Online ISSN: 1097-0177 Print ISSN: 1058-8388
Developmental Dynamics
Volume 218, Issue 1, 2000.
Pages: 160-174
Copyright © 2000 Wiley-Liss, Inc.
| Online ISSN: 1097-0177 Print ISSN: 1058-8388 Copyright © 2000 Wiley-Liss, Inc. |
| Identification of a novel mouse Iroquois homeobox gene, Irx5, and chromosomal localisation of all members of the mouse Iroquois gene family |
| Antje Bosse 1, Anastassia Stoykova 1, Kay Nieselt-Struwe 1, Kamal Chowdhury 1, Neal G. Copeland 2, Nancy A. Jenkins 2, Peter Gruss 1 * |
| 1Department of Molecular Cell Biology, Max Planck Institute of Biophysical Chemistry, Göttingen, Germany 2Mammalian Genetics Laboratory, ABL-Basic Research Program, NCI-Frederick Cancer Research and Development Center, Frederick, Maryland |
*Correspondence to Peter Gruss, Department of Molecular Cell Biology, Max Planck Institute of Biophysical Chemistry, Am Fassberg, D-37077 Göttingen, Germany
Funded by:
Deutsche Forschungsgemeinschaft; Grant Number: SFB 271 European Union National Cancer Institute, DHHS
| Keywords |
| Iroquois; prepattern gene; homeobox; mouse; Xenopus; chick; human; neurogenesis; otic vesicle; branchial cleft; heart; limb; scerotome |
| Abstract |
 |  | | | | | |
| The Drosophila genes of the Iroquois-Complex encode homeodomain containing transcription factors that positively regulate the activity of certain proneural Achaete/Scute-C (AS-C) genes during the formation of external sensory organs (J. L. Gomez-Skarmeta and J. Modolell, EMBO J 17:181-190, [1996]). Previously, we have identified three highly-related genes of the mouse Iroquois gene family that exert specific expression patterns in the central nervous system (A. Bosse et al., Mech Dev 69:169-181, [1997]). In the present paper, we report the identification of a novel member of the Iroquois gene family, Irx5, that shows a restricted spatio/temporal expression during early mouse embryogenesis, distinct from the expression of Irx1-3. An extensive sequence analysis of 20 Iroquois-like genes from seven organisms reveals a high conservation of the homeodomain. Phylogenetic tree reconstruction showed a clustering of the members of the Iroquois gene family into groups of orthologous genes. Together, with the data obtained from the chromosomal mapping analysis, the results indicate that these genes have appeared in vertebrates during evolution as a result of gene duplication. Dev Dyn;218:160-174. © 2000 Wiley-Liss, Inc. |
| Digital Object Identifier (DOI) |
| Article Text |
 | | | | | | |
The development of the nervous system includes four phases: neural induction, regionalisation and patterning, proliferation and differentiation, survival and terminal differentiation.
In Drosophila, the differentiation of primary neurons within the neural ectoderm is thought to be controlled by a network of proteins, encoded by the proneural and neurogenic genes (for reviews, see Campuzano and Modolell, [1992]; Ghysen et al., [1993]). The proneural genes, achaete and scute, are expressed in highly-restricted patterns that prefigure the positions of sensory organs in the fly imaginal discs (reviewed in Campuzano and Modolell, [1992]). Neurogenic genes act as lateral inhibitors preventing the remaining cells of the clusters to adopt a neural fate (see Campos-Ortega and Jan, [1991]; Simpson, [1990]) for reviews. Key neurogenic genes are the Delta, Notch, Hairy, Suppressor of Hairless, Enhancer of split complex, Extramacrochaete, and Pannier (Botas et al., [1982]; Moscoso del Prado and Garcia-Bellido, [1984]; Muskavitch, [1994]; Parks et al., [1997]; Ramain et al., [1993]).
It is assumed that the highly localized expression of proneural genes is regulated by a combination of factors, which have been termed 
prepattern genes.
The Drosophila genes araucan, caupolican, and mirror are members of the Iroquois-complex (Iro-C), and are considered to be such prepattern genes (Gomez-Skarmeta and Modolell, [1996]). Araucan and caupolican encode highly related homeodomain proteins that govern the spatially-restricted expression of the proneural genes achaete and scute that in term determine the loci where neural precursors arise (Gomez-Skarmeta and Modolell, [1996]; Leyns et al., [1996]).
Vertebrate homologs of the fly proneural and neurogenic genes have been identified. Increasing evidence suggests that they act within evolutionarily conserved regulatory networks, determining specific neural cell fates as in the fly (Allende and Weinberg, [1994]; Ferreiro et al., [1994]; Guillemot, [1995]; for reviews, see Bellefroid et al., [1998]; Gomez-Skarmeta et al., [1998]; Kageyama et al., [1995]; Lee, [1997]). For instance, studies in Xenopus have implicated that most of the proneural genes such as Xash3 promote neurogenesis by conveying neuronal potential to the cells in which they are expressed (Ferreiro et al., [1994]; Ma et al., [1996]; Turner and Weintraub, [1994]).
The previously-identified Xenopus Iroquois-like genes, Xiro1-3 are expressed in the neural plate and later during neurulation, the expression is confined to the prospective rhombencephalon and spinal cord (Bellefroid et al., [1998]; Gomez-Skarmeta et al., [1998]). Evidence from experiments with ectopic expression of Xenopus Iroquois genes shows that they specify neural precursor cells and are indeed involved in the positive regulation of certain proneural genes such as Xash3 (Bellefroid et al., [1998]; Gomez-Skarmeta et al., [1998]). Thus, prepattern genes being involved in the differentiation of the neural plate are existing in vertebrates and ectopic experiments suggest the conservation of at least parts of the genetic pathway to regulate proneural genes from fruitfly down to vertebrates.
To study the extent to which molecular processes of neurogenesis have been conserved between fly and mammals, we have recently identified and characterized three mouse Iroquois-related genes, Irx1, -2, and -3. Their predicted protein sequences contain as highly-conserved homeodomains as those of the Drosophila Iro-C and Xenopus Iroquois-like genes (Bosse et al., [1997]). Based on their extensive sequence conservation, we considered these genes as a separate homeobox gene family. The comparative expression analysis of Irx1, -2, and -3 revealed that these genes have distinct spatio/temporal patterns during early mouse embryogenesis (Bosse et al., [1997]). Based on the strikingly high sequence conservation between the mouse Irx genes and their fly and vertebrate homologs and their similar expression profiles, we concluded that the neurogenic pathway where Iroquois proteins act as activators, seems to be conserved in such distantly related species, as the fruitfly and the mouse.
