Gib2, A Novel Gβ-like/RACK1 Homolog, Functions as a Gβ Subunit in cAMP Signaling and Is Essential in Cryptococcus neoformans*

Canonical G proteins are heterotrimeric, consisting of α, β, and γ subunits. Despite multiple Gα subunits functioning in fungi, only a single Gβ subunit per species has been identified, suggesting that non-conventional G protein signaling exists in this diverse group of eukaryotic organisms. Using the Gα subunit Gpa1 that functions in cAMP signaling as bait in a two-hybrid screen, we have identified a novel Gβ-like/RACK1 protein homolog, Gib2, from the human pathogenic fungus Cryptococcus neoformans. Gib2 contains a seven WD-40 repeat motif and is predicted to form a seven-bladed β propeller structure characteristic of β transducins. Gib2 is also shown to interact, respectively, with two Gγ subunit homologs, Gpg1 and Gpg2, similar to the conventional Gβ subunit Gpb1. In contrast to Gpb1 whose overexpression promotes mating response, overproduction of Gib2 suppresses defects of gpa1 mutation in both melanization and capsule formation, the phenotypes regulated by cAMP signaling and associated with virulence. Furthermore, depletion of Gib2 by antisense suppression results in a severe growth defect, suggesting that Gib2 is essential. Finally, Gib2 is shown to also physically interact with a downstream target of Gpa1-cAMP signaling, Smg1, and the protein kinase C homolog Pkc1, indicating that Gib2 is also a multifunctional RACK1-like protein.

In eukaryotic organisms, GTP-binding proteins transduce external signals to regulate a wide array of cellular functions. In fungi, such regulatory functions include the control of cellular growth and differentiation, asexual and sexual reproduction, the production of secondary metabolites, and/or virulence (for review, see Refs. [1][2][3]. Whereas G␣ subunits are known to exert most regulatory functions, free G␤(␥) subunits also regulate activities of effectors either independently or in concert with G␣s (4,5). G␤ subunits contain a distinct seven-bladed propeller structure with each blade composed of a conserved core of ϳ40 amino acids flanked by Trp-Asp (WD) (6,7). Unlike the higher eukaryotes in which multiple G␤ subunits have been identified, only a single G␤ per species has been found in fungi that regulates functions such as mating (8,9), growth and sporulation (10 -12), infection-related morphogenesis (13), and virulence (10).
Similar to other fungi, there exists an elaborate G proteinsignaling network in Cryptococcus neoformans that senses the environmental and host-imposed cues to regulate growth, differentiation, and virulence (26 -28). C. neoformans is an encapsulated yeast-like basidiomycetous fungus and the main cause of meningoencephalitis in individuals with a compromised immune system (29). Previous studies have shown that the G␣ subunit Gpa1 functions in a conserved cAMP-dependent signaling pathway and this pathway regulates a variety of cellular functions, including specialized processes such as the production of the antioxidant melanin pigment and the antiphagocytic capsule, two well established virulence factors in this pathogen (30 -32).
In contrast to Gpa1, the G␤ subunit Gpb1 functions to regulate pheromone-responsive mating and haploid differentiation upstream of a conserved mitogen-activated protein kinase cascade, and Gpb1 is not involved in virulence (9). The completed C. neoformans genome project has revealed two additional G␣ subunits, Gpa2 and Gpa3, whose functions have not yet been elucidated (33). Whereas either Gpa2 or Gpa3 could couple to Gpb1 and a G␥ subunit(s) to form a functional heterotrimeric G protein complex, no G␤ subunit that could couple to Gpa1 in cAMP signaling was found. Thus, answering whether G protein signaling is indeed unique and thereby explaining the presence of a single G␤ subunit would represent a significant achievement in the understanding of G protein signaling in this or other fungal organisms. We now report that a novel G␤-like/ RACK1 protein, Gib2, functions as an atypical G␤ in Gpa1-cAMP signaling. We also show that Gib2 encodes RACK1-like multifunctions, including an essential one in C. neoformans. Our studies suggest that a similar signaling mechanism could exist in other organisms where multiple G␣ subunits but only a single G␤ subunit has been found.
pGBKT7-GIB2 was used to screen the H99 cDNA library by the same method described above. Full-length SMG1 cDNA was identified from the H99 data base using positive cDNA as trace and amplified by PCR with primers PW357 (5Ј-AAGAA-TTCATGGTTCACGCTGCTACTCACCCC-3Ј) and PW358 (5Ј-AAGGATCCTTACTTTGTCTCTTTGTAAAGGTC-3Ј). The sequence was cloned into pGADT7 at EcoRI-PstI sites, creating pGADT7-SMG1.
Protein Structures Modeling-The protein modeling tool SWISS-MODEL was used to model the structures of Gpb1 and Gib2. Gpb1 and the partial Gib2-1 clone containing five C terminus WD repeats were analyzed through the "first approach mode" using bovine G␤ as the model template (39), whereas the full-length Gib2 was analyzed by comparative modeling using Schizosaccharomyces pombe Git5 (GenBank AAD09020) and Candida albicans Tup1 (AAB63195) as additional templates (40 -42). Protein structures were colored according to B-factor and displayed through the SWISS PDB viewer, DeepView (43)(44)(45).
