The FGDY-specific X;8 translocation breakpoint within the Xp11.21 region was used as a molecular signpost to positionally clone the X-linked Aarskog-Scott syndrome gene. Molecular characterization of the FGDY-specific translocation breakpoint permitted construction of a physical map of the translocation breakpoint region and assignment of the FGDY locus to a specific interval within this map.25 DNA markers flanking the disease-specific breakpoint were used to assemble a regional yeast artificial chromosome (YAC) contig of the breakpoint region. DNA clones derived from this contig were used to identify regional transcripts including the FGD1 cDNA25 (Fig. 247-2). A number of lines of evidence indicated that FGD1 was the gene responsible for X-linked FGDY: (1) FGD1 mapped to Xp11.21, the region known to contain the FGDY disease locus, (2) the FGD1 gene was directly disrupted by the FGDY-specific t(X;8) breakpoint, (3) an insertional mutation, predicted to result in a severely abbreviated and nonfunctional FGD1 protein, segregated with the phenotype in an affected FGDY family, and (4) FGD1 mRNA was expressed in tissues involved in the disease phenotype, including fetal craniofacial bones.25
A schematic representation of the Aarskog-Scott syndrome region within Xp11.21 showing the FGD1 gene positional cloning strategy. The FGDY-specific X;8 translocation breakpoint was mapped to the region between loci ALAS2 and DXS323;23 locus order was determined previously.24 Bars indicate the relative X chromosome content of somatic cell hybrid lines used to map DNA markers to specific intervals within the Xp11.21 region; all bars extend to Xqter.24,26 A detailed composite long-range restriction map of the FGDY breakpoint region derived from YAC clones 21G3 and 29D4 is shown below; bars indicate the relative clone content. LE and RE indicate the left and right ends of clone 21G3, respectively; the interrupted bar indicates a chimeric clone segment. A composite restriction map of the FGD1 cDNA is shown below the YAC diagram; bars indicate individual clone content. (From Pasteris et al.25 Used with permission of Cell.)
FGD1 Encodes a Rho Guanine Nucleotide Exchange Protein
An analysis of the FGD1 gene sequence provided insights into FGDY pathogenesis. The gene encodes a 761-amino-acid protein that displays strong evolutionary conservation. Fgd1, the mouse FGD1 homologue, is 95 percent identical to FGD1.26 A Caenorhabditis elegans FGD1 homologue also has been identified.27 Comparative sequence analysis suggested that FGD1 encoded a guanine nucleotide exchange factor, or activator, for a member of the Rho family of p21 GTPases.25 The Rho GTPases form a subgroup of the Ras superfamily of 20- to 30-kDa GTP-binding proteins. As a group, Rho proteins have been shown to play crucial roles in a wide spectrum of cellular functions including regulation of the actin cytoskeleton, membrane trafficking and vesicular transport, transcriptional regulation, cell growth control, and embryonic morphogenesis.28 At least 10 mammalian Rho-like GTPases are known: RhoA, -B, -C, -D, and -E; Rac1 and -2; RacE; Cdc42, and TC10.28 Sequence analysis shows that the Rho proteins from various species are conserved in structure and are about 50 percent homologous to each other.