The mouse proneural gene, Mash1, exerts overlapping expression domains with the mouse Irx3 gene (Bosse et al., [1997]). Furthermore, we found that Irx3 is expressed earlier and in a broader domain in the developing nervous system as Mash1, implicating that Mash1 is a putative target gene of Irx3.
Upstream regulators for the Drosophila Iro-C genes such as cubitus interuptus (ci), that act together with dpp, have been reported (Gomez-Skarmeta and Modolell, [1996]). Ci is a zinc-finger domain containing gene also identified in C. elegans (tra1) and mouse (Gli1-3) (Hui et al., [1994]; Schimmang et al., [1992]; Vortkamp et al., [1991]). The mouse Gli genes exert expression profiles similar to the murine Irx genes. In addition, the phenotype of a naturally-occurring Gli3 mutation, the mouse extra-toes mutant, implies parallels to the Irx genes (Hui et al., [1994]). Thus, it was assumed that the Gli genes act upstream of the Iroquois genes within the genetic cascade (Bosse et al., [1997]).
In this paper, we describe the identification of a further member of the mouse Iroquois family designated as Irx5. Within the mouse Iroquois family, Irx5 is most closely-related to Irx2 with 100% amino acid identity within the homeodomain and with similarities in their expression patterns. The results from our chromosomal mapping experiments reveal that the four known murine Iroquois genes are clustered on two chromosomes, 8 and 13. The comparative sequence analysis of the murine Iroquois-related genes and the homologous genes in six other organisms suggest that these genes cluster into subgroups of orthologous genes.
| | | | | | |
Isolation of Irx5, a New Member of the Murine Iroquois Gene Family
When characterizing the genomic regions of Irx3, we identified two overlapping genomic clones that apparently correspond to a novel member of the mouse Irx gene family. The overlapping region of both genomic clones comprised the whole coding region. Both clones contained 5
and 3 from this overlapping region about 6kb and 2kb untranslated sequences, respectively. The genomic organization of exons and the position of the homeodomain was analysed (Fig. 1B). Subsequently, one part of clone 6 upstream of the homeobox was used to screen for cDNA-fragments (for details see Materials and Methods).
 | Figure 1. The nucleotide sequence and deduced amino acid sequence of Irx5, and the physical map of the corresponding genomic regions. (A) Complete nucleotide sequence of the mouse Irx5 gene and the deduced amino acid sequence of the Irx5 gene product are depicted. The sequence was assembled from the two overlapping cDNA fragments as described in the Materials and Methods section. The amino acid residues of the homeodomain and the Iro-Box are underlined, respectively. The double underline marks a putative polyadenylation signal starting at position 2104. The nucleotide and amino acid sequences of the Irx5 cDNA and deduced protein will be available from the National Center for Biotechnology Information (NCBI) under acc. no. AF230074. (B) Physical map of the corresponding genomic regions of Irx5. The structure of the overlapping Irx5 genomic clones is outlined. Coding regions are represented by the three boxes in black, containing the ATG, homeobox and stop codons, respectively. Restriction sites are: B, BamHI; Bs, BstXI; EI, EcoRI; N, NotI, SI, SacI; SII, SacII; X, Xhol. [ Normal View 162K | Magnified View 241K ] |
Sequence analysis confirmed that a further member of the mouse Iroquois homeobox family was cloned. Recently, a chick Iroquois4 (cIrx4) gene has been identified (Bao et al., [1999]). Because cIrx4 is not homologous to the newly-isolated mouse Irx gene, we named this gene Iroquois-homeobox-5, Irx5. Two overlapping cDNA clones comprised about 2.2 kb of the Irx5 sequence with an open reading frame coding for 484 amino acids (Fig. 1A). Four out of the six bases preceding the presumptive initiation codon are identical to the Kozak consensus sequence (Kozak, [1987]). The predicted amino acid sequence of the Irx5 homeodomain contains a nonpolar alanine residue at position nine of the recognition helix (position 53 of the homeodomain). This residue is found throughout all known Iroquois homeodomains (Fig. 2A).
 | Figure 2. (A) Alignment of the homeodomains and flanking regions of all Iroquois-type homeodomains described to date and four members of the TALE superfamily as well as Drosophila Antennapedia (Antp). The sequences shown include 45 amino- and carboxyterminal of the homeodomain to illustrate the extended homologies. Identical residues in the alignment are colored in black boxes, similar ones in grey boxes. Gaps for optimal alignment are displayed by dots. The regions of the four  -helices in the homeodomain are indicated at the top. The alanine residue at position nine of the recognition helix, which is important for the sequence specificity of the DNA binding of the homeodomain (Treisman et al., [1989]) is marked by an asterix. The alignment of the amino acid sequences was done using the algorithm CLUSTAL-X (Thompson et al., [1997]). (B) Amino acid identity of mouse Irx5 to the other species in the alignment of (A). The first column shows the values for the whole alignment region, the second for the homeodomain only. In the case of unknown sequence information (HIRX1, -3, -4, -5), the reduced sequence length was used. (C) Phylogenetic tree of the Iroquois family of the alignment from (A). The tree is rooted with the outgroup Antp, the branch lengths are proportional to the percent divergence between the nodes. A scale of 0.1 amino acids relative divergence is shown at the bottom. The numbers at the branches indicate bootstrap percent values (1,000 resampling steps).[ Normal View 177K | Magnified View 264K ] |
In addition to the recognition helix, Irx5 contains several conserved amino acids that are potential phosphorylation sites for mitogen-activated protein kinase (MAPK) and acidic residues downstream of the homeodomain. Such acidic regions are assumed to be involved in DNA-binding. Furthermore, the 12 amino acids comprising the Iro-Box motif in the C-terminal portion is also found in the Irx5 predicted protein sequence (Fig. 1A).