Disruption of the GIB2 Gene-Because of sequence divergence between serotype A and D strains, two serotype-specific gib2::URA5 gene knock-out alleles were constructed by a twostep process. For serotype A, a 1.6-kb region encompassing the GIB2 gene was first amplified from H99 by PCR into two partially overlapping fragments that were each ϳ0.8 kb in length. The first fragment was amplified using primers PW166 (5Ј-CTTCGTTCATCTTTCACTGTTC-3Ј) and PW226 (5Ј-TGG-AAAAGTTGGTCTCATCCCGGGTCTGTTTCACCCCTCT-3Ј), and the second fragment was amplified using primers PW225 (5Ј-AGAGGGGTGAAACAGACCCGGGATGAGA-CCAACTTTTCCA-3Ј) and PW167 (5Ј-ACGAGACGACTC-TGATTCTGAC-3Ј). A SmaI restriction site was incorporated into primers PW225 and PW226, and a second PCR amplification with primers PW166 and PW167 resulted in a 1.6-kb frag-ment containing the SmaI restriction site. This fragment was cloned into the TOPO TA vector (Invitrogen).
The URA5 gene was inserted into the SmaI site generating a gib2::URA5 disruption allele. The mutant allele was introduced into the F99 strain by biolistic transformation according to the published protocol (46). The same primer set was used to create the serotype D-specific gib2::URA5 disruption allele using JEC21 as the template. This allele was also transformed into JEC34, JEC43, and RAS009 strains by biolistic transformation.
All resulting transformants were screened by PCR. Heterozygous RAS009 transformants (gib2/GIB2) were verified by Southern hybridization analysis before subjecting to ploidy reduction. The smg1::URA5 knock-out allele was generated by overlapping PCR (47) and was used to generate smg1::URA5 mutant strains by transformation of both F99 and F99a strains.
Overexpression and Suppression of Gib2-The GIB2 gene was inserted into the pCnTel1 vector containing the C. neoformans galactose inducible GAL7 promoter (48) at the EcoRI site, and the resulting construct (sense orientation) was introduced into JEC43, JEC34, and BAC20-1 (gpa1 ura Ϫ ) strains by biolistic transformation. The BAC20-1 strain was also transformed with the GAL7-GIB2 construct in which the orientation of GIB2 was reversed (antisense construct). Transformants were verified by PCR, diluted serially, and spotted on YPD, YNB, and YNB containing either 2% galactose or 2% glucose. Cells were grown on filament agar to promote conjugation tube formation and on Niger seed agar for melanin pigmentation. Cells grown on Niger seed medium were also collected and stained with India ink to observe the capsule. RNA and protein were extracted from cells first grown overnight in YPD and then re-suspended in minimal YNB medium containing 2% galactose for up to 4 h (36). RNA and protein were also extracted from strains grown continually in YNB containing 2% galactose, or otherwise as stated. The GIB2 transcript and Gib2 protein were detected by Northern and Western blot analysis accordingly (36). A Gib2-specific oligopeptide, PW3 (NЈ-CPD-FDGLSDKARKPE-CЈ), was synthesized and the affinity purified Gib2-specific antibody was obtained through a commercial source (GenScript Corp., Piscataway, NJ).
Protein "Pulldown" Assays-The Gpa1 protein was expressed by an in vitro transcription and translation kit using the TNT T7 coupled reticulocyte lysate system (Promega). The Gib2 protein was produced in E. coli (Rosetta DE3, Novagen) as a gluta-thione S-transferase (GST) fusion protein using pET-41a(ϩ) also from Novagen. Expression of the GST fusion protein was induced for 5 h at room temperature and extracted in lysis buffer (50 mM HEPES, pH 7.6, 150 mM NaCl, 10 mM MgCl 2 , 0.5% Nonidet P-40, 1 mM EDTA, 1 mM EGTA, 25 mM ␤-glycophosphate, 100 M Na 3 VO 4 , and protease inhibitor mixture tablets (Roche) and 0.5 mM phenylmethylsulfonyl fluoride) as described (36,49). The Gpa1 protein (10 l) was conditioned in lysis buffer with 10 M GDP for 1 h at room temperature. Bacterial extracts containing GST-Gib2 were mixed with glutathione-Sepharose resin (Pierce) for 1.5 h at 4°C. The resin was washed four times with lysis buffer and mixed with Gpa1 in the same lysis buffer plus 10 M GDP overnight at 4°C. The gel was precipitated, washed three times with lysis buffer, and bound proteins were analyzed by SDS-PAGE and Western blotting using the anti-c-Myc peroxidase conjugates (Roche, number 1814150) and anti-GST antibody (Santa Cruz, SC-138).
