OUP user menu

Characterization of α-factor pheromone and pheromone receptor genes of Ashbya gossypii

Jürgen Wendland, Alexander Dünkler, Andrea Walther
DOI: http://dx.doi.org/10.1111/j.1567-1364.2011.00732.x 418-429 First published online: 1 August 2011


The genome of Ashbya gossypii contains homologs of most of the genes that are part of the Saccharomyces cerevisiae pheromone-signal transduction cascade. However, we currently lack understanding of a potential sexual cycle for this pre-whole genome duplication hemiascomycete. The sequenced strain bears three identical copies encoding MATa. We show that the syntenic A. gossypii homolog of MFα1 (AFL062w) does not encode a mature α-factor peptide, but identified another gene, AAR163c, which encodes a candidate α-specific mating pheromone and is thus reannotated as AgMFα2. The expression of the AgSTE2α-factor receptor in an Scste2 S. cerevisiae MATa strain resulted in dosage-dependent growth arrest upon exposure to A. gossypiiα-factor, which indicated that the pheromone response was effectively coupled to the S. cerevisiae signal transduction cascade. Comparison of α-pheromones and α-pheromone receptors showed greater conservation between Eremothecium cymbalariae and S. cerevisiae than between A. gossypii and E. cymbalariae. We constructed A. gossypii strains deleted for the STE2 and STE3 pheromone receptors. These strains showed no phenotypic abnormalities and an ste2, ste3 double mutant is still able to sporulate. The deletion of STE12 as the downstream target of pheromone signalling, however, led to a hypersporulation phenotype.

  • mating pheromone
  • pheromone receptor
  • growth arrest
  • signal transduction


Ashbya gossypii, a known overproducer of riboflavin/vitamin B2, is a filamentous ascomycete belonging to the family of Saccharomycetaceae, thus being closely related to yeasts, for example Saccharomyces cerevisiae (Kurtzman & Robnett, 2003). The life cycle of A. gossypii starts with the germination of a spore, generation of germ tubes, hyphal filaments and lateral branches resulting in mycelium formation. Under nutrient limitation conditions, septate hyphal compartments develop into sporangia that produce haploid uninucleate endospores. The genome of A. gossypii has been sequenced (Dietrich, 2004). The sequenced strain harbors three mating-type loci albeit on three different chromosomes, all resembling MATa. However, in contrast to S. cerevisiae, the A. gossypii mating-type loci contain two divergently transcribed genes, termed MATa1 and MATa2 (Wendland & Walther, 2005). Currently, we do not know of a sexual cycle in A. gossypii. The lack of a described MATα strain leads to the inability to test for the occurrence of classical genetics and sexual crosses with A. gossypii. The molecular characterization of genes solely relies on the highly efficient homologous recombination machinery that allowed the introduction of PCR-based gene targeting methods (Steiner, 1995; Wendland, 2000). The spores that are produced from the A. gossypii MATa mycelium could thus either be derived from mitotic or meiotic processes. Any potential ascosporogenesis in A. gossypii may, therefore, resemble a process known as homokaryotic or haploid fruiting (Esser & Meinhardt, 1977, Wickes, 1996).

In S. cerevisiae, mating of haploid a and α cells starts with the perception of secreted mating pheromones. MATa cells produce and secrete the farnesylated a-factor, while MATα cells produce and secrete α-factor. MATa cells express the seven transmembrane domain-spanning α-factor receptor, Ste2, while MATα cells express the a-factor receptor Ste3. This ensures cell type specificity and the recognition of compatible mates. Specific binding of the pheromone to its receptor triggers a common downstream signalling cascade through a heterotrimeric G-protein, composed of Gpa1, Ste4 and Ste18, and a MAP-kinase cascade consisting of Ste11, Ste7, the MAP-kinases Fus3 and Kss1 as well as the scaffold protein Ste5. This leads to the activation of the Ste12 transcription factor in a and α cells. Ste12 binds to a specific DNA motif in the promoters of regulated genes, the so-called pheromone-response element. Key events regulated by pheromone signalling are a cell-cycle arrest in G1 to prepare cells for mating, shmoo formation that leads to cell–cell contact of the mating partners, cell fusion and karyogamy (Bardwell, 2005).

The pheromone response pathway is highly conserved and components of this pathway have been analyzed in a number of fungi (Lengeler, 2000; Magee, 2002; Mayrhofer, 2006; Muller, 2008; Rispail, 2009). Because the sequenced type strain of A. gossypii is of MATa genotype, we decided to initiate the characterization of the A. gossypii mating pathway by studying the α-factor pheromone–receptor interaction. In this study, we describe the analysis of a novel gene encoding the A. gossypiiα-factor and its coupling to the AgSte2 α-factor receptor and the S. cerevisiae pheromone signal transduction cascade. Mutants of the STE2, STE3 and STE12 genes were generated in A. gossypii. However, these strains were unaltered in their growth characteristics and retained the ability to sporulate. Indeed, the deletion of STE12 resulted in a pronounced hypersporulation phenotype. Thus, pheromone signalling in A. gossypii may play a regulatory role in sporulation.

