The SCF^Fbx23 complex is among the interaction partners of transcription factor PoAce1

All strains used in this study are listed in Table 1. Colonies of P. oxalicum wild-type (WT) strain and strains constructed in this study are shown in Fig 1A.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Strains used in this study.

https://doi.org/10.1371/journal.pgen.1011539.t001

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. The results of TAP-MS using transcription factor PoAce1 as bait.

The Ace1-TAP strain was generated by fusing the TAP tag (HA-FLAG) at the C-terminus of the PoAce1 (PDE_01988, Uniprot Entry S8AYJ0). The WT strain 114–2 containing native PoAce1 was used as a control. The biological triplicates were used for the TAP-MS experiment. (A) Colonies of P. oxalicum wild-type (WT) strain and strains constructed in this study. 1 μl conidia suspension (10⁷ conidia ml⁻¹) of various strains was point-inoculated onto Vogel’s minimal medium (VMM) plus 2% glucose (VMMG) agar and then cultivated at 30°C for three days. (B) Silver staining after SDS-PAGE of Ace1-TAP eluate. The arrow represents the protein PoAce1. (C) Western blot of Ace1-TAP eluate using anti-HA antibody. (D) Intersection proteins among Ace1-TAP biological triplicates (Ace1-TAP-1/2/3) (The detailed information of proteins is shown in Table 2, S1 Spreadsheet and S1 Data). (E) The models of heterodimer Fbx23-Skp1 (left) and ubiquitin ligase complex SCF^Fbx23 (right) according to the TAP-MS results. The bait PoAce1 is the substrate. The PoFbx23 (ranked 2nd in Table 2) is the F-box-containing protein that interacts with the substrate PoAce1 and determines substrate specificity. The PoSkp1 (ranked 1st in Table 2) is an adapter protein of the SCF complex. PoCul1 (ranked 10th in S1 Spreadsheet, sheet1, row 13, yellow background) is the scaffolding protein interacting with the SKP1-F-box complex and RBX1. PoCul1 (ranked 73rd in S1 Spreadsheet, sheet1, row 75, yellow background) is the RING-finger protein that recruits the E2 ubiquitin-conjugating enzyme.

https://doi.org/10.1371/journal.pgen.1011539.g001

Tandem affinity purification coupled with mass spectrometry (TAP-MS) is a valuable technique for identifying interacting partners of the protein of interest in the studies of yeast, Aspergillus, Trichoderma, and Penicillium [36,39,40]. The eluate of the Ace1-TAP strain displayed a specific band at about 100 kDa in comparison to the control strain, as shown by the SDS-PAGE result (Fig 1B, dark arrow, PoAce1’s theoretical molecular weight with the TAP tag added is 88.4 kDa). The specific band was then cut out of the gel and was confirmed to be the PoAce1 protein through MS/MS assay (S1 Document). This finding was supported by Western blot using anti-HA antibody (Fig 1C, red arrow). In addition to the signal at about 100 kDa, the Western blot result revealed several low-molecular-weight signals at about 70-, 45- kDa, and 32-kDa (Fig 1C, green arrows).

The proteins in the third section of the eluent were analyzed using LC-MS/MS to identify the bait and interacting proteins. The PepCount identified in biological triplicates (Ace1-TAP-1, Ace1-TAP-2, and Ace1-TAP-3) through TAP-MS were listed in S1 Spreadsheet (Sheet 2 ~ Sheet 4). The proteins observed in any three Ace1-TAP samples but not in any of the controls were listed in S1 Spreadsheet (Sheet 1). The proteins were ranked by exponentially modified protein abundance index (emPAI), which was used to estimate protein abundance in the sample [41]. 237, 264, and 134 proteins were identified in the samples Ace1-TAP-1, Ace1-TAP-2, and Ace1-TAP-3, respectively. Nine proteins were observed in all three Ace1-TAP samples but not in any of the controls (Fig 1D). They were considered reliable proteins that have putative interaction with the PoAce1 and were listed in Table 2.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. The top proteins interacting with PoAce1 identified through TAP-MS.

https://doi.org/10.1371/journal.pgen.1011539.t002

As expected, the bait PoAce1 has the highest emPAI. Among the other eight proteins, the protein with the highest emPAI is PDE_02249 (Uniprot Entry S7ZF79). We name PDE_02249 as PoSkp1 because its homolog in S. cerevisiae is protein SKP1, which is an invariable, core component of the SCF E3 ubiquitin ligase complex that functions as the adaptor protein responsible for binding CUL1 and recruiting various F-box proteins for SCF complex formation [8].

The protein with the second-highest emPAI is PDE_08467 (Uniprot Entry S7ZS07), an F-box/WD repeat protein which is always a variable component of SCF complex that binds to SKP1 through the F-box domain [9]. The homologous protein of PDE_08467 is not detected in S. cerevisiae. However, it was identified in Schizosaccharomyces pombe to be the F-box/WD repeat protein Pof11 (Genbank: NP_594559.1) and in A. nidulans to be the F-box containing protein AnFbx23 (ANIA_05593, Genbank: XP_663197.1). We name PDE_08467 as PoFbx23. Like AnFbx23, PoFbx23 possesses an F-box domain and seven copies of WD repeats (S1 Fig.), which act as a platform for protein-protein or protein-DNA interaction [42].

The proteins ranked third (PDE_09763), seventh (PDE_00208), and eighth (PDE_05998) are uncharacterized proteins. The proteins ranked 4th (PDE_05651) and 6th (PDE_05666) are named PoPyc2 and PoMub1 because their homologs in S. cerevisiae are PYC2 and MUB1, respectively. The protein ranked 5th (PDE_01581) shares similarities (36%) with Set2 in S. cerevisiae. As P. oxalicum contains a protein (PDE_02255, Uniprot Entry: S8AZ99) that shares a higher identity (45%) with the S. cerevisiae Set2 and has been named PoSet2 [43], the PDE_01581 is named PoAsh1 because its homolog in Homo sapiens and Mus musculus is histone methyltransferase Ash1. We also observe PDE_02289, a RING-type domain-containing protein, and PDE_00128, a Cullin family domain-containing protein (Table 2). PDE_02289 and PDE_00128 are present in two but not three biological replicates (S1 Spreadsheet, sheet 1, row 13 and row 75). The PDE_02289 is named PoRbx1 because its homolog is protein RBX1, a subunit of the SCF complex that tethers the E2 and Cullin subunits [8]. The PDE_00128 is named PoCul1 because its homolog is the protein Cullin-1, which functions as a molecular scaffold that simultaneously interacts with the SKP1-F-box complex and with RBX1 or RBX2 [11].