The GPG2 gene was inserted in pET-41a and the GST-Gpg2 fusion protein was expressed similarly as above. Gib2 was also expressed in S. cerevisiae using the pYES2NT vector (Invitrogen). The pYES2NT-GIB2 plasmid was transformed into yeast strain INVS.c1 (Invitrogen), transformants were selected on SD-ura agar plates, and grown in liquid SD-ura medium containing 2% glucose for 24 h at 30°C. Cells were precipitated, washed with induction medium (SD-ura medium containing 2% galactose and 1% raffinose), and grown in the same medium overnight in a 30°C shaker, cells were then collected and stored at Ϫ80°C. To extract proteins, cells were resuspended in lysis buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 0.5% Nonidet P-40, and a protease inhibitor mixture (Roche)) and homogenized in a Fast-Prep cell disrupter (Bio 101 Inc., Vista, CA) six times for 40 s at 4°C, with an equal volume of acid-washed glass beads. Homogenized samples were centrifuged at 13,000 ϫ g for 30 min at 4°C, and supernatants were recovered. For protein purification, cell extracts were incubated with Co 2ϩ affinity resins (TALON, Clontech) for 1.5 h, washed, and proteins eluted with buffer containing 200 mM imidazole. Proteins were then concentrated, buffer exchanged using a centrifugal filter device (Millipore), and concentration determined by BCA protein assay (Pierce). The pulldown assay was carried out similarly as above (Gpa1-Gib2) in binding buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM MgCl 2 0.2% Nonidet P-40, 1 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride). After precipitation, resin was washed five times and bound proteins were analyzed by SDS-PAGE and Western blotting using the anti-Gib2 or anti-GST antibodies.

RESULTS
Identification of Gib2-The completed genome of C. neoformans var. neoformans has confirmed the presence of three G␣ subunits in C. neoformans: Gpa1, Gpa2, and Gpa3. Whereas Gpa1 is known to function in a conserved cAMP signaling pathway, functions of Gpa2 and Gpa3 remain unreported. We have recently identified a regulator of the G protein signaling protein, Crg1, as one of the key regulators for mating and virulence in C. neoformans (51,52) and have shown that Crg1 functions as a GTPase activating protein specific to Gpa2. 4 To expand our studies on regulators of G protein signaling, we performed a yeast two-hybrid screen using Gpa1 as bait.
Among 12 positive clones of 1 ϫ 10 8 yeast transformants, a single clone was found to contain, in pGADT7-REC, an insert of 743 nucleotides predicted to encode a peptide of 227 amino acids (GenBank AY954369). This clone, designated as GIB2-1 for Gpa1-interacting-␤, allowed interactiondependent growth of host strain AH109 only in the presence of Gpa1, but not Gpa2 or Gpa3 (Fig. 1A). Examination of the peptide sequence encoded by GIB2-1 revealed five WD repeats, and using the sequence to search C. neoformans sequence databases, an open reading frame containing an additional 87 N-terminal amino acids, for a total of 314 amino acids, was identified. The deduced amino acid sequence of Gib2 contains seven WD-40 repeats and shares 78% amino acid identity with the G␤ subunit G␤1 of shiitake mushroom Lentinula edodes (AAP13580), 72% identity with Neurospora crassa Cpc2 protein (CAA57460), and 70% identity with mammalian G␤-like/RACK1 protein GNB2L1 (AAH32006). In comparison, Gib2 shares 25% amino acid sequence identity with Gpb1 and 27% with S. pombe Git5 (supplementary Fig. S1). Apparently, Gib2 is more related to G␤-like/RACK1 proteins than G␤ subunit proteins (Fig.  2).
We named the full-length gene GIB2. A single copy of the GIB2 gene was found in both serotypes A (GenBank AY907679) and D (GenBank AY907680) strains that encodes an identical protein.
To confirm the Gib2-Gpa1 interaction, GIB2 cDNA was switched into the BD vector pGBT9 and co-transformed with pGAD424-GPA1. Only yeast cells carrying both genes were able to grow in SD medium lacking Leu, Trp, His, and Ade, plus 3-aminotrizole that was added to suppress leaky HIS3 expression of the reporter strain (Fig. 1B). A GPA1 Q284L mutant allele 4 L. Li and P. Wang, unpublished observation. . The DNA activation domain (AD) vector pGADT7-REC has an insert of GIB2-1 encoding the C-terminal partial sequence and the construct was recovered from the original yeast screen. GPA1, GPA2, and GPA3 were cloned into the DNA-binding domain (BD) vector pGBKT7. The AD and BD plasmids were co-transformed into AH109 and yeast transformants were patched on SD-Leu-Trp for plasmid-dependent growth (3 days) and transferred to SD-Leu-Trp-His-Ade plus 3-aminotrizole (3-AT) for protein interaction-dependent growth (3 days). B, two-hybrid assay confirming the positive interaction between Gib2 and Gpa1. The full-length cDNA (GIB2) was cloned into pGBKT7, and GPA1 cDNA was cloned into pGAD424. C, Gib2 interacts with Gpa1 in vitro. The Gib2-GST fusion and GST proteins were absorbed to glutathione gel, washed, and mixed with the Gpa1 protein. Bound proteins were separated by SDS-PAGE in duplicate and analyzed by Western blotting. Blotted membranes were incubated with either anti-c-Myc-peroxidase conjugate (left panel) or anti-GST monoclonal antibody (right panel). D, Gib2 is constitutively expressed in C. neoformans serotype A strains and high levels can be induced by switching cells to nutrient-limiting media such as YNB. The Gib2 protein is detected using an affinity purified Gib2-specific antibody as described in text. Novel Function of C. neoformans Gib2 OCTOBER 27, 2006 • VOLUME 281 • NUMBER 43 encoding a constitutively activated Gpa1 (30) was also tested for interaction with Gib2. A weak interaction was still observed in yeast cells expressing Gpa1 Q284L and Gib2, as growth can occur on selective medium in the absence of 3-aminotrizole (data not shown).
