Single and multiple CH (calponin homology) domain containing multidomain proteins in Arabidopsis and Saccharomyces: an inventory.

Genes for individual domains such as CH, lim, ankyrin, PH and RhoGAP, IQ motif, Ig_FLMN, spectrin, and EF hand probably existed in early evolution before there were plants, fungi or animals so that when we examine multidomain proteins in Arabidopsis, Saccharomyces, Dictyostelium or Homo Sapiens we encounter various combinations of such domains. While all of these four species express Fimbrin and EB1, the lists of CH containing multidomain proteins, however, differ in number and in type for each of them. There was no further great increase in the number of new single domain proteins. Still many new multidomain genes evolved--but far more so in metazoans--than in plants or fungi. In both plants and fungi only singlet CH domains but no doublets (other than those forming the Fimbrin quadruplet) were incorporated. That is in these two branches one finds no alpha actinin, dystrophin or filamin even though the individual building blocks (i.e. domains such as spectrin or IG-FLMN) were available in Arabidopsis. Possibly transposons create new chimeric multidomain genes by mixing and matching genes or gene fragments.

Duplex (or quadruplet) CH domain containing human multidomain proteins: an inventory.

In this paper, the inventory presented for singlet CH (calponin homology/actin binding) domain containing human multidomain proteins is extended to several duplex and one quadruplet CH containing forms. Invariably, the duplexes are located at the begin of the molecules. The regions connecting the two CH units suggest amino acid conservations which allows the placing of 18 duplex containing molecules into six groups wherein the gene for one member in each group created the others more recently by gene duplication. The ancient multidomain proteins, possibly, were primarily the result of an exon shuffling (transposition) mechanism that also guided the placing of the CH singlet or duplex domain at the amino end of the newly created proteins. A mechanism that creates pseudogenes could conceivably produce genes that encode multi-domain proteins. Intragenomic duplications (slippage) might have facilitated the occurrence of encoding repeats, thus allowing for the creation of multiple identical domains within one molecule. Gene duplication with subsequent modification and small domain gene recombination which formed multidomain proteins are important forces driving evolution.

Single and multiple CH (calponin homology) domain containing multidomain proteins in Dictyostelium discoideum: an inventory.

We present an inventory of single or multiple calponin homology (CH) domain containing proteins of Dictyostelium discoideum. A multiple alignment and a phylogenetic tree of all 60 CH domains found in 36 proteins showed that most CH domains can be assigned to one of 6 types. We have then distributed the proteins into several classes according to the type and arrangement of the CH domains. Most proteins belong to the class of ABD (actin-binding domain)-forming CH tandems (CH1-CH2) of the alpha-actinin and fimbrin families or to the class of CH3 domain-bearing proteins. There are a few examples of proteins with a single CH1 or CH2 domain, one with a CH1-CH1 doublet and a single representative of the CHe class of microtubule-binding proteins. A comparison with CH domain proteins in Homo sapiens suggests that while the individual domains are available in both species, the existence of identical multidomain proteins in toto is rare. Fimbrin 1, alpha-actinin and EB1 appear as perfect orthologs in both species, whereas filamin and interaptin may represent ancestral forms of human filamin and nesprins. In four more cases (NAV/Unc-53-, smoothelin-, transgelin- and Gas2-related proteins) functional data are needed in order to establish a potential relationship with a human counterpart. Although extensive data exist for a few of the D. discoideum CH proteins, most remain to be characterized and our analysis may help predicting some of their properties.

Singlet CH domain containing human multidomain proteins: an inventory.

The actin cytoskeleton presents the basic force in processes such as cytokinesis, endocytosis, vesicular trafficking and cell migration. Here, we list 30 human singlet CH (calpononin homology/actin binding) containing multidomain molecules, each encoded by one gene. We show the domain distributions as given by the SMART program. These mosaic proteins organize geographically the placement of selected proteins in proximity within the cell. In most instances, their precise location, their actin binding capacity by way of the singlet CH (or by other domains?) and their physiological functions need further elucidation. A dendrogram based solely on the relationship for the human singlet CH domains (in terms of AA sequences) for the various molecules that possess the domain, implies that the singlet descended from a common ancestor which in turn sprouted three main branches of protein products. Each branch bifurcated multiple times thus accounting for a cornucopia of products. Wherever, additional (unassigned), highly homologous regions exist in related proteins (e.g., in LIM and LMO7 or in Tangerin and EH/BP1), these unrecognized domain regions await assignment as specific functional domains. Frequently genes coding multidomain proteins duplicated. The varying modular nature within multidomain proteins should have accelerated evolutionary changes to a degree not feasible to achieve by means of mere post-duplication mutational changes.

