparations were hybridized in situ with probes labeled with biotin nick translation, essentially as described (34), with minor modifications. Briefly, 500 ng of labeled probe were used for the FISH experiments; hybridization was performed at 37°C in 2× SSC, 50% (v/v) formamide, 10% (w/v) dextran sulfate, 5 µg COT1 DNA (Boehringer Mannheim) and 3 µg sonicated salmon sperm DNA, in a volume of 10 µl. Post-hybridization washing was three times at 42°C in 2× SSC-50% formamide followed by three washes in 0.1× SSC at 60°C. Biotin-labeled DNA was detected with Cy3-conjugated avidin (Amersham, Uppsala, Sweden). Chromosome identification was obtained by simultaneous DAPI staining that produces a Q-banding pattern. Digital images were obtained using a Leica DMRXA epifluorescence microscope equipped with a cooled CCD camera (Princeton Instruments, Princeton, NJ). Cy3 and DAPI fluorescence signals, detected using specific filters, were recorded separately as grey scale images. Pseudocoloring and merging of images were performed using the Adobe Photoshop software.
In situ hybridization of mouse tissue sections
Radioactive in situ hybridization (ISH) (35) was carried out as follows: 7 µm paraffin serial sections from a single embryo were displaced in four to six adjacent series: two alternative series were used for each probe. Embryos were studied at 10.5, 12.5 and 14.5 days of embryonic development. Slides were deparaffinated in xylene, rehydrated through an alcohol series, treated with 4% paraformaldehyde and proteinase K, acetylated and dehydrated through an ethanol series. One microliter (3 × 106 cpm) of Mab21l2 riboprobe, labeled with [35S]UTP (Amersham) in the hybridization mix was added to each slide. Both sense and antisense probes were used but only antisense probes produced a detectable signal (data not shown). Clear differences between expression domains revealed by the Mab21l1 and Mab21l2 probes indicate complete specificity of each probe. Hybridization was carried out overnight at 60°C. Slides were washed under stringent conditions (60°C, 2× SSC, 50% formamide) and treated with RNase A. Autoradiography was performed with Kodak NT/B2 emulsion. Exposure times were 15 days. Sections were examined and photographed on dark and bright fields using a ZEISS SV11 microscope.
Whole-mount in situ hybridization
E9.5 and E10.5 embryos were isolated and immediately fixed with 4% paraformaldehyde, dehydrated, rehydrated with a methanol series, bleached with 6% hydrogen peroxide and then treated with proteinase K. One microgram of each linearized plasmid was transcribed in vitro in the presence of 0.8 µl 10 mM Dig-11-UTP (Boehringer Mannheim). Both sense and antisense probes were used. Embryos were hybridized overnight at 65°C and washed at the same temperature with 2× SSC, 50% formamide and treated with RNase A. Signal was revealed through an alkaline phosphatase-conjugated anti-digoxigenin antibody (Boehringer Mannheim) with BM-purple (Boehringer Mannheim) as substrate. Embryos were examined and photographed on bright fields using a ZEISS SV11 microscope.
Cell transfection. Constructs encoding the MAB21L1 or MAB21L2 proteins tagged at the N-terminus with a c-myc-derived epitope (36) obtained by overlap-extension PCR (37) were subcloned into pcDNA3 (Invitrogen, Carlsbad, CA) and used to transfect NIH 3T3 cells by a standard calcium phosphate precipitation protocol (29). Two micrograms of DNA were transfected in each 60 mm dish cell culture. Cells were harvested 24 h after transfection and grown on coverslips for an additional 24 h before the immunofluorescence assay. Transfection efficiency was ~15%.
Cellular immunofluorescence
Cells (105) grown on glass coverslips were fixed with 3% paraformaldehyde and permeabilized with 0.15% Triton X-100 in PBS for 5 min, and immunostained with the 9E10 anti-myc mAb followed by FITC- or TRITC-conjugated F(ab[prime])2 fragments of goat anti-mouse IgG (Jackson Immunoresearch Laboratories). Transfected cells were photographed using an Olympus AX70 microscope.
ACKNOWLEDGMENTS
We thank Anna Corradi for Mab21l1 probes. We are grateful to Laura Croci and Alessandro Bosetti for their help in this project. We are grateful to Lucy Rowe and Mary Barter (Jackson Laboratory) for sharing DNAs from the BSS interspecific backcross, and to the YAC Screening Center (San Raffaele Science Park) for genomic megaclones. This project and the work of D.B. is mainly supported by Italian Telethon grant E740 to N.M. In addition, we gratefully acknowledge the financial contribution of EU Biomed grant BMH4-CT96-0777 and of the Armenise Harvard Foundation to the laboratory of G.G.C. M.M. is supported by a fellowship from Fondazione Rita Levi Montalcini, Rome.
REFERENCES
+To whom correspondence should be addressed at: II A1, Room 7, San Raffaele Scientific Institute, Via Olgettina 58, I-20132 Milan, Italy. Tel: +39 02 26434838; Fax: +39 02 26434855; Email: g.consalez{at}hsr.it
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A S Brooks, P A Leegwater, G M Burzynski, P J Willems, B de Graaf, I van Langen, P Heutink, B A Oostra, R M W Hofstra, and A M Bertoli-Avella A novel susceptibility locus for Hirschsprung's disease maps to 4q31.3-q32.3. J. Med. Genet., July 1, 2006; 43(7): e35 - e35. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Yamada, Y. Mizutani-Koseki, T. Hasegawa, N. Osumi, H. Koseki, and N. Takahashi Cell-autonomous involvement of Mab21l1 is essential for lens placode development Development, May 1, 2003; 130(9): 1759 - 1770. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Michaut, S. Flister, M. Neeb, K. P. White, U. Certa, and W. J. Gehring Analysis of the eye developmental pathway in Drosophila using DNA microarrays PNAS, April 1, 2003; 100(7): 4024 - 4029. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||