The gene coding for the p68 calcium-binding protein is localised to bands q32-q34 of human chromosome 5, and to mouse chromosome 11

Annexin A6-Linking Ca(2+) signaling with cholesterol transport

Annexin A6 (AnxA6) belongs to a conserved family of Ca(2+)-dependent membrane-binding proteins. Like other annexins, the function of AnxA6 is linked to its ability to bind phospholipids in cellular membranes in a dynamic and reversible fashion, in particular during the regulation of endocytic and exocytic pathways. High amounts of AnxA6 sequester cholesterol in late endosomes, thereby lowering the levels of cholesterol in the Golgi and the plasma membrane. These AnxA6-dependent redistributions of cellular cholesterol pools give rise to reduced cytoplasmic phospholipase A2 (cPLA(2)) activity, retention of caveolin in the Golgi apparatus and a reduced number of caveolae at the cell surface. In addition to regulating cholesterol and caveolin distribution, AnxA6 acts as a scaffold/targeting protein for several signaling proteins, the best characterized being the Ca(2+)-dependent membrane targeting of p120GAP to downregulate Ras activity. AnxA6 also stimulates the Ca(2+)-inducible involvement of PKC in the regulation of HRas and possibly EGFR signal transduction pathways. The ability of AnxA6 to recruit regulators of the EGFR/Ras pathway is likely potentiated by AnxA6-induced actin remodeling. Accordingly, AnxA6 may function as an organizer of membrane domains (i) to modulate intracellular cholesterol homeostasis, (ii) to create a scaffold for the formation of multifactorial signaling complexes, and (iii) to regulate transient membrane-actin interactions during endocytic and exocytic transport. This article is part of a Special Issue entitled: 11th European Symposium on Calcium.

Annexin A6 inhibits Ras signalling in breast cancer cells

Overexpression of epidermal growth factor receptor (EGFR) is associated with enhanced activation of wild-type (hyperactive) Ras in breast cancer. Little is known about the regulation of Ras inactivation and GTPase-activating proteins (GAPs), such as p120GAP, in cells with hyperactive Ras. Recently, we showed that in EGFR-overexpressing A431 cells, which lack endogenous Annexin A6 (AnxA6), ectopic expression of AnxA6 stimulates membrane recruitment of p120GAP to modulate Ras signalling. We now demonstrate that, AnxA6 is downregulated in a number of EGFR-overexpressing and estrogen receptor (ER)-negative breast cancer cells. In these cells, AnxA6 overexpression promotes Ca(2+)- and EGF-inducible membrane targeting of p120GAP. In ER-negative MDA-MB-436 cells, overexpression of p120GAP, but not CAPRI or a p120GAP mutant lacking the AnxA6-binding domain inhibits Ras/MAPK activity. AnxA6 knockdown in MDA-MB-436 increases Ras activity and cell proliferation in anchorage-independent growth assays. Furthermore, AnxA6 co-immunoprecipitates with H-Ras in a Ca(2+)- and EGF-inducible manner and fluorescence resonance energy transfer (FRET) microscopy confirmed that AnxA6 is in close proximity of active (G12V), but not inactive (S17N) H-Ras. Thus, association of AnxA6 with H-Ras-containing protein complexes may contribute to regulate p120GAP/Ras assembly in EGFR-overexpressing and ER-negative breast cancer cells.

Calcium- and pH-dependent localization of annexin A6 isoforms in Balb/3T3 fibroblasts reflecting their potential participation in vesicular transport