Sequence Comparisons with Iroquois-Related Genes from Different Species
To study the phylogenetic relationship of the Iroquois gene family, we performed a comprehensive database search. We analysed 20 Iroquois-like genes from seven organisms together with several representatives of the TALE superfamily and Antp as an outgroup gene. So far, there have been several Iroquois-like genes identified in a single species, such as three Drosophila genes, three Xenopus genes, three chick genes, four mouse, and five human genes.
Alignments of the deduced amino acid sequences of the homeodomains and flanking regions (up to 25 amino acids upstream of the homeodomain and 18 amino acids downstream) of all known Iroquois-like genes, show a highly conserved homeodomain, but they are quite diverged outside this region (Fig. 2A). However, the 12 amino acid comprising Iro-Box motif in the C-terminal portion is greatly conserved among all species (Fig. 2A, bottom). Within the mouse family, Irx1, Irx2, Irx3, and Irx5 show 90-100% sequence identity within their homeodomains. All of the Iroquois family members with the exception of the C. elegans homeodomain sequence, have six amino acids, TLKAWL, which have been conserved in helix 1. In addition, there is complete agreement of all sequences between helix 1 and helix 2 (position 20-27), and between helix 2 and helix 3/4, (except for a glutamine at position 37 in C. elegans). Helix 2 itself is strongly conserved in all genes, except for C. elegans, which differs in five of the 11 amino acids. Helix 3 and 4 are completely identical among all members of the Iroquois family. Thus, the Iroquois-like gene isolated from C. elegans is quite diverged, suggesting that the worm Iroquois-like gene separated at an early stage from the common ancestor of the Iroquois homeobox genes.
A pairwise comparison of the amino acid identity of Irx5 with the other aligned sequences indicates that the human HIRX2 is the closest related gene to Irx5 (100% identity over whole alignment region) (Fig. 2B). Inside the mouse family, Irx5 and Irx2 reveal the highest degree of amino acid identity (100%). Furthermore, the pair Irx1 and Irx3 displays a striking similarity (82% in 110aa, Fig. 7). In contrast, the other four pairs share significantly less amino acid identity (68-71% in Fig. 7). Within the family of Drosophila Iroquois-like genes, the predicted protein sequences of ara and caup are less similar to Irx5 (65% and 71% amino acid identity, respectively) than mirr (73% to Irx5).
| Figure 3. Detection of Irx5 transcripts by whole mount in situ hybridization at early mouse neurulation stage, E8.5. A-B are arranged in a rostrocaudal order. (A) Dorsolateral view on an E8.5 embryo. Anterior is to the left. It shows Irx5 expression in the cephalic neural fold and the neural plate (the future spinal cord). (B, C, D, E) transverse section of the E8.5 embryo in (A). The plane of sections are illustrated in (A). (B) shows Irx5 expression in the presumptive hindbrain neuroepithelium and in the cephalic mesoderm. In (C), Irx5 expression is detectable in the epithelium of the first branchial arch and pharyngeal pouch as well as in the anlage of the hindbrain. (D) Irx5 transcripts are confined into the intra-embryonic coelomic cavity. (E) Irx5 expression is found in the entire neural plate avoiding the most ventral region. Bar 1, first branchial arch; cem, cephalic mesoderm; HB, hindbrain; icc, intra-embryonic coelomic cavity; NP, neural plate; pp, pharyngeal pouch. Scale bars =  50  m of (B-E).[ Normal View 38K | Magnified View 141K ] |
| Figure 4. Overview of the Irx5 expression profile of E9.5 (A, B) and E10.5 (C-H) whole mount in situ hybridized embryos. The embryo in (A) shows Irx5 expression in the midbrain, hindbrain, spinal cord, and epithelium of the first branchial arch and foregut. (B) transverse section at the hindbrain level of the embryo in (A) revealing the Irx5 expression in the neuroepithelium of the hindbrain. At E10.5 midgestation stage, Irx5 expression is detected in the midbrain, hindbrain, spinal cord, forelimb, and foregut (C). (D-H) are vibratome transverse sections of the embryo in (C). At the level of the rostral hindbrain (D), Irx5 transcripts are restricted to the pons and hypothalamus. Irx5 expression in the otic vesicle is restricted to the neuroepithelium adjacent to the hindbrain (E). The tegmentum of the telencephalon displays Irx5 expression (F). At the level of the forelimb, Irx5 is expressed in the proximal portion of the limb but as a ventral (strong) to dorsal (faint) gradient. Within the spinal cord, the expression is confined to the ventricular zone and intermediate columns of the basal plate as well as into the motoneuronal columns (G, H). Additionally in (H), Irx5 transcripts are found in the bronchial epithelium and the atrium. (I) represents a close up of a saggital section of an E11.5 embryo through the somites. Irx5 expression is confined to a cell sheet in the most caudal portion of the sclerotome. A, atrium; bare, epithelium of the branchial arch; cem, cephalic mesoderm; d, dorsal; Die, diencephalon; dt, distal; HB, hindbrain; InC, intermediate column; MB, midbrain; mot, motoneuronal columns; FG, foregut; HB, hindbrain; HT, hypothalamus; ov, otic vesicle; p, proximal; PN, pons; sc, sclerotome; som, somite; SP, spinal cord; tr, bronchial epitheilum; TG, tegmentum; v, ventral. Scale bars = 175 m in (B) and 30 m in (G). [ Normal View 67K | Magnified View 249K ] |
| Figure 5. Radioactive in situ analysis of Irx5 expression at E12.5 embryos. (A) saggital section presenting expression in the cortex, mammillary region of the hypothalamus (arrows), pretectum, midbrain, pons, myelencephalon, spinal cord, lower jaw, lung, epithelium, and vertebraes. (B, C) are transverse sections displaying Irx5 expression in the hypothalamus and the ventral zone of the hindbrain, dark, and bright field, respectively. A transverse section at the level of the inner ear shows expression in the vestibulocochlear ganglion, cochlea, utricle, and semicircula canals, dark, and bright field, respectively (D, E). (F) saggital section with Irx5 expression in the most superficial zone of the cortex. The expression in the posterior portion of the cortex is stronger than this in the anterior one. (G) represents a section through the eye at E12.5 with faint Irx5 signals in the neural retina. The signals obtained in the lens we consider as backround, because the same result was observed in the sense control. Arrows in A, mammillary areas of the hypothalamus; a, anterior; CX, cortex; co, cochlea; DT, dorsalthalamus; HB, hindbrain; L, lung; MB, midbrain; MN, mandibular aspect of the lower jaw; MY, myelencephalon; p, posterior; PN, pons; NR, neural retina: PT, pretectum; sec, semicircular canals; SP, spinal cord; ut, utricle; VB, vertebraes; vz, ventricular zone. [ Normal View 89K | Magnified View 279K ] |
