Mouse Annexin V Chromosomal Localization, cDNA Sequence Conservation, and Molecular Evolution

GENOMICS

31, 151–157 (1996) 0026

ARTICLE NO.

Mouse Annexin V Chromosomal Localization, cDNA Sequence Conservation, and Molecular Evolution M. ISABEL RODRIGUEZ-GARCIA, CHRISTINE A. KOZAK,* REGINALD O. MORGAN, AND M. PILAR FERNANDEZ1 Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Oviedo, E-33006 Oviedo, Spain; and *National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892 Received July 17, 1995; accepted November 2, 1995

A full-length cDNA encoding mouse annexin V (ANX5) was cloned, sequenced, and utilized for chromosomal mapping. The gene lies on mouse chromosome 3 in close linkage with the fibroblast growth factor 2 (basic) gene and is syntenic with other genes known to have orthologous counterparts on human chromosome 4q. The open reading frame encoded a protein of 319 amino acids (aa), with 92–96% identity to ANX5 in other species. Internal repeat 3 of mouse ANX5 exhibited the highest level of nonconservative aa replacements with respect to other annexin subfamilies, but the greatest sequence conservation among ANX5 species members. This region may thus contain features that distinguish ANX5 from other annexins in properties or function. Phylogenetic analysis and homology testing of ANX5 members indicated that the 34-kDa annexin from Torpedo marmorata may also belong to this subfamily. Comparison of nine species of ANX5 led to an estimation of the unit evolutionary mutation rate at 1% aa replacements every 8 million years, comparable to other annexins. q 1996 Academic Press, Inc.

INTRODUCTION

Annexin V (ANX5)2 is the common name, after a history of nomenclature changes, for a prototypical member of a multigene family with diverse biological actions and an uncertain physiological role (Raynal and Pollard, 1994). ANX5 was first described in primates as a major placental anticoagulant protein (Bohn, 1979) and independently recognized as a collagen-binding protein in chick cartilage (Mollenhauer and von der Mark, 1983; Fernandez et al., 1988). Human and chick Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under Accession No. U29396. 1 To whom correspondence should be addressed. Telephone: (34 8) 510 4214. Fax: (34 8) 510 3534. E-mail: [email protected]. uniovi.es. 2 Abbreviations used: aa, amino acid(s); ANX5, human annexin V gene or locus; Anx5, nonhuman annexin V gene or locus; ANX5, annexin V protein; anx5, annexin V cDNA or probe; bp, basepair(s); Myr, million years; nt, nucleotide(s); PCR, polymerase chain reaction; RACE, rapid amplification of cDNA ends.

ANX5 have been characterized as housekeeping genes based on their structures (Fernandez et al., 1994a,b) and their broad expression pattern compared to other annexins (van Heerde et al., 1995). Their unique 5*termini presumably account for distinct regulation by growth stimuli such as the c-fos protooncogene (Braselmann et al., 1992). ANX5 proteins have relatively short N-termini of 4–6 aa lacking phosphorylation sites, a feature that has made them attractive for the study of structure–function relationships within the four homologous repeats that harbor calcium- and phospholipid-binding domains and calcium channel function common to other annexins (Demange et al., 1994; van Heerde et al., 1995). Characteristic annexin properties such as collagen binding and the inhibition of phospholipases A2 and protein kinase C are more likely to have a molecular basis in the tetrad core region but these functional domains have not yet been localized (Raynal and Pollard, 1994). The human gene has been mapped to chromosome 4q26–q28 (Tait et al., 1991) and the chick gene to linkage group C11 (Bumstead et al., 1994). No specific genetic defects nor characteristic phenotypes have yet been identified. The comparative study of distinct annexin subfamilies in a range of species can help to identify conserved domains responsible for a common function(s) as well as variable domains that may signify structural or regulatory divergence from other subfamily members (Barton et al., 1991). An understanding of annexin evolutionary history can also be useful for tracing their origin(s), relative mutation rates, and functional diversification (Morgan and Fernandez, 1995). We have therefore cloned and sequenced the cDNA for mouse annexin V, primarily to characterize ANX5 sequence conservation and to generate a probe for gene mapping. This has also led to the identification of a related homologue and assessment of the gene’s mutation rate. MATERIALS AND METHODS cDNA cloning and sequencing. A cDNA library prepared from the mouse monocyte–macrophage cell line RAW 264.7 (American Type Culture Collection, Rockville, MD) was provided by Dr. F. Segade

151

/ m4752f3838

01-03-96 05:26:51

gnma

0888-7543/96 $12.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

AP-Genomics

152

RODRIGUEZ-GARCIA ET AL.