To test Gib2-Gpa1 interaction in vitro, a protein pulldown assay was performed. The c-Myc-tagged Gpa1 (ϳ48 kDa), produced by in vitro transcription and translation, was pulled down by the GST-Gib2 fusion protein (ϳ60.5 kDa) (Fig. 1C, left  panel), but not by the GST protein alone (ϳ26 kDa) (Fig. 1C,  right panel). Gib2 is constitutively expressed in serotype A cells grown in YPD, however, higher levels of expression were observed when cultures were switched to nutrient limiting YNB minimal medium (Fig. 1D).
Gib2 Is Modeled to Have a Seven-bladed ␤ Propeller-like Structure-The crystal structures of seven WD repeats containing proteins such as ␤ transducin and Tup1 transcription factor have shown that these proteins have a seven-bladed ␤ propeller structure with each of the seven WD repeats forming a discrete ␤ propeller fold (39,53). The mammalian RACK1 protein contains a seven WD repeat motif and has also been modeled to show a similar structure (40). Through SWISS MODEL, we have shown that the sequence of previously characterized G␤ protein Gpb1 is compatible with that of a ␤ transducin in forming a seven-bladed ␤ propeller-like structure (Fig.  3, left panel). The modeling also supported a partial structure of five blades for Gib2-1 containing five WD repeats (Fig. 3, right   upper panel). However, the same modeling did not support the sequence of the full-length Gib2 in forming a seven-bladed structure when the mammalian G␤ was provided as the only template. Nevertheless, Gib2 could be modeled to exhibit a seven-bladed ␤ propellerlike structure when sequences of C. albicans Tup1 (GenBank AAB63195, 30% homologous to Gib2) and Git5 were provided as additional templates and through comparative modeling by combining the model with that of Gib2-1 (Fig. 3, right lower panel).
Gib2 Binds to G␥ Subunits Gpg1 and Gpg2-Conventional G␤ subunits exhibit a high affinity for G␥ subunits and function as G␤␥ heterodimers to bind and stabilize GDP-bound G␣ subunits. In addition, a G␤ can associate with multiple individual G␥ subunits (4). To test whether Gib2 could physically associate with a G␥ subunit(s), two G␥ homologs were identified from C. neoformans by using the amino acid sequence of the mushroom fungus L. edodes G␥ subunit homolog G␥1 (GenBank Q870G5) to search the H99 data base. Two open reading frames each encoding a putative G␥ subunit were obtained and named Gpg1 (GenBank AY907677) and Gpg2 (GenBank AY907678). Gpg1 consists of 81 amino acids and shares 62% amino acid identity with L. edodes G␥1, whereas Gpg2 consists of 87 amino acids and shares 59% sequence identity with G␥1 (supplementary Fig.  S2). In comparison, Gpg1 and Gpg2 are less similar to S. cerevisiae Ste18 (28 and 30% identity, respectively) (supplementary Fig. S2). Both Gpg1 and Gpg2 exhibit common characteristics of G␥ subunits, short peptides with the C-terminal CAAX (C ϭ cysteine, A ϭ aliphatic, X ϭ cysteine, methionine, serine, etc.) motif marked for protein prenylation, which is necessary for membrane association and G␥ function (54).
cDNA for GPG1 and GPG2 was obtained and inserted into pGBKT7, creating pGBKT7-GPG1 and pGBKT7-GPG2. Using the yeast two-hybrid assay, both Gpg1 and Gpg2 were found to interact with Gib2 (Fig. 4A), suggesting that Gib2 forms heterodimers with Gpg1 or Gpg2. Similarly, Gpg1 and Gpg2 interacted with Gpb1 when expressed, respectively, as pGBKT7-GPG1, pGBKT7-GPG2, and pGADT7-GPB1. Previously, such an interaction could not be detected when proteins were expressed using the low-level expression plasmids pGBT9 and pGAD424 (data not shown).
A protein pulldown assay was performed to verify the interaction between Gib2 and Gpg2. Gpg2 was expressed as a bacterial GST-Gpg2 fusion protein and Gib2 was expressed in yeast as a His-tagged protein that was purified through immobilized metal affinity chromatography. Consistently, Gib2 bound to Gpg2 after five stringent buffer washes (Fig. 4C). No binding was detected between GST and Gib2 (Fig. 4C).
