- Email: [email protected]
Human annexin 31 genetic mapping and origin
Gene 227 (1999) 33–38
Human annexin 31 genetic mapping and origin Reginald O. Morgan a,*, Daphne W. Bell b, Joseph R. Testa b, Maria-Pilar Fernandez a a Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Oviedo, E-33006 Oviedo, Spain b Department of Medical Oncology, Fox Chase Cancer Center, Philadelphia, PA 19111, USA Received 9 September 1998; received in revised form 19 November 1998; accepted 20 November 1998; Received by S. Yokoyama
Abstract The cDNA encoding novel human annexin 31 was utilized for chromosomal mapping, structural comparison, and phylogenetic analysis to clarify its genetic relationship to other annexins. The ANX31 gene locus was mapped by fluorescence in situ hybridization to human chromosome 1q21, remote from ten other paralogous human annexins on different chromosomes but near the epidermal differentiation gene complex, the S100A gene cluster and a breast-cancer translocation region. Protein homology testing and characterization of incompletely processed expressed sequence tags identified annexin 2 as the closest extant homologue. Maximum likelihood analysis confirmed its most recent common ancestor with vertebrate annexin 2 and validated its classification, in order of discovery, as annexin 31. This subfamily was formed approx. 500–600 million years ago, subsequent to the gene duplication that produced annexin 1. It has diverged rapidly and extensively, especially in the well-conserved and functionally critical type II calcium-binding sites. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Chromosomal localization; Fluorescence in situ hybridization; Gene duplication; Molecular evolution; Molecular sequence data; Nucleotide substitution rates
1. Introduction Phylogenetic and genetic studies have confirmed that the annexin tetrad is a unique, ancient and highly conserved structure (Morgan and Fernandez, 1995, 1997a) with some membrane-related role involving calcium channels, cellular signaling, vesicular transport or extracellular matrix (Raynal and Pollard, 1994). Vertebrate annexins are known to have emanated from a common eukaryotic ancestor since the emergence of metazoa around 800 million years ago (Mya) and their independent evolution at dispersed genomic locations implies differentiated, non-redundant functions (Morgan and Fernandez, 1997b; Morgan et al., 1998). The structural and evolutionary interrelationships of annexin genes provide vital information for establishing prospective links to functionally related genes, shared * Corresponding author. Tel: +34-98-510-4214; Fax: +34-98-510-3157; e-mail [email protected] Abbreviations: aa, Amino acid(s); ANXn, human annexin gene or locus (suffixed by subfamily number); DAPI, diamidino2-phenylindole; EST, expressed sequence tag; Myr/Mya, million years/ago.
phenotypes and, ultimately, to specific hereditary diseases. A complete knowledge of all extant annexins and their systematic analysis are essential for precisely identifying conserved regions that have a common, critical function and variable regions that distinguish the individual members by their regulation, properties and subcellular roles. Ten human annexins have been discovered on the basis of diverse functional assays and sequenced between the years 1986 and 1992 (Raynal and Pollard, 1994). The recent identification of an eleventh human annexin featuring lost calcium-binding sites, an RGD-like cell attachment motif and a distinctive expression pattern has raised key questions about the role of membrane interactions in annexin function and the actual diversity of this multigene family (Morgan and Fernandez, 1998). An examination of this gene’s chromosomal location, structural organization and phylogenetic place in relation to other annexins was therefore undertaken here to further characterize its unique identity and to resolve its true genetic origin. Such analyses could reveal some evolutionary pattern in annexin divergence and link this gene to novel, specialized functions or hereditary traits.
0378-1119/99/$ – see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 9 8 ) 0 0 59 7 - 6
34
R.O. Morgan et al. / Gene 227 (1999) 33–38
2. Materials and methods
3. Results and discussion
2.1. Chromosomal mapping by fluorescence in situ hybridization
3.1. Genetic mapping of ANX31
The A31H hybridization probe for chromosomal mapping was the 1.3 kb insert of expressed sequence tag ( EST ) clone ID 111373 from the I.M.A.G.E. (Integrated Molecular Analysis of Genome Expression) Consortium (LLNL) (Lennon et al., 1996). It was sequence verified to represent most of the coding and 3∞ untranslated region of a novel human annexin cDNA (Morgan and Fernandez, 1998; gb: AJ009985). Metaphase spreads were prepared from phytohemaglutinin-stimulated human lymphocytes of a healthy female donor (Fan et al., 1990). Fluorescence in situ hybridization and detection of immunofluorescence followed described protocols (Bell et al., 1995). The probe was labeled with biotin-16–dUTP (Boehringer Mannheim) by nick translation and hybridized overnight to denatured chromosomes at 37°C under stringent conditions. Hybridization sites were detected with fluoresceinlabeled avidin (Oncor) and amplified by addition of anti-avidin antibody (Oncor) and a second layer of fluorescein-labeled avidin. Chromosome preparations were counterstained with diamidino-2-phenylindole (DAPI ) and observed with a Zeiss Axiophot epifluorescence microscope equipped with a cooled charge coupled device camera (Photometrics, Tucson AZ ) operated by a Macintosh computer workstation. Digitized images of DAPI staining and fluorescein signals were captured, pseudocolored and merged using Oncor version 1.6 software.
The A31H cDNA probe hybridized to human metaphase spreads with specific labeling on chromosome 1 ( Fig. 1A). Fluorescence signals were detected on chromosome 1 in each of 23 metaphase spreads scored. Among 122 signals observed, 54 (44%) were on 1q with the following distribution: one chromatid (five cells), two chromatids (eight cells), three chromatids (seven cells), four chromatids (three cells). Among the 92 chromosome 1 chromatids scored, 54 (59%) hybridized to 1q. All chromosome-specific signals were localized to 1q21. This chromosomal band is contained within human–mouse homology group 4 (DeBry and Seldin, 1996), which is represented by more than 40 orthologous gene pairs also localized to human chromosome 1q21 and homologous with the mouse chromosome 3 region 42–49 cM from the centromere (Fig. 1B). Two potentially relevant associations are the neighboring S100 gene cluster and the epidermal differentiation complex (Scha¨fer et al., 1995). The former group consists of at least 13 genes for type I calcium-binding proteins, defined by the EF-hand conformation of their protein a-helices, and comprising members known to bind the amino-termini of certain annexins and influence their function (namely S100A3, S100A6-calcyclin, and S100A10-p11). The latter group includes loricrin, involucrin, filaggrin and trichohyalin with defined roles in skin growth and differentiation and a possible functional or regulatory relationship to annexin 31. 3.2. Structural characterization of anomalous ESTs
2.2. Molecular systematics Pairwise sequence comparisons employed programs from the FASTA package version 3.1 (Pearson, 1990) and multiple alignments were assembled using CLUSTALW version 1.6 ( Thompson et al., 1994). Database search and retrieval were done with the BLAST and RETRIEVE netservers of the National Center for Biotechnology Information (NCBI, Bethesda, MD) on the database of expressed sequence tags (dbEST, Boguski et al., 1993). Phylogenetic analyses utilized programs from the Phylogeny Inference Package (PHYLIP release 3.57) ( Felsenstein, 1989) for statistical analysis of branching order by sequence alignment bootstrapping and for tree reconstruction under various evolutionary models. Protein maximum likelihood analyses used the PUZZLE program version 4.0 (Strimmer and von Haeseler, 1996) to resolve trees by quartet puzzling under a discrete c-distribution rate model (a-parameter 1.89±0.11). The branch length test ( Takezaki et al., 1995) was used to verify rate linearity among taxa for annexin genes.