| Figure 6. Murine chromosomal location of the four Irx genes. The segregation patterns of the Irx loci and their flanking genes in 144 and 121 backcross animals, respectively, that were typed for all loci are shown at the top of the figure. For individual pairs of loci, more animals were typed (see text). Each column represents the chromosome identified in the backcross progeny that was inherited from the (C57BL/6J ò M. spretus) F1 parent. The shaded boxes represent the presence of a C57BL/6J allele and white boxes represent the presence of a M. spretus allele. The number of offspring inheriting each type of chromosome is listed at the bottom of each column. Partial chromosome 8 and 13 linkage maps showing the location of the Irx genes in relation to linked genes are shown at the bottom of the figure. Recombination distances between loci in centiMorgans are shown to the left of the chromosome and the positions of loci in human chromosomes, where known, are shown to the right. References for the human map positions of loci cited in this study can be obtained from GDB (Genome Data Base), a computerized database of human linkage information maintained by the William H. Welch Medical Library of The Johns Hopkins University (Baltimore, MD). [ Normal View 16K | Magnified View 22K ] |
| Figure 7. Schematic summary of the chromosomal mapping analysis of the four mouse Iroquois genes. The numbers indicate percent amino acid identities between the respective pairs of Irx genes over the whole alignment region (108 aa) and/or the homeodomain. [ Normal View 9K | Magnified View 12K ] |
A phylogenetic tree was inferred from the alignment of the 20 Iroquois-like genes and Antp (Fig. 2C). This tree displays several clusters. There are five groups with one vertebrate gene and homologous genes from other vertebrates in each group. Thus, the genes of each vertebrate Iroquois family are less similar to each other than to their homologous ones in other species. In contrast, the genes of the fly Iroquois-Complex and the Iroquois-like gene of C. elegans form distinct clusters. We propose the following arrangement of vertebrate Iroquois-related genes from different species into five groups: (a) Irx1, xIro1, HIRX5; (b) Irx2, cIrx2, xIro2; (c) Irx3, cIrx3, xIro3, zIro3, HIRX1; (d) cIrx4, HIRX4; and (e) HIRX2, Irx5. Furthermore, two clusters can be identified: one groups the Iroquois genes of Irx1, xIro1, HIRX5 together with Irx3, cIrx3, xIro3, zIro3, HIRX1. The second cluster links the group Irx2, cIrx2, xIro2 with Irx5, HIRX2. These clusters may represent paralogous groups of genes (i.e., genes that might have derived from duplication events).
This grouping was confirmed by the bootstrap method with 1,000 resampling steps which was used to evaluate the significance level of the inferred phylogeny. In our phylogenetic analysis, HIRX3 groups distant from the cIrx4- and HIRX4-group and very distant from the group of Irx2, Xiro2, and cIrx2, thus it is obviously not the human orthologous gene of Irx2. It remains to be clarified whether HIRX3 is a splicing variant of HIRX4 or could represent a further human cluster.
Based on the occurrence of three extra amino acids between helix 1 and helix 2, the Iroquois family was classified as a member of the TALE superclass with the PBC class as the next closest relative (Bürglin, [1997]). Our sequence database search revealed that outside of the IRO class, the closest relative to the protein sequence of Irx5 is the homeodomain-containing protein of the BEL1 gene of Arabidopsis thaliana (44% amino acids identity within the homeodomain). The next relatives are the homeodomain protein of the mouse TGIF gene (40%), the mouse homeodomain protein of the PKNOX1 gene (40%) as well as the mouse homeodomain protein of the PBX1 gene (38%). All four genes are members of the TALE superclass.
The homeodomain of the members of the TALE superfamily (including the Iroquois-like genes) contain certain clearly conserved positions that are distinct from the Antp sequence as for example the PYP between helix 2 and helix 3/4 (Fig. 2A). However, outside of the homeodomain the sequences are much more diverged. The Iro-Box motif in the C-terminal portion of all Iroquois-like sequences, for example, is not found in any of the other TALE superfamily members.
Taken together, the results of the sequence analysis show that Irx5 is an additional member of the mammalian Iroquois gene family with structural similarities to homologous genes in Drosophila, Xenopus, Zebrafish, chick, and human.
In Situ Hybridization Analysis of Irx5 During Early Embryogenesis (E6.5-E8.5)
During gastrulation, no Irx5 expression was detectable. At the beginning of neurulation, at E8.0, Irx5 transcripts were confined to the neuroepithelium of the cephalic neural fold and neural plate (Fig. 3A-E), thus outlining the anlage of the midbrain, hindbrain, and spinal cord. The early expression of Irx5 was similar to that of Irx1 and Irx3; however, the Irx5 expression in mesencephalon was found to be much weaker as compared to Irx1 and Irx3 (Bosse et al., [1997]). Further, the cephalic mesenchyme under the prospective neural fold, the first branchial arch and the pharyngeal pouch epithelium showed specific hybridization signals only with the Irx5 probe (Fig. 3B and C). At the level of the caudal hindbrain, Irx5 expression was detectable throughout the entire neural plate except for the most ventral portion including the floorplate and in the intra-embryonic coelomic cavity (future pericardio-peritoneal canal) (Fig. 3D). Caudally, at the level above the tailbending, Irx5 expression was detected in the entire basal plate (Fig. 3E).
Irx5 Expression During Later Phases of Nervous System Development
Following the closure of the neural tube at E9.5, high accumulation of Irx5 transcripts was detected in the hindbrain and rostral spinal cord (Fig. 4A). Similar to Irx2, the expression of Irx5 in the midbrain was fainter and restricted to the tectum. In the tegmentum of mesencephalon, the Irx5 expression started between E9.5 and E10.0, parallel to Irx2 expression (Fig. 4F). Likewise, Irx5 and Irx2 expression was detected at the level of the hindbrain, where transcripts of both genes were restricted to the dorsal portion of the neural tube (Fig. 4B). At this early stage, all Iroquois genes including Irx5 were not expressed in the midbrain/hindbrain border. The rostral limit of Irx5 expression in the brain at E9.5 was delineated by the boundary between prosomere 2 (pretectum) and prosomere 3 (dorsal thalamus). This is also a characteristic feature of all other mouse Iroquois family members.