Materials and methods

Strains and media

The strains used in this study are listed in Table 1. Yeast cells were grown in YPD (1% yeast extract, 2% peptone, 2% glucose) or a complete supplement medium (CSM) at 30 °C. Ashbya gossypii was grown in AFM (1% yeast extract, 2% caseine peptone, 2% glucose) or CSM at 30 °C. Antibiotics (G418 or clonat) were added using a final concentration of 200 μg mL−1 for the selection of transformants. Plates were incubated 3–7 days at 30 °C before photography.

View this table:

Strains used in this study

Ashbya gossypii
ATCC10895Ashbya gossypii wildtypeLab collection
Agleu2leu2Lab collection
AWA26aleu2, ste3kanMX/STE3This study
AWA27aleu2, ste12kanMX/STE12This study
AWA28leu2, ste2kanMX/STE2This study
AWA31aleu2, ste3kanMXThis study
AWA32aleu2, ste12kanMXThis study
AWA33leu2, ste2kanMXThis study
AWE67leu2, ste3kanMX, ste2NAT1/STE2This study
AWE73leu2, ste3kanMX, ste2NAT1This study
Eremothecium cymbalariae DBVPG 7215DBVPG
Saccharomyces cerevisiae
BY4741his3Δ1; leu2Δ0; met15Δ0; ura3Δ0,Euroscarf
BY4741-ste2his3Δ1; leu2Δ0; met15Δ0; ura3Δ0, ste2kanMXEuroscarf
WYE106his3Δ1; leu2Δ0; met15Δ0; ura3Δ0, ste2NAT1This study


Ashbya gossypii was transformed by electroporation as described (Wendland, 2000). The transformation of A. gossypii generates primary transformants that are heterokaryotic in that they carry both wild-type and mutant nuclei in one mycelium. Via clonal selection of uninucleate and haploid spores, homokaryotic mycelia can be generated that solely carry mutant nuclei. These homokaryotic mutants were used for phenotypic characterization. Saccharomyces cerevisiae was transformed using the lithium acetate procedure (Gietz & Schiestl, 2007). The marker exchange in the S. cerevisiae ste2kanMX4 strain was carried out by transforming a restriction fragment from pFA-NAT1 (#486) that contains the nourseothricin resistance gene ORF under control of the A. gossypii TEF promoter and terminator into the Scste2 strain. This made use of the A. gossypii TEF promoter and terminator sequences as homology regions for the marker replacement. Transformants were selected on clonat.

Pheromone arrest assay

Synthetic A. gossypiiα-factor was obtained from GenScript (Piscataway, NJ). The peptides were dissolved in 10% dimethyl sulfoxide at a concentration of 10 mg mL−1. Appropriate dilutions from the stock solution were carried out in H2O. A lawn of yeast cells from an overnight culture (approximately 105 cells per plate) was spread and, when dry, 5 μL of α-pheromone solution was spotted on the plate. Pictures were taken after 2 days of growth at 30 °C.

Generation of plasmids

The A. gossypii STE2-ORF with its terminator was amplified as a 1553-bp fragment from genomic DNA using primers #3640 and #3641 and cloned into pBluescriptSK+ using XhoI and XbaI (C304, see Table 2 for a list of plasmids and Table 3 for a list of the primers used in this study). This fragment was then subcloned as an XhoI/SacI-fragment into pRS169. This generated plasmid C311, in which AgSTE2 is under control of the S. cerevisiae TEF1 promoter.

View this table:

Plasmids used in this study

Plasmid numberFeatureSource
121pFA-kanMX6Lab collection
486pFA-NAT1Lab collection
C169pRS417-GEN3-ScTEF1promThis study
C304PSK+AgSTE2-ORF-termThis study
C311pRS417-ScTEF1prom-AgSTE2This study
C343PSK+Agste2∷kanMX6This study
C344PSK+Agste2∷NAT1This study
C345PSK+AgSTE3This study
C351PSK+Agste12∷kanMXThis study
C353PSK+Agste3∷kanMXThis study
View this table:

Oligonucleotide primers used in this study

Primer numberSequence of primer
  • * In longer primers, upper case sequences correspond to genomic DNA sequences used as homology regions for amplification, whereas lower case sequences correspond to 5′-terminal additions of restriction sites (in bold). All sequences are written from 5′ to 3′.