The dbEST provides sequence and expression data for cDNAs from transcribed genes and can also be a valuable genomic resource for normal and alternative exon splicing patterns, because it contains many incompletely processed transcripts ( Wolfsberg and Landsman, 1997). The characterization of anomalous ESTs is especially important for resolving identification errors in public databases, such as Unigene clusters and the Human Gene Map (NCBI ), and such discrepancies apply to annexin 31. We estimate, from our analyses of annexin ESTs (Morgan and Fernandez, 1998), that approx. 5% of these transcripts contain unspliced introns and have confirmed their authenticity in all cases where annexin gene structures have been determined. Two EST clone IDs 188508 and 137481 were tentatively identified as partially processed genomic transcripts of human annexin 31 ( Fig. 2), analogous to those detected previously for mouse annexin 3 ( Fernandez et al., 1998a). The former EST contained 237 bp of the 3∞ penultimate intron plus 25 bp of coding exon after the splice ‘cccctactag/GAA ACT’. The latter was represented by
R.O. Morgan et al. / Gene 227 (1999) 33–38
35
of all known vertebrate annexins except annexin 7 (Morgan and Fernandez, 1997b), and this has been corroborated by genomic polymerase chain reaction sequencing as part of initial efforts to characterize the gene structure. The segment of annexin 31 cDNA corresponding to exon 8 differed in size from that in annexin 6b (54 bp) and annexins 1, 3, 5 and 6a (57 bp), but was congruous with annexin 2 in which exon 8 is 60 bp due to a single codon insertion (Spano et al., 1990). This similarity with annexin 2 in the regions of both exons 8 and 12 provided preliminary evidence for homologous gene structures, and protein comparisons further strengthened their association. The characteristic amino-terminal tail of annexin proteins comprised a 25-aa unique region in the 338-aa complete annexin 31 (Morgan and Fernandez, 1998; gb: AJ009985) and this was most similar in size to the corresponding segment in annexin 2 (24 aa) and annexin 1 (33 aa), in contrast to other human annexins with much shorter or longer amino termini (range 6–191 aa). Homology testing of the 338-aa, full-length annexin 31 protein by sequence shuffling in PRSS-FASTA (Pearson, 1990) again identified annexin 2 with the highest Smith–Waterman alignment score of 842 and lowest probability of random occurrence of 4.78×10−55, while annexin 13 had the lowest homology index of 547. Since these latter two annexins lie at extreme branch tips of the annexin evolutionary tree (Morgan et al., 1998), it was worthy of testing whether annexin 31 might be directly related to annexin 2, both phylogenetically and genetically. 3.3. Phylogenetic classification of novel annexins
Fig. 1. Physical mapping of human annexin 31. (A) A human metaphase spread shows specific fluorescence in situ hybridization signals from a fluorescein-labeled pA31H probe at 1q21. The photograph represents computer-enhanced, merged images of fluorescein signals (arrowheads) and DAPI-stained chromosomes. (B) The cytogenetic idiogram of human chromosome 1 shows the band position containing the ANX31 locus, with a partial list of gene symbols and locus positions for neighboring genes known to be linked in the expansive human– mouse homology group 4.
two overlapping ESTs with 533 bp of intron containing a 5∞ dimeric Alu–Sxc insert followed by a typical 3∞ splice consensus ‘ttgtctgcag/GAT GCA’ and 70 bp coincident with the final exon of most vertebrate annexins. Annexin 31 was thus inferred to possess a 123 bp exon identical in size and location to the penultimate exon 12
The evolutionary position of annexin 31 was determined by statistical analysis of its alignment with other annexins to establish its clade association. The 309-aa tetrad core sequence of annexin 31 was aligned with representative members of the 27 previously classified subfamilies (Morgan and Fernandez, 1997a,b) plus other full-length proteins recently characterized in plants and invertebrates. Randomized bootstrap alignments of 143 informative sites (Morgan et al., 1998; Fernandez et al., 1998b) were first generated for analysis by maximum parsimony and by the distance matrix algorithms for neighbor-joining and least-squares ( Felsenstein, 1989). The resulting consensus phylogram was consistent with previous cladistic analysis of annexin subfamilies (Morgan and Fernandez, 1997a, 1998) and with maximum likelihood analysis of all 311 sites by the quartet puzzle method (Strimmer and von Haeseler, 1996). The protein maximum likelihood tree (Fig. 3) was therefore computed using the branching order topology of the consensus tree to estimate branch lengths and confidence values for each node. We extended our previous classification of annexin
36
R.O. Morgan et al. / Gene 227 (1999) 33–38
Fig. 2. Characterization of anomalous ESTs for annexin 31. The annexin 31 cDNA structure (top outline, gray means untranslated, checkered denotes matching exons 8 and 12) is aligned with sequenced segments of EST clone IDs 188508 and 137481 below. The latter represent partially processed genomic transcripts derived from coding exons with unspliced introns (dashed lines), one containing tandem Alu–Sx short, interspersed, repetitive elements (SINEs). The remaining outlines describe the four internally homologous domains of the annexin tetrad core conserved region (between vertical boundaries), and the exon blocks (boxed segments with nucleotide number) summarize the known, minor variations in different annexin gene structures. Annexin 7 is shown with several distinct exons, including exon 12, while other characterized genes differ by only a singlecodon variation in the size of exon 8. The complete human annexin 31 cDNA sequence has been deposited in EMBL/GenBank/DDBJ sequence databases under accession number AJ009985.
