Global Survey of Canonical Aspergillus flavus G Protein-Coupled Receptors

A. flavus encodes 15 putative GPCRs.We began our search for A. flavus GPCRs by examining the set of 16 GPCRs reported for A. nidulans (7). In an effort to focus on the canonical GPCRs, our study did not include the PTH11-like membrane-spanning proteins, of which A. nidulans is predicted to encode at least 25 (7, 8). A BLAST search of the amino acid sequences of each of the 16 A. nidulans GPCRs against the A. flavus genome yielded 77 total hits, 57 of which were unique. GPCRs pass the membrane seven times, so two protein topology prediction tools, TMHMM 2.0 (13, 14) and TopPred 1.10 (15, 16), were used to determine which of these 57 proteins were likely GPCRs. Twenty-nine of the fifty-seven sequences were not predicted to encode any transmembrane (TM) domains, so they were eliminated. Of the remaining 28 proteins, 13 were predicted to have 7 TM domains by at least one of the prediction tools. Two additional proteins that were homologous to A. nidulans GprF and GprO were predicted to have fewer than 7 TM domains, but since their A. nidulans orthologs were predicted to have 7 TM domains, they were also included with the rest of the A. flavus GPCRs, bringing the total to 15 (Table 1).

We also compared the A. flavus genome to the reported A. oryzae GPCRs (7) since these two species are believed to be the same species, with A. oryzae constituting a domesticated clade (17). Similar to what was reported for A. oryzae, no orthologs for A. nidulans GprE, GprI, or GprN were found in the A. flavus genome. A novel A. oryzae GPCR not found in A. nidulans was identified and named GprQ. Its A. flavus ortholog (AFLA_132040; 99% identity, with an E value of 7e−115) had only four (TMHMM) or three (TopPred) predicted TM domains. Since A. oryzae GprQ was reported to encode only five predicted TM domains, the A. flavus homolog was not included in the set of GPCRs.

The BLAST search also identified two GPCRs that were not previously identified in the aspergilli, GprR and GprS. Like GprK, GprR harbors an RGS (regulator of G protein signaling) domain at the C-terminal end. This type of GPCR was first discovered in Arabidopsis thaliana and called A. thaliana RGS1 (AtRGS1). AtRGS1 regulates cell proliferation as well as sensitivity to sugars via its functional RGS domain. Unlike canonical GPCRs, it does not trigger the exchange of GTP for GDP on a G protein. Upon sensing glucose, AtRGS1 interacts with the constitutively active Gα subunit A. thaliana GPA1 (AtGPA1), resulting in hydrolysis of GTP and subsequent deactivation of the G protein (18–20). GPCR-RGS hybrids have since been found in several species of filamentous ascomycete fungi in addition to the aspergilli, including N. crassa, M. grisea, Fusarium graminearum, and several Trichoderma species (7, 10, 12). GprS contains a PQ loop repeat, grouping it with the class IV receptors GprF, GprG, and GprJ. The function of the PQ loop repeat domain is unknown, although this class contains the Schizosaccharomyces pombe nitrogen starvation sensor Stm1 (7).

Based on their domain structures, A. flavus’ 15 putative GPCRs were assigned to nine different classes. These classes have been previously described (7, 21), and a few notes on each class are included in Table 1. The amino acid sequences for all 15 GPCRs, as well as the sequences for GPCRs from A. nidulans, A. fumigatus, A. oryzae, F. graminearum, M. grisea, N. crassa, Trichoderma reesei, and V. dahliae, were aligned using the Clustal Omega software tool (22), and a phylogenetic tree was generated using the Phylogeny.fr software tool (23) (see Fig. S1 in the supplemental material). From this analysis, it was apparent that the A. flavus GPCRs GprF, GprH, GprS, and GprR are not orthologous to any of A. oryzae’s predicted GPCRs, despite the high degree of genetic similarity between the two fungi. Interestingly, of the two GPCRs that were not previously found in the aspergilli, GprS was most closely related to two GPCRs from V. dahliae, while GprR was unique among the fungi analyzed here.

In order to verify the coding sequences of the A. flavus GPCRs, 13 of the 15 GPCRs were amplified, cloned, and sequenced from A. flavus mRNA. cDNA for gprJ and gprM could not be amplified. Two GPCRs, gprF and gprO, appeared to be misannotated. In both cases, the actual sequence improved the proteins’ alignments with their A. nidulans orthologs, and for GprO, it changed the number of predicted TM domains from six to seven (TopPred), with the updated prediction reported in Table 1. GprF was only predicted to encode four (TMHMM) or five (TopPred) TM domains, but since A. nidulans GprF was predicted to encode seven TM domains, it was included with the set of A. flavus GPCRs.