At E10.5, the anterior expression border expanded into the basal plate including the posterior hypothalamus and the mammillary anlage (Fig. 4D). This is different from the expression of Irx1 and Irx2 that extended only into the dorsal thalamus. Within the caudal spinal cord, Irx5 mRNA was detected in the ventricular and mantle zone of the basal plate as well as in differentiating motoneuronal columns (Fig. 4G). Irx5 was also specifically expressed during inner ear development. Its expression was restricted to the otic vesicle neuroepithelium adjacent to the hindbrain in a region that was smaller when compared to Irx3 (Fig. 4E). Within the developing limb bud Irx5 expression was restricted with highest level in the proximoventral domain of the limb, thus displaying an opposite expression patterns as compared with Irx3.
The expression of Irx5 outside of the nervous system, in the cephalic mesenchyme, first branchial arch and bronchial epithelium was similar to the expression patterns of the other Irx genes. In addition, Irx5 was specifically expressed in the atrium of the developing heart.
At stage E12.5, the expression of Irx5 was maintained in the spinal cord, the ventricular zone of the hindbrain, mesencephalon, pretectum, and mammillary region of the hypothalamus (Fig. 5A and B). It is noteworthy, that although not expressed within the anlage of pons and cerebellum at early stages, at E12.5 Irx5 showed expression in this region as well. This late expression within the midbrain-hindbrain border has also been observed for Irx3 (own observation) and for Irx1, Irx2 as well (M.-B. Becker and A. Zülch, personal communication).
It is noteworthy that at E12.5 in the telencephalon, Irx5 showed a rostral (faint) to caudal (stronger) gradient of expression, being confined to the most superficial zone that consists of postmitotic preplate neurons (Fig. 5F). The expression of Irx3 in the cortical anlage was wider including possibly mitotically active progenitors as well (own observation).
Outside of the CNS, the developing inner ear displayed Irx5 expression in the vestibulocochlear ganglion, cochlea, and utriculus and weak in the semicircular canals (Fig. 5D). Furthermore, a faint Irx5 expression was detected in the internal ganglionic layer of the retina (Fig. 5G). In addition to the nervous system, the nasal mesenchyme, lung epithelium, mandibular region, and the proximal aspect of the limb buds showed the presence of Irx5 transcripts. Interestingly, whereas at E10.5 Irx1 and Irx2 were expressed in the whole somites, Irx5 transcripts are confined at E11.5 to a cell layer in the most caudal portion of the sclerotome (Fig. 4I). At E12.5, however, Irx5 mRNA signals were found in the whole caudal portion of the sclerotome (Fig. 5A).
The Iroquois Gene Family Is Clustered
The chromosomal location of Irx1-3 and Irx5 was determined by interspecies backcross analysis using progeny derived from matings of (C57BL/6J ò Mus spretus) F1 ò C57BL/6J) mice. To identify informative restriction fragment length polymorphisms (RFLPs), Southern blot hybridization analysis (of C57BL/6J and M. spretus genomic DNA) using specific cDNA fragments without homeobox sequences for the different Iroquois genes were performed. The 3.4 kb SacI and 6.2 kb BglII M. spretus RFLPs (see Materials and Methods) were used to follow the segregation of the Irx1 and Irx2 loci, respectively, in backcross mice. Irx1 and Irx2 are linked and mapped to the central region of mouse chromosome 13 adjacent to the genes Ntrk2 and Slc9a3. According to the ratios of the total number of mice exhibiting recombinant chromosomes to the total number of mice analysed for each pair of loci the most likely gene order is: centromere - Ntrk2 - 11/148 - Irx1 - 0/146 - Irx2 - 2/155 - Slc9a3. The recombinationfrequencies (expressed as genetic distances in centiMorgans (cM) ù the standard error) are: Ntrk2 - 7.4 +/- 2.2 - (Irx1, Irx2) - 1.3 +/- 0.9 - Slc9a3. No recombinants were detected between Irx1 and Irx2 in 146 animals typed in common suggesting that the two loci are within 2.1 cM of each other (upper 95% confidence limit).
The 2.2 kb SacI and the 9.1 kb BglII M. spretus RFLPs (see Materials and Methods) were used to follow the segregation of the Irx3 and Irx5 loci, respectively, in backcross mice. The mapping results indicated that Irx3 and Irx5 are linked and are located in the central region of mouse chromosome 8 adjacent to the genes Siah1a and Scyd1. For the chromosomal region including Irx3 and Irx5, the most likely gene order is: centromere - Siah1a - 6/167 - Irx3 - 2/167 - Irx5 - 5/167 - Scyd1. The recombination frequencies are: centromere - Siah1a - 3.6 +/- 1.4 - Irx3 - 1.2 +/- 0.8 - Irx5 - 3.0 +/- 1.3 - Scyd1. Including the known biologically relevant mutation, oligosyndactyly, the most likely linkage order is: Siah1a - os - Irx3 - Irx5 - Scyd1.
| | | | | | |
Comparison of Murine Irx Gene Expression
A common feature of vertebrate and invertebrate Iroquois family members is their early and dynamic expression during initial phases of neural development. At the end of gastrulation, the neural plate consists of undifferentiated precursor cells. During gastrulation, Irx1 and Irx3 are expressed in patches of cells in the presumptive headfold region (Bosse et al., [1997]) whereas Irx2 and Irx5 transcripts are not detectable. During early neurulation, the neural keel arises by convergence of the neural plate cells from lateral positions toward the midline (Smith and Schoenwolf, [1991]). At these early stages of neural development, the spinal cord represents a pseudostratified epithelium of mitotically active cells that are radially arranged surrounding its lumen (Sauer, [1959]; Sidman et al., [1959]). At the beginning of neurulation, Irx5 starts to be expressed in the neuroepithelium of the cephalic neural fold and entire neural plate. This is consistent with the expression of Irx1, Irx2, and Irx3. However, the expression of Irx5 in the developing mesencephalon is much weaker than the expression of Irx1 and Irx3. A dorsoventral polarity within the neural epithelium arises during the process of final mitosis and differentiation where the cells migrate out of the ventricular zone (Jacobson, [1985]). During this process, motoneurons, interneurons, and neurons of the sensory pathway appear from the primarily ventral and dorsal portions of the neural tube (Altman, [1984]).