subfamilies (Morgan and Fernandez, 1997a) by including other recent database entries with complete coding regions in the phylogenetic analysis ( Fig. 3). Tobacco cDNA clones X511 and X671 from Nicotiana tabacum (Schantz et al., unpublished; gb: Y14972, Y14973) were confirmed with 97% probability to belong with Capsicum annum (green pepper) (Proust et al., 1996) and tomato p34 (Lim et al., 1998) in the previously classified annexin 24 subfamily. Cotton annexin AnnGh2 from Gossypium hirsutum (Potikha and Delmer, 1997) separated together with annexins 18 and 22 from other plant annexins with 99% probability, but its poor bootstrap association (58%) with either of these subfamilies led to its assignment with tomato annexin p35 (Lim et al., 1998) to the new annexin 28 subfamily. The paralogous cotton annexin AnnGh1 (Potikha and Delmer, 1997) and annexin AnxLt1 from leaves of Lavatera thuringiaca ( Vazquez-Tello and Uozumi, unpublished; gb: AF006197), clearly separated from known plant annexins (92%) and together (100%) formed a distinct annexin 29 subfamily. A fourth annexin gene C37H5.1 from the nematode Caenorhabditis elegans ( Wilson et al., 1994; gb: U88315) had a unique exon splicing pattern and no clear association with known invertebrate annexin subfamilies, so it was designated annexin 30 to represent yet another new subfamily. The novel human annexin was assigned to subfamily 31, based on its bifurcation with high confidence level from a peripheral node of one vertebrate annexin clade (Fig. 3) and its distinct chromosomal locus ( Fig. 1), which confirmed paralogy with ten other human annexins. Its nomenclature followed customary subfamily enumeration in the order of discovery of full-length sequences (including those above), consistent with cladistic and genetic analyses, and adopting Arabic num-
bering in place of the now cumbersome Roman numerals. Annexin 31’s common ancestry with annexin 2 was supported by an approximate maximum likelihood value of 62% and a consensus bootstrap frequency of 79% (cf. 88% by parsimony, 74% by neighbor-joining and 76% by least squares analyses) that may have been partially compromised by the inclusion of some misinformative sites, the protein’s extensive sequence divergence and its atypical aa composition. More rigorous testing was desirable to confirm its direct origin. 3.4. Common genetic origin of annexins 2 and 31 The relationship of annexins in the divergent 1–2–31 subclade was re-examined with a focus on ascertaining their respective origins and estimating the dates of probable gene duplication. Rather than presuppose informative sites (above), a reduced number of vertebrate sequences comprising the relevant clade (i.e., annexins 11, 6b, 3, 1, 2 and 31) and phylogenetically more primitive species were analyzed to better delineate the branching order. These now included teleost fish species representing previously unclassified and unpublished members of the annexin 1 and 11 (Osterloh et al., 1998), annexin 2 (Fujiki et al., unpublished; gb: C88389, C88439) and annexin 3 ( Elgar, 1996) subfamilies. Site rate heterogeneity was taken into account under a discrete c model, evolution parameters were estimated precisely and exact maximum likelihood values were calculated using the quartet PUZZLE program (Strimmer and von Haeseler, 1996). The resulting maximum likelihood tree ( Fig. 4) confirmed the branching order of divergent annexins 1, 2 and 31 among the most recently formed vertebrate subfamilies, although their possible origin from annexins 3, 6b or 11 could not be
R.O. Morgan et al. / Gene 227 (1999) 33–38
Fig. 3. Phylogram of the annexin superfamily containing human annexin 31. Protein tree topology was first determined from the consensus of analyses by neighbor-joining, least squares distance and maximum parsimony of 600 bootstrap alignments at 143 protein informative sites in representatives from all known annexin subfamilies (Morgan et al., 1998). Protein maximum likelihood analysis by quartet puzzling (Strimmer and von Haeseler, 1996) confirmed the overall tree topology ( log likelihood −16392) and computed the maximum likelihood branch lengths according to the scale of estimated aa replacements/site. Statistical support for the branching order is given by the percentage values above 50% at resolved nodes for maximum likelihood (bootstrap consensus values in parentheses). Full-length annexins recently described in plants and C. elegans were assigned to subfamilies 24, 28, 29 and 30 (starred labels) and the novel human annexin was sequentially assigned to subfamily 31 as ANX31-Hsa. Genus-species (Gsp) names and references for the newly classified annexins are given in Results Section 3.3, others have been described previously (Morgan and Fernandez, 1997a)
ascertained to a maximum likelihood value greater than 50. The most recent common ancestor of annexins 1 and 2 was previously estimated to have existed about 500 Mya, based on gene mutation rate determinations from rodent–mammal sequence comparisons (Morgan et al., 1998). Values of non-synonymous nucleotide substitutions (Li, 1993) recalculated from fish–mammal comparisons of those particular annexins and calibrated to those species separation times now predicted evolution rates half of the previous estimate for annexin 1 and double that for annexin 2. Since the separation time between fish and humans spans a much greater period of these genes’ lifetime, these were judged to be more
37
Fig. 4. Phylogenetic confirmation of common ancestors for annexins 1, 2 and 31. Parameter estimation and site rate heterogeneity were evaluated by PUZZLE to obtain exact maximum likelihood values (i.e., numbers at nodes) for the homologous protein pairs (312 aa) from each subfamily associated with the annexin 1–2–31 subclade. The evolutionary order of annexins 11 (assigned root), 3 and 6b was not resolved, but annexins 1, 2 and 31 were confirmed with high probability to be subsequent descendents, and the latter two genes shared the most recent common ancestor in this clade. Individual branch lengths reflect the numbers of estimated aa replacements per site and the average distance between all taxa was 1.16. Genus–species abbreviations for taxa include Cca, Cyprinus carpio (carp); Fru, Fugu rubripes (Japanese pufferfish); Gga, Gallus gallus (chicken); Hsa, Homo sapiens (human); and Ola, Oryzias latipes (Japanese medaka fish).
suitable for recalibrating the date of the annexin 2<1 gene duplication closer to 750 Mya. The application of a branch length test ( Takezaki et al., 1995) to annexin 6b, 3, 1 and 2 sequences from Fig. 4 further confirmed that these taxa exhibited a uniform rate of change according to a molecular clock. However, fish annexin sequences need to be completed and the rapid mutation rate of human annexin 31 (cf. branch length disparity between human annexins 2 and 31) requires assessment against more distant orthologues for accurate molecular dating. The current phylogenetic data and limited rate uniformity permitted the inference that annexin 31 duplicated from a common ancestor with annexin 2 prior to the separation of fish from eukaryotic lineage, possibly 500–600 Mya, and this would be substantiated by the eventual identification of annexin 31 in fish. Partial mouse annexin 31 sequences have been obtained (unpublished ) that appear to confirm both the rapid mutation rate and obliteration of type II calcium binding sites in this unusual annexin. The relatively recent fixation of annexin 31 in vertebrate genomes represents a significant evolutionary step in the structural, regulatory and presumably functional diversification of this superfamily.