Alternative splicing for members of human mosaic domain superfamilies. I. The CH and LIM domains containing group of proteins.

In this paper we examine (restricted to homo sapiens) the products resulting from gene duplication and the subsequent alternative splicing for the members of a multidomain group of proteins which possess the evolutionary conserved calponin homology CH domain, i.e. an "actin binding domain", as a singlet and which, in addition, contain the conserved cysteine rich double Zn finger possessing Lim domain, also as a singlet. Seven genes, resulting from gene duplications, were identified that code for seven group members for which pre-mRNAs appear to have undergone multiple alternative splicing: Mical 1, 2 and 3 are located on chromosomes 6q21, 11p15 and 22q11, respectively. The LMO7 gene is present on chromosome 13q22 and the LIMCH1 gene on chromosome 4p13. Micall1 is mapped to chromosome 22q13 and Micall2 to chromosome 7p22. Translated Gen/Bank ESTs suggest the existence of multiple products alternatively spliced from the pre-mRNAs encoded by these genes. Characteristic indicators of such splicing among the proteins derived from one gene must include containment of some common extensive 100% identical regions. In some instances only one exon might be partly or completely eliminated. Sometimes alternative splicing is also associated with an increased frequency of creation of an exon or part of an exon from an intron. Not only coding regions for the body of the protein but also for its N- or -C ends could be affected by the splicing. If created forms are merely beginning at different starting points but remain identical in sequence thereafter, their existence as products of alternate splicing must be questioned. In the splicings, described in this paper, multiple isoforms rather than a single isoform appear as products during the gene expression.

Centrin isoforms in mammals. Relation to calmodulin.

In mammals, three calmodulin (CaM) genes code for 100% identical proteins. In these species, four centrin (Cetn) genes have been reported to exist. They are examined in this paper. While the gene for Cetn 1 contains no introns and appears to be derived from Cetn 2 by retroposition, a gene product for Cetn 1 is expressed. Cetn 2, 3, and 4 represent bona fide genes. The major difference between the members of the CaM and the Cetn subfamilies is the presence (usually) in Cetn of an approximately 23 amino acids long (but occasionally much longer) protruding amino acid end. In all members of these two subgroups, four EF hand motifs (in this paper taken as loops containing 12 amino acids) are separated by 24, 25 and 24 amino acids (each a helix-loop-helix) positioned between motifs 1and 2, 2 and 3, and 3 and 4, respectively. This rule applies not only to CaM and Cetn in mammals but also to these two subfamilies in simpler eukaryotes such as Saccharomyces cerevisiae and Giardia lamblia. The various mRNA products can be identified most readily by their characteristic 3' UTRs. While CaM is an ancient molecule that is expressed in all cells and is ubiquitous within these cells and interacts therein with almost 100 different proteins, many of which display the IQ or related binding motifs, the distribution and function of Cetn (an equally ancient molecule) is restricted mostly to basal bodies (e.g. in rods of the retina), axonemes, flagella, cilia and centrosomes. Are these two subclasses of calcium carriers (each molecule possessing four EF hands which possibly interact with different association constants)-if they are both present within a cell-randomly chosen for their service to the specific proteins with which they interact?

Parvalbumin isoforms in zebrafish.

By using an analysis of existing genomic information it is concluded that in zebrafish nine genes encode parvalbumin (PV). These genes possess introns that differ in size and show nucleotide variability but they contain the same number of exons, and for each corresponding exon, the number of nucleotides therein are identical in all the paralogs. This rule also applies to the multiple PV genes of other species e.g. mammals. Each of these genes displays, however, characteristic 5' and 3' UTRs which appear highly conserved between closely related species (so that orthologs among these species can be readily identified) but which show larger numbers of mutations between species that are more distant in evolution. A tree is presented which suggests that the traditional classification of PVs as alpha or beta (based mainly on charge of the protein molecule) is not sustainable. Numbers 1-9 are assigned to the various isoforms to facilitate their identification in future studies. A bifurcation of isoforms into 1 and 4; 2 and 3; 6 and 7; 8 and 9 appears to have occurred simultaneously in more recent time, i.e. perhaps approximately 60 mys ago when primates and rodents branched.