Individual GPCRs deleted in A. flavus.All 15 of the GPCRs were individually disrupted in the CA14 ΔpyrG Δku70 strain of A. flavus by replacing the GPCR-encoding gene with pyrG. Deletion mutants were confirmed by PCR (data not shown) and Southern blotting (see Fig. S2 in the supplemental material), and confirmed isolates were checked for marker gene effects by growing on media with and without uridine and uracil (UU) (see >Fig. S3 to S5; also data not shown) (24). The ΔgprG, ΔgprK, ΔgprM, and ΔgprR mutants all exhibited growth defects that were at least partially remediated by supplementation of UU.

To investigate the impact of UU supplementation on a phenotypic assay, a subset of mutants was examined in greater detail. First, gprR was disrupted in strain NRRL3357 of A. flavus. Like the CA14 ΔgprR mutant, the NRRL3357 ΔgprR mutant was severely restricted in growth in the absence of UU, demonstrating that the need for pyrimidine supplementation was conserved in different genetic backgrounds (see Fig. S3 in the supplemental material). Strains were then exposed to two different stressors: alkaline pH (pH 8) and cell wall stress (via Congo red), with and without UU. The wild types for both the CA14 and NRRL3357 backgrounds were inhibited to similar degrees regardless of UU supplementation, whereas the addition of UU was required to observe stress-induced inhibition for both ΔgprR mutants (see Fig. S3). Multiple independent isolates of a mutant in the same genetic background were similarly examined. The ΔgprM mutant had shown a marker gene effect for all three confirmed transformants, and without UU, the effects of the stresses were more dramatic (see Fig. S4). As a final test of the effects of UU supplementation, multiple isolates of two mutants that had not shown marker gene effects, the ΔgprD and ΔgprH mutants, were analyzed. Independent isolates of both mutations showed similar patterns of inhibition regardless of UU supplementation (see Fig. S5). The impacts of alkaline pH and cell wall stress were quantified for all of the isolates of the ΔgprR, ΔgprM, ΔgprD, and ΔgprH mutants, revealing that multiple isolates of the same mutant behaved similarly (see Table S1). Therefore, a single isolate for each GPCR disruption was chosen, and UU was included in the media for all subsequent experiments.

GPCRs important in basic A. flavus development and aflatoxin synthesis.Various fungal developmental processes have previously been linked to G protein signaling. To elucidate which GPCRs may be involved in mediating these signals, we measured the germination rate, AF biosynthesis, sporulation under both light and dark conditions, and production of spores and sclerotia at high and low cell densities, described below.

(i) Germination.Germination studies in A. nidulans have revealed several components of G protein signaling to be important in that process. The heterotrimeric G protein GanB-SfaD-GpgA activates a cAMP pathway in response to glucose to drive germination, and this is abrogated by RGS RgsA (25–27). There is also evidence of a small GTPase, RasA, controlling germination independently of cAMP (25, 28). The only GPCRs implicated in germination thus far are A. nidulans GprD (21) and A. fumigatus GprC and GprD (29). Therefore, germination of the entire set of A. flavus Δgpr mutants was monitored from 4 to 9 h postinoculation. During this time period, the wild type went from 1% to 77% germinated spores (data not shown). The germination rates of the Δgpr mutants were measured, and a heat map was generated with the percentage of mutant germination versus wild-type germination for each time point (Table 2). Aside from a slightly enhanced germination rate at 6 h by the ΔgprO mutant, the rest of the mutants were either equivalent or impaired in germination compared to the wild type. The most severely deficient mutants were the ΔgprA and ΔgprK strains, with <50% of the wild-type germination rate at nearly all time points measured. Many of the other mutants, including the ΔgprB, ΔgprC, ΔgprH, ΔgprJ, ΔgprP, ΔgprR, and ΔgprS strains, were impaired in germination as well.

(ii) Light-induced sporulation.Rhodopsin is well characterized as a receptor of light, and A. flavus encodes one GPCR, NopA, that belongs to the class IX bacteriorhodopsin-like GPCR family. Since light is an inducer of sporulation, the fungi were exposed to continuous light or continuous dark, and their spores were counted. The wild type and most of the mutants produced significantly more spores in the light versus production in the dark. Two mutants, the ΔgprF and ΔgprR mutants, showed no significant difference between the two conditions. Notably, the ΔnopA mutant exhibited the wild-type response to light (Table 2).