The discoveries of PoSkp1, PoFbx23, PoRbx1, and PoCul1 suggest there exist four components of the SCF^Fbx23 E3 ubiquitin ligase complex in the Ace1-TAP eluent. However, it is noteworthy that the top three proteins, bait (PoAce1), PoSkp1, and PoFbx23, which were observed in all three biological samples, had high emPAI values (12.15, 9.00, and 3.58, respectively). In contrast, the other two subunits (PoRbx1 and PoCul1) of the SCF complex, which only appeared in two of the three biological samples, had low emPAI values (0.77 and 0.07, respectively) (Table 2). As previous studies have shown that substrate adaptors do not always associate with Cul1-Rbx1 heterodimers [44]; assembly of F-box proteins into SCF complexes ranges from 0–70% (median of ~19%) [45], we speculate that the majority of PoFbx23 binds PoAce1 by forming a heterodimer PoFbx23-Skp1 (Fig 1E, left), while a small portion of PoFbx23 assembles with the other three components to form the SCF^Fbx23 complex. In the SCF^Fbx23 complex, PoCul1 acts as a molecular scaffold, PoRbx1 binds to ubiquitin-loaded E2, PoSkp1 serves as an adaptor, and PoFbx23 is a substrate receptor protein determining the substrate specificity. The PoFbx23 recognizes PoAce1 as the substrate (Fig 1E, right). In addition, the discovery of PoMub1 is also noteworthy. PoMub1, as a component of the Ubr2/Mub1 ubiquitin ligase complex, implies that PoAce1 may not be confined to just one ubiquitin ligase system.

PoAce1 and PoFbx23 are localized in the nucleus and interact directly with each other there

Transcription factors can be localized within the nucleus or shuttle from the cytoplasm into it in response to different signals [46]. Moreover, SCF E3 can also be localized in the nucleus and cytoplasm [47,48]. Even for a particular type of FBP, such as FBXW7, its different isoforms can still be localized to the nucleus, cytoplasm, and nucleolus [49]. In light of this, even though TAP-MS results indicate that PoFbx23 might recognize and bind PoAce1, the precise location of this recognition remains unknown. In addition, the putative collaborators identified by TAP-MS contain not only the true interactors of the bait protein but also the proteins that indirectly interact with the bait protein mediated by other proteins when multiple proteins form complexes. Therefore, we first identified the subcellular localization of PoAce1 and PoFbx23 and then determined whether there is a direct interaction between PoAce1 and PoFbx23 and the subcellular location where the interaction occurs using the methods of yeast two-hybrid (Y2H) and bimolecular fluorescence complementation (BiFC).

The strains Ace1-GFP and Fbx23-GFP were constructed. Due to the various signals that can trigger different subcellular localization of transcription factors and FBPs, we observed the subcellular localization of PoAce1 and PoFbx23 on Vogel’s minimal medium (VMM) plus 2% glucose (VMMG) and VMMC (plus 2% cellulose), respectively. The VMM is a universal and convenient minimal medium for studying various fungi such as Neurospora, Aspergillus, and Penicillium in the lab [27,50,51]. The glucose in VMMG inhibits cellulolytic gene transcription due to carbon catabolite repression (CCR) [52]. The VMMC is a suitable medium for studying regulatory proteins for cellulolytic gene transcription due to the inducer cellulose. An overlap of green fluorescence (Fig 2 A & 2B, bottom left) and nuclear staining (Fig 2A & 2B, top right) was observed on the merged image (Fig 2A & 2B, bottom right), regardless of whether the strains Fbx23-GFP and Ace1-GFP were cultured on VMMG or VMMC (Fig 2A & 2B). To observe whether the subcellular localization of PoAce1 is affected by PoFbx23, the mutant ΔF-Ace1-GFP (the Ace1 protein fused the GFP in the Δfbx23 background) was constructed. PoAce1 was localized in the nucleus in the absence of PoFbx23, whether under glucose or cellulose signal (Fig 2C). The results indicate that PoFbx23 and PoAce1 are both localized in the nucleus, whether under glucose or cellulose signal; the localization of PoAce1 is not affected by PoFbx23.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Subcellular localization of PoFbx23A and PoAce1.

(A) Subcellular localization of PoFbx23 on VMMG (left) and VMMC (right), respectively. (B) Subcellular localization of PoAce1 on VMMG (left) and VMMC (right), respectively. (C) Subcellular localization of PoAce1 in the Δfbx23 background on VMMG (left) and VMMC (right), respectively. The image is divided into four parts: upper left, white light; upper right, Hoechst 33342 was used to stain nuclei in blue; bottom left, green fluorescence; bottom right, merged image of green fluorescence and nuclear staining.

https://doi.org/10.1371/journal.pgen.1011539.g002

Y2H results showed two fragments of PoAce1 had evident interaction with PoFbx23, suggesting both fragments may contain regions that interact with PoFbx23 (Fig 3A). The Ace1-YFP-Fbx23 strain for BiFC was constructed. An overlap of yellow fluorescence (Fig 3B, bottom left) and nuclear staining (Fig 2, top right) was observed on the merged image (Fig 3B, bottom right), regardless of whether the strain Ace1-YFP-Fbx23 was cultured on VMMG or VMMC (Fig 3B). These results strengthen the reliability and certainty of the direct interaction between the PoFbx23 and PoAce1 and demonstrate that the interaction occurs inside the nucleus regardless of the glucose or cellulose signal.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. The interaction between PoFbx23A and PoAce1.

(A) Yeast two-hybrid strategy and hybridization results of Y2H Gold-BD-PoFbx23 and Y187-AD-PoAce1 strains on QDO (quadruple-dropout, SD-Ade/-His/-Leu/-Trp) and QDO/x-α-gal/Aba (QDO supplemented with X-α-gal and aureobasidin A). DNA-binding domain (397–483) was removed from PoAce1 as the intact protein showed autoactivation. (B) Microscopy of BiFC strain Ace1-YFP-Fbx23 on VMMG (left) and VMMC (right), respectively. The image is divided into four parts: upper left, white light; upper right, Hoechst 33342 was used to stain nuclei in blue; bottom left, yellow fluorescence; bottom right, merged image of yellow fluorescence and nuclear staining.

https://doi.org/10.1371/journal.pgen.1011539.g003

PoFbx23 absence under the glucose signal resulted in decreased transcription of the brlA gene and spore pigmentation genes, along with impaired conidiation

TAP-MS, Y2H, and BiFC results demonstrate that the PoFbx23 recognizes transcription factor PoAce1 as the substrate in the nucleus. Given that SltA (the homolog of Ace1) is known to be crucial in fungal growth, development, and secondary metabolite biosynthesis in various fungi [29,30,31], it is reasonable to assume that PoFbx23 could indirectly influence these biological processes. However, the nature of this influence, whether positive (activating) or negative (repressing), remains uncertain. Therefore, we constructed the Pofbx23 gene deletion (Δfbx23) mutant and conducted phenotype, RT-qPCR, and transcriptome analyses to explore this assumption.