(Segade et al., 1995). It was screened using a rat anx5 DNA probe generated by reverse transcription of rat liver RNA, followed by polymerase chain reaction (RT-PCR). Reverse transcription was primed with an antisense oligonucleotide corresponding to rat cDNA positions 1010–1036 (Pepinsky et al., 1988). PCR was carried out with the same antisense primer and the upstream primer corresponding to rat cDNA positions 62–88. Positive hybridization plaques of the insert containing LambdaZAPII clones were identified with the radiolabeled rat cDNA probe and purified by standard protocols (Sambrook et al., 1989). Excision and recircularization of the cloned insert contained within the LambdaZAPII vector were performed according to instructions accompanying the cDNA synthesis kit (Stratagene). DNA sequencing of selected clones was performed by the dideoxy chain termination method with [a-35S]dATP (Amersham) and the Sequenase 2.0 kit (US Biochemical). Determination of the transcription initiation site. The 5* end of anx5 mRNA from mouse liver was defined by primer extension analysis and by Rapid Amplification of cDNA Ends (RACE). Primer extension was performed according to standard protocols (Sambrook et al., 1989) using a 32P end-labeled antisense oligonucleotide (5*-ATGGCCTTCCGAAGGACTTC) corresponding to mouse cDNA sequence positions 158–177. The size of the primer-extended product was determined on a sequencing gel. 5*-RACE was carried out to isolate and identify the 5* terminal sequence of mouse liver anx5 mRNA. The first strand of cDNA was obtained by reverse transcription using the same primer as above. It was poly(A)-tailed using terminal transferase (GIBCO-BRL) and subjected to PCR with oligo(dT) as upstream primer and antisense oligonucleotide corresponding to cDNA positions 158–177 or 52–72 as downstream primer. PCR products were cloned and sequenced. Genetic mapping. The progeny of two genetic crosses were typed for Anx5 polymorphisms by Southern blotting: (NFS/N or C58/J 1 Mus musculus musculus) 1 M. m. musculus (NMM; Kozak et al., 1990) and (NFS/N 1 Mus spretus) 1 M. spretus or C58/J (NSS, NSC; Adamson et al., 1991). These crosses have been typed for inheritance of over 850 markers, including the chromosome 3 markers Fgf2 (fibroblast growth factor 2, basic), Il2 (interleukin 2), Evi1 (ecotropic viral integration site-1), Mtv56 (mammary tumor virus 56), Tpi-rs5 (triosephosphate isomerase, related sequence 5), and Gba (glucocerebrosidase). Gba, Fgf2, and Evi1 were typed as previously described (Pandey et al., 1994; Kozak et al., 1995). Mtv56 was typed as a 4.2kb PvuII spretus fragment reactive with an MMTV envelope probe (Majors and Varmus, 1981) obtained from Dr. R. Callahan (NCI, NIAID, Bethesda, MD). Tpi-rs5 was identified as a ScaI fragment of 5.0 kb in spretus using a pHTPI-5A (Brown et al., 1985) obtained from Dr. L. E. Maquat (Roswell Park Cancer Institute, Buffalo, NY) as probe. Il2 was typed following digestion with PvuII or HindIII in the M. spretus cross (Yokota et al., 1985) using pMUT-1 obtained from the American Type Culture Collection as probe. Computer sequence analyses. Genetic data were stored and analyzed using the program LOCUS developed by C. E. Buckler (NIAID, NIH, Bethesda, MD). Recombinational distances and standard errors were calculated according to Green (1981), and loci were ordered by minimizing recombinants. Current databases of gene loci chromosomal maps were retrieved from network servers of the human Genome DataBase (GDB; Fasman et al., 1994) and the Mouse Genome Database (MGD; Prins et al., 1994; http://www.informatics.jax.org/ mgd.html). Mouse annexin V cDNA and protein sequences were compared with other annexin sequences retrieved from current GenBank, EMBL, PIR, and SwissProt sequence libraries, using the RETRIEVE server of the National Center for Biotechnology Information (Bethesda, MD). These included genomic DNA sequences for human ANX5 and chick anx5 (Fernandez et al., 1994a,b), rat anx5 cDNA (Pepinsky et al., 1988), bovine ANX5 protein isoforms 33 and 37 kDa (Learmonth et al., 1992), and ANX5 peptide fragments from rabbit (Okabe et al., 1993), pig (Baldwin et al., 1991), hamster (Ideta et al., 1995), and the 34-kDa ‘‘calelectrin’’ of Torpedo marmorata (Geisow et al., 1986). Protein sequences were analyzed by PATMAT for similarity to alignments of functional motifs contained in Release 8.0 of the BLOCKS database (Henikoff and Henikoff, 1994). The latter is

/ m4752f3838

01-03-96 05:26:51

gnma

derived from the PROSITE pattern database and contains short, conserved regions of alignment of multiple proteins with a common defined property or function. A search window of 30 residues was used and match scores over 1000 for a particular block with a probe sequence signified similarity better than 99.5% of spurious matches with remaining proteins in the SwissProt protein database. Molecular phylogeny analysis of annexin peptide fragments was performed with distance (PROTDIST, FITCH, and CONSENSE) and parsimony (PROTPARS) programs from the Phylogeny Inference Package, PHYLIP (Felsenstein, 1989). Bootstrap sampling of input data was used for statistical validation, and distance calculations were based on the Dayhoff substitution matrix.