An STE2-disruption cassette was constructed based on C304. This plasmid was cleaved with EcoRI and EcoRV, which removes 929 bp from the 1377-bp STE2-ORF. Plasmid 121-pFA-kanMX6 was used to provide the kanMX selectable marker as an EcoRI/SmaI fragment, which was ligated into C304/EcoRI–EcoRV to generate pSK+Agste2∷kanMX6 (C343). Similarly, plasmid 486-pFA-NAT1, which contains the nourseothricin resistance gene ORF under control of the A. gossypii TEF-promoter and terminator, was used to provide the NAT1 selectable marker as an EcoRI/PvuII-fragment, which was also ligated into C304 cleaved with EcoRI and EcoRV to generate pSK+Agste2∷NAT1 (C344). The disruption cassettes from C343 and C344 were released from the vector backbone by cleaving with XhoI/XbaI before transformation. Flanking homology regions were thus 258 bp at the 5′-end and 368 bp at the 3′-end.

Cloning of an STE3-disruption cassette: A 1637-bp PCR-product containing AgSTE3 was generated from genomic A. gossypii DNA using primers #3706 and #3707, cleaved with KpnI/NsiI and cloned as a 1430-bp fragment into pBluescriptSK+ cleaved with KpnI/PstI, which generated pSK+AgSTE3 (C345). C345 was cleaved with EcoRI to delete a 403-bp fragment of the STE3-ORF. Into this position the kanMX selectable marker was ligated, generating pSK+Agste3,∷kanMX (C353). The release of the disruption cassette was achieved by KpnI/SacI-digest before its use in transformation. Flanking homology regions were thus 614 bp at the 5′-end and 418 bp at the 3′-end.

Cloning of an STE12-disruption cassette was performed by a fusion PCR reaction (Noble & Johnson, 2005). To this end, flanking homology regions were amplified using the primer pairs #3727/#3728 and #3729/#3730. The 370-bp 5′-flank encompasses the 5′end of the STE12-ORF, while the 419-bp 3′-flank is in the STE12-terminator. The kanMX cassette was amplified from plasmid #121 using primers #3725 and #3726. Primers #3728 and #3729 carry extensions that allow annealing to the kanMX PCR product. Thus, in a fusion-PCR, both flanks can be added to the selectable marker. The fusion-PCR product was precipitated and resuspended in cleavage buffer. This PCR product was then cleaved with EcoRI and XhoI using the terminal restriction sites provided by primers #3727 and #3729 and cloned into pBluescriptSK+generating pSK+Agste12∷kanMX (C351). The disruption cassette was released by digesting again with EcoRI and XhoI before transformation.

Reverse transcriptase PCR (RT-PCR)

Total A. gossypii RNA was isolated using the RNAgents kit (Promega, Madison, WI) followed by cDNA synthesis using ThermoScript (Invitrogen) and PCR. In brief, 2 mL of mycelium of an A. gossypii culture grown in CSM was harvested by filtration and resuspended in 500 μL denaturing buffer. This mixture was vortexed three times with glass beads, placed in liquid nitrogen and vortexed again. Then 50 μL 2 M sodium acetate and 500 μL phenol : chloroform : isoamyl alcohol were added, the solution was vortexed again and then incubated on ice for 15 min. After centrifugation (20 min at 8000 g at 4 °C), the water phase was used and the RNA precipitated by isopropanol. RNA was converted to cDNA by gene-specific primers and then amplified by PCR.

Eremothecium cymbalariae

The E. cymbalariae strain DBVPG 7215 was obtained from the Industrial Yeast Collection, Perugia, Italy (http://www.agr.unipg.it/dbvpg/). The genome of this strain has been sequenced recently, is currently being annotated and will be presented elsewhere. Eremothecium cymbalariae sequences used in this study have been deposited with GenBank under the accession numbers HQ266574HQ266578.


All microscopy was performed on an AxioImager microscope (Zeiss, Jena and Göttingen, Germany) using metamorph software tools (Molecular Devices Corp., Downington, PA) and a MicroMax1024 CCD camera (Princeton Instruments, Trenton, NJ).


Components of the pheromone response pathway in A. gossypii

The annotated A. gossypii genome contains homologs of the central components of the pheromone response pathway of S. cerevisiae (Fig. 1 and Table 4). The function of this pathway has not been elucidated in A. gossypii so far. The genome sequence indicates that the sequenced strain contains three mating-type loci on chromosomes 4, 5 and 6, respectively. Two of the mating-type loci are telomeric, and the mating-type locus on chromosome 6 shows a conserved syntenic gene order of adjacent genes to other ascomycetous mating-type loci (Dietrich, 2004). All three mating-type loci harbor two divergently transcribed transcriptional regulators with similarity to Kluyveromyces lactisa1 and a2, respectively (Wendland & Walther, 2005). Therefore, this A. gossypii strain is of MATa genotype. To perform an analysis of the pheromone response pathway in A. gossypii we thus wanted to elucidate a potential role of an A. gossypiiα-factor.


Schematic cartoon of central elements of the Saccharomyces cerevisiae mating pheromone response pathway (see text for details).

