The role of type 3 CH domains in signaling and cytoskeletal proteins

We have recently analyzed the sequences of all CH domain-containing proteins currently available and defined three groups of CH domain containing molecules: i) the fimbrin family of monomeric actin cross-linking molecules containing two actin binding domains (ABDs): ii) dimeric cross-linking proteins ?like??-Actinin, ?-Spectrin, filamin etc.) and F-actin binding proteins (dystrophin, utrophin) each containing one ABD, and iii) proteins containing only a single (type 3) CH domain. The amino terminal (type 1) CH domains of proteins containing 1 ABD are clearly more similar to each other than they are to the carboxy terminal (type 2) CH domains of the motif. Type 3 CH domains are distinctly different from the individual type 1 and type 2 CH domains found in tandem. It is well established that type 1 and type 2 CH domains have largely different affinities for actin, although their three dimensional structure may be very similar. Thus, the biological function of a type 3 CH domain cannot be deduced from either that of a type 1 or a type 2 CH domain. Replacement of a type 3 CH domain with any of the other type 3 CH domains, or even a complete ABD, fails to functionally replace the original module. Thus, the contribution of type 3 CH domains to the function of regulatory molecules of the Rho-signaling pathway like Vav and IQGAP, but also to that of cytoskeletal components such as CaP and SM22, whose biological function is still under debate, needs to be resolved. Elucidating the biological function(s) of the CH domain is a challenging task since this new protein module is present in both actin binding proteins and molecules involved in the regulation of small GTPases. To fully understand the function of single CH domains a broad range approach towards identifying the factors that determine the functional specificity of the individual CH domains is required. In this project we wish to address the question as to which single CH domains are functionally similar and, where function is different, we plan to investigate to which extent the conserved amino acid residues define their specific function in each subgroup, using site directed mutagenesis and domain swap experiments. This will entail a detailed cell biological analysis of the subcellular localization, targeting potential and localization dynamics of the individual isolated CH domains, but also of their contribution to the function of the whole molecule. Specific attention will be paid to the comparative analysis of the CH domain surface properties as determinants of the specificity for interaction with binding partners. Exploiting our CH domain sequence database we will also attempt to search all available databases to identify new type 3 CH domain-containing molecules. Employing co-immunoprecipitation studies in parallel with yeast two-hybrid system analyses we will further aim at identifying new, specific interaction partners for the diverse type 3 CH domains.

The aim of this project was to shed light on the biological function of single type3 CH domains in cytoskeletal and signaling molecules. We addressed this question by site directed mutagenesis and domain swap experiments followed by detailed cell biological analysis of the subcellular localization, targeting potential and localization dynamics of the individual isolated CH domains, and their contribution to the function of the whole molecule. Employing co-immunoprecipitation and the yeast two-hybrid system we identified new, specific interaction partners. For both Vav and Calponin we obtained a reliable yeast two-hybrid interaction with the stathmin variant SCLIP. We have confirmed this molecular interaction by immunoprecipitation in mammalian cells. Stathmin-family members regulate the polymerization dynamics of microtubules by interaction with a/ß protomers. This result is of particular relevance since the EB family of microtubule plus-end binding proteins also contains a single CH domain (of a novel type), and binding to tubulin requires the presence of the CH domain. Thus, our results indicate that single CH domains can function as potential microtubule anchoring modules, linking the actin cytoskeleton to the microtubule system. For SM22 we have identified the LIM domain-containing protein CRP2 as a binding partner for the CH domain. SM22 and CRP2 colocalize in transfected A7r5 smooth muscle cells and co-sediment in vitro with purified F-actin. Our studies support SM22 as a prime candidate for a component of the "mechanosensor-complex", which drives smooth muscle-specific gene-expression via the serum response factor SRF and the formin mDia, and which is regulated by the fluctuations in the levels of G and F-actin in the cell. We have determined the localization of isolated, GFP-tagged CH domains from Vav, Calponin and SM22 in fixed cells and under live conditions. In all cases, the CH domain associated with a particulate membrane fraction, reminiscent of endocytotic vesicles. Our domain swap experiments confirmed our initial hypothesis that CH domains from different subfamilies within the type3 CH domain family are not functionally interchangable. Exchange of the Vav CH domain with the homologous region from either Calponin or SM22 failed to reduce the GEF activity of Vav. Based on sequence alignments we can now attribute the functional differences to the loop region between helices A and B in the CH domain fold. These findings are in good agreement with the recently solved NMR structure of the Calponin CH domain, which highlights significant structural differences between type1, type2 and type3 CH domains in this region. A striking result was also obtained with a chimera of the C.elegans protein UNC-87 and h1CaP. The actin binding modules of both proteins are highly conserved, but UNC-87, which lacks an N-terminal CH domain is by about 4-fold more efficient in its stabilizing function than h1CaP. Fusion of the h1CaP CH domain onto the N- terminus of UNC-87 significantly reduced the effect of UNC-87 and the mutant protein displayed an activity similar to that seen with h1CaP. These findings support once more that the type3 CH domains perform regulatory functions on the diverse proteins. The work performed within the framework of this project helped to underscore the hypothesis of "protein lingusitics", a rapidly growing field aiming at understanding the use of protein modules as building blocks for protein diversity in evolution.