Cultivated on VMMG agar, the Δfbx23 mutant showed a reduced rate of colony expansion than that of the WT (Fig 4A). The hyphal elongation speed of the Δfbx23 colony is about 0.69 cm per day, a 21% decrease from the WT of about 0.90 cm per day. However, the growth patterns of the strains were almost identical when they were cultivated in VMMG liquid, as shown by the biomass measurement (Fig 4B). When strains WT, Δfbx23, and Refbx23 were grown on VMMC agar, there was no significant difference in their radial growth. However, the colony diameters of these three strains were significantly smaller on VMMC than on VMMG. For example, on the 7^th day of incubation, the colony diameters of WT and Δfbx23 on VMMC agar were only 40.6% and 59.6% of those on VMMG agar, respectively (Fig 4A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Effects of PoFbx23 protein on fungal growth, conidiation, and the transcription of genes involved in conidiation and spore pigment synthesis under glucose signal.

(A) Colony diameter determination. 1 μl conidia suspension (10⁷ conidia ml-¹) of various strains was point-inoculated onto the VMMG or VMMC agar and cultivated at 30°C. G, VMMG. C, VMMC. All P values of radial growth of the Δfbx23 strain compared to the WT when they were cultivated in VMMG are < 0.05. (B) The biomass measurement of different strains. (C) Conidia quantification. The WT, Δfbx23, and Refbx23 strains were incubated on VMMG agar at 30°C for 5 days. Then, 5-mm diameter colony agar plugs were harvested with the physiological salt (0.2% w/v Tween 80 and 0.8% w/v NaCl). Conidia quantification was determined based on a direct count by using a hemocytometer. (D) Transcription level of the brlA gene assayed by qRT-PCR. (E) The expression of the brlA gene and six genes abrB, abrA, aygA, arpA, arpB, and albA involved in the DHN-melanin pathway of pigment synthesis revealed by transcriptome data after the strains were incubated in VMMG liquid. The differential expression folds change and the Q-value of each gene between the mutant and the WT are shown in S2 Spreadsheet (Shee1). All Q-values are < 0.05. The copy number of unambiguous tags mapping to genes was normalized to fragments per kilobase transcriptome per million mapped reads (FPKM). (F) The transcription level of four genes, abrB, abrA, arpB, and albA assayed by qRT-PCR. The values were normalized to the actin gene levels considered 1. *p < 0.05, **p < 0.01, ***p < 0.001.

https://doi.org/10.1371/journal.pgen.1011539.g004

The conidiation in Δfbx23 was reduced to approximately 63.2% of that in the WT (Fig 4C). As transcription factor BrlA is a master regulator for asexual development and its regulatory role cannot be bypassed during conidiation of Penicillium and Aspergillus [34,35], brlA expression was assayed using RT-qPCR after the strains were cultivated in VMMG liquid. The deletion of Pofbx23 gene downregulated the expression of the brlA gene significantly (Fig 4D). Transcriptome data from Δfbx23 cultured in the same medium also supported the decrease in expression of the brlA gene; the transcription level of the brlA gene in the Δfbx23 reduce to only 9.1% of that in the WT (Fig 4E).

Interestingly, in addition to the downregulation of the brlA gene, downregulated expression of genes related to spore pigment synthesis was observed when we investigated the transcriptome data. Melanin, the critical spore wall pigment, provides the characteristic dark green pigment for many fungal spores. Melanin is synthesized through the dihydroxynaphthalene (DHN)-melanin pathway that includes six proteins (Abr2→Abr1→Ayg1→Arp1→Arp2→Alb1) in A. fumigatus [53]. The homologs of the above six proteins in P. oxaliucm are laccase gene AbrB/yA (PDE_09492, Genbank No. EPS34528.1), pigment biosynthesis protein brown 1 AbrA (PDE_09453, Genbank No. EPS 34489.1), pigment biosynthesis protein AygA (PDE_09452, Genbank No. EPS34488.1), tetrahydroxy-naphthalene reductase ArpA (PDE_04496, Genbank No. EPS29546.1) and ArpB (PDE_04495, Genbank No. EPS29545.1), and polyketide synthase gene AlbA/wA (PDE_09491, Genbank No. EPS34527.1), respectively. The transcription of all six pigment biosynthesis-related genes was significantly downregulated in the Δfbx23. The transcription level of abrB, abrA, aygA, arpA, arpB, and albA gene in the Δfbx23 reduced to 2.0%, 46.3%, 2.8%, 2.3%, 2.8%, and 4.1% of that in the WT, respectively (Fig 4E, right). The RT-qPCR assay of the first two genes (abrB and abrA) and the last two genes (arpB and albA) in the DHN-melanin pathway supported the transcriptome results; the transcription level of abrB, abrA, arpB, and albA gene in the Δfbx23 reduce to 2.0%, 36.8%, 0.6%, and 4.5% of that in the WT at cultivation time of 24 h (Fig 4F and S2 Data).

Deleting Pofbx23 gene also affected the expression of secondary metabolic gene clusters in addition to fungal growth and asexual development. There are 28 secondary metabolism gene clusters predicted in the P. oxalicum genome (S1 Table) [32]. Among the 28 secondary metabolism gene clusters, four secondary metabolism gene clusters, which correspond to clusters 1, 2, 26, and 28, respectively (S1 Table, yellow background), were downregulated in the Δfbx23 mutant (S2A~D Fig). Among four secondary metabolism gene clusters, the two clusters presented in S2B Fig (Cluster 2) and S2C Fig (Cluster 26) have cluster synteny with Penicillium chrysogenum. Cluster 2 is involved in the biosynthesis and secretion of the mycotoxins roquefortine C and meleagrin [54]. Interestingly, all four secondary metabolism gene clusters, silenced in the Δfbx23, were also silenced in the previously reported brlA deletion mutant [35]. This finding implies that the four secondary metabolic gene clusters are indirectly downregulated due to the Pofbx23 deletion-induced downregulation of brlA expression.

In conclusion, the involvement of PoFbx23 in hyphal development, conidiation, and secondary metabolite gene expression is comparable to the identification of SltA/Ace1 in these same processes in various fungi [29–31]. In addition, the phenotype of the Poace1 gene deletion mutant (Δace1) of P. oxalicum was observed (S3 Fig). The Δace1 mutant showed a slower hyphal elongation speed than the WT (S3A&S3B Fig). It also showed a significant decrease in conidiation and the transcription levels of the brlA and abrB genes (S3C~E Fig). The results further suggested a correlation between PoFbx23 and PoAce1 in regulating hyphal development and conidation.