38
R.O. Morgan et al. / Gene 227 (1999) 33–38
4. Conclusions (1) Human annexin 31 resides on chromosome band 1q21, remote from ten other annexins but in proximate linkage to S100A and skin differentiation genes with which it may have some regulatory or functional relationship. (2) Its coding region and protein structure most closely resemble annexin 2, but its unique lack of calciumbinding sites and lower expression level contrast markedly with the ubiquitous annexin 2 and imply their distinct subcellular location and significant functional divergence. (3) Phylogenetic analyses established that annexin 31 was a mutual gene duplication product with annexin 2 approx. 500–600 Mya and that it has been evolving at a comparatively accelerated rate. (4) Unspliced introns in dbEST transcripts can be informative for gene splicing patterns, and annexin sequences from early diverging species can provide a more reliable index of average gene mutation rates. Acknowledgements This work was supported by grant PM95-0152 from D.G.E.S. of Spain, National Institutes of Health Grant CA-06927 ( USA), and by an appropriation from the Commonwealth of Pennsylvania.
References Bell, D.W., Taguchi, T., Jenkins, N.A., Gilbert, D.J., Copeland, N.G., Gilks, C.B., Zweidler-McKay, P., Grimes, H.L., Tsichlis, P.N., Testa, J.R., 1995. Chromosomal localization, in man and rodents, of a gene (GFI1) encoding a novel zinc finger protein reveals a new syntenic region between mouse and man. Cytogenet. Cell Genet. 70, 263–267. Boguski, M.S., Lowe, T.M., Tolstoshev, C.M., 1993. dbEST — database for expressed sequence tags. Nat. Genet. 4, 332–333. DeBry, R.W., Seldin, M.F., 1996. Human/mouse homology relationships. Genomics 33, 337–351. Elgar, G., 1996. Quality not quantity: the pufferfish genome. Hum. Mol. Genet. 5, 1437–1442. Fan, Y.S., Davis, L.M., Shows, T.B., 1990. Mapping small DNA sequences by fluorescence in situ hybridization directly on banded metaphase chromosomes. Proc. Natl. Acad. Sci. USA 87, 6223–6227. Felsenstein, J., 1989. PHYLIP — Phylogeny inference package. Cladistics 5, 164–166. Fernandez, M.P., Copeland, N.G., Gilbert, D.J., Jenkins, N.A., Morgan, R.O., 1998a. Mouse annexin III cDNA, genetic mapping and evolution. Gene 207, 43–51. Fernandez, M.P., Copeland, N.G., Gilbert, D.J., Jenkins, N.A.,
Morgan, R.O., 1998b. The genetic origin of mouse annexin VIII. Mamm. Genome 9, 8–14. Lennon, G.G., Auffray, C., Polymeropoulos, M., Soares, M.B., 1996. The I.M.A.G.E Consortium: an integrated molecular analysis of genomes and their expression. Genomics 33, 151–152. Li, W.H., 1993. Unbiased estimation of the rates of synonymous and nonsynonymous substitution. J. Mol. Evol. 36, 96–99. Lim, E.K., Roberts, M.R., Bowles, D.J., 1998. Biochemical characterization of tomato p35. Independence of calcium binding and phosphatase activities. J. Biol. Chem. 273, 34920–34925. Morgan, R.O., Fernandez, M.P., 1995. Molecular phylogeny of annexins and identification of a primitive homologue in Giardia lamblia. Mol. Biol. Evol. 12, 967–979. Morgan, R.O., Fernandez, M.P., 1997a. Distinct annexin subfamilies in plants and protists diverged prior to animal annexins and from a common ancestor. J. Mol. Evol. 44, 178–188. Morgan, R.O., Fernandez, M.P., 1997b. Annexin gene structures and molecular evolutionary genetics. Cell. Mol. Life Sci. 53, 508–515. Morgan, R.O., Fernandez, M.P., 1998. Expression profile and structural divergence of novel human annexin 31. FEBS Lett. 434, 300–304. Morgan, R.O., Bell, D.W., Testa, J.R., Fernandez, M.P., 1998. Genomic locations of ANX11 and ANX13 and the evolutionary genetics of human annexins. Genomics 48, 100–110. Osterloh, D., Wittbrodt, J., Gerke, V., 1998. Characterization and developmentally regulated expression of four annexins in the killifish medaka. DNA Cell Biol. 17, 835–847. Pearson, W.R., 1990. Rapid and sensitive sequence comparison with FASTP and FASTA. Meth. Enzymol. 183, 63–98. Potikha, T.S., Delmer, D.P., 1997. cDNA clones for annexin AnnGh1 (accession no. U73746) and AnnGh2 (accession no. U73747) from Gossypium hirsutism (cotton). Plant Physiol. 113, 305 Proust, J., Houlne, G., Schantz, M.L., Schantz, R., 1996. Characterization and gene expression of an annexin during fruit development in Capsicum annum. FEBS Lett. 383, 202–212. Raynal, P., Pollard, H.B., 1994. Annexins: the problem of assessing the biological role for a gene family of multifunctional calcium- and phospholipid-binding proteins. Biochim. Biophys. Acta 1197, 63–93. Scha¨fer, B.W., Wicki, R., Engelkamp, D., Mattei, M.G., Heizmann, C.W., 1995. Isolation of a YAC clone covering a cluster of nine S100 genes on human chromosome 1q21: rationale for a new nomenclature of the S100 calcium-binding protein family. Genomics 25, 638–643. Spano, F., Raugei, G., Palla, E., Colella, C., Melli, M., 1990. Characterization of the human lipocortin-2-encoding multigene family: its structure suggests the existence of a short amino acid unit undergoing duplication. Gene 95, 243–251. Strimmer, K., von Haeseler, A., 1996. Quartet puzzling: a quartet maximum likelihood method for reconstructing tree topologies. Mol. Biol. Evol. 13, 964–969. Takezaki, N., Razhetsky, A., Nei, M., 1995. Phylogenetic test of the molecular clock and linearized trees. Mol. Biol. Evol. 12, 823–833. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680. Wilson, R., Ainscough, R., Anderson, K., Baynes, C., Berks, M., Bonfield, J. et al., 1994. 2.2 Mb of contiguous nucleotide sequence from chromosome III of C. elegans. Nature 368, 32–38. Wolfsberg, T.G., Landsman, D., 1997. A comparison of expressed sequence tags (ESTs) to human genomic sequences. Nucleic Acids Res. 25, 1626–1632.
