taxonID	type	description	language	source
F34966167B0DFFBAFF45FA82FDC66D1A.taxon	description	As seen in Figure 2, P. anserina appears unable to use glycerol as a sole carbon source for its vegetative growth as mycelium growth speed and density was not different on M 0 medium containing 0.01 M to 0.04 M (= 0.9 g / L to 3.7 g / L) of glycerol compared to M 0 medium lacking a carbon source. With higher glycerol concentrations (0.05 M to 0.3 M; i. e., 4.5 g / L to 27.6 g / L), growth of P. anserina was inhibited and even completely abolished at the highest concentrations. At 0.05 M glycerol, growth speed was inhibited by 15 % (6.5 mm / day instead of 7.5 mm / day on M 0 without glycerol) and at 0.1 M by 66 % (2.5 mm / day instead of 7.5 mm / day). On the same media containing glycerol with concentrations up to 0.3 M, N. crassa could grow and form a profuse mycelium (Fig. 2). Interestingly, the plates containing 0.01 to 0.04 M of glycerol developed mature perithecia producing ascospores (Fig. 2), while the ones lacking glycerol did not (Fig. 2). These matured slowly and required three weeks to produce ripe ascospores instead of one week on the standard M 2 medium. They were mainly located within the medium, were small and contained few ascospores. At higher concentrations, perithecium production was inhibited and mostly small protoperithecia were observed on the M 0 medium. This indicated that P. anserina, while unable to use glycerol for its vegetative growth could use it to produce mature fruiting bodies, albeit inefficiently. The presence of small amounts of metabolites (0.1 g / L of glucose, galactose, glutamine, peptone and yeast extract) did not improve the growth of P. anserina on M 0 medium containing 0.05 M glycerol unlike what has been reported for other fungi. Indeed, the cultures with 0.05 M glycerol displayed the same growth speed, spindly mycelium morphology, and fertility as the cultures without glycerol and having only 0.1 g / L of added metabolites. Overall, the data showed that P. anserina was unable to use glycerol as a carbon source for its vegetative growth even in the presence of metabolites, unlike what had been reported for other fungi, including its close relative N. crassa. Nevertheless, albeit inefficiently, P. anserina could use glycerol to make mature fruiting bodies. GLYCEROL IS TOXIC TO P. ANSERINA Growth and fertility of P. anserina on medium with glycerol as sole carbon source was inhibited at high concentrations. We thus explored whether these inhibitions were also observed on the standard M 2 medium containing dextrin as a carbon source that is routinely used to grow P. anserina. As a control, we tested sorbitol at the same osmolarity as glycerol to check whether inhibition was due to high osmolarity. As seen in Figure 3, glycerol inhibited the growth of P. anserina in M 2 medium like it did in M 0 medium (by 15 % at 0.05 M and 69 % at 0.1 M), while sorbitol did not even at the highest concentrations. We tested whether sorbitol could be used as sole carbon source and whether in these conditions high amounts of sorbitol had an inhibitory effect on P. anserina growth and fertility (Fig. 3). Sorbitol was used by P. anserina as sole carbon source and, even at the highest concentration, sorbitol did not impair the growth of the fungus and was efficiently used as a sole carbon source. This indicated that glycerol was toxic to P. anserina and that toxicity was not due to an effect on osmolarity. MINING THE GENOME OF P. ANSERINA FOR GENES INVOLVED IN GLYCEROL CATABOLISM The P. anserina genome was mined to identify CDS possibly involved in glycerol uptake and catabolism (Table 1; Fig. 1). We could detect by “ BEST-BEST hit ” using BLAST an orthologue of the S. cerevisiae STL 1 transporter. However, BLAST analyses failed to uncover homologues for S. cerevisiae FPS 1 or A. glaucus AgglpF channels. We could detect orthologues of GUT 1 / AN 5589.2 encoding GK and GUT 2 / AN 1396.2 encoding G 3 PDH (encoded by Pa _ 4 _ 7610 and Pa _ 6 _ 5500, respectively), indicating that P. anserina likely has a functional G 3 P pathway. The presence of the DHA pathway in P. anserina needed to be clarified. Indeed, five proteins with significant levels of similarity to S. cerevisiae GCY 1 and