View this table:

Key components of the mating pheromone response pathway

S. cerevisiae gene nameS. cerevisiae protein length (aa)A. gossypii homologA. gossypii protein length (aa)blastpE-value
MATa1124AER456w AFR643C ADL394c1026e-10
a-factor MFA238ABL196c332e-05
α-factor MFα1165AFL062w1151e-04
α-factor MFα2120AAR163c964e-07
  • * As determined by blastp analysis of the yeast protein sequence against the Ashbya proteins.

Identification of a- and α-factor genes in Eremothecium

The A. gossypii genome sequence annotates ABL196c and AFL062w– denoted herein as AgMFa1 and AgMFα1 – as a- and α-factor encoding genes, respectively, based on synteny. Comparison of the translated A. gossypiia-factor ORF with MFa1 and MFa2 of S. cerevisiae and with MFa1 of more closely related filamentous fungus E. cymbalariae indicates the presence of conserved processing sites. Processing at the conserved CAAX site at the C-terminus should result in truncated and farnesylated mature A. gossypiia-factor peptides (Fig. 2a). The A. gossypii and E. cymbalariaea-factor peptides are more similar to each other than to the S. cerevisiaea-factor. Using multiple sequence alignment tools, AgMFα1 shows some degree of conservation at the amino terminal end of the prepropeptide. However, inspection of the preproprotein of AgMFα1 does not provide evidence for a conserved α-factor peptide with similarity to the S. cerevisiaeα-factor. The corresponding region has been resequenced and verified (not shown). In S. cerevisiae, MFα1 encodes a pheromone precursor containing four mature α-factor peptides and the MFα2 precursor protein contains two mature peptides. We therefore searched both the A. gossypii genome sequence and the E. cymbalariae genome sequence, which is currently being established in our group, for pheromone encoding genes. In this way, we could identify the A. gossypii AAR163c gene – now annotated as AgMFα2 – and two E. cymbalariae genes, which encode candidate α-specific mating pheromones (Fig. 2b). Both of the E. cymbalariae genes are found in syntenic positions to the A. gossypii AFL062c/AgMFα1 and AAR163c loci/AgMFα2 genes (not shown).


Comparison of predicted pheromone genes of Ashbya gossypii and Eremothecium cymbalariae with Saccharomyces cerevisiaea- and α-factor. (a) The predicted amino acids sequences of the a-factor precursors are shown. Ashbya gossypii and E. cymbalariae possess only a single MFa gene. The CAAX-farnesylation consensus motif is shown. The predicted mature peptides are boxed. Identical amino acid residues in all four precursor peptides are shaded in black. (b) The predicted amino acid sequences of the α-factor precursors are shown. Conserved amino acid residues in the N-terminal part of the preproproteins are shaded in black. The predicted mature peptides are shaded in gray. Conserved basic dipeptide Kex2p-cleavage sites shown in bold. Recognition sites for Ste13p are underlined. AAR163C=AgMFα2; AFL062w=AgMFα1; YPL187w=ScMFα1; YGL089C=ScMFα2. (c) Comparison of predicted mature α-factor peptides. Identical amino acids are in bold and conserved amino acids occurring either in A. gossypii or E. cymbalariae and in S. cerevisiae are in gray.

Analysis of Eremotheciumα-factor peptides: unique properties of AgMFα2

EcMFα1, the AFL062w homolog, encodes four peptide repeats and EcMFα2, the AAR163c homolog, encodes three repeats of a mature α-factor pheromone. In contrast, A. gossypii AgMFα1 harbors none and AgMFα2 encodes a peptide that contains only one sequence with similarity to a mature pheromone. Thus, apparently, in A. gossypii, an evolutionary reduction in MFα genes has taken place. The S. cerevisiae pheromone repeats are preceded by the basic dipeptide KR (Kex2-cleavage site), followed by two to three repeats of an xA dipeptide (with x being either D, E or N) (Ste13-cleavage site; Julius, 1983). These sites are necessary for processing of the mature pheromone. This structure is conserved in the E. cymbalariaeα-factor encoding genes, only that there are two to four xA repeats, with x being either D, S, V or N. In contrast, the A. gossypii AgMFα2 gene encodes a pheromone precursor that does not show either of these dipeptide signals. Instead, directly preceding the candidate α-factor is a QK-dipeptide. By homology, we inferred the duodecapeptide ‘WFRLSLHHGQSM’ derived from AgMFα2 to represent the potential active A. gossypiiα-factor pheromone. This peptide shows six conserved residues with S. cerevisiaeα-factor peptides and only four with E. cymbalariae. In contrast, E. cymbalariaeα-factor peptides are highly similar to S. cerevisiaeα-factor (best match with eight out of 13 identical amino acids), substantially differing only in the central loop region of the pheromone (Fig. 2c). The AgMFα1 translated protein does show a KR-dipeptide near the C-terminus. This could represent a Kex2-cleavage site, which is, however, not followed by xA-dipeptides. This conceptually generates a matured peptide with the sequence ‘LMEYGHRRIFPAIDL’. However, such a peptide does not bear any similarity to known α-factor peptides. Thus, we find it very unlikely that AgMFα1 encodes an A. gossypiiα-factor.