RESULTS

Mouse Annexin V cDNA and Protein Sequences Annexin V cDNA clones were isolated from a mouse macrophage cDNA library. Three independent LambdaZAPII clones were identified by hybridization to the rat anx5 cDNA probe and sequenced. The 1498-bp cDNA sequence for mouse anx5 (GenBank Accession No. U29396) extended from the 5*-end, determined by primer extension and RACE, to the poly(A) tail. It contained 933 bp in the tetrad core region, homologous with other annexins, flanked by regions of sequence similarity limited to other anx5 species. The open reading frame encoded a 319-aa protein (Fig. 1) with a calculated molecular weight of 35,712 Da and showed 93.9% nucleotide (nt) identity and 95.9% aa identity with its rat homologue. Acidic Glu residues at positions 72, 89, and 106 (Fig. 1), believed to be critical determinants of calcium channel activity (Demange et al., 1994), and the type II calcium-binding sites in repeats 1, 2, and 4 were all fully preserved. Rodent anx5 differed slightly from human, chick, and bovine sequences in their 5*-termini, where the third and fourth codons were missing, and in the 3*-termini, where the human sequence had a unique deletion of the fourth from last codon for Gly-312 (Fig. 1). The core tetrad region had a unique, nonconservative aa replacement of Arg to Ser at position 117 in repeat 2. Three motifs that could represent potential protein kinase C phosphorylation sites at Thr-37, Thr-68, and Ser-237 coincided with the same internal residues in rat or bovine ANX5, which are known to not be cellular targets of this enzyme (Learmonth et al., 1992). The deduced protein sequence for mouse ANX5 was compared with other known annexins to identify distinguishing features. Two consensus sequences were obtained by residue frequency analysis of 20 distinct subfamilies (10 human and 10 invertebrate ANX) and 9 species representatives within the ANX5 subfamily. Among the aa replacements in mouse ANX5 compared to the general ANX consensus, some conferred a change in identity or physicochemical property that was also preserved in other ANX5 subfamily members. The aligned, 69-aa core of ANX5 repeat 3 exhibited the highest number of nonconservative aa replacements (28%) with respect to other annexins (reverse-shaded residues in Fig. 1), yet it also represented the region

AP-Genomics

MOUSE ANNEXIN V cDNA

153

FIG. 1. Sequence comparison of mouse annexin V protein with consensus sequences for other annexin subfamilies (ANX) and for members of the same subfamily (ANX5). The ANX consensus comprised the most frequent residues in the homologous tetrad core region of 20 distinct annexin subfamilies (Morgan and Fernandez, 1995). Position numbers refer to this 311-aa sequence, displayed with alignment of the 68- to 69-aa segments exhibiting similarity across individual repeats. The ANX5 consensus represents nine species in this subfamily, and the mouse protein sequence (ANX5 mouse) was deduced from the new cDNA (GenBank Accession No. U29396). Mouse and consensus ANX5 sequences were compared with the general ANX consensus, and identical aa are represented by dots, differences by the standard single-letter abbreviation for aa, and gaps by dashes. Stippled residues denote aa replacements in the mouse or consensus ANX5 at positions of variability or constituting a conservative change in physiochemical property from the general ANX consensus. Reverse-shaded residues represent a significant change in physicochemical property with greater than 80% conservation among ANX5 subfamily members. Regions identified below the mouse ANX5 sequence refer to analogous domains and their PATMAT similarity score (in parenthesis) based on comparison with known motifs in other protein families (Henikoff and Henikoff, 1994).

with highest mean sequence identity (97%) within the ANX5 subfamily. These characteristics were reminiscent of annexin N-terminal regions that are conserved exclusively within subfamilies and contrasted with the known variablility of repeat 3 across different subfamilies (Morgan and Fernandez, 1995). One particular segment at aa 207–229 (Fig. 1) showed absolute conservation of aa and corresponding nt identity in all known species of ANX5 and corresponded to the initial 23 codons of exon 10 in human and chick genomic sequences (Fernandez et al., 1994a,b). The mouse ANX5 sequence was compared to 2884 entries in the BLOCKS database of protein functional domains in an attempt to identify regions of structural similarity with other proteins that might possess ‘‘analogous’’ properties or functions. PATMAT ranked known functional motifs according to their similarity with regions of mouse ANX5, and scores over 1000 were regarded as potentially significant. Among the highest scores were regions of mouse ANX5 that aligned with the nucleotide binding sites of ABC transporters, histone H2B, RNP-1 RNA binding, and DNA topoisomerase II, respectively (Fig. 1). Other segments showed similarity to the lipid-binding site of fatty-acid-binding proteins, to calsequestrin or the cyclin signature in re-