PoFbx23 absence under the glucose signal does not affect Poace1 gene transcription but results in an accumulation of PoAce1 protein

Then, we investigated how PoFbx23 absence affected PoAce1 under glucose signal (Fig 5). The Pofbx23 gene was deleted in the Ace1-TAP strain to construct the TAP-Δfbx23 mutant. Using the TAP method, PoAce1 was obtained from mycelial cells of the same biomass of WT, Ace1-TAP, and TAP-Δfbx23 strains after they were cultured in VMMG liquid. When PoFbx23 is present in the Ace1-TAP strain, PoAce1 exists as a full-length version (about 100 kDa, theoretical molecular weight is 88.4 kDa), and several low-molecular-weight degraded versions (about 70-, 60-, 45-, and 32-kDa). The abundance of intracellular PoAce1 was significantly increased in the TAP-Δfbx23 strain. While the band of PoAce1 was nearly invisible on the SDS-PAGE of the eluate of the Ace1-TAP strain, they were visible at ~100 kDa on the TAP eluate of the TAP-Δfbx23 strain (Fig 5A, left). The Western Blot results also demonstrated a stronger PoAce1 signal at about 100-, 70-, 60-, 45-, and 32-kDa of the TAP-Δfbx23 strain than that of Ace1-TAP (Fig 5A, right). The findings demonstrate that PoAce1 protein is more stable and accumulates in the cell in the absence of PoFbx23.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. The effect of PoFbx23 on the protein abundance, transcription level, and DNA binding ability of PoAce1 under glucose signal.

The strains were grown in VMMG. (A) The results of TAP using PoAce1 as bait in the Ace1-TAP and TAP-Δfbx23 strains. Left, silver staining after SDS-PAGE of Ace1-TAP eluate (The dark arrow represents the PoAce1 protein (about 100 kDa on the gel, theoretical MW: 88.4 kDa). Right, the Western blot of Ace1-TAP eluate using an anti-HA antibody. The arrow represents the PoAce1 protein. (B) The transcription level of the Poace1 gene in the WT, Δfbx23, Refbx23, Ace1-TAP, and TAP-Δfbx23 strains, assayed by RT-qPCR. The values were normalized to the actin gene levels considered 1. (C) The transcription level of Poace1 gene in the WT, Δfbx23, Refbx23, and revealed by transcriptome data. The copy number of unambiguous tags mapping to genes was normalized to FPKM. (D) The PoAce1 protein enrichment in the specific regions of the abrB gene assayed by ChIP-qPCR. Top subgraph, Strategy of ChIP-qPCR. The transcription start point (TSP) is designated as +1. The initiator (Inr) and TATA box are illustrated. Six specific regions (R1 to R6) of each gene are designed for qPCR assay (as shown at the top of the subgraph). Region 1 (R1), region 2 (R2), and region 3 (R3) are located upstream of the TSP. Region 4 (R4) covers the TATA-box and the initiator (Inr) (R4*, no appropriate primers for qPCR could be designed within this region). Region 5 (R5) is located in the 5′ region of the CDS of the abrB gene. Region 6 (R6) is located in the middle of the CDS. Bottom subgraph, the results of ChIP-qPCR for abrB gene. (E) The PoAce1 protein enrichment in the specific regions of the brlA gene assayed by ChIP-qPCR. Top subgraph, Strategy of ChIP-qPCR. The TATA box and Inr were not covered by the region design, because no appropriate primers for qPCR could be designed within this region. Bottom subgraph, the results of ChIP-qPCR for brlA gene. Purple triangles indicate the putative DNA-binding sites of PoAce1. The orientation of the triangle represented the orientation of the binding motif. The relative enrichment of IP DNA was calculated by % of input. All values are means from measurements of biological triplicates. The error bars indicate standard deviations *p < 0.05, **p < 0.01, ***p < 0.001.

https://doi.org/10.1371/journal.pgen.1011539.g005

Initially, we hypothesized that the increase in PoAce1 protein abundance was due to the increased transcription of its genes. Therefore, RT-qPCR was used to determine the transcription levels of the Poace1 gene in WT, Δfbx23, Refbx23, Ace1-TAP, and TAP-Δfbx23 grown in VMMG liquid. Neither the Δfbx23 mutant nor the TAP-Δfbx23 mutant showed a change in the transcript levels of the Poace1 gene compared to their respective parent strains, the WT and Ace1-TAP (Fig 5B). Additionally, the transcriptome data of the WT and Δfbx23 mutant grown in VMMG also showed that the transcription levels of Poace1 gene were not significantly different between the WT and the Δfbx23 mutant (Fig 5C). The results demonstrate that the deletion of Pofbx23 had no effect on Poace1 gene transcription but led to the accumulation of PoAce1 protein under glucose signal. Given the biological function of the SCF complex, it is reasonable to assume that PoAce1 protein accumulates in the cell due to PoFbx23’s absence which prevents PoAce1 protein from being degraded via the nuclear UPS pathway.

Accumulated PoAce1 no longer binds pigmentation-related gene abr2

The Pofbx23 deletion strain exhibited an accumulation of PoAce1 protein, accompanying downregulated expression of the brlA gene, pigmentation-related genes, and impaired conidiation. This finding contradicts earlier research that suggested a positive correlation between the transcription factor Ace1 and brlA expression and conidiation [29,30]. The activating activity of PoAce1 seems lost due to Pofbx23 deletion. According to previous research on SltA/Ace1 in Aspergillus sp. and Trichoderma sp., SltA/Ace1 regulates the expression of target genes by directly binding to the conserved 5′-AGGCA-3′ motif in their promoter regions [26,55]. Our analysis of the promoter region of the pigmentation-related gene abr2, which encodes the first enzyme in the DHN-melanin pathway, revealed a significantly higher frequency of motif occurrence (5 sites of 5′-AGGCA-3′) than the theoretical frequency (2 sites) (Fig 5D, upper). This finding suggests that PoAce1 might bind directly to the abrB gene. To confirm this, we used Chromatin immunoprecipitation (ChIP)-qPCR to detect the degree of enrichment of PoAce1 in the promoter of the abrB gene in the WT, Δfbx23, Ace1-TAP, and TAP-Δfbx23 strains (Fig 5D&5E and S3 Data). In addition, as the transcription of brlA gene was downregulated in the Δfbx23 mutant, the enrichment of PoAce1 in the promoter of brlA gene was also determined.