Human annexin 31 genetic mapping and origin Reginald O. Morgan a,*, Daphne W. Bell b, Joseph R. Testa b, Maria-Pilar Fernandez a a Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Oviedo, E-33006 Oviedo, Spain b Department of Medical Oncology, Fox Chase Cancer Center, Philadelphia, PA 19111, USA Received 9 September 1998; received in revised form 19 November 1998; accepted 20 November 1998; Received by S. Yokoyama
Abstract The cDNA encoding novel human annexin 31 was utilized for chromosomal mapping, structural comparison, and phylogenetic analysis to clarify its genetic relationship to other annexins. The ANX31 gene locus was mapped by fluorescence in situ hybridization to human chromosome 1q21, remote from ten other paralogous human annexins on different chromosomes but near the epidermal differentiation gene complex, the S100A gene cluster and a breast-cancer translocation region. Protein homology testing and characterization of incompletely processed expressed sequence tags identified annexin 2 as the closest extant homologue. Maximum likelihood analysis confirmed its most recent common ancestor with vertebrate annexin 2 and validated its classification, in order of discovery, as annexin 31. This subfamily was formed approx. 500–600 million years ago, subsequent to the gene duplication that produced annexin 1. It has diverged rapidly and extensively, especially in the well-conserved and functionally critical type II calcium-binding sites. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Chromosomal localization; Fluorescence in situ hybridization; Gene duplication; Molecular evolution; Molecular sequence data; Nucleotide substitution rates
1. Introduction Phylogenetic and genetic studies have confirmed that the annexin tetrad is a unique, ancient and highly conserved structure (Morgan and Fernandez, 1995, 1997a) with some membrane-related role involving calcium channels, cellular signaling, vesicular transport or extracellular matrix (Raynal and Pollard, 1994). Vertebrate annexins are known to have emanated from a common eukaryotic ancestor since the emergence of metazoa around 800 million years ago (Mya) and their independent evolution at dispersed genomic locations implies differentiated, non-redundant functions (Morgan and Fernandez, 1997b; Morgan et al., 1998). The structural and evolutionary interrelationships of annexin genes provide vital information for establishing prospective links to functionally related genes, shared * Corresponding author. Tel: +34-98-510-4214; Fax: +34-98-510-3157; e-mail [email protected] Abbreviations: aa, Amino acid(s); ANXn, human annexin gene or locus (suffixed by subfamily number); DAPI, diamidino2-phenylindole; EST, expressed sequence tag; Myr/Mya, million years/ago.
phenotypes and, ultimately, to specific hereditary diseases. A complete knowledge of all extant annexins and their systematic analysis are essential for precisely identifying conserved regions that have a common, critical function and variable regions that distinguish the individual members by their regulation, properties and subcellular roles. Ten human annexins have been discovered on the basis of diverse functional assays and sequenced between the years 1986 and 1992 (Raynal and Pollard, 1994). The recent identification of an eleventh human annexin featuring lost calcium-binding sites, an RGD-like cell attachment motif and a distinctive expression pattern has raised key questions about the role of membrane interactions in annexin function and the actual diversity of this multigene family (Morgan and Fernandez, 1998). An examination of this gene’s chromosomal location, structural organization and phylogenetic place in relation to other annexins was therefore undertaken here to further characterize its unique identity and to resolve its true genetic origin. Such analyses could reveal some evolutionary pattern in annexin divergence and link this gene to novel, specialized functions or hereditary traits.
0378-1119/99/$ – see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 9 8 ) 0 0 59 7 - 6
34
R.O. Morgan et al. / Gene 227 (1999) 33–38
2. Materials and methods
3. Results and discussion
2.1. Chromosomal mapping by fluorescence in situ hybridization
3.1. Genetic mapping of ANX31
The A31H hybridization probe for chromosomal mapping was the 1.3 kb insert of expressed sequence tag ( EST ) clone ID 111373 from the I.M.A.G.E. (Integrated Molecular Analysis of Genome Expression) Consortium (LLNL) (Lennon et al., 1996). It was sequence verified to represent most of the coding and 3∞ untranslated region of a novel human annexin cDNA (Morgan and Fernandez, 1998; gb: AJ009985). Metaphase spreads were prepared from phytohemaglutinin-stimulated human lymphocytes of a healthy female donor (Fan et al., 1990). Fluorescence in situ hybridization and detection of immunofluorescence followed described protocols (Bell et al., 1995). The probe was labeled with biotin-16–dUTP (Boehringer Mannheim) by nick translation and hybridized overnight to denatured chromosomes at 37°C under stringent conditions. Hybridization sites were detected with fluoresceinlabeled avidin (Oncor) and amplified by addition of anti-avidin antibody (Oncor) and a second layer of fluorescein-labeled avidin. Chromosome preparations were counterstained with diamidino-2-phenylindole (DAPI ) and observed with a Zeiss Axiophot epifluorescence microscope equipped with a cooled charge coupled device camera (Photometrics, Tucson AZ ) operated by a Macintosh computer workstation. Digitized images of DAPI staining and fluorescein signals were captured, pseudocolored and merged using Oncor version 1.6 software.
The A31H cDNA probe hybridized to human metaphase spreads with specific labeling on chromosome 1 ( Fig. 1A). Fluorescence signals were detected on chromosome 1 in each of 23 metaphase spreads scored. Among 122 signals observed, 54 (44%) were on 1q with the following distribution: one chromatid (five cells), two chromatids (eight cells), three chromatids (seven cells), four chromatids (three cells). Among the 92 chromosome 1 chromatids scored, 54 (59%) hybridized to 1q. All chromosome-specific signals were localized to 1q21. This chromosomal band is contained within human–mouse homology group 4 (DeBry and Seldin, 1996), which is represented by more than 40 orthologous gene pairs also localized to human chromosome 1q21 and homologous with the mouse chromosome 3 region 42–49 cM from the centromere (Fig. 1B). Two potentially relevant associations are the neighboring S100 gene cluster and the epidermal differentiation complex (Scha¨fer et al., 1995). The former group consists of at least 13 genes for type I calcium-binding proteins, defined by the EF-hand conformation of their protein a-helices, and comprising members known to bind the amino-termini of certain annexins and influence their function (namely S100A3, S100A6-calcyclin, and S100A10-p11). The latter group includes loricrin, involucrin, filaggrin and trichohyalin with defined roles in skin growth and differentiation and a possible functional or regulatory relationship to annexin 31. 3.2. Structural characterization of anomalous ESTs
2.2. Molecular systematics Pairwise sequence comparisons employed programs from the FASTA package version 3.1 (Pearson, 1990) and multiple alignments were assembled using CLUSTALW version 1.6 ( Thompson et al., 1994). Database search and retrieval were done with the BLAST and RETRIEVE netservers of the National Center for Biotechnology Information (NCBI, Bethesda, MD) on the database of expressed sequence tags (dbEST, Boguski et al., 1993). Phylogenetic analyses utilized programs from the Phylogeny Inference Package (PHYLIP release 3.57) ( Felsenstein, 1989) for statistical analysis of branching order by sequence alignment bootstrapping and for tree reconstruction under various evolutionary models. Protein maximum likelihood analyses used the PUZZLE program version 4.0 (Strimmer and von Haeseler, 1996) to resolve trees by quartet puzzling under a discrete c-distribution rate model (a-parameter 1.89±0.11). The branch length test ( Takezaki et al., 1995) was used to verify rate linearity among taxa for annexin genes.