Gib2 Positively Regulates cAMP Signaling-Previously, Gpb1 was found to play a positive role in mating and haploid differentiation, because the gpb1 mutant strain was sterile and pheromone-induced conjugation tube formation, a response leading to mating, was diminished (9). In addition, overproduction of Gpb1 induced conjugation tube formation in the absence of pheromone stimulation (9). A similar approach was used to test Gib2 function by fusing GIB2 with the galactose-inducible GAL7 promoter (48) and by introducing the GAL7-GIB2 fusion gene into JEC34 and JEC43 strains, as well as the BAC20-1 strain. When grown in filament agar medium containing 0.5% galactose, no conjugation tube formation was observed in strains transformed with the GAL7-GIB2 construct (Fig. 5A, two left panels), in contrast to the same strain overexpressing Gpb1 (Fig. 5A, far right panel).
The gpa1 mutant strain is defective in cAMP signaling and mutant cells are unable to produce melanin and capsule. To determine whether Gib2 has a role in cAMP signaling, the gpa1 cells with the GAL7-GIB2 fusion gene construct were placed on melanin-inducing Niger seed agar. On medium containing 0.1% galactose, cells overexpressing Gib2 partially restored the production of melanin (Fig. 5B). The result is similar to that when cells were grown on medium supplemented with exoge-nous cAMP (5 mM final concentration, Fig. 5B). Intriguingly, colonies overexpressing Gib2 also exhibited a mucoid appearance indicating a restoration in capsule formation (Fig. 5B). Indeed, cells collected from the medium containing galactose showed capsules similar to or even larger than those of the wild type strain (Fig. 5C). No production of melanin or capsule was observed in the control gpa1 mutant strain (Fig. 5, B and C). Consistent with the observation that supports a positive regulatory role of Gib2 in cAMP signaling, the levels of intracellular cAMP in the Gib2 overexpression strain were similar to that of the wild type strain, again in sharp contrast to the gpa1 mutant strain (Fig.  5D).
Gib2 Encodes an Essential Function-To further dissect the function of Gib2, we employed an antisense suppression approach by transforming strains with a GIB2 antisense construct driven by the same GAL7 promoter. Modulation of gene expression by antisense suppression (or repression) occurs at the level of mRNA rather than at the level of DNA and this approach has been successfully employed in this organism previously (55). Suppression of GIB2, instead of gene disruption, provided us with an alternate experimental approach to assess Gib2 functions. In general, transformation of the BAC20-1 (gpa1 ura Ϫ ) strain with the GIB2 antisense construct yielded few transformants that were characterized by slow growth. The deleterious effect of Gib2 suppression was apparent in assays for melanin, capsule, and cAMP (Fig. 5,  B-D). Furthermore, the strains with the antisense construct exhibited a severe growth defect when cultured in YNB containing 2% galactose (Fig. 6A, right panel), in contrast to growth on nutrient rich YPD (Fig. 6A, left panel). Due to the observation that YNB induces Gib2 expression, as well as the leaky nature of the GAL7 promoter, strains with the GIB2 antisense construct also exhibited a partial growth defect on YNB (Fig.  6A, middle panel).
Suppression of Gib2 is synthetically lethal, indicating that Gib2 encodes an essential function in cellular growth. To test this hypothesis, serotype-specific gib2::URA5 mutant alleles were introduced into F99, JEC34, and JEC43 strains. Previous studies have shown that the knock-out efficiency via biolistic transformation in either serotypes is 2-40% (47). However, no knock-out strains were recovered from a total of 247 transformants, suggesting that Gib2 is required for cellular viability.
To confirm that Gib2 is essential, the laboratory derived serotype D diploid strain RAS009 (35) was employed. First, the gib2::URA5 allele was used by biolistic transformation to obtain gib2::URA5/GIB2 heterozygous transformants. Four such heterozygous transformants were obtained out of 198 transfor- . Gib2 associates with G␥ proteins Gpg1 and Gpg2. A, Gpg1 and Gpg2 interact with Gib2 in a yeast two-hybrid assay. GIB2 cDNA was cloned into pGADT7 (AD) and GPG1 and GPG2 cDNA was cloned into pGBKT7 (BD). Yeast transformation, selection of positive interaction transformants, and incubation conditions were the same as described in the legend to Fig. 1. B, ␤-galactosidase (gal) activities were measured to quantify bindings between G␤/G␤-like proteins and G␥ subunits. The assay was performed in duplicate and data were expressed as the mean Ϯ S.E. of two independent experiments. C, Gib2 interacts with Gpg2 in vitro. Gib2 was expressed in yeast and purified by chromatography. Gpg2 was expressed as a GST fusion protein in bacteria. Bound proteins were analyzed with the anti-Gib2 antibody (top panel). GST and GST-Gpg2 were visualized using the anti-GST monoclonal antibody (lower panel). Gib2 protein input (20 l) is shown in the first lane (top panel), whereas 300 l of purified proteins were used in each immunoprecipitation experiment. 3-AT, 3-aminotrizole. mants. Once heterozygosity was verified by Southern blot hybridization analysis (Fig. 6B), transformant number 65 was allowed to undergo meiosis and sporulation. Basidiospores were microdissected and genotypes of surviving progenies were examined. Among seven segregants recovered (2 MAT␣ and 5 MATa), all contained the wild type GIB2 gene, confirming that Gib2 is essential in serotype D strains (Fig. 6C).