/ m4752f3838

01-03-96 05:26:51

gnma

peat 1, or to GTP-binding protein receptors in repeat 4. Borderline scores were also obtained with alignments to cytoskeletal protein-binding sites, including those of vinculin, neuraxin, intermediate filaments, cadherins, and integrins. These results were consistent with known properties of annexins, but the assignment of specific functional domains was precluded by the modest scores and apparent multiple locations. Chromosomal Localization Southern hybridization employed a KpnI–HindIII cDNA fragment of 1116 bp as probe. This was used to identify PvuII fragments of 4.4 and 1.5 kb in m. spretus and 9.8, 8.6, 7.0, 6.4, 5.0, 4.8, 3.0, 2.8, and 1.5 kb in NFS/N. BstEII fragments of 14.2 and 12.4 kb were identified in m. m. musculus and NFS/N DNAs, respectively. Inheritance of the variant fragments in the progeny of the crosses was compared with inheritance of over 700 markers previously typed in these crosses. The gene encoding mouse ANX5 (Anx5) was mapped to chromosome 3 (Fig. 2). Closest linkage was observed in the m. spretus crosses between Anx5 and Fgf2, for which no recombinants were identified in the 94 mice typed for both markers. Thus, at the 95% confidence

AP-Genomics

154

RODRIGUEZ-GARCIA ET AL.

could be excluded by the absence of a similar highmass 68-kDa homologue in Torpedo marmorata (Sudhof et al., 1984) and by the resulting anomaly in ANX6 species branching order (Fig. 3). Anx5 Mutation Rate

FIG. 2. Map location of Anx5 on mouse chromosome 3. The map to the left was derived from analysis of the cross (NFS/N 1 M. spretus) X C58/J, and the map to the right from the crosses (NFS/N 1 M. spretus) 1 M. spretus and (NFS/N or C58/J 1 M. m. musculus) 1 M. m. musculus. Recombination fractions are provided for adjacent markers, and the numbers in parentheses represent the recombinational distances and standard errors calculated according to Green (1981). To the extreme left are the map locations for human genes orthologous to the underlined mouse genes.

The extent of aa replacement in the 311-aa core region between human and mouse ANX5s (5.8%) was used to calibrate the gene’s relative and absolute mutation rates. For the same species pair, ANX6, 7, and 11 showed a similar mean change of 5.4%, while ANX1 was significantly higher (12.5%) and ANX2 lower (2.5%). Constancy of the ANX5 mutation rate across species was assessed by the ‘‘relative rate test,’’ which computed aa replacement as a distance measure between all pairs of species for which sequence data were available (Fig. 4, top). Comparison of each species with all earlier diverging species showed variation in the mean absolute values, but with a standard error (SE) less than 1.4% that indicated narrow precision and relative rate constancy. Thus, chick ANX5 showed a mean

level, these markers are within 3.1 cM. This map location places Anx5 close to Fgf2, Ccna, and Il2 markers previously mapped to human chromosome 4q25–q31 and is consistent with the earlier assignment of human ANX5 to 4q26–q28 (Tait et al., 1991). Molecular Phylogeny Two previously unclassified peptide fragments comprising 32 aa from the 34-kDa annexin of marbled electric ray Torpedo marmorata (Geisow et al., 1986) showed greatest similarity with conserved regions of rodent ANX5 in repeats 1 and 2. Their 75% aa identity with ANX5 proteins yielded optimal alignment zscores of 56 – 60 that were 9 – 11 standard deviations above the mean score for shuffled sequences, using the RSS program from the FASTA package (Pearson and Lipman, 1988). These data and other similarities in molecular mass, biochemistry, and immunological properties (Sudhof et al., 1984) provided a basis for phylogenetic determination of their relatedness to individual annexin subfamilies. The 32-aa fragments were aligned with corresponding fragments from all known human and invertebrate annexins, and 200 bootstrapped samples were analyzed by the protein distance matrix programs PROTDIST and FITCH. The resulting tree (Fig. 3) matched that obtained using maximum parsimony methods to determine the minimum mutation tree and established that the 34kDa calelectrin was most closely related to the ANX5 subfamily. Although bootstrap values did not achieve statistical significance, the closest alternative, ANX6,

/ m4752f3838

01-03-96 05:26:51

gnma

FIG. 3. Protein distance tree showing the evolutionary relationship of electric ray 34-kDa calelectrin (Anx5-Tma) to other annexins. Analyzed data consisted of a limited sequence alignment of the 32-aa fragments from Anx5-Tma (Geisow et al., 1986) with corresponding 32-aa segments from representatives of all known human and invertebrate annexin subfamilies. Statistical sampling by SEQBOOT generated 200 bootstrap alignments, and numbers at the branches specify the percentage of samples (probability) in which taxa to the right appeared in the output. Other programs from the PHYLIP package, viz. PROTDIST, FITCH, and CONSENSE, were used to compute the unrooted, consensus distance tree from the aa replacement values between all pairwise sequence comparisons. Labels identify annexin subfamily classifications and genus/species, where Bta refers to Bos taurus; Cel, Caenorhabditis elegans; Ddi, Dictyostelium discoideum; Dme, Drosophila melanogaster; Gga, Gallus gallus; Hsa, Homo sapiens; Hvu, Hydra vulgaris; Mmu, Mus musculus; Msa, Medicago sativa; Rno, Rattus norvegicus; and Tma, Torpedo marmorata.