The dbEST provides sequence and expression data for cDNAs from transcribed genes and can also be a valuable genomic resource for normal and alternative exon splicing patterns, because it contains many incompletely processed transcripts ( Wolfsberg and Landsman, 1997). The characterization of anomalous ESTs is especially important for resolving identification errors in public databases, such as Unigene clusters and the Human Gene Map (NCBI ), and such discrepancies apply to annexin 31. We estimate, from our analyses of annexin ESTs (Morgan and Fernandez, 1998), that approx. 5% of these transcripts contain unspliced introns and have confirmed their authenticity in all cases where annexin gene structures have been determined. Two EST clone IDs 188508 and 137481 were tentatively identified as partially processed genomic transcripts of human annexin 31 ( Fig. 2), analogous to those detected previously for mouse annexin 3 ( Fernandez et al., 1998a). The former EST contained 237 bp of the 3∞ penultimate intron plus 25 bp of coding exon after the splice ‘cccctactag/GAA ACT’. The latter was represented by
R.O. Morgan et al. / Gene 227 (1999) 33–38
35
of all known vertebrate annexins except annexin 7 (Morgan and Fernandez, 1997b), and this has been corroborated by genomic polymerase chain reaction sequencing as part of initial efforts to characterize the gene structure. The segment of annexin 31 cDNA corresponding to exon 8 differed in size from that in annexin 6b (54 bp) and annexins 1, 3, 5 and 6a (57 bp), but was congruous with annexin 2 in which exon 8 is 60 bp due to a single codon insertion (Spano et al., 1990). This similarity with annexin 2 in the regions of both exons 8 and 12 provided preliminary evidence for homologous gene structures, and protein comparisons further strengthened their association. The characteristic amino-terminal tail of annexin proteins comprised a 25-aa unique region in the 338-aa complete annexin 31 (Morgan and Fernandez, 1998; gb: AJ009985) and this was most similar in size to the corresponding segment in annexin 2 (24 aa) and annexin 1 (33 aa), in contrast to other human annexins with much shorter or longer amino termini (range 6–191 aa). Homology testing of the 338-aa, full-length annexin 31 protein by sequence shuffling in PRSS-FASTA (Pearson, 1990) again identified annexin 2 with the highest Smith–Waterman alignment score of 842 and lowest probability of random occurrence of 4.78×10−55, while annexin 13 had the lowest homology index of 547. Since these latter two annexins lie at extreme branch tips of the annexin evolutionary tree (Morgan et al., 1998), it was worthy of testing whether annexin 31 might be directly related to annexin 2, both phylogenetically and genetically. 3.3. Phylogenetic classification of novel annexins
Fig. 1. Physical mapping of human annexin 31. (A) A human metaphase spread shows specific fluorescence in situ hybridization signals from a fluorescein-labeled pA31H probe at 1q21. The photograph represents computer-enhanced, merged images of fluorescein signals (arrowheads) and DAPI-stained chromosomes. (B) The cytogenetic idiogram of human chromosome 1 shows the band position containing the ANX31 locus, with a partial list of gene symbols and locus positions for neighboring genes known to be linked in the expansive human– mouse homology group 4.
two overlapping ESTs with 533 bp of intron containing a 5∞ dimeric Alu–Sxc insert followed by a typical 3∞ splice consensus ‘ttgtctgcag/GAT GCA’ and 70 bp coincident with the final exon of most vertebrate annexins. Annexin 31 was thus inferred to possess a 123 bp exon identical in size and location to the penultimate exon 12
The evolutionary position of annexin 31 was determined by statistical analysis of its alignment with other annexins to establish its clade association. The 309-aa tetrad core sequence of annexin 31 was aligned with representative members of the 27 previously classified subfamilies (Morgan and Fernandez, 1997a,b) plus other full-length proteins recently characterized in plants and invertebrates. Randomized bootstrap alignments of 143 informative sites (Morgan et al., 1998; Fernandez et al., 1998b) were first generated for analysis by maximum parsimony and by the distance matrix algorithms for neighbor-joining and least-squares ( Felsenstein, 1989). The resulting consensus phylogram was consistent with previous cladistic analysis of annexin subfamilies (Morgan and Fernandez, 1997a, 1998) and with maximum likelihood analysis of all 311 sites by the quartet puzzle method (Strimmer and von Haeseler, 1996). The protein maximum likelihood tree (Fig. 3) was therefore computed using the branching order topology of the consensus tree to estimate branch lengths and confidence values for each node. We extended our previous classification of annexin
36
R.O. Morgan et al. / Gene 227 (1999) 33–38
Fig. 2. Characterization of anomalous ESTs for annexin 31. The annexin 31 cDNA structure (top outline, gray means untranslated, checkered denotes matching exons 8 and 12) is aligned with sequenced segments of EST clone IDs 188508 and 137481 below. The latter represent partially processed genomic transcripts derived from coding exons with unspliced introns (dashed lines), one containing tandem Alu–Sx short, interspersed, repetitive elements (SINEs). The remaining outlines describe the four internally homologous domains of the annexin tetrad core conserved region (between vertical boundaries), and the exon blocks (boxed segments with nucleotide number) summarize the known, minor variations in different annexin gene structures. Annexin 7 is shown with several distinct exons, including exon 12, while other characterized genes differ by only a singlecodon variation in the size of exon 8. The complete human annexin 31 cDNA sequence has been deposited in EMBL/GenBank/DDBJ sequence databases under accession number AJ009985.