Gib2 Interacts with Smg1, a Downstream Target of Gpa1-cAMP Signaling, and PKC-Our studies have provided evidence suggesting that Gib2 functions as an atypical G␤ subunit in Gpa1-cAMP signaling, and that it also encodes an essential function. We reasoned that an effector(s)/target(s) of Gib2 might provide additional clues for revealing the true identity of Gib2 and have thus performed a yeast two-hybrid screen using Gib2 as bait. A screen of 10 6 yeast transformants yielded several positive clones, including a peptide containing an Src homology 3 (SH3) domain and a peptide that encodes a glucose-methanol-choline oxidoreductase domain (GenBank AY954370). Whereas the prospect of Gib2 binding to an SH3 domain would support Gib2 being a G␤and RACK1-like protein because the mammalian RACK1 is known to negatively regulate the activity of the Src kinase through binding, we chose to examine the clone containing the oxidoreductase domain, as it represents an internal region of Smg1 (GenBank AF349949), a protein previously identified as a multicopy suppressor of gpa1 mutation. Smg1 has been implicated as one of the downstream targets of the Gpa1-cAMP signaling pathway in C. neoformans (56,57). Physical association between Gib2 and the full-length Smg1 was also confirmed (Fig. 7A).
Whereas a chromosome translocation and segmental duplication event resulted in two SMG1 alleles in serotype D strain JEC21 (58), only a single SMG1 allele was found in serotype A strains H99 and KN99a (data not shown). smg1::URA5 mutant strains of both mating types were generated to test the role of Smg1 in melanin production. Surprisingly, all five (three MAT␣ and two MATa) smg1::URA5 mutants exhibited levels of melanization similar to that of the FIGURE 5. Gib2 regulates cAMP signaling. A, Gib2 does not play a role in pheromone responsive mating. JEC34 transformants containing GAL7-GPB1 and GAL7-GIB2 constructs were patched on filament agar containing either 0.5% galactose or 0.5% glucose to induce or suppress the expression of Gpb1 or Gib2. Plates were incubated at room temperature (25°C) for 2 days before being photographed (ϫ25 magnification). B, overexpression of Gib2 suppresses the melanization defect of gpa1 mutation. The gpa1 ura Ϫ strain was transformed with forward (sense) and reverse (antisense) GAL7-GIB2 constructs, respectively. Cells were spotted on Niger seed agar containing 0.1% glucose (suppression), 0.1% galactose plus 0.1% raffinose (induction), or 0.1% glucose with cAMP. Plates were incubated at 30°C for 3 days before being photographed. C, overexpression of Gib2 suppresses the capsule defect of gpa1 mutation. Cells were collected from colonies on Niger seed agar with galactose (shown in Fig. 4B), washed in sterile water, and stained with India ink. Images were documented using an Olympus microscope (BX51) equipped with a digital camera. Scales of magnification are as indicated. D, Gib2 positively regulates intracellular cAMP. Cells were grown in YNB (with 0.2% galactose) overnight before being subjected to glucose starvation. Because of poor growth in medium containing galactose, 0.2% raffinose was added to cells with the GAL7-GIB2 antisense construct to improve growth. Each sample was assayed in quadruplet and mean values were plotted. WT, wild type. FIGURE 6. Gib2 is essential in serotype D strains. A, the BAC20-1 transformant strains with sense and antisense GAL7-GIB2 constructs were serially diluted and spotted onto YPD, YNB, and YNB with 2% galactose. Plates were incubated at 30°C for 2 days. B, Southern blot hybridization indicates the presence of a heterozygous GIB2/gib2::URA5 allele in diploid transformant number 65. Genomic DNA was digested with XbaI and XmnI, and the full-length GIB2 was used as the probe. The presence of the wild type GIB2 gene is indicated by a 1.6-kb DNA fragment, whereas the mutant gib2::URA5 allele is indicated by a 3.5-kb fragment. wild type strain (data not shown), bringing the role of Smg1 in melanin formation and virulence into question.
Because the mammalian RACK1 protein was originally identified and named for its ability to bind PKC, a protein pulldown assay was performed to test for the physical interaction between Gib2 and Pkc1. Pkc1 is a diacylglycerol-activated protein kinase C homolog that is a part of the inositol phosphorylceramide (Ipc1)-diacylglycerol-Pkc1 signaling pathway and this pathway is linked to the production of the virulence factor melanin in C. neoformans (50). Consistently, Gib2 binds to Pkc1 in vitro and the binding occurred even without the presence of the common PKC activator 1-oleoyl-2-acetyl-sn-glycerol (Fig. 7B). It is not known whether the heterologously expressed Pkc1 protein is already partially activated or whether Gib2 can bind to Pkc1 regardless of its activation state. The conserved nature of the interaction between these two proteins could suggest that Gib2 encodes a RACK1-like function important in PKC signaling of C. neoformans.