Rho GTPases function as molecular switches, cycling between an inactive GDP-bound state and an active GTP-bound state. Rho guanine nucleotide exchange factors (RhoGEFs), including FGD1, constitute a rapidly growing family of diverse proteins that activate the GTPase Ras-like family of Rho proteins by catalyzing the exchange of bound GDP for free GTP. As shown schematically in Fig. 247-3, the ratio of the two states of a Rho GTPase is regulated by the opposing effects of guanine nucleotide exchange factors (GEFs), which catalyze the exchange of bound GDP for GTP, and the GTPase-activating proteins (GAPs), which increase the intrinsic rate of hydrolysis of bound GTP.29 At least 20 different RhoGEFs are known.30 The structural organization of several RhoGEFs is shown in Fig. 247-4. Like FGD1, all RhoGEF family members contain a 200-amino-acid RhoGEF domain. A comparison of RhoGEF domain sequences from various species shows that they are conserved in primary structure and are 25 to 30 percent homologous.30 Dbl, the prototype RhoGEF, was isolated originally by its ability to induce focus formation and tumorigenicity when expressed in NIH-3T3 cells.31 Lbc, ect2, and most of the isolated RhoGEFs also were isolated by virtue of their transforming capability through gene transfer experiments.30 In contrast, RhoGEFs including Cdc24, Bcr, mSos1, RasGRF, and Vav were identified by their role in cell growth regulation.30 The recognition that Dbl contained a 29 percent sequence identity with the Saccharomyces cerevisiae cell division cycle protein Cdc24, a known yeast Cdc42 activator, provided the first clue that Dbl was a RhoGEF.32 Biochemical analysis showed that Dbl was able to catalyze the release of GDP from the human homologue of Cdc42. Deletion analysis showed that the Dbl RhoGEF domain was essential and sufficient for the Cdc42 exchange activity and Dbl oncogenicity.32-34 It remains to be determined as to how the RhoGEF domain catalyzes the exchange of GDP for GTP. For the RasGEF Son of sevenless (Sos), structural studies indicate that Sos catalyzes GDP-GTP exchange by binding to Ras to alter the structure of its nucleotide switch regions, thereby reducing the affinity of the Ras molecule for GDP.35 However, among GEFs with a known molecular structure (i.e., Sos and ARNO, the GEF for the Arf small GTPase), although the GEF domains are primarily composed of α-helixes, the three-dimensional structure of ARNO does not resemble that of a RasGEF.36,37 These results imply that the structural mechanisms of nucleotide exchange may differ among GEF proteins.
RhoGEFs activate a Rho protein GTPase by catalyzing the exchange of GDP for GTP. A variety of stimuli lead to the activation of Rho protein family members via RhoGEFs, including the p21 GTPase Ras, receptor protein tyrosine kinases, and G protein-coupled receptors.28 Activated Rho leads to modified cell morphology by a reorganization of the actin cytoskeleton, a modulation of gene transcription by the activation of MAPK cascade, and the sequential activation of other Rho family member proteins. RhoGAP facilitates the hydrolysis of GTP and Rho protein inactivation. (From Gorski.5 Used with permission of Humana Press.)
A schematic representation of the molecular structure of the FGD1 protein compared with other RhoGEF proteins. Structural domains are drawn approximately to scale. FGD1 contains at least four distinct domains including a RhoGEF domain and a PH domain, motifs common to all RhoGEF family members.30 FGD1 also contains a cysteine-rich FYVE domain, a PtdIns(3)P-binding domain, and a putative SH3-binding (Grb2-binding) region. Vav contains SH3, Src-homology 2 (SH2), and a putative diacylglycerol/phorbol ester-binding (DAG/PE) zinc butterfly motif. Bcr contains a Rho GTPase activator protein (RhoGAP) domain. RasGRF and mSos1 contain a Ras guanine nucleotide exchange factor (RasGEF) domain. (Adapted from Gorski.5 Used with permission of Humana Press.)
Cellular microinjection experiments and biochemical studies show that FGD1 is a specific activator for the p21 GTPase Cdc42.38,39 Studies showed that the FGD1 GEF domain specifically complexed to Cdc42 but that it did not bind to other Rho proteins.38 In addition, in a reconstituted in vitro system, the FGD1 GEF domain stimulated [3H]GDP dissociation from and [35S]GTP binding to Cdc42.38 When microinjected into cultured cells, the FGD1 GEF domain induced fibroblasts to form filopodia, actin-associated membrane complexes generated by activated Cdc42 (Fig. 247-5). Studies showed that the FGD1 GEF domain specifically interacted with the Cdc42 protein and that FGD1-dependent filopodia formation was blocked by complexing Cdc42 to other Cdc42-binding proteins.39 In FGD1-expressing fibroblasts, c-Jun N-terminal kinase/stress-activated protein kinase (JNK/SAPK) activity was stimulated in a manner similar to that obtained with constitutively activated Cdc42.38,39 In addition, like constitutively expressed Cdc42, FGD1 stimulated the S6 kinase signaling cascade38 and stimulated the passage of fibroblasts through the G1 phase of the cell cycle.39 Together these results showed that FGD1 is a specific Cdc42 activator and a component of the Cdc42 signaling pathway.