Calmodulin genes in zebrafish (revisited).

Calmodulin (CaM), a ubiquitous protein, ancestral in early eukaryotes, regulates a large number of physiologically important functions by activating other proteins, some of them enzymes, usually in response to changes in the local concentration of calcium ions. Invertebrates possess one gene that codes for CaM. Among vertebrates, mammals display three genes that code for a 100% identical CaM molecule, while for zebra fishes etc., a non-mammalian vertebrate, we reported earlier the existence of four such genes. The number of multiple genes coding for a 100% identical CaM molecule present in the zebra fish genome, however, is corrected here, from the four, as previously suggested, to six (alpha, alpha2, beta, beta2, gamma and gamma 2). Identification of each of these genes is readily achieved upon examination of the characteristic 5' and 3' UTRs within their respective mRNAs even though we do not know at present what role these UTRs might play. A scanning of the 3' UTRs for short homology elements among the six genes (and a comparison with the human type I, II, and III CaM 3' UTRs) also suggests that duplication processes for three genes resulted in the formation of six such genes. As they become available, the promoter regions for these six genes should be scanned for possible identification of putative regulatory elements if we are to understand the need for the uniquely rigid evolutionary maintenance of these six genes. A comparison of the promoter regions for the beta and beta 2 genes is presented in this paper. A few common short homologous elements appear to be retained in these generally highly variant two regions, but conclusions about differential expression controls must be delayed until the promoter regions for all the other CaM genes have been examined.

Multiple calmodulin genes in fish.

In mammals, identical calmodulin (CaM) proteins are encoded by three nonallelic genes that differ in their promoter regions and untranslated regions (UTRs). The UTRs of each of these three genes are specific for each gene and are highly conserved. In this study, sequences obtained from the GenBank and EST databases and sequencing were examined for several species of fish to ascertain whether this multi-gene one protein system exhibited in mammals extends to other vertebrates. Three genes in zebrafish (Danio rerio) designated alpha, beta, and gamma were identified. As in mammals, these genes differ in the 3'-UTR region but encode completely identical CaMs. PCR primers spanning the coding and the 3'-UTR regions were designed based on the assembled sequences and used to confirm the presence of each gene in the cDNA library. Other species of fish were also found to contain homologous genes that were closely related as indicated by phylogenetic analysis. The 3'-UTR of the alpha, beta and particularly the gamma CaM gene of fish were not found to be as conserved as the corresponding genes of mammalian species possibly due to the span of evolutionary time. Only a few short elements in the 3'-UTR were observed to be similar in fish and mammals. These short regions of identity are shared primarily between the mammalian CaM II and CaM I and the alpha gene and beta gene of fish, respectively. Thus, the multi-gene one protein system occurs among fish as well as among mammals.

Age and detection of retroprocessed pseudogenes in murine rodents.

Retroprocessed pseudogenes, calmodulin II (psi1, psi2, and psi3 CALMII), psi alpha-tubulin, pi-glutathione S-transferase (psi pi-GST) from rat, lactic acid dehydrogenase (psi LDH) from mouse, and heat shock protein 60 chaperonin (psi HSP60) from Chinese hamster, were examined for their presence in these species by polymerase chain reaction (PCR). Pseudogenes of these murine rodents were detected by PCR only in those species in which the genes were originally identified, suggesting that the selected pseudogene of one species arose too recently to be detected in the genomes of the other rodent species. The calculated ages of the rodent pseudogenes ranged from 1.7 Myr (psi alpha-tubulin) to 7.5 Myr (psi3 CALMII) when employing a homologous functional gene of the taxon as a reference in the relative rate test with the mouse or rat as the outgroup. Given the high rate of divergence of the genes of rodents relative to other species, selection of an outgroup with similar mutation rates seems warranted. To justify further the conclusion that the selected pseudogenes were indeed retroprocessed after these three taxa diverged, the presence of the pseudogenes in the genome of different rat species was examined. The existence of psi3 CALMII and psi alpha-tubulin pseudogenes of Rattus norvegicus among species belonging to Rattus sensu stricto is evidence for the common ancestry of this group.