Investigation of Aip1p Function in S. cerevisiae
We first discovered Aip1p as an actin interaction protein from a two-hybrid screen with yeast actin. The AIP1 gene encodes a 68 kDa, WD repeat containing protein. Proteins with WD repeats adopt a common structural fold called the ß-Propeller and tend to be found in multi-protein complexes that display dynamic behaviors. We have also found that Aip1p interacts with yeast cofilin (Cof1p) in a two-hybrid assay and that these proteins cooperate to affect actin dynamics. Both cofilin and Aip1p localize to the actin cortical patches. Aip1p localization to the cortical patches requires it be able to interact with actin and requires an ill defined function of cofilin.
The AIP1 gene is not essential in S. cerevisiae but the knockout strain does have distinct defects. Cortical patches appear to be larger and/or to stain with rhodamine-phalloidin more brightly. The knock out allele is synthetic lethal with several cofilin alleles including: cof1-4, cof1-5, and cof1-22, it is synthetic lethal with the actin filament stabilizing mutant act1-159 and synthetic sick with deletion alleles of sac6, cap2 and sla1. Note that the products of all of these genes are also found in the actin cortical patch. In addition, the aip1D strain has interesting defects in cofilin localization. Normally, cofilin is found in cortical patches and the cytosol, but in the aip1Dstrain cofilin also accumulates on actin cables.
Biochemically, Aip1p alone does not form a stable association with actin filaments. In contrast, it causes the very efficient destruction of actin filaments and only actin filaments optimally decorated with cofilin. In contrast, substoichiometric amounts of Aip1p can efficiently catalyze the destruction of cofilin decorated actin filaments. The Aip1p induced destruction of cofilin decorated actin filaments can also be observed in both dilution induced and Latrunculin A induced diassembly assays (unpublished) suggesting that Aip1p is not acting as a capping protein but is likely inducing cofilin severing activity. This is supported by studies of both Xenopus and Dictyostelium Aip1p using electron microscopy to show apparant severing of cofilin decorated actin filaments by Aip1p.
We have also footprinted the binding sites of cofilin and Aip1p on actin using a collection of actin alanine scan mutants and the two-hybrid system. The data resulting from this analysis is modeled on the actin structure at right. The mutations rendered in red disrupted the cofilin-actin two-hybrid interaction. This description of the cofilin binding site is in very good agreement with the modeling approach of Wriggers and Janmey and is consistent with the binding site of the related actin binding/severing protein gelsolin. The mutations rendered in both red and blue disprupted the Aip1p-actin two-hybrid interaction. We believe the apparant overlap of the Aip1p and cofilin binding sites likely means that Aip1p binding to actin is dependent on cofilin binding to actin.
More recently, our collaborators in the lab of David Sept, have derived a model for cofilin-Aip1p interaction that is in complete agree with our mutagenesis data. Their model is shown at right:
Aip1p is rendered in green and cofilin is rendered in orange. The model predicts that cofilin is cradeled between the two ß-propeller domains of Aip1p.
Extensive mutagenesis of Aip1p has allowed us to uncover two separate actin binding sites on Aip1p and to map the cofilin interaction on Aip1p as well. Either propeller domain can interact with actin and, as shown at right and in red, mutations at comparable locations on each ß-propeller domain disrupt their respective interactions with actin. The other mutants that are highlighted are disruptive to the Aip1p cofilin interaction and lie within the contact surface predicted for the Aip1p-cofilin interaction displayed above. More recently, we have data supporting the existence of another contact sight for Aip1p on actin allowing us to start to think about how the entire complex assembles onto an actin filament. The figure at right is a hand placed model in which we imagine that Aip1p reaches across cofilin to contact two actin subunits via binding sites on its N- and C-terminal propeller domains. We theorize that Aip1p binding induces additional torsional stress on the filament leading to severing.
Our papers on Aip1p:Amberg DC, Basart E, Botstein D. (1995) Defining Protein Interactions with Yeast Actin in vivo. Nature Structural Biology; 2: 28-35.
Rodal AA, Tetrault JW, Lappalainen P, Drubin G and Amberg DC. (1999) Aip1p interacts with cofilin to disassemble actin filaments. J Cell Biol.; 145: 1251-1264.
Clark MG, Teply J, Haarer BK, Viggiano SC, Sept D and Amberg DC. (2006) A genetic dissection of Aip1p's interactions leads to a model for Aip1p-cofilin cooperative activities. Molec. Biol. Cell; 17: 1971-1984.
Clark MD and Amberg DC. (2007) Biochemical and genetic analyses provide insight into the structural and mechanistic properties of actin filament disassembly by the Aip1p-cofilin complex in Saccharomyces cerevisiae. Genetics. Jul;176(3):1527-39. Epub 2007 May 4.