AP-Genomics

155

MOUSE ANNEXIN V cDNA

the linear regression line for ANX5 was consistent with its previous identification as an ANX5 subfamily member based on sequence homology and phylogenetic analysis. DISCUSSION

FIG. 4. Annexin V protein mutation rate. (A) The relative rate test was applied to calculate the percentage of aa replacements between all pairs of ANX5 from nine different species, independent of their divergence times. The pair values and range of these distances are shown for each species compared with all others to the left. Chick and ray ANX5, to the right, comprised the greatest number of comparisons (seven each), and precision for all means was characterized by standard errors of 1.4% or less. (B) The linear regression plot (bold line) describes the percentage aa replacement in ANX5 of eight species, indicated with respect to the human protein as a function of their respective, estimated divergence times from human (Gould, 1993). The line equation y Å 0.127x (r Å 0.994) was used to calculate a unit evolutionary rate for ANX5 protein mutation at 1% aa replacements every 7.9 { 0.5 Myr. It is bordered by dashed lines describing the 95% confidence limits of the slope for ANX5 and by a shaded region showing the calculated confidence limits of all annexins for which comparison with a human ortholog was possible (n Å 51).

difference of 22.5 { 0.8% (SE) compared to seven mammalian species, and Torpedo ANX5 had a mean 41.0 { 1.4% aa replacement compared to seven earlier diverging species, shown to the left in Fig. 4A. Reliable estimates of absolute evolutionary rates depend on accurate estimates of divergence times as well as distance calculations. To the extent that the fossil record has provided reasonably valid times for familiar species (Gould, 1993), we observed a linear relationship for aa change with estimated divergence time between the human protein and ANX5 of pig, cow, hamster, rabbit, mouse, rat, chick, and electric ray (Fig. 4, bottom). The inverse of the slope could be used to calculate a unit evolutionary rate for ANX5 proteins of 1% aa replacements every 7.9 { 0.5 Myr. This mutation rate was in the upper-normal range for other annexins. The position of the marbled electric ray was based on 28% aa replacement from human ANX5 in 32-aa fragments and an estimated divergence time between primates and ray of 210 Myr (Gould, 1993). Its coincidence with

/ m4752f3838

01-03-96 05:26:51

gnma

Annexins are regarded as a unique class of calciumdependent, phospholipid-binding proteins because their homologous repeat regions have no known evolutionary precursor nor apparent functional analogues. We have isolated a new species of anx5 to assess its genomic location, characterize its similarities and differences with other homologues, and gain insight into its evolutionary history. The Anx5 locus was closely linked with Fgf2 on mouse chromosome 3 and syntenic with at least seven other genes previously mapped to orthologous loci on human chromosome 4. Genes for annexins I– VIII have been mapped to chromosomes 9, 15, 4, 2, 4, 5, 10, and 10, respectively, in human, while annexins I, II, IV, V, VI, and VII have been mapped to mouse chromosomes 19, 9, 6, 3 (this study), 11, and 14, respectively. These data confirm the broad dispersal of annexins in mammalian genomes (O’Brien, 1993) and may eventually contribute to a better understanding of the mechanisms of gene duplication. The primary structure of mouse annexin V reflected clear homology with other annexins and strong identity with the ANX5 subfamily. Key similarities included the preservation of acidic residues relevant to calciumbinding and calcium channel function (Demange et al., 1994). Similarities with analogous domains from other protein families suggested that mouse ANX5 may not be exclusively a calcium- and phospholipid-binding protein but may also have the capacity to interact with nucleotides, cytoskeletal proteins, or fatty acids. Such diverse properties could apply to other annexins as well, since PATMAT analysis of the ANX consensus yielded many of the same matches. Similarity of annexin repeat 1 to the ATP-binding domain of ABC transporter proteins is noteworthy because such a property for annexins was originally suggested by Geisow (1986), and recent evidence appears to confirm an overlap of calcium-, phospholipid-, and DNA-binding domains in annexin II (Boyko et al., 1994). The concept of fatty acids as annexin cofactors in exocytosis (Creutz, 1992) could involve physical interaction via the fattyacid-binding region identified in repeat 1. This type of analysis offers a new approach to detecting functional domains for experimental testing. The structural differences of mouse ANX5 from other annexins were of special interest because they identified regions with the potential to contribute specificity for evolutionary selection and functional differentiation of the ANX5 subfamily. Its short N-terminal and consequent lack of phosphorylation sites were consistent with the unique activity of human ANX5 as a nonsubstrate inhibitor of protein kinase C (Schlaepfer et al., 1992). A comparative study between various

AP-Genomics

156

RODRIGUEZ-GARCIA ET AL.

