Genomic organization and expression profile of the parvin family of focal adhesion proteins in mice and humans.

We have characterized the genomic organization and the expression pattern of alpha-, beta- and gamma-parvin, a novel family of focal adhesion proteins, in mice and humans. alpha-Parvin is nearly ubiquitously expressed, beta-parvin is preferentially expressed in heart- and skeletal muscle, and gamma-parvin in lymphoid tissues. Parvins display diverse patterns of developmental regulation. The alpha-form is present throughout mouse development, beta-parvin is gradually upregulated and gamma-parvin is downregulated at embryonic day 11. The human alpha-parvin gene (PARVA), extending over 160 kb, is located on chromosome 11. Both, the human beta-parvin gene (PARVB), which is over 145 kb long, and the gamma-parvin gene (PARVG) of a total length of about 25 kb are positioned on chromosome 22 with PARVG located about 12 kb downstream of the 3' end of PARVB. Multiple tissue array analysis indicates that parvins are expressed at reduced levels in cancer as compared to the corresponding normal tissues. Analysis of ESTs and PCR-amplified fragments reveals alternatively spliced and alternatively polyadenylated gene products. Mammalian parvins are likely to have arisen late in evolution from gene duplication as they share a remarkably similar exon/intron organization, which is different from the organization of the single genes encoding parvin-like proteins in Drosophila and Caenorhabditis.

Parvin, a 42 kDa focal adhesion protein, related to the alpha-actinin superfamily.

We have identified and cloned a novel 42-kDa protein termed alpha-parvin, which has a single alpha-actinin-like actin-binding domain. Unlike other members of the alpha-actinin superfamily, which are large multidomain proteins, alpha-parvin lacks a rod domain or any other C-terminal structural modules and therefore represents the smallest known protein of the superfamily. We demonstrate that mouse alpha-parvin is widely expressed as two mRNA species generated by alternative use of two polyadenylation signals. We analyzed the actin-binding properties of mouse alpha-parvin and determined the K(d) with muscle F-actin to be 8.4+/-2.1 microM. The GFP-tagged alpha-parvin co-localizes with actin filaments at membrane ruffles, focal contacts and tensin-rich fibers in the central area of fibroblasts. Domain analysis identifies the second calponin homology domain of parvin as a module sufficient for targeting the focal contacts. In man and mouse, a closely related paralogue beta-parvin and a more distant relative gamma-parvin have also been identified and cloned. The availability of the genomic sequences of different organisms enabled us to recognize closely related parvin-like proteins in flies and worms, but not in yeast and Dictyostelium. Phylogenetic analysis of alpha-parvin and its para- and orthologues suggests, that the parvins represent a new family of alpha-actinin-related proteins that mediate cell-matrix adhesion.

Mutational analysis of Ser14 and Asp157 in the nucleotide-binding site of beta-actin.

This paper compares wild-type and two mutant beta-actins, one in which Ser14 was replaced by a cysteine, and a second in which both Ser14 and Asp157 were exchanged (Ser14-->Cys and Ser14-->Cys, Asp157-->Ala, respectively). Both of these residues are part of invariant sequences in the loops, which bind the ATP phosphates, in the interdomain cleft of actin. The increased nucleotide exchange rate, and the decreased thermal stability and affinity for DNase I seen with the mutant actins indicated that the mutations disturbed the interdomain coupling. Despite this, the two mutant actins retained their ATPase activity. In fact, the mutated actins expressed a significant ATPase activity even in the presence of Ca2+ ions, conditions under which actin normally has a very low ATPase activity. In the presence of Mg2+ ions, the ATPase activity of actin was decreased slightly by the mutations. The mutant actins polymerized as the wild-type protein in the presence of Mg2+ ions, but slower than the wild-type in a K+/Ca2+ milieu. Profilin affected the lag phases and elongation rates during polymerization of the mutant and wild-type actins to the same extent, whereas at steady-state, the concentration of unpolymerized mutant actin appeared to be elevated. Decoration of mutant actin filaments with myosin subfragment 1 appeared to be normal, as did their movement in the low-load motility assay system. Our results show that Ser14 and Asp157 are key residues for interdomain communication, and that hydroxyl and carboxyl groups in positions 14 and 157, respectively, are not necessary for ATP hydrolysis in actin.

The role of profilin in actin polymerization and nucleotide exchange.

Properties of human profilin I mutated in the major actin-binding site were studied and compared with wild-type profilin using beta/gamma-actin as interaction partner. The mutants ranged in affinity, from those that only weakly affected polymerization of actin to one that bound actin more strongly than wild-type profilin. With profilins, whose sequestering activity was low, the concentration of free actin monomers observed at steady-state of polymerization [Afree], was close to that seen with actin alone ([Acc], critical concentration of polymerization). Profilin mutants binding actin with an intermediate affinity like wild-type profilin caused a lowering of [Afree] as compared to [Acc], indicating that actin monomers and profilin:actin complexes participate in polymer formation. With a mutant profilin, which bound actin more strongly than the wild-type protein, an efficient sequestration of actin was observed, and in this case, the [Afree] at steady state was again close to [Acc], suggesting that the mutant profilin:actin had a greatly lowered ability to incorporate actin subunits at the (+)-end. The results from the kinetic and steady-state experiments presented are consonant with the idea that profilin:actin complexes are directly incorporated at the (+)-end of actively polymerizing actin filaments, while they do not support the view that profilin facilitates polymer formation.

Characterization of a mutant profilin with reduced actin-binding capacity: effects in vitro and in vivo.

We are investigating structure-function relationships in profilin and actin by site-specific mutagenesis using a yeast, Saccharomyces cerevisiae, expression system to produce wild-type and mutant proteins. This paper shows that deleting proline 96 and threonine 97, which are located close to the major actin binding site on profilin, did not significantly alter the interaction between profilin and phosphatidylinositol 4,5-bisphosphate, nor did it affect the profilin:poly(L-proline) interaction. The mutant protein, however, had a lower capacity to bind to actin in vitro than wild-type profilin, though it showed a slightly increased profilin-enhanced nucleotide exchange on the actin. When microinjected into Swiss 3T3 mouse fibroblasts or porcine aortic endothelial cells, the mutant profilin did not change the organization of the microfilament system like the wild-type profilin did. This provides further evidence that profilin controls microfilament organization in the cell by interacting directly with actin.

Isolation and characterization of two mutants of human profilin I that do not bind poly(L-proline).

A simple procedure for the isolation of profilin mutants having a reduced capacity to bind poly(L-proline) is used to isolate two mutants of human profilin I, W3N and H133S. Binding of the mutants to poly(L-proline), actin, and phosphatidylinositol (4,5)-bisphosphate (PIP2) was studied. Both mutations abolished the poly(L-proline)-binding activity of profilin. This suggests that the arrangement of the N- and C-terminal helices forming the poly(L-proline)-binding site depends on the stabilizing interaction between W3 and W31 in the underlying beta-strand, and that the H133S mutation in the C-terminal helix also must have distorted the arrangement of the terminal helices. Both mutations caused a reduced affinity for actin, with the W3N replacement having the most pronounced effect. This shows that structural changes in the poly(L-proline)-binding region of profilin can affect the distantly located actin-binding site. Thus, ligands influencing the structure of the poly(L-proline)-binding site may regulate actin polymerization through profilin. This is consonant with the finding that PIP2, which changes the tryptophan fluorescence in wild-type profilin and dissociates the profilin:actin complex in vitro, binds more strongly to the W3N mutant profilin. Thus, the poly(L-proline)-binding surface represents a crucial regulatory site of profilin function.