During neurulation, all four mouse genes keep their predominant expression patterns along the A/P-axis of the CNS. Striking is their common rostral sharp limit of expression at the border between pretectum and mesencephalon. This boundary corresponds to a well-studied neuromeric border between the diencephalon and mesencephalon (Puelles et al., [1987]). In general, the expression of all mouse Irx genes differ more prominently during the early stages (E7.5-9.5). Later, their profiles overlap extensively. Noteworthy, from E10.5 onward, although the expression of all four Irx genes extends into the dorsal thalamus, only transcripts of Irx3 and Irx5 are detected in the posterior hypothalamus and mammillary area. The rostral limit of Irx1 and Irx2 expression is at the level of the zona limitans intrathalamica (Bosse et al., [1997]), a neuromeric border located between the ventral and dorsal thalamus (Puelles et al., [1987]). Interestingly, in the developing telencephalon, Irx3 and Irx5 are differentially expressed with Irx5 being expressed predominantly in the postmitotic preplate neurons whereas Irx3 mRNA, in addition, was located in mitotically active cells in the upper part of the ventricular zone. In the hindbrain and diencephalon, Irx5 expression is mainly confined to the mitotic active cells in the ventricular zone.
The recently isolated chick cIrx4 gene is specifically expressed in the ventricle of the developing heart (Bao et al., [1999]). Different from that, from E10.5 onward, Irx5 transcripts are restricted to the atria of the heart, suggesting that cIrx4 and Irx5 might be involved in the development of the different cardiac chambers. Most of the chamber-differentially distributed genes encode isoforms of contractile proteins, including the myosin heavy chains and light chains (Chien et al., [1993]; Lyons, [1994]). Recently, cIrx4 was shown to alter the myosin heavy chain expression profile and therefore was assumed to be involved in maintenance or determination during cardiac chamber development (Bao et al., [1999]). Based on the molecular relationships between predicted Iroquois protein sequences, one can also expect an evolutionarily conserved function during heart development for Irx5.
Here we report the selective Irx5 expression in a subset of cells in the caudal halves of the somitic sclerotome of E11.5 embryos. Molecules expressed in either half of the somitic sclerotome are possible candidates to be involved in the rostrocaudal polarity of this structure. The condensing caudal sclerotome participates in the formation of the vertebral bodies and gives rise to the proximal ribs and neural arches. Only few genes are known to show a specific distribution within the caudal portion of the sclerotome like Uncx4.1 (Mansouri et al., [1997]), T-cadherin (Ranscht and Bronner-Fraser, [1991]), and the peanut lectin binding glycoproteins (Davies et al., [1990]; Stern et al., [1989]).
In view of the evolutionarily conserved neurogenic pathway, it was of particular interest, that the expression of the putative proneural target gene, Xash3, correlates with the expression of the Xenopus Iroquois-like genes. Accumulating evidences from experiments with ectopic expression of Xenopus Iroquois-like genes show that they specify neural precursor cells and are indeed involved in the positive regulation of certain proneural genes (Bellefroid et al., [1998]; Gomez-Skarmeta et al., [1998]). These findings suggest the conservation of the genetic cascade regulating the activity of proneural genes from the fruitfly to vertebrates. Furthermore, the mouse Irx3 and the frog Xiro3 display overlapping expression such as in the ventricular zone of the developing spinal cord containing the uncommitted neural precursor cells (Bellefroid et al., [1998]) implying that the neurogenic pathway has been conserved even from insects down to mammals.
Conclusions of the Sequence and Mapping Analysis
The thorough sequence analysis of all Iroquois-related genes and the mapping experiments of the mouse Irx genes revealed several new characteristics of these genes.
The most striking feature of all members of the Iroquois family is the nonpolar alanine residue at position nine of the recognition helix. The recognition helix is supposed to make contacts to bases present in their target genes, thus being the major determinant for the DNA-binding specificity (Treisman et al., [1989]). Outside of the Iroquois family, no other known homeodomain has an alanine at this specific position. Polar residues are found in most homeodomains, and the fact that the nonpolar alanine in Iroquois homeodomains is conserved, suggests that DNA-protein interaction could be of a different nature (Bürglin, [1997]). For example, the Paired-class proteins are subclassified according to the residue at position 50 of the homeodomain, which can be one of the polar residues, serine, glutamine, or lysine, underlining the key role of that residue for DNA-binding specificity (Galliot et al., [1999]).
Recently, the Iroquois family was classified as a subclass of the TALE superclass of homeobox-containing genes because of its atypical homeodomain that has three extra amino acids between helix 1 and helix 2 (Bürglin, [1997]). The unique Iro-Box motif in the C-terminal portion of all Iroquois-like sequences is not found in any of the other TALE superfamily members. It can be speculated that this motif is also involved in the interaction of DNA and proteins. The existence of the Iro-Box motif and the nonpolar residue in the recognition helix of the homeodomain confirm that the Iroquois genes represent a separate family. Because of the high sequence similarity of the Iroquois homeodomains and the BEL1 homeodomain of Arabidopsis thaliana, one can speculate about a putative common ancestral gene of the Iroquois family and BEL1. However, because the other members of the TALE group differ only marginally less compared to the Iroquois family, more data are required to evaluate this hypothesis. Nonetheless, the close relationship of the Iroquois family and the Arabidopsis thaliana BEL1 gene support the view that a common ancestor of the TALE homeobox genes was present prior to the separation of plant and animal kingdoms (Bürglin, [1997]).
It is generally accepted that the homeobox genes have undergone duplication during evolution, which resulted in the appearance of multiple genes in vertebrates in comparison to a single gene found in Drosophila (Holland et al., [1994]). These duplicated copies can either stay together in gene clusters, duplicate as clusters later on again or drift apart to different chromosomal locations (Kappen et al., [1993]; Schughart et al., [1989]).