(iii) Quorum sensing.A. flavus undergoes a quorum-sensing-mediated developmental shift in which at low inoculum density, it produces many sclerotia and very few spores. The opposite is observed at high inoculum density (30). Because an exogenous signal mediates this phenomenon, GPCRs are plausible targets for relaying the signal. In fact, two GPCRs in the A. flavus strain NRRL3357, encoded by gprC and gprD, have already been implicated as quorum-sensing receptors based on their inability to respond to a high-density extract (31). Surprisingly, all 15 Δgpr mutants produced more sclerotia and fewer spores at high density than the wild type (data not shown). However, the trend of decreasing sclerotia and increasing spore production as inoculum density increased was still observed for most of the mutants, with the exceptions of the ΔgprB and ΔgprF mutants There was no significant difference in the amount of spores produced at high and low densities for the ΔgprB mutant, nor was there a significant difference in sclerotia production at the two densities for the ΔgprF mutant (Table 2).

(iv) Aflatoxin.G protein signaling has been shown to regulate synthesis of AF and its precursor, sterigmatocystin (ST). The Gα subunit FadA and its RGS FlbA are part of an adenylate cyclase/cAMP/protein kinase A (PKA) (PkaA) pathway that controls ST production in A. nidulans via transcriptional and posttranscriptional regulation of aflR (32–34). The orthologous pathway regulates AF biosynthesis in Aspergillus parasiticus (35) and A. flavus (J. W. Bok and N. P. Keller, unpublished). Nothing is known about which GPCR(s) might be initiating this pathway, so the panel of Δgpr mutants was tested for their ability to make AF. AF was extracted and measured following growth on AF-inducing and AF-repressing media (YES [2% yeast extract and 6% sucrose, pH 5.8] and YEP [2% yeast extract and 6% peptone, pH 5.8], respectively). No mutant gained the ability to produce detectable AF on YEP. However, on YES medium, the ΔgprA mutant made ~1.5× more AF than the wild type, and the ΔgprP mutant made ~2.3× more AF than the wild type (Table 2). None of the other mutants were affected in their ability to produce AF.

GPCRs impact growth on various carbon and nitrogen sources.In addition to developmental cues, GPCRs are able to detect the presence or absence of important nutrients. GprC and GprD belong to class III, which is the group that also contains the sugar receptors S. cerevisiae Gpr1 and S. pombe Git3. Gpr1 senses glucose and sucrose (with a stronger affinity for sucrose) and transmits the signal through a G protein/cAMP/PKA pathway (36–38). Other Gpr1 homologs are also carbon source sensors, such as S. pombe Git3 (reviewed in reference 39) and N. crassa GPR-4 (40), although the role of C. albicans Gpr1 in carbon sensing is ambiguous (41, 42). While a carbon-sensing GPCR has not been identified in the aspergilli, it has been shown that carbon sensing is mediated by the heterotrimeric G protein GanB-SfaD-GpgA in A. nidulans (27). To identify A. flavus GPCRs involved in carbon sensing, strains were grown on a variety of carbon and nitrogen sources. These included glucose, galactose, xylose, sucrose, and corn oil for carbon sources. Corn oil is mainly composed of C18 and C16 fatty acids (43). The ΔgprA, ΔgprC, ΔgprJ, ΔgprK, and ΔgprR mutants were impaired in growth on several of the carbon sources tested (Table 3). The ΔgprR mutant was restricted in growth on every carbon source tested except galactose and corn oil.

Class IV receptors are grouped together based on their relatedness to the S. pombe nitrogen starvation sensor Stm1, and A. flavus has four receptors in this group. Nitrogen starvation results in cell cycle arrest in S. pombe, which allows the yeast to remain viable either until nutrition becomes available or mating partners are present. stm1 is expressed at low, moderate, and high levels on peptone, ammonium, and proline, respectively. Deletion of stm1 causes premature cell cycle arrest under nitrogen starvation, while overexpression of stm1 represses vegetative growth under nitrogen-replete conditions (44). Therefore, the mutants were tested for their ability to grow on peptone, ammonium chloride, and proline. While no mutant was restricted on peptone, mutants carrying both ΔgprC and ΔgprD deletions—the class III receptors—were inhibited on proline. The ΔgprR mutant was impaired on ammonium chloride and proline (Table 3).

GPCRs contribute to fungal stress responses.Fungi transmit the myriad stress signals they encounter through several conserved, interconnected signal transduction pathways. Multiple stimuli trigger each pathway, and the pathways themselves engage in cross talk with one another. For example, the cell wall integrity (CWI) MAPK pathway is triggered in response not only to cell wall stress but to osmoregulatory and oxidative stresses as well, among others (reviewed in reference 45). In addition to the CWI pathway, osmotic stress is also managed by another MAPK pathway, the high-osmolarity glycerol (HOG) pathway (reviewed in reference 46). pH tolerance is maintained by the Pal/Rim alkaline response pathway. It transmits its signal via PalH, a 7-TM-domain protein that interacts with an arrestin protein to initiate endocytosis, ultimately activating the PacC transcription factor (reviewed in reference 47). Besides their overlapping roles in stress tolerance, what these pathways have in common is the prevalence of G protein signaling components. To examine whether any of A. flavus’ GPCRs mediate stress responses, the mutants were exposed to a variety of stressors that inhibit the radial growth of the wild type. The percent inhibition of growth in the presence versus absence of the stressor was compared between the wild type and each mutant, and the changes in inhibition between wild type and the mutants are presented as a heat map (Table 4).