The detailed strategy for ChIP-qPCR design is shown in Fig 5D&5E (top). ChIP-qPCR results demonstrated PoAce1 binding in the promoter and 5’-CDS of the abrB gene in the Ace1-TAP strain. Enrichment of PoAce1 was observed at all assayed regions (R1/2/3/5/6). However, despite the PoAce1 protein accumulating in the TAP-Δfbx23 strain, PoAce1 binding of the target gene was defective; the levels of PoAce1 binding DNA at all assayed regions (R1/2/3/5/6) of the abrB gene reduced remarkably in TAP-Δfbx23 compared with those in the Ace1-TAP strain (Fig 5D, bottom). PoAce1 was somewhat enriched in two (R3/4) of the six detected regions of the brlA gene (Fig 5E, bottom). ChIP-qPCR results showed that the PoFbx23 is essential for PoAce1’s binding DNA activity under glucose signal. PoAce1 binds to the regions adjacent to the core promoter of the target gene within the normal cells; the active form of PoAce1 is required in abrB expression. In the absence of PoFbx23, despite PoAce1 protein being more stable and accumulating, it no longer binds to the target gene and takes on an inactive form; the expression of the target gene is reduced due to the dysfunctional PoAce1.

PoFbx23 absence under the cellulose signal results in an accumulation of PoAce1 protein without affecting the Poace1 gene transcription

An individual F-box protein can perform different functions in response to different signals [10]. In cellulase-producing fungi like T. reesei and P. oxalicum, Ace1 is of interest as a negative regulator of (hemi)cellulase gene expression under (hemi)cellulose induction signal [26,27]. We investigated how PoFbx23’s absence affected PoAce1 under the cellulose signal.

Using the TAP method, PoAce1 was obtained from mycelial cells of the same biomass of WT, Ace1-TAP, and TAP-Δfbx23 strains after they were cultured in VMMC liquid. When PoFbx23 is present in the Ace1-TAP strain, PoAce1 exists as a full-length version (~100 kDa, theoretical molecular weight is 88.4 kDa), and several low-molecular-weight degraded versions (about 70-, 45-, and 32-kDa). The abundance of intracellular PoAce1 was significantly increased in the TAP-Δfbx23 strain. While the band of PoAce1 was nearly invisible on the SDS-PAGE of the eluate of the Ace1-TAP strain, they were visible at ~100 kDa on the TAP eluate of the TAP-Δfbx23 strain (Fig 6A, left). The Western Blot results also demonstrated a stronger PoAce1 signal at about 100-, 70-, 45-, and 32-kDa of the TAP-Δfbx23 strain than that of Ace1-TAP (Fig 6A, right). The results show that PoAce1 protein is more stable after the deletion of Pofbx23; PoAce1 protein is accumulated in the cell under cellulose signal.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. The effect of PoFbx23 on the PoAce1 protein abundance and transcription levels of genes encoding plant cell wall-degrading enzymes under cellulose signal.

the strains were grown in VMMC. (A) The results of TAP using PoAce1 as bait in the Ace1-TAP and TAP-Δfbx23 strains. Left, silver staining after SDS-PAGE of TAP eluate. The dark arrow represents the protein PoAce1 (about 100 kDa on the gel, theoretical MW: 88.4 kDa). Right, the Western blot of TAP eluate using an anti-HA antibody. (B) The transcription level of the Poace1 gene in the WT, Δfbx23, Refbx23, SltT-TAP, and TAP-Δfbx23 strains, assayed by RT-qPCR. The values were normalized to the actin gene level considered 1. (C) The transcription level of Poace1 gene in the WT, Δfbx23, Refbx23 revealed by transcriptome data. The copy number of unambiguous tags mapping to genes was normalized to FPKM. (D) Transcription levels of cellulase genes cbh1 and eg1 were determined by qRT-PCR. The values were normalized to the actin gene level considered 1. (E) Transcription levels of 13 cellulase genes in the Δfbx23 mutant compared with WT revealed by transcriptome data. (F) Clustering analysis of 115 genes that encode plant cell wall-degrading enzymes in the Δfbx23 mutant compared with WT using Genesis [56]. The color of each block represents the log₂ (fold change) in gene transcription of Δfbx23 vs. WT. The genes marked with red fonts are directly related to cellulose degradation. The genes marked with green fonts are directly related to hemicellulose degradation. The genes marked with blue fonts encode the protein/enzyme assisting the breakdown of plant cell walls.

https://doi.org/10.1371/journal.pgen.1011539.g006

Initially, we hypothesized that the increase in PoAce1 protein abundance was due to the increased transcription of its genes. Therefore, RT-qPCR was used to measure the transcript levels of the Poace1 gene in WT, Δfbx23, Refbx23, Ace1-TAP, and TAP-Δfbx23 grown in VMMC liquid. It was found that neither the Δfbx23 mutant nor the TAP-Δfbx23 mutant showed a significant change in the transcript levels of the Poace1 gene compared to their respective parent strains, the WT and Ace1-TAP (Fig 6B). Additionally, the transcriptome data of WT and Δfbx23 grown in VMMC also showed that the transcription levels of Poace1 gene were not significantly different (Fig 6C). The results demonstrate that the deletion of Pofbx23 does not affect Poace1 gene transcription but results in the accumulation of PoAce1 protein under cellulose signal, which is the same observation as under glucose signal (Fig 5).

PoFbx23 absence results in increased transcription levels of the (hemi)cellulase gene in response to cellulose induction signal

The absence of PoFbx23 results in an accumulation of PoAce1 protein under the cellulose signal. As transcription factor Ace1 has been demonstrated to be a negative regulator for transcription of the (hemi)cellulase gene in various fungi [26–28], it was initially assumed that the expression of (hemi)cellulase genes could be repressed in the Pofbx23 deletion mutant. However, the RT-qPCR and transcriptome results of (hemi)cellulase genes proved that the initial assumption was wrong. PoFbx23’s absence increases transcription levels of the (hemi)cellulase gene under the cellulose induction signal.

RT-qPCR was used to determine the expression of the two essential cellulase genes cbh1 (encoding exocellulase/cellobiohydrolase, PDE_07945/Cel7A) and eg1 (encoding endocellulase/(endo-β-1,4-glucanase, PDE_07929/Cel7B) in the WT and Δfbx23 cultivated in VMMC liquid. These two cellulases accounted for more than 70% of all the cellulases secreted by cell cultures in response to cellulose signals [32]. There was a significant increase in both cbh1 and eg1 expression in Δfbx23 comparison to WT (Fig 6D).

Then, the transcriptome data of WT and Δfbx23 were also examined following their cultivation in VMMC liquid. The expression levels of 1332 genes between the WT and the Δfbx23 significantly differed (fold change ≥ 2, Q-Value < 0.05) (S2 Spreadsheet, Sheet 3). Among the regulated genes, 780 genes (58.6%) were downregulated, and 552 genes (41.4%) were upregulated in the Δfbx23 compared with those in the WT (S2 Spreadsheet, Sheet 3). GO enrichment analysis revealed that genes related to cell wall polysaccharide metabolic/catabolic process, cellulose metabolic/catabolic process, xylan metabolic/catabolic process, and β-glucan metabolic/catabolic process (S3 Spreadsheet, Sheet1 Biological Process), and cellulose binding, monooxygenase activity, polygalacturonase activity, xylanase activity, cellulase activity (S3 Spreadsheet, Sheet2 Molecular Function) were enriched in the genes that were differentially expressed in the Δfbx23 compared to the WT. This result suggests that the deletion of Pofbx23 does affect the gene expression involved in the process of cell wall polysaccharide metabolic/catabolic, specifically the transcription of cellulase and xylanase (hemicellulase) genes.

