| |
(c) copyright 2004 David Siderovski & Francis Willard The
spectrum of RGS protein structure and function
Founding
members of the RGS protein superfamily were discovered in 1996 in a
wide
spectrum of species: “supersensitivity to pheromone-2” (Sst2) in the
budding
yeast Saccharomyces cerevisiae
[5, 19, 20], FlbA in the
aspergillus
Emericella
nidulans [9], EGL-10 in the nematode
worm Caenorhabiditis elegans [7],
and RGS1 and RGS2 from human B- and T-lymphocytes, respectively [6, 8].
Nearly
a decade later, new RGS-box-containing proteins are still being
identified in
mammalian species (e.g., humans in Fig. 3) and even in plants (Fig.
4C).
For example, RGS21 was recently identified as a putative component of
taste bud
signal transduction that is transacted in lingual epithelium via the
T1R2/T1R3
sweet taste and T2R bitter taste 7TM receptors [21]. At a mere
152 amino-acids, RGS21 is the
shortest RGS protein known to date, with no obvious structural domain
N- or
C-terminal to the central RGS-box. The singular nature of the RGS21
domain
structure is typical of the A/RZ and B/R4 subfamilies (Figs. 2&3)
but is
unlike other superfamily members that either contain two or more
RGS-boxes in
tandem (i.e., the PKA regulatory subunit binding partner D‑AKAP2 [22]
and the testes-specific protein PRTD-NY2 [a.k.a. RGS22; Willard &
Siderovski, unpublished observations]) or possess one or more
functional
modules beyond the defining RGS-box (Fig. 2). Several recent findings
as to the
functions of these multi-domain RGS proteins are described below.
![]() Figure 2. Multidomain architecture of representative members from all subfamilies of the mammalian RGS protein superfamily. Two alternative nomenclatures have been proposed for several RGS protein subfamilies [23, 24]. RZ- or A-subfamily members such as RGS17 [25] are characterized by an N-terminal poly-cysteine region (“Cys”) thought to be reversibly palmitoylated [26]. R4- or B-subfamily members include RGS2 and RGS21 that are described in the text. RGS11 was the first member of the R7- or C-subfamily to be shown to bind Gbeta5 via its Ggamma-like or “GGL” domain [27]. Of the three members of the R12- or D-subfamily (RGS10, RGS12, RGS14), RGS10 is the smallest and comprises little more than an RGS-box [11], whereas both RGS12 and RGS14 have tandem Ras-binding domains (RBDs) and a C-terminal Galpha-i/o-Loco interaction (GoLoco) motif [28], and RGS12 additionally has N-terminal PDZ (PSD95/Dlg/ZO-1 homology) and PTB (phosphotyrosine-binding) domains [29]. Axin and Axin2 (a.k.a. Axil) are negative regulators of the Wnt signaling pathway and comprise the RA- or E-subfamily; neither protein has been shown to interact with Galpha subunits, but rather interact with the tumor suppressor protein adenomatous polyposis coli (APC) using the RGS-box [30]. Axin and Axil also contain other domains that interact with beta-catenin, the kinase GSK3beta, the phosphatase PP2A, and the protein Dishevelled (“DIX” domain) [31]. The GEF- or F-subfamily includes three RhoA-specific guanine nucleotide exchange factors (“GEFs”) with canonical Dbl-homology (DH) and pleckstrin-homology (PH) domains: p115-RhoGEF, PDZ-RhoGEF, and leukemia-associated RhoGEF (LARG); the latter two RhoGEFs each possess an N-terminal PDZ domain, as described in the text. In 1996, we were the first group to identify [8] an N-terminal RGS-box within each member of the G protein-coupled receptor kinase family (known as the GRK- or G-subfamily in the context of the RGS protein superfamily). At least three sorting nexins (SNX13, SNX14, SNX25) have RGS-boxes between phosphatidylinositol-binding (PX) and PX-associated (PXA) domains and thus comprise the SNX- or H-subfamily of RGS proteins. Zheng and colleagues reported that SNX13 (a.k.a. “RGS-PX1”) could act as a GAP for the adenylyl-cyclase-stimulatory isoform of Galpha (Galpha-s) [32]; however, this report has yet to be confirmed in the literature. TM, putative transmembrane regions. The multiple RGS-box family members D-AKAP2 and RGS22 fall outside the eight established subfamilies; the superscript designations of their RGS-boxes match that used in Figure 3. ![]() Figure 3. Relationship between RGS-box sequences of all 37 human RGS proteins identified to date. Unrooted dendrogram was generated by Clustal-W [33] and TreeView [34] using sequences identified by the SMART profile [35] for RGS-boxes as well as those identified by protein-fold recognition algorithms [36]. Subfamily designations and identification of isolated RGS-box sequences from multi-RGS-containing proteins D-AKAP2 and RGS22 are as described for Figure 2. Note that there is no RGS15, contrary to an early report [7]. |