subfamilies (Morgan and Fernandez, 1997a) by including other recent database entries with complete coding regions in the phylogenetic analysis ( Fig. 3). Tobacco cDNA clones X511 and X671 from Nicotiana tabacum (Schantz et al., unpublished; gb: Y14972, Y14973) were confirmed with 97% probability to belong with Capsicum annum (green pepper) (Proust et al., 1996) and tomato p34 (Lim et al., 1998) in the previously classified annexin 24 subfamily. Cotton annexin AnnGh2 from Gossypium hirsutum (Potikha and Delmer, 1997) separated together with annexins 18 and 22 from other plant annexins with 99% probability, but its poor bootstrap association (58%) with either of these subfamilies led to its assignment with tomato annexin p35 (Lim et al., 1998) to the new annexin 28 subfamily. The paralogous cotton annexin AnnGh1 (Potikha and Delmer, 1997) and annexin AnxLt1 from leaves of Lavatera thuringiaca ( Vazquez-Tello and Uozumi, unpublished; gb: AF006197), clearly separated from known plant annexins (92%) and together (100%) formed a distinct annexin 29 subfamily. A fourth annexin gene C37H5.1 from the nematode Caenorhabditis elegans ( Wilson et al., 1994; gb: U88315) had a unique exon splicing pattern and no clear association with known invertebrate annexin subfamilies, so it was designated annexin 30 to represent yet another new subfamily. The novel human annexin was assigned to subfamily 31, based on its bifurcation with high confidence level from a peripheral node of one vertebrate annexin clade (Fig. 3) and its distinct chromosomal locus ( Fig. 1), which confirmed paralogy with ten other human annexins. Its nomenclature followed customary subfamily enumeration in the order of discovery of full-length sequences (including those above), consistent with cladistic and genetic analyses, and adopting Arabic num-
bering in place of the now cumbersome Roman numerals. Annexin 31’s common ancestry with annexin 2 was supported by an approximate maximum likelihood value of 62% and a consensus bootstrap frequency of 79% (cf. 88% by parsimony, 74% by neighbor-joining and 76% by least squares analyses) that may have been partially compromised by the inclusion of some misinformative sites, the protein’s extensive sequence divergence and its atypical aa composition. More rigorous testing was desirable to confirm its direct origin. 3.4. Common genetic origin of annexins 2 and 31 The relationship of annexins in the divergent 1–2–31 subclade was re-examined with a focus on ascertaining their respective origins and estimating the dates of probable gene duplication. Rather than presuppose informative sites (above), a reduced number of vertebrate sequences comprising the relevant clade (i.e., annexins 11, 6b, 3, 1, 2 and 31) and phylogenetically more primitive species were analyzed to better delineate the branching order. These now included teleost fish species representing previously unclassified and unpublished members of the annexin 1 and 11 (Osterloh et al., 1998), annexin 2 (Fujiki et al., unpublished; gb: C88389, C88439) and annexin 3 ( Elgar, 1996) subfamilies. Site rate heterogeneity was taken into account under a discrete c model, evolution parameters were estimated precisely and exact maximum likelihood values were calculated using the quartet PUZZLE program (Strimmer and von Haeseler, 1996). The resulting maximum likelihood tree ( Fig. 4) confirmed the branching order of divergent annexins 1, 2 and 31 among the most recently formed vertebrate subfamilies, although their possible origin from annexins 3, 6b or 11 could not be
R.O. Morgan et al. / Gene 227 (1999) 33–38
Fig. 3. Phylogram of the annexin superfamily containing human annexin 31. Protein tree topology was first determined from the consensus of analyses by neighbor-joining, least squares distance and maximum parsimony of 600 bootstrap alignments at 143 protein informative sites in representatives from all known annexin subfamilies (Morgan et al., 1998). Protein maximum likelihood analysis by quartet puzzling (Strimmer and von Haeseler, 1996) confirmed the overall tree topology ( log likelihood −16392) and computed the maximum likelihood branch lengths according to the scale of estimated aa replacements/site. Statistical support for the branching order is given by the percentage values above 50% at resolved nodes for maximum likelihood (bootstrap consensus values in parentheses). Full-length annexins recently described in plants and C. elegans were assigned to subfamilies 24, 28, 29 and 30 (starred labels) and the novel human annexin was sequentially assigned to subfamily 31 as ANX31-Hsa. Genus-species (Gsp) names and references for the newly classified annexins are given in Results Section 3.3, others have been described previously (Morgan and Fernandez, 1997a)
ascertained to a maximum likelihood value greater than 50. The most recent common ancestor of annexins 1 and 2 was previously estimated to have existed about 500 Mya, based on gene mutation rate determinations from rodent–mammal sequence comparisons (Morgan et al., 1998). Values of non-synonymous nucleotide substitutions (Li, 1993) recalculated from fish–mammal comparisons of those particular annexins and calibrated to those species separation times now predicted evolution rates half of the previous estimate for annexin 1 and double that for annexin 2. Since the separation time between fish and humans spans a much greater period of these genes’ lifetime, these were judged to be more
37
Fig. 4. Phylogenetic confirmation of common ancestors for annexins 1, 2 and 31. Parameter estimation and site rate heterogeneity were evaluated by PUZZLE to obtain exact maximum likelihood values (i.e., numbers at nodes) for the homologous protein pairs (312 aa) from each subfamily associated with the annexin 1–2–31 subclade. The evolutionary order of annexins 11 (assigned root), 3 and 6b was not resolved, but annexins 1, 2 and 31 were confirmed with high probability to be subsequent descendents, and the latter two genes shared the most recent common ancestor in this clade. Individual branch lengths reflect the numbers of estimated aa replacements per site and the average distance between all taxa was 1.16. Genus–species abbreviations for taxa include Cca, Cyprinus carpio (carp); Fru, Fugu rubripes (Japanese pufferfish); Gga, Gallus gallus (chicken); Hsa, Homo sapiens (human); and Ola, Oryzias latipes (Japanese medaka fish).
suitable for recalibrating the date of the annexin 2<1 gene duplication closer to 750 Mya. The application of a branch length test ( Takezaki et al., 1995) to annexin 6b, 3, 1 and 2 sequences from Fig. 4 further confirmed that these taxa exhibited a uniform rate of change according to a molecular clock. However, fish annexin sequences need to be completed and the rapid mutation rate of human annexin 31 (cf. branch length disparity between human annexins 2 and 31) requires assessment against more distant orthologues for accurate molecular dating. The current phylogenetic data and limited rate uniformity permitted the inference that annexin 31 duplicated from a common ancestor with annexin 2 prior to the separation of fish from eukaryotic lineage, possibly 500–600 Mya, and this would be substantiated by the eventual identification of annexin 31 in fish. Partial mouse annexin 31 sequences have been obtained (unpublished ) that appear to confirm both the rapid mutation rate and obliteration of type II calcium binding sites in this unusual annexin. The relatively recent fixation of annexin 31 in vertebrate genomes represents a significant evolutionary step in the structural, regulatory and presumably functional diversification of this superfamily.