ANX5s and other subfamilies should permit localization and identification of the specific domain responsible for this C-kinase inhibition. Within the homologous tetrad, repeat 3 showed the most changes in physicochemical property from other annexin subfamilies and the greatest sequence conservation within the ANX5 subfamily. Comparison with the BLOCKS database identified the latter half of this repeat as having structural similarity with cytoskeletal protein-binding sites, which could be related to the characteristically strong collagen-binding property of ANX5 (Mollenhauer and von der Mark, 1983). Further studies of gene structure and regulation will be required to identify other distinguishing features of ANX5 responsible for its unique expression pattern, which is contrary to annexin I in its response to glucocorticoids and cell proliferation (Raynal and Pollard, 1994). Some annexins have been identified in plants and invertebrates, and a more extensive knowledge of ANX5 evolutionary history could provide perspective on its basic role in different organisms. After finding that the unclassified 34-kDa annexin from Torpedo marmorata had the highest identity with mouse ANX5, its close relatedness was supported by a statistical test of homology, phylogenetic analysis, and determination of a mutation rate consistent with other ANX5. Note that a more accurate picture of the relatedness between individual annexins can be achieved using complete, homologous sequences (Morgan and Fernandez, 1995), rather than the limited alignment corresponding to available Torpedo ANX5 fragments applied in the present classification. The results are, nevertheless, corroborated by previous studies of the intact protein, which had a similar molecular mass and immune cross-reactivity to 33-kDa mammalian annexins (Sudhof et al., 1984). The association of 34-kDa calelectrin with cholinergic synapses points to a possible involvement of annexin V in nerve function. It is therefore relevant that annexin V has been localized to glial cells of the nervous system (Spreca et al., 1992) and is a recognized neurotrophic factor (Takei et al., 1994). Interspecies sequence comparisons indicated that ANX5 had a mutation rate that was relatively stable in different species and in the upper range of normal for other annexins (Fig. 4). The linear relationship between ANX5 mutations and estimated species divergence times is consistent with the concept of a molecular clock (Wilson et al., 1977), whereby homologous regions of the gene and protein are changing at a constant rate independent of speciation. Comparisons of different annexins in similar species have indicated relatively atypical mutation rates for only annexins I, II, and III (Morgan and Fernandez, 1995). Absolute estimates, based on divergence times given by the fossil record (Gould, 1993), place annexins in the midrange of protein mutation rates, between conserved histones with a unit evolutionary period of 8–400 Myr and immunoglobulins with a 1% change in aa every 0.7–1.7 Myr (Wilson et al., 1977). An approximate unit evolu-

/ m4752f3838

01-03-96 05:26:51

gnma

tionary period of 8 Myr for ANX5 can be used to predict that a species that diverged from humans 600 Myr ago might still have a recognizable sequence retaining at least 25% identity to human ANX5. A constant genetic clock can also be used to infer divergence times for any pair of species from gene sequences for which a rate has been established. Since annexins have origins dating back to the earliest eukaryotes over 1000 Myr ago (Morgan and Fernandez, 1995), it is evident that studies of structure conservation and phylogenetic analysis will have great utility in identifying the immediate progenitor gene and species for ANX5. The availability of the mouse anx5 cDNA sequence now enables more detailed study of its genomic structure, tissue expression, and protein function in relation to other annexins. We anticipate that knowledge of its location in the genome will prove useful in the study of hereditary anomalies involving annexins and in the development of transgenic animal models. ACKNOWLEDGMENTS This work was supported by Grant PB92-1000 from DGICYT of Spain. We thank F. Segade for supplying the cDNA library, and M. R. Fernandez and M. T. Carcedo for helpful discussions.