The transcription levels of all genes encoding cellulases were examined. It is predicted that the P. oxalicum genome contains 13 cellulases [32]: 5 members of the glycoside hydrolase (GH) 5 family (PDE_02093/eg2/cel5A, PDE_03711/cel5E, PDE_05193/cel5D, PDE_09226/cel5B, PDE_09969/cel5C), one member of the GH6 family (PDE_07124/cbh2/cel6A), three members of the GH7 family (PDE_05445, PDE_07945/cbh1/cel7A, PDE_07929/eg1/cel7B), three members of the GH12 family (PDE_02886/cel12B, PDE_06439/cel12A, PDE_09041), and one member of the GH45 family (PDE_07928/eg5/cel45A). Of the 13 genes, 12 are upregulated in the Δfbx23 mutant (Fig 6E red bars), except for only one PDE_05445 downregulated (Fig 6E, green bar).

The transcription levels of genes encoding plant cell wall-degrading enzymes were also examined. The genome of P. oxalicum is predicted to contain 115 genes that encode plant cell wall-degrading enzymes, which comprise most of the extracellular proteins secreted by the cells under the cellulose signal [32]. The clustering analysis was performed according to the transcription data of 115 genes, identifying 4 co-regulated gene clusters, I/II/III/IV (Fig 6F). The transcription levels of genes in clusters I and III were upregulated, whereas genes in clusters II and IV were downregulated in the Δfbx23 mutant. Nearly all of the genes encoding the key exocellulases, endocellulases, hemicellulases, β-glucosidase and lytic polysaccharide monooxygenases (LPMO) that assist the breakdown of plant cell wall are found in Clusters I and III. Cluster I include the key cellulase genes such as cellobiohydrolase genes cbh1 and cbh2, and the endo-β-1,4-glucanase genes eg1 and eg2 (Fig 6F cluster I, red fonts), the key hemicellulase genes such as xylanase genes xyn10A (PDE_08094) and xyn11A (PDE_02101) as well as the xylosidase gene xyl3A (PDE00049) and the mannosidase gene man5A (PDE06023) (Fig 5C cluster I, green fonts). Cluster III also includes cellulase genes such as endo-β-1,4-glucanase genes cel5C and cel5D, β-glucosidase gene bgl1 (Fig 6C cluster III, red fonts), some hemicellulase genes (Fig 6C cluster III, green fonts), and LPMO gene (PDE_05633) and swollenin gene (PDE_02102) which assist the breakdown of plant cell wall (Fig 6C cluster III, blue fonts) (Fig 6F).

The results of RT-qPCR and transcriptome support a negative correlation between the regulatory function of PoFbx23 protein and the transcription of (hemi)cellulase genes. Since (hemi)cellulase synthesis is usually controlled at the transcriptional level, increased transcription of the (hemi)cellulase gene in the Δfbx23 mutant may result in a rise in (hemi)cellulase synthesis. Fungal (hemi)cellulases are secreted extracellularly. The filter paper activity (FPA) indicates the synthesis of total cellulase; xylanase activity indicates the synthesis of main hemicellulase, and carboxymethyl cellulose activity (CMCase) indicates the synthesis of endocellulase (endo-β-1,4-glucanase). These enzymatic activities were determined for WT, Δfbx23, and Refbx23 incubated in VMMC liquid for 2 to 6 days. There was a significant increase in FPA, xylanase activity, and CMCase activity in the Δfbx23 compared to the WT. For example, on the fourth day of cultivation, the FPA, xylanase activity, and CMCase activity of Δfbx23 increased by 39.2%, 64.8%, and 16.9% compared to WT, respectively (S3G~I Fig). Cellulases and hemicellulases account for a large proportion of the extracellular proteins of P. oxalicum [32]. An assay of the extracellular protein concentration the Δfbx23’s extracellular proteins showed a significantly higher protein content than the WT (S3F Fig). This result also supports the higher ability of the Δfbx23 to synthesize cellulase and hemicellulase. In addition, the Δace1 mutant strain showed increased levels of protein secretion and (hemi)cellulase production (S3F~I Fig). The results suggested a correlation between PoFbx23 and PoAce1 in regulating (hemi)cellulase production.

Accumulated PoAce1 under cellulose signal no longer binds the (hemi)cellulase genes

The Pofbx23 deletion strain showed accumulation of PoAce1 protein under cellulose signal but increased transcription of the (hemi)cellulase gene. This finding contradicts earlier research, which indicates that Ace1 is a negative regulator of the transcription of the (hemi)cellulase gene in P. oxalicum and other cellulase-producing fungi [26–28]. The repressing activity of PoAce1 seems lost due to Pofbx23 deletion. Then, ChIP-qPCR was used to detect the degree of enrichment of PoAce1 protein in the promoter of the key cellulase gene cbh1 and eg1, and the key xylanase (hemicellulase) gene xyn11A in the WT, Δfbx23, Ace1-TAP, and TAP-Δfbx23 strains grown in VMMC. The schemes and results of ChIP-qPCR are shown in Fig 7 and S3 Data. Five typical regions (Regions 1–5) that cover upstream sequences and CDS were studied for each gene.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. PoAce1 enrichment in the specific regions of (hemi)cellulase genes assayed by ChIP-qPCR.

The strains were grown in VMMC. (A) Cellulase (cellobiohydrolase) gene cbh1 (B) Cellulase (endo-β-1,4-glucanase) gene eg1 (C) Hemicellulase (xylanase) gene xyn11A. Top subgraphs, Strategies of ChIP-qPCR for each gene. The TSS is designated as +1. The initiator (Inr) and TATA box are illustrated. For genes cbh1 and eg1, five specific regions (R1 to R5) of each gene were designed for qPCR assay. Region 1 (R1), region 2 (R2) and region 3 (R3) are located upstream of the TSS. Region 4 (R4) covers the TATA-box and the Inr. Region 5 (R5) is located in the 5′ region of the CDS of each gene. For the xyn11A gene, R1, R2, and R3 are located upstream of the TSS. R2 and R3 overlap for the appropriate design of primers. No appropriate primers for qPCR could be designed close to the Inr region. R4 is located in the 5′ region of the CDS of xyn11A. R6 is located in the middle of the CDS of xyn11A. Purple triangles indicate the putative DNA-binding sites of PoAce1. The orientation of the triangle represents the orientation of the binding motif. Bottom subgraphs, the results of ChIP-qPCR for each gene. The relative enrichment of IP DNA was calculated by % of input. All values are means from measurements in biological triplicates. The error bars indicate standard deviations, *p < 0.05, **p < 0.01, ***p < 0.001.