A. nidulans AN 7193.2 NDGDs were encoded in the P. anserina genome. To identify which one was a potential NDGD, a phylogenetic tree was constructed with the five P. anserina potential NDGDs, the S. cerevisiae GCY 1 protein and its paralogue resulting from the duplication of the S. cerevisiae genome YPR 1, the putative S. pombe NDGDs SPBC 8 E 4.04 and SPAC 26 F 1.07, as well as all the proteins of A. nidulans similar to AN 7193.2 (Appendix 2). As seen in Appendix 2, GCY 1 and AN 7193.2 did not appear to have true orthologues in P. anserina. However, most nodes are poorly supported owing to the small sizes of the proteins and the poor conservation of their primary sequences, making it difficult to draw a final decision as to the presence of actual NDGD in P. anserina. We tentatively propose that Pa _ 1 _ 3840, Pa _ 1 _ 21080, and Pa _ 6 _ 7100 may be NDGDs owing to their similarity to AN 7193.2 or GCY 1, while no conclusion may be reached for Pa _ 3 _ 2850 and Pa _ 3 _ 4920. Note that a homologue for the Gld 1 NDGD of S. pombe could not be found in the genomes of P. anserina, (nor in the genomes of S. cerevisiae and A. nidulans). On the contrary, two clear homologues of the Dak 1 and Dak 2 DHAK of S. pombe, DAK 1 and DAK 2 DHAK of S. cerevisiae, and AN 0034.2 DHAK of A. nidulans were encoded in the P. anserina genome (Pa _ 2 _ 11570 and Pa _ 6 _ 1010). As stated in the introduction, the key GO enzyme of the GA pathway has yet to clearly be identified in fungi. The only known enzyme endowed with a weak GO activity has been identified in P. chrysosporium (Nguyen et al. 2018) and belongs to the GMC oxidoreductase family. The genome of P. anserina contains 28 genes encoding such enzymes, none of which has been characterized to date (Ferrari et al. 2021). Similarly, no GDHK sequence appears available for fungi, only for mammals (https: // www. brenda-enzymes. org / enzyme. php? ecno = 2.7.1.28). Using sequences of GDHK from mammals, BLAST analysis identified Pa _ 2 _ 11570 and Pa _ 6 _ 1010 putative DHAKs as closely related to these enzymes from animals; due to the close chemical structure of dihydroxyacetone and D-glyceraldehyde (Fig. 1), both may be used as substrates by Pa _ 2 _ 11570 and Pa _ 6 _ 1010. Finally, P. anserina has two potential ADHs encoded by Pa _ 6 _ 4990 and Pa _ 7 _ 1930 and one potential G 3 K encoded by Pa _ 1 _ 23800. Regarding a potential regulator of glycerol metabolism, we could not detect in the genome of P. anserina an obvious orthologue of the S. cerevisiae ADR 1 transcription factor involved in the regulation of GUT 1 and GUT 2 expression. Mining the genome of A. nidulans and N. crassa also failed to retrieve an orthologue of ADR 1, suggesting that, even though A. nidulans glycerol genes are controlled by a transcription factor binding a consensus similar to the one bound by ADR 1, a different transcription factor is likely involved in glycerol gene regulation in the Pezizomycotina. We also mined the transcriptomic data generated by Silar et al. (2019), which provided an estimate of the expression of the genes in non-germinated ascospores, in ascospores eight hours after germination trigger, in 1 - day-old and 4 - day-old mycelia and in 2 - day-old and 4 - day-old perithecia (Appendix 5). All glycerol genes were expressed at all stages of the lifecycle, including the glycerol symporter encoded by Pa _ 1 _ 5200, the GK encoded by Pa _ 4 _ 7610, and the G 3 PDH encoded by Pa _ 6 _ 5500. Overall, genome and transcriptome mining suggested that P. anserina can import glycerol and metabolize it through the G 3 P pathway. The presence of the DHA and GA pathways is more dubious; however, P. anserina appears to have enzymes from both pathways. DELETION OF THE P. ANSERINA PAGUT 1 AND PAGUT 2 GENES Because the G 3 P pathway appears to be the major pathway for glycerol uptake in the Pezizomycotina, including for P. anserina, we deleted the genes encoding the putative GK Pa _ 4 _ 7610 and G 3 PDH Pa _ 6 _ 5500; these were renamed PaGUT 1 and PaGUT 2, respectively, and the deletion mutants PaGUT 1 ∆ and PaGUT 2 ∆. We also retrieved the PaGUT 1 ∆ PaGUT 2 ∆ double mutant in the progeny of the cross between the PaGUT 1 ∆ and PaGUT 2 ∆ single mutants. The mutants exhibited no phenotype on M 0 and M 