REFERENCES Adamson, M. C., Silver, J., and Kozak, C.A. (1991). The mouse homologue of the gibbon ape leukemia virus receptor: Genetic mapping and a possible receptor function in rodents. Virology 183: 778– 781. Baldwin, G. S., Moritz, R. L., Rubira, M., Seet, K. L., Weinstock, J., and Simpson, R. J. (1991). Purification and partial amino acid sequence of annexin V from porcine gastric mucosal membranes. Comp. Biochem. Physiol. 100B: 661–665. Barton, G. J., Newman, R. H., Freemont, P. S., and Crumpton, M. J. (1991). Amino acid sequence analysis of the annexin supergene family of proteins. Eur. J. Biochem. 198: 749–760. Bohn, H. (1979). Placental and pregnancy proteins. In ‘‘Carcinoembryonic Proteins’’ (F. G. Lehman, Ed.), Vol. 1, pp. 289–299, Elsevier/North Holland Biomedical Press, Amsterdam. Boyko, V., Mudrak, O., Svetlova, M., Negishi, Y., Ariga, H., and Tomilin, N. (1994). A major cellular substrate for protein kinases, annexin II, is a DNA-binding protein. FEBS Lett. 345: 139–142. Braselmann, S., Bergers, G., Wrighton, C., Graninger, P., Supertifurga, G., and Busslinger, M. (1992). Identification of fos target genes by the use of selective induction systems. J. Cell Sci. 16(Suppl.): 97–109. Brown, J. R., Daar, I. O., Krug, J. R., and Maquat, L. E. (1985). Characterization of the functional gene and several processed pseudogenes in the human triosephosphate isomerase gene family. Mol. Cell. Biol. 5: 1694–1706. Bumstead, N., Fernandez, M. P., and Sillibourne, J. (1994). An XbaI RFLP at the locus encoding the chicken homologue of ANX5. Animal Genet. 25: 290. Creutz, C. E. (1992). The annexins and exocytosis. Science 258: 924– 931. Demange, P., Voges, D., Benz, J., Liemann, S., Gottig, P., Berendes, R., Burger, A., and Huber, R. (1994). Annexin V: The key to understanding ion selectivity and voltage regulation? Trends Biochem. Sci. 19: 272–276. Fasman, K. H., Cuticchia, A. J., and Kingsbury, D.T. (1994). The

AP-Genomics

MOUSE ANNEXIN V cDNA GDBTM Human Genome Data Base Anno 1994. Nucleic Acids Res. 22: 3462–3469. Felsenstein, J. (1989). PHYLIP—Phylogeny inference package (version 3.2). Cladistics 5: 164–166. Fernandez, M. P., Fernandez, M. R., and Morgan, R. O. (1994a). Structure of the gene encoding anchorin CII (chick annexin V). Gene 141: 179–186. Fernandez, M. P., Morgan, R. O., Fernandez, M. R., and Carcedo, M. T. (1994b). The gene encoding human annexin V gene has a TATA-less promoter with a high G / C content. Gene 149: 253– 260. Fernandez, M. P., Selmin, O., Martin, G. R., Yamada, Y., Pfaffle, M., Deutzmann, R., Mollenhauer, J., and von der Mark, K. (1988). The structure of anchorin CII, a collagen binding protein isolated from chondrocyte membrane. [See also erratum correction for Fig. 4 in J. Biol. Chem. (1990) 265(14):8344] J. Biol. Chem. 263: 5921– 5925. Geisow, M. J. (1986). Common domain structure of Ca2/ and lipidbinding proteins. FEBS Lett. 203: 99–103. Geisow, M. J., Fritsche, U., Hexham, J. M., Dash, B., and Johnson, T. (1986). A consensus amino-acid sequence repeat in Torpedo and mammalian Ca2/-dependent membrane-binding proteins. Nature 320: 636–638. Gould, S. J. (Ed.) (1993). ‘‘The Book of Life,’’ pp. 256, Random Century House, London. Green, E. L. (1981). ‘‘Genetics and Probability in Animal Breeding Experiments,’’ MacMillan Co., New York. Henikoff, S., and Henikoff, J. G. (1994). Protein family classification based on searching a database of blocks. Genomics 19: 97–107. Ideta, R., Seki, T., and Adachi, K. (1995). Sequence analysis and characterization of FAR-17c, an androgen-dependent gene in the flank organs of hamsters. J. Dermatol. Sci. 9: 94–102. Kozak, C. A., Peyser, M., Krall, M., Mariano, T. M., Kumar, C. S., Pestka, S., and Mock, B. A. (1990). Molecular genetic markers spanning mouse chromosome 10. Genomics 8: 519–524. Kozak, C. A., Adamson, M. C., Buckler, C. E., Segovia, L., Paralkar, V., and Wistow, G. (1995). Genomic cloning of mouse MIF (macrophage inhibitory factor) and genetic mapping of the human and mouse expressed gene and nine mouse pseudogenes. Genomics 27: 405–411. Learmonth, M. P., Howell, S. A., Harris, A. C. M., Amess, B., Patel, Y., Giambanco, I., Bianchi, R., Pula, G., Ceccarelli, P., Donato, R., Green, B. N., and Aitken, A. (1992). Novel isoforms of CaBP 33/ 37 (annexin V) from mammalian brain: Structural and phosphorylation differences that suggest distinct biological roles. Biochim. Biophys. Acta 1160: 76–83. Majors, J. E., and Varmus, H. E. (1981). Nucleotide sequences at host–proviral junctions for mouse mammary tumor virus. Nature 289: 253–258. Mollenhauer, J., and von der Mark, K. (1983). Isolation and characterization of a collagen–binding glycoprotein from chondrocyte membrane. EMBO J. 2: 45–50. Morgan, R. O., and Fernandez, M. P. (1995). Molecular phylogeny of annexins and identification of a primitive homologue in Giardia lamblia. Mol. Biol. Evol. 12: 967–979. O’Brien, S. J. (Ed.) (1993). ‘‘Genetic Maps. Locus Maps of Complex Genomes,’’ 6th ed., Cold Spring Harbor Laboratory Press, Plainview, New York. Okabe, M., Kishi, Y., Ying, X., Kohama, Y., Mimura, T., and Li, S. S. (1993). Characterization of capacitation inhibitory protein