DISCUSSION
Gib2 Is an Atypical G␤-The discovery of multiple G␣ subunits, but only a single G␤ subunit, suggests that G protein signaling is unique in fungi: certain G␣s could function as monomeric proteins independent of G␤␥ heterodimers, a single G␤(␥) heterodimer could couple to multiple G␣s to transduce different signals, or certain proteins could substitute for functions of a conventional G␤. Whereas studies supporting the first two hypotheses are lacking, recent studies have shown that S. cerevisiae proteins Gpb1 and Gpb2 could couple to Gpa2 in cAMP signaling (49,59). The yeast Gpb1 and Gpb2 contain a seven-Kelch repeat motif mimicking the structure of the ␤ transducin. No similar proteins have been found in C. neoformans (33) or any other fungi, however.
The C. neoformans cAMP signaling pathway consists of several conserved components such as Gpa1, Cac1, AcaI, Pka1, and Pkr1 (30 -32, 60). Recently, a G protein coupled receptor homologue, Gpr4, has been found to activate Gpa1 by sensing ligands such as methionine (61), further highlighting the conserved mechanism of this pathway. Without a conventional G␤ subunit as a G␤␥ heterodimer, how Gpa1 accomplishes such a signaling role remains elusive.
The observation that Gib2 binds to Gpa1 may enable us to elucidate the mechanism by which Gpa1 functions as a G␣. Several lines of evidence suggest that Gib2 functions as an atypical G␤ by coupling to Gpa1. First, Gib2 was identified by virtue of its binding to Gpa1, which appeared to be specific, as Gib2 did not physically associate with Gpa2 or Gpa3. Second, the sequence of Gib2 is compatible with that of the ␤ transducin in forming a seven-bladed propeller-like structure. Third, Gib2 bound to Gpg1 and Gpg2 in the yeast two-hybrid and protein pulldown (Gpg2) assays. Fourth, overexpression of Gib2 restored cAMP signaling by suppressing defects in melanin and capsule formation due to gpa1 mutation. Finally, Gib2 interacts with Smg1, a multicopy suppressor of gpa1 mutation and a presumptive downstream target of cAMP signaling. Despite previous studies reporting that the mammalian RACK1 protein can bind to G␣(t)␤␥ and G␤␥ but not the G␣ subunit alone (21)(22)(23), our study does support a model in which Gib2 binds directly to Gpa1 as a G␤-like protein to likely stabilize Gpa1 and facilitate its activation and inactivation cycle, and to regulate cAMP signaling, in conjunction with Gpa1 (Fig. 8).
Gpg1 and Gpg2 Are Conserved G␥ Subunits-Our studies have demonstrated that Gpg1 and Gpg2 are G␥ subunit homologs that bind to Gpb1 and Gib2 to form protein heterodimers. It has been suggested that the G␥ folds into two linked ␣-helical domains that interact with the G␤ (39). Indeed, a truncated form of Gpg1 retaining only the N-terminal 43 amino acid residues still exhibited binding with Gib2 (data not shown). Binding between Gpb1 and Gpg1/Gpg2 could not be established initially using the low protein expression plasmids FIGURE 7. Gib2 interacts with multiple proteins. A, Smg1 was identified from a yeast two-hybrid screen using Gib2 as bait. pGADT7-REC-SMG1-1 was recovered from a positive yeast culture and full-length SMG1 cDNA was cloned into pGADT7. Yeast transformation, selection of transformants, and incubation conditions were the same as described in the legend to Fig. 1, except that the plate was incubated 5 days for interaction-dependent growth. B, in vitro GST pulldown assay showing the interaction between Gib2 and Pkc1. GST and GST-Gib2 fusion proteins were absorbed to glutathione gel, washed, and mixed with Pkc1. Bound proteins were separated by SDS-PAGE and filter membranes were blotted either with an anti-Xpress antibody (upper panel) or an anti-GST antibody (lower panel). FIGURE 8. Two conserved G protein signal transduction pathways functioning in C. neoformans. Crg1 and Gpb1 are two key regulators of the pheromone-responsive mating pathway, whereas Gpa1 and Gib2 in conjunction govern the cAMP signaling pathway to regulate the production of two key virulence factors melanin and capsule formation. Gpa2 is the cognate G␣ for Gpb1 (L. Li and P. Wang, unpublished observation), whereas Gpr4 is a G protein-coupled receptor homolog recently implicated in Gpa1-cAMP signaling (61).
Novel Function of C. neoformans Gib2 OCTOBER 27, 2006 • VOLUME 281 • NUMBER 43 JOURNAL OF BIOLOGICAL CHEMISTRY 32603 pGAD424 and pGBT9, but were possible with plasmids pGADT7 and pGBKT7 that allow a high level of protein expression.