Fibroblasts microinjected with RhoGEF expression constructs showing the promotion of actin polymerization through the activation of different Rho proteins. FGD1, Dbl, Lbc, and Vav promote the polymerization of actin in quiescent cells; serum-starved Swiss 3T3 cells were microinjected with plasmid DNA encoding epitope-tagged RhoGEF proteins.39 FGD1 stimulates the formation of filopodia (activation of Cdc42). In contrast, Dbl and Vav stimulate the formation of lamellipodia (activation of Rac), and Lbc stimulates the formation of actin stress fibers (activation of Rho). A construct containing an alternatively spliced form of FGD1 that lacks 36 amino acids near the N-terminal end of the GEF domain, FGD1Δ, fails to stimulate filopodial formation. Actin cytoskeletal structures in cells expressing the protein constructs were visualized with TRITC-conjugated phalloidin. Scale bar in A represents 20 μm and refers to all panels. (From Olsen et al.39 Used with permission of Current Biology.)
Immediately adjacent to the RhoGEF domain, all RhoGEF proteins, including FGD1, contain a region of sequence homology of approximately 120 amino acids termed the pleckstrin homology (PH) domain 25 (see Fig. 247-4). This domain was first identified as a duplicated and conserved domain in the protein pleckstrin, the major substrate of protein kinase C in platelets.40 PH domains form a diverse family of signaling domains.41 Some PH domains have been shown to bind to the second messenger molecule phosphatidylinositol-4,5-bisphosphate.42 Alternatively, other PH domains have been shown to bind to the βγ-subunits of the G-proteins43 or to proteins containing a phosphotyrosine binding (PTB) domain.44 No specific ligand has yet to be identified for the PH domain associated with the RhoGEF family of proteins. However, in the absence of an identified ligand, the PH domain has been shown to be essential for proper cellular localization of RhoGEF proteins.45
A comparative analysis of FGD1 showed that it contained at least two additional conserved signaling motifs that potentially could regulate the localization and/or activity of the Cdc42GEF domain. First, the FGD1 N-terminal proline-rich region was found to contain at least two putative Src-homology 3 (SH3) domain binding sites that exhibited strong similarity to the functionally significant regions of several proteins with demonstrated SH3 domain binding, including mSos1, a RasGEF known to bind to the SH3 domain of Grb2.25 SH3 domains have been shown to specifically bind short (9- to 10-amino-acid) proline-rich structurally conserved motifs.46-48 Grb2, a component of the Ras signaling pathway, was shown to selectively bind to the proline-rich motifs of the mSos1 protein to form a link in a signal-transduction pathway that functionally ties tyrosine kinase receptors to Ras.46,48 Among the identified Ras and RhoGEF family members, mSos1 is unique in containing an SH3-, or Grb2-, binding domain. The identification of a putative proline-rich SH3-binding domain in FGD1 infers that, like Sos, the location and/or activity of the FGD1 protein may be modified by proteins containing an SH3 domain. Second, immediately downstream from the PH domain, FGD1 was found to contain a cysteine-rich evolutionarily conserved zinc-finger motif, termed the FYVE domain 25 (see Fig. 247-4).
At least 30 different proteins of yeast, nematode, plant, insect, and mammalian origin have been found to contain this domain.49 Most of the proteins containing the FYVE domain are known to be involved in membrane trafficking, including the yeast vacuolar sorting proteins Vac1p50 and Vps27p,51 the yeast phophatidylinositol-4-phosphate-5-kinase Fab1,52 and the mammalian ATPase Hrs-2.53 Recent results have illuminated the role the FYVE domain plays in cellular signaling. Stenmark and coworkers have shown that the FYVE domains of both EEA1, a mammalian early endosomal protein, and Hrs-2 selectively bind to phosphatidylinositol-3-phosphate (PtdIns-3-P) and that this binding is necessary for the proteins to localize to the early endosomal membranes.49,54 Similar studies showed that Vac1p and Vps27 also selectively bind to PtdIns(3)P.55 The observation that the predicted FGD1 sequence contains an EEA1-like FYVE domain suggests that the FGD1 protein is likely to bind to and interact with phosphatidylinositol second-messenger molecules. Therefore, it is likely that the FGD1 protein interacts with or is regulated by the components of multiple signal-transduction pathways.