2 media, i. e., their growth speed, mycelium morphology and fertility were indistinguishable from that of the wild type (Fig. 4). On the contrary, on M 0 and M 2 medium in the presence of 0.05 M of glycerol, the vegetative growth of PaGUT 1 ∆, PaGUT 2 ∆ and PaGUT 1 ∆ PaGUT 2 ∆ was slightly more severely affected than that of the wild type because growth of the mutants was reduced by 20 % instead of 15 % at 0.05 M glycerol in the medium (speed of c. 6 mm / day instead of c. 6.5 mm / day for the wild type; see Figure 4 for pictures of the growth plates). However, at 0.1 M glycerol in M 0 or M 2, the growth of the mutants was as much impaired as that of the wild type (Fig. 4). Importantly, unlike the wild type, PaGUT 1 ∆, PaGUT 2 ∆ and PaGUT 1 ∆ PaGUT 2 ∆ did not produce mature perithecia after three weeks incubation on medium containing 0.02 M glycerol (Fig. 5). Longer incubation time did not result in the production of mature perithecia by these mutants. Finally, their fertility was like that of the wild type on an M 2 medium containing glycerol. Also, on M 0 and M 2 media with sorbitol, the mutants grew and were fertile like the wild type. Reintroduction of the PaGut 2 wild-type allele into the mutant PaGUT 2 ∆ restored a wild-type phenotype (i. e., the production of perithecia on M 2 + 0.02 M glycerol). Unfortunately, we could not complement the PaGUT 1 ∆ mutant because we failed to recover in bacteria a plasmid carrying the PaGut 1 wild-type allele without sequence alteration (a recombination event eliminated systematically the end of the gene) and direct co-transformation with the PaGut 1 wild-type allele PCR amplification product with a resistance marker yielded few transformants for an unknown reason despite several attempts; none of these had a wild-type phenotype. Nonetheless, in crosses between PaGUT 1 ∆ and the wild type the lack of perithecium production on glycerol medium co-segregated with the resistance used for deleting PaGut 1, arguing that this phenotype was actually due to the deletion of PaGut 1 and not to an additional unlinked mutation. THE PRESENCE OF BACTERIA GREATLY IMPROVES THE GROWTH AND FERTILITY OF P. ANSERINA ON GLYCEROL MEDIA We serendipitously discovered that the presence of bacteria in co-culture with P. anserina significantly improves growth and fertility of the fungus on M 0 medium added with glycerol as sole carbon source. Indeed, we observed that, in a series of plates contaminated by bacteria, the bacterial colonies were surrounded by what looked like a denser hyphal thallus bearing mature perithecia (Fig. 6). We identified the contaminating bacterium by sequencing its 16 S barcode and found it to be Paenibacillus validus (Bacillota formerly Firmicutes). To check whether the improved growth and fertility were specific to P. validus or could be achieved by other bacteria, we tested several other species including Microbacterium proteolyticum (Actinomycetota), Massilia sp. (Pseudomonadota – Betaproteobacteria), Paracoccus sp. (Pseudomonadota – Alphaproteobacteria), Sphingobacterium kitahiroshimense (Bacteroidota – Sphingobacteriia), Pseudomonas putida (Pseudomonadota – Gammaproteobacteria) and Escherichia coli T. Escherich DH 5 α (Pseudomonadota – Gammaproteobacteria). All were found to improve the growth and fertility of P. anserina on glycerol as sole carbon source (see Figures 6 and 7 for improvement of growth and fertility of P. anserina by E. coli). As it is a well-known model and grows well on high glycerol concentrations (Appendix 3), we chose E. coli DH 5 α for further testing. We especially tested whether the growth and fertility of the PaGUT 1 ∆, PaGUT 2 ∆ and PaGUT 1 ∆ PaGUT 2 ∆ mutants were also improved in the presence of the bacteria. As seen in Figure 7, E. coli improved the vegetative growth of all mutants and the fertility of all mutants except for PaGUT 2 ∆, for which only growth was improved, as PaGUT 2 ∆ did not produce perithecia in the presence of bacteria.	en	Gautier, Valérie, Nguyen, Tinh-Suong, Silar, Philippe (2024): Fungal / bacterial syntrophy of glycerol utilization. Cryptogamie, Mycologie 20 (6): 53-69, DOI: 10.5252/cryptogamie-mycologie2024v45a6, URL: https://sciencepress.mnhn.fr/sites/default/files/articles/pdf/mycologie2024v45a6.pdf