Our studies have indicated that Gpg1 and Gpg2 could bind to both Gib2 and Gpb1, which is not surprising because a single G␥ could interact with various G␤, and different G␤␥ combinations often confer different signal strength and specificity (62). It is interesting to note that mutant strains of gpg1 and gpg2 each exhibited distinct dysfunction in mating and that gpg1 gpg2 mutants were nonviable. 5 On-going studies should provide further understanding for any signaling roles of Gpg1 and Gpg2 in mating and cAMP signaling.
Gib2 Is a Multifunctional Protein-The S. cerevisiae Asc1/ Cpc2 protein is a core 40 S ribosomal protein that represses gene expression (24,25). The S. pombe Cpc2/RACK1 homolog is required for efficient translation, cell cycle progression, and meiotic development of the cell (63,64). The high sequence similarity shared between Gib2 and Asc1/Cpc2/RACK1 could suggest that Gib2 has functions similar to these proteins. Indeed, cells grown, although poorly, under the condition for depletion of Gib2, appeared to have a higher protein content (data not shown), analogous to the asc1 null mutant strain (25).
Gib2 is shown to interact with Pkc1 and Smg1, and one of the positive two-hybrid clones also contains a SH3 domain that is known to interact with the RACK1 protein. In addition, Gib2 is a unique protein. Unlike RACK1 that binds to heterodimeric G␤␥ or heterotrimeric G␣␤␥ (21)(22)(23), Gib2 appears to bind to Gpa1 directly and it also positively regulates cAMP signaling.
Finally, Gib2 appears to encode an essential function whereas Asc1/Cpc2/RACK1 do not.
We took into consideration the possibility that yeast G␤␥ (Ste4/Ste18) might mediate the interaction between Gib2 and Gpa1 or Gib2 and Gpg1/Gpg2. In addition to protein pulldown assays indicating direct bindings, we also reasoned that if Ste4 mediated the binding, Gpa3 or Gpa2 would be identifiable from the yeast screen, not Gpa1, because Gpa3 shared the highest sequence homology with the yeast Gpa1 (40%) that form a heterotrimeric G protein with Ste4 and Ste18, and Gpa2 that is functionally analogous to yeast Gpa1 couples to Gpb1 in mating. 4 Finally, we failed to demonstrate any direct interactions between Ste4 and Gib2 using the yeast assay (data not shown).
Gib2 Is Involved in Fungal Virulence-Gib2 plays a positive role in cAMP signaling of C. neoformans, in contrast to S. cerevisiae Gpb1 and Gpb2 that play a negative role. Gpb1 and Gpb2 were suggested to have a target(s) either upstream of the cAMP-dependent protein kinase (PKA) signaling pathway or downstream of a pathway that requires PKA for function (49,65). The mammalian RACK1 has been shown to modulate the level of cAMP through its interaction with the phosphodiesterase Pde4D5 (20) via a N-terminal RACK1-interacting domain (RAID1), however, yeast phosphodiesterases, including C. neoformans var. grubii Pde1 and Pde2 homologs (66), do not appear to contain the unique N-terminal RAID1 region. Interestingly, a phosphodiesterase open reading frame (Cnb01170) with a putative RAID1 was identified from C. neoformans var.
neoformans whose study may reveal the mechanism by which overexpressing Gib2 increased the level of cAMP in BAC20-1, a var. neoformans gpa1 mutant strain.
To find a target(s) of Gib2, we have resorted to the interaction cloning strategy and have identified Smg1. This is rather intriguing because Smg1 was previously identified as a multicopy suppressor of gpa1 mutation for melanin deficiency and is implicated in melanin formation and virulence (56,57). Results of our gene disruption studies have contradicted this proposition, because smg1 mutant strains did not exhibit any altered levels of melanin pigmentation. It is feasible, however, to speculate that Smg1 may play a certain role in melanin biosynthesis and that Gib2 could recruit additional targets to modulate the Gpa1-cAMP signaling pathway that affects melanin production. Conversely, the interaction between Gib2 and Smg1 could be mediated through PKC signaling, independent of the Gpa1-cAMP pathway, as Pkc1 has been demonstrated to also play a regulatory role in melanin formation and virulence (50). Therefore, further examination of interactions between these proteins and their resulting biological significance would provide meaningful insights into functions of Gib2, either as an atypical G␤ or a RACK1 protein, in virulence.
Taken together, our studies have identified the novel G␤-like/RACK1 protein homolog Gib2 that functions as an atypical G␤ in cAMP signaling, in conjunction with Gpa1. Additionally, Gib2 is a multifunctional protein that encodes essential functions for survival. Our studies suggest that fungal G␤-like/RACK1 protein homologs may have diverged from their mammalian counterparts in evolution by adapting a role as an atypical G␤ in G protein signaling, whereas maintaining a RACK1-like function. Given the remarkable conservation in signaling mechanisms among fungal and other eukaryotic organisms, similar proteins may also exist in those systems that constitute unique signaling networks controlling cellular growth, differentiation, and/or virulence.