Identification of a- and α-factor pheromone receptor genes in Eremothecium

Thus, we conclude that there is a striking dissimilarity of α-factor peptides found in the two Eremothecium species and a much higher resemblance of α-factor peptides between E. cymbalariae and S. cerevisiae. To explore an underlying structure/function relationship at the pheromone receptor level, we identified the STE2 and STE3 pheromone receptor genes in E. cymbalariae and compared the translated proteins with the corresponding homologs of S. cerevisiae and A. gossypii (Fig. 3). Using protein alignment tools and pair-wise comparisons, we found 40.1% identity between the A. gossypiiS. cerevisiae Ste2 pair. The A. gossypii and E. cymbalariae showed a slightly higher level of amino acid identity, with 43.9%. However, the E. cymbalariae Ste2 pheromone receptor shared 51.2% amino acid sequence identity to the S. cerevisiae Ste2p (Fig. 3a). This demonstrates a greater similarity of the E. cymbalariae and S. cerevisiaeα-factor receptors and thus provides a structural aspect on the pheromone receptor level that correlates with the more closely related α-factor peptides recognized by these receptors. The greater level of similarity between EcSte2p and ScSte2p is particularly pronounced in the first extracellular loop of the receptors, which provides an exposed surface for pheromone–receptor interactions (Fig. 3b). In contrast, the Ste3p pairs of E. cymbalariae and A. gossypii are more similar to each other than to ScSte3, which may reflect the greater similarity of their a-factors compared with S. cerevisiae (Fig. 3c).


Alignment of pheromone receptors. (a) The percentage of amino acid sequence identity between pairs of receptors is shown for the Ste2 homologs (left) and Ste3 homologs (right). The Ste2 (b) and Ste3 (c) pheromone receptors of Ashbya gossypii, Eremothecium cymbalariae and Saccharomyces cerevisiae were aligned using the clustalw algorithm of the lasergene8 software package. Identical residues in at least two receptors are shaded in black. The putative seven transmembrane domains are indicated. Conserved residues in the first and second extracellular loops of EcSte2 and ScSte2 are marked with an asterisk.

Functional interaction of A. gossypiiα-factor with its α-factor receptor in S. cerevisiae

The predicted A. gossypiiα-factor encoding gene AgMFα2 is rather unique in that it only encodes one putative mature peptide. Given that AgMFα1 does not encode an α-factor peptide at all, it may be questionable with regard to the noncanonical cleavage sites whether the deduced mature α-factor peptide of AgMFα2 is produced and functional. To test A. gossypiiα-factor peptide efficacy in vivo, we obtained a chemically synthesized α-factor. In plate assays using A. gossypii spores challenged with high concentrations of α-factor, we were not able to discern a morphological response, for example growth arrest or directed growth towards the pheromone source (Fig. 4a). Therefore, we decided to establish a functional relationship of A. gossypiiα-factor with its cognate α-factor-receptor in S. cerevisiae.


Pheromone growth arrest assay in Ashbya gossypii and Saccharomyces cerevisiae. Ashbya gossypii spores (a) or a lawn of S. cerevisiae (WYE106; Scste2:NAT1, AgSTE2) (b) cells were plated and challenged with A. gossypiiα-factor at the indicated concentrations. The A. gossypiiα-factor was diluted in 10% dimethyl sulfoxide (DMSO) and DMSO controls did not produce any halo. Plates were incubated at 30°C for 2 days before photography.

To this end, we used an S. cerevisiae MATa strain carrying a deletion in ScSTE2 and expressed a plasmid-encoded AgSTE2 under control of the S. cerevisiae TEF1 promoter. Cells of this strain were plated and challenged with different concentrations of A. gossypiiα-factor in a growth arrest assay (Fig. 4b). As shown, the S. cerevisiae cells expressing AgSTE2 were responsive to the A. gossypiiα-factor at concentrations that are within the known range of responsiveness to the endogenous S. cerevisiaeα-factor. These results suggest that the α-factor peptide identified is indeed active and that heterologous expression of AgSTE2 results in the coupling of AgSte2 to the S. cerevisiae pheromone response pathway. The yeast strain used was no longer responsive to S. cerevisiaeα-factor and generally S. cerevisiae MATa cells did not respond to A. gossypiiα-factor. The pheromone response of S. cerevisiae not only included cell-cycle arrest as was obvious from the halo assay, but also mating projection formation and the ability to form zygotes (not shown). This identifies the designated peptide derived from AgMFα2 as A. gossypiiα-factor, which excludes the possibility that the potential peptide derived from AgMFα1 could interact with AgSte2.