Our results from the chromosomal mapping experiments revealed that the mouse Iroquois genes are clustered, with Irx1 and Irx2 located on chromosome 13 within 2.1cM of each other and Irx3 and Irx5 on chromosome 8 within 1.2 cM of each other (Fig. 6). The data obtained from Drosophila mapping experiments showed also a clustered genomic organization of the genes araucan, caupolican, and mirr, of the Iro-Complex (Gomez-Skarmeta and Modolell, [1996]; McNeill et al., [1997]). Three Drosophila, one Zebrafish, three Xenopus, three chick, four mammalian, and five human Iroquois genes have been identified so far, and this may not represent their entire complements. Because the amino acid sequences between Irx1 and Irx3 as well as between Irx2 and Irx5 are significantly more identical than the other respective pairs (Fig. 7), one can speculate about a possible scenario for the evolutionary origin of the various Irx genes. They could have arisen by an early independent gene duplication, which was followed by a tandem duplication event. This series of gene duplications could also have happened in the other organisms.
Based on the advanced sequencing effort, the genome of C. elegans most likely contains only one Iroquois-like gene. In addition, the C. elegans Iroquois-like gene shows the highest degree of divergence within the Iroquois gene family suggesting that this worm gene separated at an early stage from the common ancestor of the Iroquois homeobox genes. Consequently, the first duplication might have occurred following the separation of nematoda.
Because the phylogenetic tree shows a clustering of the individual Iroquois-like genes into species subgroups (regarded as orthologous genes), one can imagine that (at least one) duplication occurred prior to the radiation of the chordata. Furthermore, their grouping into two or even three paralogous clusters as deduced from our molecular phylogenetic reconstruction corresponds to our chromosomal mapping results and reflects our gene duplication hypothesis. This hypothesis is supported by the assumption that the TALE homeobox genes constitute an old, distinct group, which diverged long ago from typical homeoboxes (Bürglin, [1997]).
Thus, it was not surprising to find two pairs of mouse Iroquois genes in close vicinity within the mouse genome as well as five currently known Iroquois-like sequences within the human genome.
Furthermore, our sequence analysis has revealed that the chick cIrx4 gene seems to be more diverged in comparison to all mouse Irx genes. However, cIrx4 is more closely related to Irx3 and Irx5 than to Irx1 and Irx2. Thus, one can expect that a putative mouse Irx4 gene is located on chromosome 8 possibly linked to Irx3 and Irx5. Indeed, work to examine whether the murine Iroquois gene family might consist of more than the four members found so far is in progress (T. Peters, R. Didrop, and U. Rüther, in preparation). The hypothesis of further members is supported by the fact that the chicken Irx4 gene (Bao et al., [1999]) and the human HIRX4 are clearly separated in the phylogenetic tree. The gene duplication hypothesis is further consistent with the existence of five or possibly even more mammalian family members and the fact that their predicted proteins outside of the homeodomain are relatively diverged, suggesting that genes of this family may have adopted different and overlapping functions.
The currently known human Iroquois-like genes have most likely been named according to their time of identification [database entries: HIRX1 (acc. no. U90308), HIRX2 (U90309), HIRX3 (U90307), HIRX4 (U90306), HIRX5 (U90307)]. The results of our sequence analysis and in particular the grouping of orthologous genes into clusters might be a basis for a unified nomenclature of the Iroquois genes in the future.
We have compared our chromosomal mapping result with a composite mouse linkage map that reports the map location of many uncloned mouse mutations (provided from Mouse Genome Database, a computerized database maintained at The Jackson Laboratory, Bar Harbor, ME, USA). The Irx loci mapped in regions of the composite map that lack mouse mutations with a phenotype that might be expected for an alteration in these loci. The central region of mouse chromosome 8 shares a region of homology with human chromosomes 16q (Fig. 6). Our placement of Irx3 and Irx5 in this interval suggests that they will map to 16q in humans. The proximal region of mouse chromosome 13 shares regions of homology with human chromosomes 9q and 5p, suggesting that the human homologs of Irx1 and Irx2 will reside in one of these two chromosomal regions, as well.
Interestingly, although no mutations have yet been reported for the locus of Irx1, Irx2, and Irx5, Irx3 is located near to a region where one mutation with biological relevance has occurred (data not shown). Oligosyndactyly, an irradiation mutation, leads to embryogenic lethality of homozygotes at day 5pc. Heterozygous embryos show fusions of the digits and muscle malformations on all feet (Gruneberg, [1961]; van Valen, [1966]). Because heterozygous embryos for oligosyndactyly do not show an alteration of Irx3 expression, the locus of Irx3 seems not to coincide with that of oligosyndactyly. The impact of the mutation, oligosyndactyly, on Irx3 is currently being investigated.
With the isolation of Irx5, four mouse Iroquois genes have been currently characterized. They are predominantly expressed in the nervous system and overlap with the expression of their predicted up- and downstream genes. The results obtained from the sequence analysis of all presently known Iroquois-like genes and chromosomal mapping experiments indicate that these genes have appeared in vertebrates as a result of several gene duplication events. The possibility of the existence of either further genes or an additional cluster within the mouse genome is currently being examined.
In conclusion, the current data suggest that the neurogenic pathway determining specific cell fates where Iroquois proteins act as regulators is conserved from fruitfly down to mammals during evolution.
| | | | | | |
Cloning and Analysis of Genomic and cDNA Regions of Irx5
A mouse genomic library (strain 129svj) was screened with a 760bp BamHI fragment located downstream of the homeobox from the mouse Irx3 cDNA using standard conditions except for a reduced first hybridization and washing temperature of 60÷C (Sambrook et al., [1989]). Several mouse genomic clones were isolated in parallel and partially sequenced by homeobox flanking primers. The partial sequences obtained have revealed two overlapping clones containing a diverged member of the Iroquois gene family, corresponding to Irx5. The genomic organization of exons and the position of the homeodomain was analysed (Fig. 1B).
To isolate the cDNA of the new genomic Irx5 clone, a 1-kb XhoI-fragment containing the first exon devoided from the homeobox sequences was used to screen an E14.5 randomly-primed cDNA library (Wijnholds et al., [1995]). Nonradioactive hybridization was performed using Random prime labelling and detection systems (Amersham) according to the manufacturer's instructions. A total of 103 plaques were screened and ten independent phages were isolated. During further rounds of plate purifications, high stringency conditions at 65÷C were used. Three of the ten clones were isolated by two further rounds of plate purifications. Based on restriction analysis, two of three positive clones encoded the same transcript. The remaining two inserts, of about 1.8 kb and 2 kb were subcloned into pBluescript (Stratagene), sequenced and confirmed to represent the same gene. The 1.8 kb clone extended the 2 kb cDNA clone around 316 bp in the 5 direction. The Irx5 cDNA has a total length of 2.2 kb including an open reading frame encoding a 484 amino acid protein. The sequence directly upstream of the proposed initiation methionine agrees in four of the six bases with the Kozak consensus sequence (Kozak, [1987]). The Genbank accession number for the mouse Irx5 gene is AF230074.