Hydrogen peroxide was used at a concentration of 3 mM to assay oxidative stress. Loss of gprH rendered the fungus more resistant to hydrogen peroxide, while the ΔgprG mutant was more sensitive. Cell wall stress was created by growing the fungi on Congo red, which weakens the cell wall by binding to growing chitin chains, preventing their attachment to other cell wall components (48, 49). The ΔgprK, and ΔgprS mutants were both more sensitive to Congo red. The impact of osmotic pressure was measured with 1 M sodium chloride. Three mutants, carrying the ΔgprK, ΔgprM, and ΔgprR deletions, were more sensitive than the wild type to the hyperosmotic condition. Finally, the strains were grown on media buffered to pH 4 and pH 8. The ΔgprM mutant was more sensitive to both conditions, while the ΔgprK and ΔgprR mutants were more sensitive to acidic and alkaline pHs, respectively. Alternatively, the ΔgprD, ΔgprF, and ΔgprG mutants were more resistant to acidic pH.

GPCRs are involved in response to lipids and oxylipins.It has long been speculated that oxylipins are detected by fungal GPCRs, since this is their mode of perception in mammalian cells (reviewed in references 3, 5, and 50). To address this, strains were exposed to the oxylipins methyl jasmonate (MeJA) and 13-hydroperoxy-octadecadienoic acid [13(S)-HpODE] and to the fatty acid linoleic acid. Any mutant that deviated from the wild-type response to these chemicals is indicated in Table 5 with a shaded box.

The effect of MeJA on AF production varies depending on the experimental setup and concentration of MeJA used (51). Under certain conditions, MeJA has been shown to repress AF production by A. flavus (52), but under others, it has enhanced AF synthesis by A. parasiticus (53). For this study, an AF-repressive protocol was followed (51) in which cultures were exposed to 10⁻⁴ M MeJA, and AF was measured after 3 days. AF repression was observed for the wild type and five of the Δgpr mutants: the ΔgprC, ΔgprD, ΔgprH, ΔgprJ, and ΔnopA mutants. The rest of the mutants showed no change in AF production except for the ΔgprK mutant, which actually made more AF when exposed to MeJA (Table 5). A lack of response to MeJA cannot be attributed to a basic biosynthetic defect because, as previously shown, none of the mutants were impaired in AF biosynthesis on YES media (Table 2).

13(S)-HpODE, as well as its precursor fatty acid, linoleic acid, has been shown to induce sporulation in A. flavus (54). Stimulation of conidiation by 13(S)-HpODE was lost for the ΔgprA, ΔgprC, ΔgprD, ΔgprF, ΔgprG, ΔgprJ, ΔgprO, ΔgprP, and ΔgprR mutants. To test whether this represented a general sporulation defect or a specific lack of response to 13(S)-HpODE, sporulation was also monitored in response to linoleic acid. All of the mutants except the ΔgprR mutant responded as did the wild type to linoleic acid, indicating that the lack of response to 13(S)-HpODE was specific for all of the above mutants except the ΔgprR mutant (Table 5).

GPCRs are differentially expressed under disparate conditions.To further understand potential roles of A. flavus GPCRs, their expression levels under different conditions from various published studies were investigated. The results are compiled in Table 6, which revealed several GPCRs to be differentially expressed on maize versus liquid culture mycelia, and these GPCRs do not overlap with those that are differentially expressed on maize versus wheat (55). The corn kernel germ is mainly composed of lipids, while the endosperm is predominantly starch, and A. flavus preferentially grows in the lipid-rich portions of the kernel (56). Interestingly, four GPCRs, encoded by gprC, gprH, gprM, and gprS, were differentially expressed on the germ versus the endosperm (55). Two separate studies compared genomic expression from mycelial growth versus sclerotial production, though the methods used were quite different between the two studies (55, 57). The strains of A. flavus, media, and growth conditions (i.e., liquid versus solid culture) that were used varied, as did the methods for measuring gene expression (i.e., microarray versus transcriptome sequencing [RNA-seq]). Both studies showed that many of A. flavus’ GPCRs were differentially expressed depending on whether the fungus was growing as mycelia or producing sclerotia, although the expression patterns differed greatly between the two studies. 5-Azacytidine (5-AC) is an inhibitor of AF production and sporulation (58), but application of 5-AC to cultures of A. flavus did not impact expression of any of its GPCRs (59).