Expression of A. gossypii genes of the pheromone response pathway

At present, we cannot exclude that expression of the AgMFα2 gene yields only a nonfunctional α-factor due to the potential lack of processing signals. Because the treatment of A. gossypii with the α-factor peptide did not result in strong phenotypic changes, we went on to analyze the expression of genes that are homologs of S. cerevisiaea-specific, α-specific or haploid-specific genes. To this end, we performed RT-PCR on A. gossypii total RNA (Fig. 5). As expected, genes encoding the heterotrimeric G-protein, GPA1, STE4 and STE18 were expressed. Surprisingly, we found that both α-factor and a-factor receptors encoded by STE2 and STE3, respectively, were expressed in the A. gossypii MATa strain. Furthermore, homologs of both S. cerevisiaeα-factor and a-factor genes and the genes required for processing of pheromone precursors were found to be expressed in A. gossypii. Noticeably, the expression of MFα1 was only barely detectable and that of MFα2 was rather low. Thus, our results suggest that A. gossypiia-factor and a-factor pheromone-receptor transcripts are generated within the same mycelium and genes of the pheromone response pathway may be expressed in a nonmating-type-specific manner.


Expression analysis of Ashbya gossypii homologs of Saccharomyces cerevisiaea, α and haploid-specific genes. Ashbya gossypii mycelium was grown in CSM, RNA was reverse transcribed and gene-specific primers corresponding to the 3′-end of the ORFs were used for amplification of the selected genes. The A. gossypii profilin homolog PRF1, which contains one intron, served as an RNA quality control. Genes are arranged according to their function in S. cerevisiae.

Deletion of A. gossypii STE2, STE3 and STE12 does not abolish sporulation

To further our understanding on the role of the pheromone signal transduction pathway, we decided to functionally characterize the pheromone receptor genes, STE2 and STE3, and of the downstream transcription factor STE12 in A. gossypii. To this end, deletion cassettes were generated and used for transformation (see Materials and methods). Primary, heterokaryotic transformants were verified by PCR and sporulated for the clonal selection of homokaryotic mutant strains. We initially obtained ste2, ste3 and ste12 mutants and went on to also construct an ste2/ste3 double mutant. For each mutant, at least two independent transformants were obtained and verified by diagnostic PCR. The deletion of either STE2 or STE3 did not result in an altered colony morphology. However, both mutants showed a slight increase in sporulation. Strikingly, the ste12 mutant showed a dramatic increase in sporulation – a phenotype that has not been observed previously in other A. gossypii mutants (Fig. 6). This increase in sporulation was readily visible on plates as highly enlarged zones of sporulation (i.e. the area of the mycelium that has produced spores). The increase in sporulation in an ste12 mutant was also very obvious when grown in a liquid culture, which resulted in almost quantitative conversion of hyphal segments into spore-generating sporangia. An ste2/ste3 double mutant was somewhat delayed in growth. However, the double mutant sporulated readily. Thus, we can conclude that disrupting the pheromone response pathway in A. gossypii by the deletion of both pheromone receptor genes does not abolish sporulation. On the contrary, the deletion of the transcription factor STE12 generates a hypersporulation phenotype.


Sporulation efficiency of Ashbya gossypii ste-mutants. The indicated homokaryotic mutant strains were grown on AFM for 7 days. Samples from different zones of the plates were analyzed microscopically for the presence of spores. Spores are needle shaped and form clusters by attaching via their terminal filaments. The ability to sporulate was classified as no sporulation (−), weak sporulation (+) or abundant sporulation (++).


The mating pheromone response pathway of S. cerevisiae is one of the best-known signal transduction pathways in eukaryotes. This pathway is required for the traditional mating response, but also for filamentation. The mating response is initiated by the recognition of a- and α-factor peptide pheromones secreted by cells of the opposite mating type. This requires binding of the pheromone and activation of signalling via seven-transmembrane spanning pheromone receptors encoded by STE2 and STE3. Pheromone receptors signal through a common downstream pathway that consists of a heterotrimeric G-protein encoded by GPA1 (Gα-subunit), STE4 (Gβ-subunit) and STE18 (Gγ-subunit), an Ste20p-like kinase, a MAP-kinase cascade with Ste11p, Ste7p and Fus3p scaffolded by Ste5p, finally activating the transcription factor Ste12. In S. cerevisiae, this cascade triggers growth arrest, shmoo formation, cell–cell agglutination and cell fusion leading to karyogamy and the formation of a sporulation-competent diploid cell (Herskowitz, 1995; Bardwell, 2005).

Components of this signal transduction cascade in S. cerevisiae, particularly the MAP-kinase cascade and Ste12, also participate in the regulation of morphogenesis. This includes pseudohyphal formation and unipolar budding, generating chains of cells that can thus forage for nutrients (Gimeno, 1992). In animal and plant pathogenic fungi, Ste12 homologs play an important role in pathogenicity (Rispail & Di Pietro, 2010; Wong Sak Hoi & Dumas, 2010). In these fungi, including for example Cryptococcus neoformans, Fusarium oxysporum and Aspergillus fumigatus, the Ste12 protein harbors C-terminal C2H2 zinc finger domains, which are absent in the S. cerevisiae and A. gossypii Ste12.