https://doi.org/10.1371/journal.pgen.1011539.g007

For the ChIP-qPCR design of the cbh1 gene, regions 1, 2, and 3 (R1/2/3) are located at approximately 600 bp upstream (−) of the TSS. Region 4 (R4) covers the core promoter, including the TATA box and initiator element (Inr). Region 5 (R5) is located proximately at the 5’ region of CDS. ChIP-qPCR results demonstrated PoAce1 binding in Regions 2/3/4 and 5’-CDS in the Ace1-TAP strain. However, despite the PoAce1 protein accumulating in the TAP-Δfbx23 strain, PoAce1 binding of the target gene was defective; the levels of PoAce1 binding DNA at various regions (R2-5) of the cbh1 gene reduced remarkably in the TAP-Δfbx23 compared with those in the Ace1-TAP strain (Fig 7A).

The ChIP-qPCR design of the eg1 gene is similar to that of the cbh1 gene. ChIP-qPCR results demonstrated that the levels of PoAce1 binding DNA in Regions 2/3/4 reduced remarkably in the TAP-Δfbx23 strain compared with those in the Ace1-TAP strain. However, there was no significant difference between the levels of PoAce1 binding to R1 and R5, despite two and one PoAce1 binding sites in R1 and R5, respectively (Fig 7B).

For the ChIP-qPCR design of the xyn11A gene, R1 and R2 are located at approximately 500 bp upstream of the TSS. R3 covers the core promoter, including the TATA box. R4 and R5 are located in the 5’ region and middle of the CDS region, respectively. ChIP-qPCR results demonstrated that the levels of PoAce1 binding DNA in regions 3/4/5 reduced remarkably in the TAP-Δfbx23 compared with those in the Ace1-TAP strain. However, there was no significant difference between the levels of PoAce1 binding to R1 and R2 (Fig 7C). We noticed that the “percent of input” of PoAce1 enrichment at cbh1, eg1, and xyn11A have the highest values of 3.54, 6.39, and 3.74, respectively, while the “percent of input” of PoAce1 enrichment at brlA and abrB gene is relatively low, with a maximum of 0.75 and 0.92, respectively. The results suggest that the degree of PoAce1 enrichment in (hemi)cellulase genes is much higher than that of brlA and abrB genes, indicating PoAce1’s primary direct regulatory role in regulating (hemi)cellulase gene transcription.

ChIP-qPCR results showed that the PoFbx23 protein is essential for PoAce1’s binding DNA activity under cellulose signal. The PoAce1 protein binds to the regions adjacent to the core promoter of the target gene within the normal cells; the active form of PoAce1 works as a repressor to inhibit the transcription of the (hemi)cellulase gene. In the absence of PoFbx23, despite PoAce1 accumulating, it no longer binds to the target genes and takes on an inactive form; the transcription of the (hemi)cellulase gene is derepressed due to the dysfunctional PoAce1.

PoAce1 is a ubiquitinated protein and takes the UPS pathway

The absence of PoFbx23 resulted in the accumulation of PoAce1. It is necessary to detect whether PoAce1 takes the UPS pathway.

Firstly, we detected whether PoAce1 is a ubiquitinated protein. As substrate proteins are linked to ubiquitin using seven distinct ubiquitin lysine residues (Lys6, Lys11, Lys27, Lys29, Lys33, Lys48, and Lys63), we used anti-K6-, K11-, K27-, K29-, K33-, K48-, and K-63 linkage specific polyubiquitin antibodies to detect the ubiquitination level of PoAce1 (Fig 8). The result showed that PoAce1 has K6-, K11-, K29-, and K48-linked polyubiquitin modification (Fig 8A, 8B, 8D & 8F, black arrows), but without K27-, K33-, and K63-polyubiquitin (Fig. 8C, 8E & 8G). All of the polyubiquitin signals are present where the intact PoAce1 protein is (at ~100 kDa), while the degraded versions of PoAce1 do not exhibit any polyubiquitin signals. Then, we assayed the Ace1-TAP and TAP-Δfbx23 strains treated with or without proteasome inhibitor MG132. SDS-PAGE showed that the silver-stained bands of PoAce1 were thickened after TAP-Δfbx23 was treated with MG132 (Fig 8H, left, black arrow), corresponding to a stronger signal in the Western Blot (Fig 8H, right, red arrow). It was unexpected that the PoAce1 signal at ~100 kDa was attenuated in the MG132-treated Ace1-TAP sample. However, degraded versions (~ 50kDa) of PoAce1 accumulate in both Ace1-TAP and TAP-Δfbx23 samples treated with MG132 (Fig 8I, right, green arrow). PoAce1 accumulation was promoted by MG132 treatment, but this effect was weak and significantly less than that caused by PoFbx23 absence. It should be noted that PoAce1 maintains ubiquitin modifications and accumulates with degraded versions even in the absence of PoFbx23. The result gives a hint that PoAce1 may have other UPS besides the SCF^Fbx23. Indeed, the results of Ace1-TAP show the presence of the protein PoMub1 (Table 2, ranked 6th), a component of the Mub1/Ubr2 ubiquitin ligase complex.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Ubiquitination of PoAce1 and effect of proteasome inhibitor MG132 on PoAce1.

(A~G) The ubiquitination modification of PoAce1 in WT, Ace1-TAP, and TAP-Δfbx23, detected using anti-K6-, K11-, K27-, K29-, K33-, K48-, and K-63 linkage specific polyubiquitin antibodies, respectively. The dark arrows represent the protein PoAce1 (about 100 kDa on the gel, theoretical MW: 88.4 kDa). (H) The results of TAP using PoAce1 as bait in the Ace1-TAP and TAP-Δfbx23 strains treated with and without MG132. Left, silver staining after SDS-PAGE of TAP eluate. The dark arrow represents the protein PoAce1 (about 100 kDa on the gel, theoretical MW: 88.4 kDa). Right, the Western blot of TAP eluate using an anti-HA antibody.

https://doi.org/10.1371/journal.pgen.1011539.g008

PoFbx23 mostly takes the form of free protein within cells

The PoFbx23 protein possesses seven copies of WD repeats S1 Fig. As WD repeats are known as a site for protein-protein or protein-DNA interaction and proteins containing WD repeats are known to serve as platforms for the assembly of multiple proteins [42], it is considered that PoFbx23 may recognize other substrate proteins in addition to PoAce1 as its substrate. We constructed the Fbx23-TAP strain to validate the association between PoFbx23 and PoAce1 further and identify additional potential substrates for the PoFbx23. The WT strain containing native PoFbx23 was used as a control. The eluate of the Fbx23-TAP strain was obtained by two-step tandem purification. Then, the eluate was separated into three sections: one for SDS-PAGE, one for Western blot, and one for MS-MS assay to identify proteins with interactions with protein PoFbx23.