Annexin A6 (AnxA6), calcium- and membrane-binding protein, is involved in membrane dynamics. It exists in the cell in two isoforms, AnxA6-1 and AnxA6-2, varying only by the VAAEIL sequence. In most cells, AnxA6-1 predominates. A limited number of observations suggests that both isoforms differ from each other functionally. The EGF-dependent Ca(2+) influx in A431 cells is inhibited only by AnxA6-1. Moreover, AnxA6-2 was found to exhibit higher affinity for Ca(2+). In this report we addressed the potential significance of the VAAEIL deletion in AnxA6-2. For this purpose, we expressed AnxA6 isoform cDNAs in bacteria or mouse Balb/3T3 fibroblasts. The recombinant AnxA6-2 was characterized by a less extended molecular shape than that of AnxA6-1 and required a narrower [Ca(2+)] range to bind liposomes. Upon lowering pH in the presence of EGTA recombinant AnxA6-2 became less hydrophobic than AnxA6-1 as revealed by the Triton X-114 partition. Furthermore, AnxA6-2 revealed stronger F-actin binding than that of AnxA6-1. Immunofluorescence microscopy showed that the EGFP-tagged AnxA6 isoforms expressed in Balb/3T3 fibroblasts relocate in a Ca(2+)- and H(+)-sensitive manner to the vesicular structures in a perinuclear region or in cytosol. Cell fractionation showed that in resting conditions AnxA6-1 is associated with early endosomes and AnxA6-2 with late endosomes, and an increase in [Ca(2+)] and/or [H(+)] induced their opposite distribution. These findings suggest a potentially independent regulation, localization, and function of AnxA6 isoforms in Balb/3T3 fibroblasts. More generally, our findings suggest distinct functions of AnxA6 isoforms in membrane dynamics.

A conserved nuclear element with a role in mammalian gene regulation

Mammalian genomes contain numerous fragments of DNA that are derived from inactivated transposable elements. The accumulation and persistence of these elements is generally attributed to transposase activity rather than through possession or acquisition of a function of value to the host genome. Here we describe such a repetitive element, named ALF (forannexin VILINE-2fragment), comprising 130 bp of DNA derived from a LINE-2 sequence, which functions as a potent T-cell-specific silencer. The expansion of the DNA database arising as a result of the human genome sequencing project enabled us to identify ALF in, or close to, several well characterized genes including those for annexin VI, interleukin-4 and protein kinase C-beta. A systematic analysis of the entire LINE-2 sequence revealed that ALF, and not other regions of the LINE-2 sequence, was especially highly represented in the human genome. Acquisition of a function by this repetitive element may explain its abundance. These data show that a conserved fragment of an interspersed nuclear element has the potential to modulate gene expression, a discovery that has broad implications for the way in which we view so-called 'junk' DNA and our understanding of eukaryotic gene regulation.

Novel human and mouse annexin A10 are linked to the genome duplications during early chordate evolution

We have identified and characterized a 12th subfamily of vertebrate annexins by systematic analysis of the primary structure, chromosomal mapping, and molecular evolution of unique cDNA and protein sequences from human and mouse. Distinctive features included rare expression, a codon deletion in conserved repeat 3, and an unusual ablation of the type II calcium-binding sites in tetrad core repeats 1, 3, and 4. The paralogy of novel annexin A10 (following revised nomenclature) was confirmed by FISH-mapping human ANXA10 to chromosome 4q33 and genetic linkage mapping mouse Anxa10 to midchromosome 8. Phylogenetic analysis established that the 5' and 3' halves of the annexin A6 octad are more closely related to annexins A5 and A10, respectively, than they are to each other. Molecular date estimates, paralogy linkage maps between human chromosomes 4 and 5, and annexin structural considerations led to the proposal that annexins A5 and A10 may have been the direct progenitors of annexin A6 octad formation via chromosomal duplication during the genome expansion in early chordates.

Genomic locations of ANX11 and ANX13 and the evolutionary genetics of human annexins

We have reconstructed a molecular genetic history of human annexins to chronicle their origins and dispersal throughout the genome. This involved the completion of chromosomal mapping, determination of ancestral relationships, and estimation of gene duplication dates. Fluorescence in situ hybridization localized human annexin XI (ANX11) to 10q22.3-q23.1 and annexin XIII (ANX13) to 8q24.1-q24.2. Orthologous annexins showed minor rate variation when calibrated to species separation times given by the fossil record, but paralogous subfamilies have diverged at fivefold variable rates. The rates and extents of sequence divergence were used to predict a mean separation time of 450 million years between vertebrate annexins, although their common ancestor may have emanated from invertebrate stock. Annexins XIII and VII formed a phylogenetically early clade, and annexins II and VIa were the most divergent members of two distinct clades. ANX6 may have been created by tandem duplication about 500 million years ago (Mya) and duplicated again to form ANX5 400 Mya, whereas ANX4 and ANX8 are proposed to be sequential duplication products from annexin XI. Vertebrate annexins thus proliferated via a cascade of gene duplications in higher metazoa to form at least three diverging groups of ubiquitous and structurally related genes. These can be distinguished by their dispersed genomic locations as well as their individual patterns of expression and partially differentiated functions.