Here, we have started the functional dissection of the pheromone response pathway in A. gossypii. This shall help us understanding the role of this cascade in morphogenesis and development. Homologs of essentially all the components involved in the pheromone response in S. cerevisiae can be found in the A. gossypii genome. These genes were found to be expressed in a rather non mating-type-specific fashion. This resembles Candida glabrata in which the expression of STE2 and STE3 is also not mating type specific (Muller, 2008). In A. gossypii, development can be initiated from a single, uninucleate and haploid spore. Spores will germinate and form mycelia that generate spores without the need for a mating partner. Endospores will be generated in sporangia that develop from septate hyphal segments (Wendland & Walther, 2005). Sexual endospore formation is characteristic for fungi of the Saccharomycetaceae. Thus, A. gossypii may be regarded as homothallic. Yet, we are currently unaware of a sexual cycle or the mechanism used for spore formation in A. gossypii.

Ashbya gossypii seems to have undergone an evolutionary reduction in α-factor production. Of the two α-factor-encoding genes, one does not harbor a functional peptide while the second gene encodes only one mature peptide. This is in contrast to other ascomycetes, where α-factor-encoding genes encode preproproteins that are post-translationally processed into several mature α-factor peptides. Also, the close A. gossypii relative E. cymbalariae shows this conserved α-factor gene composition. Strikingly, α-factor pheromones of E. cymbalariae resemble more closely the α-factor peptides of S. cerevisiae rather than the A. gossypiiα-factor peptide. In particular, residues at the amino-terminus of S. cerevisiae and E. cymbalariaeα-factors (WHWL) are identical. These residues have been implicated in pheromone signalling in S. cerevisiae. The carboxyl terminal domains of E. cymbalariae and S. cerevisiaeα-factor are also very similar. This region has been shown to be important for binding of the pheromone to the receptor (Naider & Becker, 2004). The structural similarity of E. cymbalariae and S. cerevisiaeα-factor pheromones is also reflected in the higher level of amino acids sequence identity of their Ste2 pheromone receptors. Nevertheless, using the heterologous host S. cerevisiae, we could demonstrate the functionality of the A. gossypiiα-factor signalling via AgSte2. This suggests conserved interactions between AgSte2 and the S. cerevisiae heterotrimeric G-protein. Similarly, the C. albicans and Sordaria macrosporaα-factor receptors were expressed and effectively coupled to the S. cerevisiae signal transduction pathway (Janiak, 2005; Mayrhofer & Pöggeler, 2005).

The deletion analysis of three components of this signal transduction cascade in A. gossypii revealed a role in sporogenesis. The deletion of STE12 resulted in a striking hypersporulation phenotype that also resulted in next-to quantitative conversion of hyphal segments into sporangia and spores in liquid minimal media. Thus, signalling through this cascade could actually be used to regulate development in A. gossypii. The deletion of STE12 homologs in other fungal systems, for example in S. macrospora, resulted in a block in ascospore formation, while the deletion of SteC (a STE11 homolog) blocked cleistothecium development in Aspergillus nidulans (Wei, 2003; Nolting & Poggeler, 2006).

In S. cerevisiae, both pheromone response and filamentation pathways rely on Ste12. To activate filamentation genes, ScSte12 interacts with ScTec1, while pheromone-responsive genes do not require Tec1 (Wong Sak Hoi & Dumas, 2010). How a filamentation pathway could regulate growth in a filamentous fungus has not been addressed. We have recently characterized the A. gossypii TEC1 homolog and found that it regulates a flocculin, encoded by FIG2, and a chitinase, encoded by CTS2, in A. gossypii (Grünler, 2010). Reanalysis of the Agtec1 mutant showed a highly similar hypersporulation phenotype as seen in Agste12. This suggests that contrasting the transcriptional circuitry in S. cerevisiae (Laloux, 1990), in A. gossypii, Tec1 and Ste12 may cooperate in the regulation of sporulation.

In S. cerevisiae, cell type specificity is conferred by the mating-type locus. In A. gossypii, organization of the MATa locus differs from S. cerevisiae in that two proteins a1 and a2 are encoded by MATa similarly as, for example in Candida albicans (Johnson, 2003). Further studies on the role of the A. gossypii mating-type locus, AgSTE12 and AgTEC1 will require DNA-array transcription profiling or similar techniques to elucidate the target genes regulated by these transcription factors. Rewiring of conserved transcription factors has been shown to be frequent in C. albicans (Ihmels, 2005; Tsong, 2006; Rokas & Hittinger, 2007; Tuch, 2008). The first example of transcriptional rewiring has been the Tec1-dependent constitutive expression of FIG2 in A. gossypii (Grünler, 2010). Fig2 is a cell wall adhesin that is specifically expressed during mating in S. cerevisiae (Zhang, 2002). However, the deletion of FIG2 in A. gossypii did not alter the colony morphology of fig2 mycelia.