Blast searches were performed using BLASTP at the NCBI with the deduced amino acid sequence of the mouse Irx5 as the query sequence. An alignment of deduced amino acid sequences of 20 Iroquois-related genes, as well as four representatives from different families within the TALE superfamily (mouse mPKNOX1, mouse mPBX1, mouse mTGIF, Arabidopsis t. AtBEL1) and Antennapedia (Drosophila Antp) as an outgroup was constructed using CLUSTAL-X (Thompson et al., [1997]). The region of the homeodomain as well as approximately 43 amino acids upstream and downstream were further analysed. A distance matrix of the relative divergence between all pairs of aligned sequences was computed. Based on this distance matrix, a phylogenetic tree was reconstructed using the neighbor joining method of Saitou and Nei as installed in CLUSTAL-X. Trees were visualized using TreeView on the Macintosh V1.5 by R. Page (Page; http://www.zoology.taxonomy.gla.ac.uk/rod/treeview.html).
Accession numbers: Sequences were retrieved from Genbank/Swissprot. The Iroquois sequences: Chick: cIrx2 (AJ237599), cIrx3 (AF157620), cIrx4 (AF091504); Xenopus: Xiro1 (AJ001834), Xiro2 (AJ001835), Xiro3 (AF027175); Zebrafish: Ziro3 (AF124095), C. elegans: ceIrx (Q93348); Drosophila: ara (X95179), caup (X95178), mirr (AF004710); Human: HIRX1 (U90308), HIRX2 (U90309), HIRX3 (U90307), HIRX4 (U90306), HIRX5 (U90307); Mouse: Irx1 (Y15002), Irx2 (Y15000), Irx3 (Y15001), mPKNOX1 (AAC15990), mPBX1 (P41778), mTGIF (P70284). Arabidopsis t. AtBEL1 (AAD21503), Antennapedia sequence: Antp (P02833).
Whole mount in situ hybridizations (Wilkinson and Nieto, [1993]) using in vitro probes derived from regions outside of the conserved Irx homeobox were performed on mouse embryos from E6.5-E11.5. Radioactive in situ analysis has been performed as described (Kessel and Gruss, [1991]). Two different probes between 0.65-1.0 kb have been used, and the detected patterns were identical. Sense RNA probes were used as negative controls in all experiments and gave no signal. Structures in the developing nervous system were named according to an atlas (Alvarez-Bolado and Swanson, [1996]).
Interspecific Mouse Backcross Mapping
Interspecific backcross progeny were generated by mating (C57BL/6J ò M. spretus) F1 females and C57BL/6J males as described (Copeland and Jenkins, [1991]). The interspecific backcross mapping panel has been typed for over 2,900 loci that are well distributed among all the autosomes as well as the X chromosome (Copeland and Jenkins, [1991]). A total of 205 N2 mice were used to map the Irx loci. C57BL/6J and M. spretus DNAs were digested with several enzymes and analysed by Southern blot hybridization for informative restriction fragment length polymorphisms (RFLPs) using mouse cDNA probes. DNA isolation, restriction enzyme digestion, agarose gel electrophoresis, Southern blot transfer, and hybridization were performed essentially as described (Jenkins et al., [1982]). All blots were prepared with Hybond - N+ nylon membrane (Amersham). The Irx1 probe, an 700 bp BsshII fragment of mouse cDNA, detected a 3.2 kb fragment in SacI digested C57BL/6J (B) DNA and a 3.4 kb fragment in SacI digested M. spretus DNA (S). The Irx2 probe, a 1.4 kb SalI fragment of mouse cDNA, detected BglII fragments of 9.7 kb (B) and 6.2 kb (S). The Irx3 probe, an 700 bp PstI fragment of mouse cDNA, detected SacI fragments of 2.7 kb (B) and 2.2 kb (S). The Irx5 probe, an 500 bp fragment of mouse cDNA, detected BglII fragments of 6.1 kb (B) and 9.1 kb (S). All probes were labelled with (32P) dCTP using a random primed labelling kit (Stratagene); washing was done to a final stringency of 0.1-1.0 ò SSCP, 0.1% SDS, 65÷C. The presence or absence of the M. spretus-specific fragments was followed in backcross mice.
For Irx1 and Irx2, 121 mice were analysed for every marker and are shown in the segregation analysis (Fig. 6), and up to 155 mice were typed for some pairs of markers. Each locus was analyzed in pairwise combinations for recombination frequencies using the additional data. The ratios of the total number of mice exhibiting recombinant chromosomes to the total number of mice analyzed for each pair of loci were determined.
For Irx3 and Irx5, 144 mice were analysed for every marker and are shown in the segregation analysis (Fig. 6), up to 167 mice were typed for some pairs of markers. Again, each locus was analysed in pairwise combinations for recombination frequencies and the ratios of the total number of mice exhibiting recombinant chromosomes to the total number of mice analysed for each pair of loci were computed.
A description of the probes and RFLPs for the loci linked to Irx3 and Irx5, including Siah1a and Scyd1 has been reported previously (Holloway et al., [1997]; Rossi et al., [1998]); those linked to Irx1 and Irx2 include Ntrk2 and Slc9a3 (Pathak et al., [1996]; Tessarollo et al., [1993]). Recombination distances were calculated using Map Manager, version 2.6.5. Gene order was determined by minimising the number of recombination events required to explain the allele distribution patterns.
| | | | | | |
We gratefully acknowledge D. Treichel for sharing his expertise in molecular and computational biology. We thank Heidemarie Wegener as well as Claus-Peter Adam for the help with the image-processing and Ralf Altschäffel for the perfect photographic work. We thank Silke Eckert and Debra J. Gilbert for excellent technical assistances. This research was funded in part by the European Union and by the National Cancer Institute, DHHS, under contract with ABL.
| References |
| | | | | | |