/ m4752f3838

01-03-96 05:26:51

gnma

157

from rabbit seminal plasma: Homology with human annexins. Biol. Pharmacol. Bull. 16: 453–456. Pandey, K. N., Adamson, M. C., Gu, Y. C., and Kozak, C. A. (1994). Genetic mapping of the gene encoding guanylate cyclase-A/atrial natriuretic factor receptor (Npra) to mouse chromosome 3. Mamm. Genome 5: 520–522. Pearson, W. R., and Lipman, D. J. (1988). Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. USA 85: 2444– 2448. Pepinsky, R. B., Tizard, R., Mattaliano, R. J., Sinclair, L. K., Miller, G. T., Browning, J. L., Chow, E. P., Burne, C., Huang, K. S., Pratt, D., Wachter, L., Hession, C., Frey, A. Z., and Wallner, B. P. (1988). Five distinct calcium and phospholipid binding proteins share homology with lipocortin I. J. Biol. Chem. 263: 10799–10811. Prins, J.-B., Meisler, M. H., and Seldin, M. F. (1994). Mouse chromosome 3. Mammal. Genome 5S: S40–S50, and http://www.informat ics.jax.org/mgd.html. Raynal, P., and 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. Reutelingsperger, C. P. M., van Heerde, W., Hauptmann, R., Maassen, C., van Gool, R. G. J., de Leeuw, P., and Tiebosch, A. (1994). Differential tissue expression of annexin VIII in human. FEBS Lett. 349: 120–124. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). ‘‘Molecular Cloning: A Laboratory Manual,’’ 2nd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Schlaepfer, D. D., Jones, J., and Haigler, H. T. (1992). Inhibition of protein kinase C by annexin V. Biochemistry 31: 1886–1891. Segade, F., Claudio, E., Wrobel, K., Ramos, S., and Lazo, P. S. (1995). Isolation of nine gene sequences induced by silica in murine macrophages. J. Immunol. 154: 2384–2392. Spreca, A., Rambotti, M. G., Giambanco, I., Pula, G., Bianchi, R., Ceccarelli, P., and Donato, R. (1992). Immunocytochemical localization of annexin-V (CaBP33), a Ca2/-dependent phospholipidbinding and membrane-binding protein, in the rat nervous system and skeletal muscles and in the porcine heart. J. Cell Physiol. 152: 587–598. Sudhof, T. C., Ebbecke, M., Walker, J. H., Fritsche, U., and Boustead, C. (1984). Isolation of mammalian calelectrins: A new class of ubiquitous Ca2/-regulated proteins. Biochemistry 23: 1103–1109. Tait, J. F., Frankenberry, D. A., Shiang, R., Murray, J. C., Adler, D. A., and Disteche, C. M. (1991). Chromosomal localization of the human gene for annexin-V (placental anticoagulant protein I) to 4q26 r q28. Cytogenet. Cell Genet. 57: 187–192. Takei, N., Ohsawa, K., Imai, Y., Nakao, H., Iwasaki, A., and Kohsaka, S. (1994). Neurotrophic effects of annexin V on cultured neurons from embryonic rat brain. Neurosci. Lett. 171: 59–62. van Heerde, W. L., de Groot, P. G., and Reutelinsperger, C. P. M. (1995). The complexity of the phospholipid binding protein annexin V. Thromb. Haemostasis 73: 172–179. Vishwanatha, J. K., and Kumble, S. (1993). Involvement of annexin II in DNA replication. Evidence from cell-free extracts of Xenopus eggs. J. Cell Sci. 105(Pt.2): 533–540. Wilson, A. C., Carlson, S. S., and White, T. J. (1977). Biochemical evolution. Annu. Rev. Biochem. 46: 573–639. Yokota, T., Arai, N., Lee, F., Rennick, D., Mosmann, T., and Arai, K. I. (1985). Use of a cDNA expression vector for isolation of mouse interleukin 2 cDNA clones: Expression of T-cell growth-factor activity after transfection of monkey cells. Proc. Natl. Acad. Sci. USA 82: 68–72.

AP-Genomics