Functional and genetic analysis of annexin VI

This study is concerned with the determination of the function of the 68kDa calcium-binding protein, annexin VI. Studies on the structure and regulation of the gene include a detailed analysis of annexin VI expressed heterologously in human A431 carcinoma cells. We have recently discovered that annexin VI is subject to a novel growth dependent post-translational modification. Interestingly, the protein exerts a negative effect on A431 cells. This effect was manifested as a partial reversal of the transformed phenotype. We are currently exploring the hypothesis that the post-translational modification of annexin VI is required for sub-cellular targeting, and that correct localisation within the cell is essential for function.

Molecular genetic approaches to the study of human craniofacial dysmorphologies

Craniofacial dysmorphologies are common, ranging from simple facial disfigurement to complex malformations involving the whole head. With the advent of gene mapping and cloning techniques, the genetic element of both simple and complex human craniofacial dysmorphologies can be investigated. For many of the dysmorphic syndromes, it is possible to find families that display a particular phenotype in either an autosomal dominant, recessive, or X-linked manner. This article focuses on a subgroup of craniofacial dysmorphologies, covering these three main inheritance patterns, that are being studied using molecular biology techniques: DiGeorge syndrome, Treacher Collins syndrome, Greig cephalopolysyndactyly syndrome, acrocallosal syndrome, amelogenesis imperfecta, and X-linked cleft palate with ankyloglossia. Once the mutated or deleted gene or genes for each syndrome have been cloned, patterns of normal and abnormal craniofacial development should be elucidated. This should enhance both diagnosis and treatment of these common and disfiguring disorders.

Annexin VI isoforms are differentially expressed in mammalian tissues

Purified annexin VI migrates as a closely spaced doublet when separated by SDS-PAGE. Immunolocalization of annexin VI in heart demonstrates staining at different defined subcellular compartments. Moss et al. identified two cDNAs, one having an insert of 18 bases encoding VAAEIL at the beginning of repeat domain seven. We have identified the splicing site of the murine annexin VI gene. It contains a single small exon of 18 bases. PCR amplification of reverse transcribed (RT) mRNA demonstrates that, in all tissues tested, the mRNA isoform containing the insert is predominant. Site-directed antibody was produced and affinity purified against peptides reflecting the insert and deletion sequences. The steady-state isoform ratio of the annexin VI protein is consistent with the RT-PCR data. Chromatographic experiments demonstrate that the annexin VI protein isoforms have biochemical differences. These differences may target the individual isoforms to unique cellular compartments or alter functional properties.

Divergent structure of the human synexin (annexin VII) gene and assignment to chromosome 10

The human synexin (annexin VII) gene occurs as a single copy at chromosome 10q21.1-21.2 and substantially deviates in size and in the location of splice junctions from the other two well-characterized members of the annexin gene family, lipocortin I (annexin I) and calpactin I (annexin II). The synexin gene contains 14 exons, including an alternatively spliced cassette exon, and spans approximately 34 kb of DNA. Only five of the fourteen splice junctions are conserved compared to other annexins, and the differences are particularly pronounced in the exons that encode the C-terminal third and fourth conserved repeats in the gene product. Although parallels between exons and protein domains were not apparent, we did observe clustering of splice junctions corresponding to either the unique N-terminal domain or the conserved C-terminal tetrad repeat domain, which is common to all annexins. Furthermore, a complete analysis of the 5' flanking region of the annexin VII gene revealed an entirely different set of cis-acting and enhancer elements compared to other annexin genes. We conclude that the annexin VII gene may have arisen by a divergence from the evolutionary pathway taken by both annexins I and II.