Homokaryotic or haploid fruiting has been described in several basidiomycetes, particularly in C. neoformans (Esser & Meinhardt, 1977, Wickes, 1996). At first, this was thought to be MATα specific in C. neoformans (Lin, 2005). However, recently, it was shown that upon overexpression of STE12a MATa strains also show haploid fruiting (Tscharke, 2003). Haploid fruiting also occurs in the basidiomycete Coprinopsis cinerea in AmutBmut strains (Kues, 2000). In ascomycetes, specifically in Neurospora, heterothallic and homothallic species have been described (Glass & Smith, 1994). Interestingly, homothallic species can either carry both mating-type loci A and a or only the A mating-type locus (Pöggeler, 1999).

Haploid fruiting has not been described in Saccharomycetes so far. It is tempting to speculate that the specialized lifestyle of A. gossypii, which grows for example on cotton bolls and citrus fruits and includes transmission by stinging–sucking insects may have led to this mode of sporulation. Yet, this will require further investigation, particularly the demonstration of meiosis in A. gossypii. Spore production in A. gossypii without meiosis could also be conceivable and would then resemble apomixis found in plants. Meiosis in mating-type homozygous strains has been demonstrated for some fungi, for example meiosis in α/α diploids occurs in C. neoformans. This is similar to meiosis in dipoid a/a or α/αS. cerevisiae cells lacking the repressor of meiosis, Rme1 (Lin, 2005). In traditional matings in S. cerevisiae, the a1/α2 heterodimer is required to repress RME1 expression to allow meiosis to occur. Interestingly, A. gossypii lacks an RME1 gene. Thus, one of the safeguards against meiosis has apparently been removed in A. gossypii. Second, the presence of homologs of S. cerevisiae meiosis-specific genes in A. gossypii could be indicative for such a process in A. gossypii. In this respect, the presence of S. cerevisiae homologs of DMC1 (a recombinase that is required for repair of double-strand breaks and pairing between homologous chromosomes, Bishop, 1992), SPO11 (a meiosis-specific endonuclease that catalyzes the formation of double-strand breaks, Keeney, 1997), as well as synaptonemal complex proteins Zip1 and Hop1 in A. gossypii provides some clues on meiosis in A. gossypii, which we will pursue in the future. However, in the A. gossypii genome, there is also a notable absence of MSH4/MSH5, which, in S. cerevisiae, are required for normal levels of cross-over during meiotic recombination. With the multinucleate mycelial compartments, A. gossypii could gain the benefits of recombination and meiosis even without mating.

In some nonmating Candida species, however, the pheromone response could also be used for alternative outputs, for example biofilm formation (Daniels, 2006; Butler, 2009).

Our results show that the disruption of the pheromone response pathway at the level of pheromone-receptors or at the level of the Ste12 transcription factor influence, but are not required for sporulation. Functional analyses for example of key meiotic genes could be useful to demonstrate the existence of bona fide meiosis in A. gossypii. At present, we have uncovered several reductionist events in the processes of pheromone-signal transduction/meiosis that may have contributed to the mechanism of reproduction in A. gossypii: (1) Truncation of AFL062w/MFα1 to not include a functional pheromone. (2) Crippling of the AAR163c/MFα2 to only harbor one active pheromone peptide that may not be processed correctly due to lack of the conserved Kex2-cleavage site. (3) Downregulation of pheromone gene expression and/or preproprotein processing resulting in insufficient amounts of mature α-factor to trigger any response. Similarly, the a-factor processing could be affected based on the weak expression we observed for STE16, which, in S. cerevisiae, is involved in farnesylation of the a-factor. These events are may not be detrimental when pheromone signalling is not used to attract a mating partner. (4) Rewiring of transcriptional regulation of A. gossypii genes whose homologs in S. cerevisiae are mating-specific, for example FIG2. (5) Deletion of RME1 to eliminate a safeguard that represses the ability of haploid cells to sporulate. (6) Adopting a multinucleate mycelial lifestyle that eliminates the need for cell fusion before karyogamy. In this respect, it is interesting to note that several of the S. cerevisiae KAR-genes are conserved in A. gossypii. (7) Elimination of genes, for example MSH4/MSH5 or ZIP2 that in S. cerevisiae participate in meiosis. These findings make the genus Eremothecium an interesting group to study sexual reproduction and sporulation and their reproductive isolation from other yeasts within the Saccharomycetes.


This research was supported in part by the European Union Marie Curie Research Training Network Signalpath. Sequence data for A. gossypii were obtained from the Ashbya Genome Database website at http://agd.vital-it.ch/index.html.


  • Editor: Teun Boekhout


View Abstract