38
R.O. Morgan et al. / Gene 227 (1999) 33–38
4. Conclusions (1) Human annexin 31 resides on chromosome band 1q21, remote from ten other annexins but in proximate linkage to S100A and skin differentiation genes with which it may have some regulatory or functional relationship. (2) Its coding region and protein structure most closely resemble annexin 2, but its unique lack of calciumbinding sites and lower expression level contrast markedly with the ubiquitous annexin 2 and imply their distinct subcellular location and significant functional divergence. (3) Phylogenetic analyses established that annexin 31 was a mutual gene duplication product with annexin 2 approx. 500–600 Mya and that it has been evolving at a comparatively accelerated rate. (4) Unspliced introns in dbEST transcripts can be informative for gene splicing patterns, and annexin sequences from early diverging species can provide a more reliable index of average gene mutation rates. Acknowledgements This work was supported by grant PM95-0152 from D.G.E.S. of Spain, National Institutes of Health Grant CA-06927 ( USA), and by an appropriation from the Commonwealth of Pennsylvania.
References Bell, D.W., Taguchi, T., Jenkins, N.A., Gilbert, D.J., Copeland, N.G., Gilks, C.B., Zweidler-McKay, P., Grimes, H.L., Tsichlis, P.N., Testa, J.R., 1995. Chromosomal localization, in man and rodents, of a gene (GFI1) encoding a novel zinc finger protein reveals a new syntenic region between mouse and man. Cytogenet. Cell Genet. 70, 263–267. Boguski, M.S., Lowe, T.M., Tolstoshev, C.M., 1993. dbEST — database for expressed sequence tags. Nat. Genet. 4, 332–333. DeBry, R.W., Seldin, M.F., 1996. Human/mouse homology relationships. Genomics 33, 337–351. Elgar, G., 1996. Quality not quantity: the pufferfish genome. Hum. Mol. Genet. 5, 1437–1442. Fan, Y.S., Davis, L.M., Shows, T.B., 1990. Mapping small DNA sequences by fluorescence in situ hybridization directly on banded metaphase chromosomes. Proc. Natl. Acad. Sci. USA 87, 6223–6227. Felsenstein, J., 1989. PHYLIP — Phylogeny inference package. Cladistics 5, 164–166. Fernandez, M.P., Copeland, N.G., Gilbert, D.J., Jenkins, N.A., Morgan, R.O., 1998a. Mouse annexin III cDNA, genetic mapping and evolution. Gene 207, 43–51. Fernandez, M.P., Copeland, N.G., Gilbert, D.J., Jenkins, N.A.,
Morgan, R.O., 1998b. The genetic origin of mouse annexin VIII. Mamm. Genome 9, 8–14. Lennon, G.G., Auffray, C., Polymeropoulos, M., Soares, M.B., 1996. The I.M.A.G.E Consortium: an integrated molecular analysis of genomes and their expression. Genomics 33, 151–152. Li, W.H., 1993. Unbiased estimation of the rates of synonymous and nonsynonymous substitution. J. Mol. Evol. 36, 96–99. Lim, E.K., Roberts, M.R., Bowles, D.J., 1998. Biochemical characterization of tomato p35. Independence of calcium binding and phosphatase activities. J. Biol. Chem. 273, 34920–34925. Morgan, R.O., Fernandez, M.P., 1995. Molecular phylogeny of annexins and identification of a primitive homologue in Giardia lamblia. Mol. Biol. Evol. 12, 967–979. Morgan, R.O., Fernandez, M.P., 1997a. Distinct annexin subfamilies in plants and protists diverged prior to animal annexins and from a common ancestor. J. Mol. Evol. 44, 178–188. Morgan, R.O., Fernandez, M.P., 1997b. Annexin gene structures and molecular evolutionary genetics. Cell. Mol. Life Sci. 53, 508–515. Morgan, R.O., Fernandez, M.P., 1998. Expression profile and structural divergence of novel human annexin 31. FEBS Lett. 434, 300–304. Morgan, R.O., Bell, D.W., Testa, J.R., Fernandez, M.P., 1998. Genomic locations of ANX11 and ANX13 and the evolutionary genetics of human annexins. Genomics 48, 100–110. Osterloh, D., Wittbrodt, J., Gerke, V., 1998. Characterization and developmentally regulated expression of four annexins in the killifish medaka. DNA Cell Biol. 17, 835–847. Pearson, W.R., 1990. Rapid and sensitive sequence comparison with FASTP and FASTA. Meth. Enzymol. 183, 63–98. Potikha, T.S., Delmer, D.P., 1997. cDNA clones for annexin AnnGh1 (accession no. U73746) and AnnGh2 (accession no. U73747) from Gossypium hirsutism (cotton). Plant Physiol. 113, 305 Proust, J., Houlne, G., Schantz, M.L., Schantz, R., 1996. Characterization and gene expression of an annexin during fruit development in Capsicum annum. FEBS Lett. 383, 202–212. Raynal, P., Pollard, H.B., 1994. Annexins: the problem of assessing the biological role for a gene family of multifunctional calcium- and phospholipid-binding proteins. Biochim. Biophys. Acta 1197, 63–93. Scha¨fer, B.W., Wicki, R., Engelkamp, D., Mattei, M.G., Heizmann, C.W., 1995. Isolation of a YAC clone covering a cluster of nine S100 genes on human chromosome 1q21: rationale for a new nomenclature of the S100 calcium-binding protein family. Genomics 25, 638–643. Spano, F., Raugei, G., Palla, E., Colella, C., Melli, M., 1990. Characterization of the human lipocortin-2-encoding multigene family: its structure suggests the existence of a short amino acid unit undergoing duplication. Gene 95, 243–251. Strimmer, K., von Haeseler, A., 1996. Quartet puzzling: a quartet maximum likelihood method for reconstructing tree topologies. Mol. Biol. Evol. 13, 964–969. Takezaki, N., Razhetsky, A., Nei, M., 1995. Phylogenetic test of the molecular clock and linearized trees. Mol. Biol. Evol. 12, 823–833. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680. Wilson, R., Ainscough, R., Anderson, K., Baynes, C., Berks, M., Bonfield, J. et al., 1994. 2.2 Mb of contiguous nucleotide sequence from chromosome III of C. elegans. Nature 368, 32–38. Wolfsberg, T.G., Landsman, D., 1997. A comparison of expressed sequence tags (ESTs) to human genomic sequences. Nucleic Acids Res. 25, 1626–1632.