Structure of the human annexin VI gene

We report the structure of the human annexin VI gene and compare the intron-exon organization with the known structures of the human annexin I and II genes. The gene is approximately 60 kbp long and contains 26 exons. Consistent with the published annexin VI cDNA sequence, the genomic sequence at the 3' end does not contain a canonical polyadenylation signal. The genomic sequence upstream of the transcription start site contains TATAA and CAAT motifs. The spatial organization of the exons does not reveal any obvious similarities between the two halves of the annexin VI gene. Comparison of the intron-exon boundary positions of the annexin VI gene with those of annexins I and II reveals that within the repeated domains the break points are perfectly conserved except for exon 8, which is one codon smaller in annexin II. The corresponding point in the second half of annexin VI is represented by two exons, exons 20 and 21. The latter exon is alternatively spliced, giving rise to two annexin VI isoforms that differ with respect to a 6-amino acid insertion at the start of repeat 7.

Chromosomal mapping of the human annexin IV (ANX4) gene

Annexin IV (placental anticoagulant protein II) is a member of the annexin or lipocortin family of calcium-dependent phospholipid-binding proteins. A cDNA for human annexin IV was isolated from a placental library that is 675 bases longer in the 3' untranslated region than previously reported, indicating the existence of alternative mRNA processing for this gene. Genomic Southern blotting with a cDNA probe indicated a gene size of 18-56 kb. Primers developed for polymerase chain reaction (PCR) allowed amplification of a 1.6-kb portion of the ANX4 gene. DNA sequence analysis showed that this PCR product contained a single intron with exon-intron boundaries in exactly the same position as in the mouse annexin I and annexin II genes. PCR analysis of a somatic cell hybrid panel mapped the ANX4 gene to chromosome 2, and in situ hybridization with a cDNA probe showed a unique locus for ANX4 at 2p13. This study provides further evidence that genes for the annexins are dispersed throughout the genome but are similar in size and exon-intron organization.

Chromosomal localization of the human annexin III (ANX3) gene

The annexins or lipocortins are a new family of calcium-dependent phospholipid-binding proteins. Annexin III has been previously identified as inositol 1,2-cyclic phosphate 2-phosphohydrolase (EC 3.1.4.36), an enzyme of inositol phosphate metabolism, and also as placental anticoagulant protein III, lipocortin III, calcimedin 35-alpha, and an abundant neutrophil cytoplasmic protein. In this study, the gene (ANX3) encoding annexin III was localized to human chromosome 4 at band q21 (q13-q22) by (1) polymerase chain reaction analysis of a human-rodent hybrid cell panel, confirmed by genomic Southern blot analysis of the same panel with a cDNA probe and (2) in situ hybridization with a cDNA probe.

Developmental expression of the endogenous TIMP gene and a TIMP-lacZ fusion gene in transgenic mice

We determined the expression pattern of the tissue inhibitor of metalloproteinase (TIMP) in the development of the mouse embryo using in situ hybridization and transgenesis. Localized TIMP RNA was first detected at 13.5 days post conceptus (p.c.) in tissues undergoing osteogenesis, such as the mandible, ribs, and calvaria. As development proceeded, TIMP RNA could be detected at additional sites, including the tooth buds, vertebrae, and long bones. To define the sequences regulating TIMP expression, we generated transgenic mice that expressed the Escherichia coli beta-galactosidase gene under control of a 5' region of the mouse TIMP gene containing -2158 to -58 bp upstream of the initiator ATG. By use of an in situ assay for beta-galactosidase activity, the TIMP-lacZ fusion gene product was localized to tissues that also expressed the endogenous TIMP gene, such as the mandible, calvaria, and vertebrae. The localization of TIMP to regions of intramembranous and endochondral bone is similar to that previously reported for TGF-beta, a growth modulator believed to be involved in regulation of extracellular matrix (ECM) formation. Thus, the expression of TIMP in these regions is consistent with it playing a role in ECM deposition and turnover in development.

Alternative splicing gives rise to two forms of the p68 Ca2(+)-binding protein

The p68 Ca2+ and phospholipid binding protein of the lipocortin/calpactin family appears to exist as two forms. These may be resolved into a closely-spaced polypeptide doublet by SDS-PAGE. The cloning and sequencing of p68 revealed an apparent 18 nucleotide alternative splice sequence, which could account for this observation. We show here that an antiserum directed against a synthetic peptide corresponding to the region containing the splice sequence, recognises only the upper band of the p68 doublet by both immunoprecipitation and Western blotting. These results are consistent with alternative splicing being responsible for the generation of the two forms of p68.