There is no specific band between the control and Fbx23-TAP on SDS-PAGE (Fig 9A, left). However, the Western blot result revealed a signal at about 105 kDa (Fig 9A, right, dark arrow, PoFbx23’s theoretical molecular weight with the TAP tag added is 97.8 kDa). The proteins in the third section of the eluent were analyzed using LC-MS/MS to identify the bait and interacting proteins. The PepCount identified through TAP-MS is listed in S4 Spreadsheet (Sheet 2 & Sheet 3) and S4 Data. The proteins observed in two Fbx23-TAP samples but not in the controls are listed in S4 Spreadsheet (Sheet 1). The proteins were ranked by emPAI [41]. 110 and 135 proteins were identified in two samples of Fbx23-TAP, respectively. 45 proteins were observed in two Fbx23-TAP samples but not the controls (Fig 9B). They are considered reliable proteins with putative interaction with the PoAce1.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. The results of TAP-MS using PoFbx23 as bait.

(A) Left, silver staining after SDS-PAGE of Fbx23-TAP eluate. Right, Western blot of Fbx23-TAP eluate using anti-HA antibody. The arrow represents the protein PoFbx23 (approximately 105 kDa on the gel, theoretical MW: 97.8 kDa). (B) Intersection proteins between two Fbx23-TAP biological samples (The detailed information about proteins is shown in S4 Spreadsheet). (C) Transcription factors interacting with PoFbx23 identified through TAP-MS. The bait PoAce1 has the highest emPAI (S4 Spreadsheet, sheet 1, row 3, green background). Three Transcription factors, PDE_04990, PDE _04678, and PDE _09448, are racked 28th (row 30), 31st (row 33), and 35th (row 37) in S4 Spreadsheet, sheet 1 (blue background), respectively.

https://doi.org/10.1371/journal.pgen.1011539.g009

As expected, the bait PoFbx23 has the highest emPAI (S4 Spreadsheet, Sheet 1, green background). Interestingly, the TAP results did not reveal the presence of PoSkp1, an adaptor that bridges the core SCF complex with the PoFbx23. This absence could be attributed to multiple FBPs existing in the cell, such as the over 50 FBPs in the genome of P. oxalicum and over 70 FBPs in A. nidulans [15]. These FBPs compete with Fbx23 for binding to SKP1. In the research about A. nidulans AnFbx23, AnSkp1 (AN2302) was pulled down by AnFbx23 only under the condition of H₂O₂ oxidative stress but not detected under the condition of normal or NaCl osmotic stress [15]. This finding suggests that most Fbx23 proteins are free forms within the cell, as there is competition among the numerous F-box proteins in cells for binding to Rbx1-Cul1-Skp1 scaffolds [23]. In other words, only a small portion of PoFbx23 is assembled into the core SCF to form an SCF^Fbx23 complex that recognizes substrate PoAce1, while the majority of PoFbx23 does not bind to Skp1 and exists in the free form of SCF-independent.

PoAce1 protein was not found in the MS results. This result is not wildly unexpected. Many previous studies have observed this phenomenon: when a transcription factor is used as bait, a multisubunit protein complex, such as a cofactor that interacts with that transcription factor, can be identified in the TAP results. Conversely, the transcription factor is not always found when a specific protein complex subunit is used as the bait protein [36,57]. The low abundance of transcription factors is one possible explanation for this phenomenon. Another possibility is that when a specific subunit is used as bait, a variety of other proteins compete with the transcription factor to bind the subunit because the subunit often contains multiple domains for protein-protein interaction. However, in the research about A. nidulans AnFbx23, 23 proteins were pulled down as potential interacting proteins with AnFbx23, including AnAce1 [15]. The finding also indirectly supported our result that the PoAce1 is the direct substrate of the SCF^Fbx23 complex.

Among the 44 potential interacting proteins, three transcriptional regulators were identified. All three transcription factors PDE_04990 (UniProt: S7ZH90), PDE_04678 (UniProt: S8AUB2), and PDE_09448 (UniProt: S7ZVN6) are Zn2Cys6 class fungal-specific transcription factors (S4 Spreadsheet, Sheet1, row 30, 33 and 37) (Fig 9C). Particularly, PDE_04678 was also found in one sample of PoAce1-TAP (S1 Spreadsheet, Sheet 1, row 480), suggesting a very high probability that this transcription factor serves as another actual substrate for PoFbx23 in addition to PoAce1. Interestingly, another E3 ubiquitin-protein ligase Bre1 homolog, (PDE_03581 (UniProt: S7ZDA7), was observed in the Fbx23-TAP result (Fig 9C).

Fungal strains and culture conditions

The wild-type (WT) P. oxalicum 114–2 (CGMCC 5302) was used as the original parent strain. The WT strain and mutants constructed in this study were grown on agar with 10% wheat bran juice at 30°C for 5 days for conidiation. For mycelial growth, the strains were grown at 30°C in 1 × Vogel’s minimal medium (VMM) (50 × Vogel’s salt: 125.0 g, Na₃Citrate‧2H₂O, 250.0 g KH₂PO₄, 100.0 g NH₄NO₃, 10.0 g, MgSO₄‧7H₂O, 5.0 g CaCl₂‧2H₂O, 0.25 mg biotin, 0.25 g citric acid, 0.25 g ZnSO₄‧7H₂O, 0.05 g, Fe(NH₄)₂(SO₄)₂‧6H₂O, 12.5 mg CuSO₄‧5H₂O, 2.5 mg MnSO₄‧H₂O, 2.5 mg H₃BO₃, 2.5 mg Na₂MoO₄‧2H₂O, and 1 L of water) [50], plus 2% glucose (w/v) (VMMG) or 2% ball-milled cellulose (VMMC) as a sole carbon source. 1.5% agar was added to the VMMG or VMMC for solid culture.

Construction of different mutants in P. oxalicum

All primers used in this study are listed in S2 Table.

The construction strategy and verification for the Ace1-TAP strain, which has PoAce1 fused with a FLAG-HA tag at the C-terminus, are shown in S4A~C Fig. For constructing the Ace1-TAP strain, primer pairs Ace1-TAP-UF/Ace1-TAP-UR and Ace1-TAP-DF/Ace1-TAP-DR were used to amplify the 5’- upstream and 3’-downstream homologous arms of the Poace1 gene using P. oxalicum WT genomic DNA as the template. Primer pairs hph-F/hph-R were used to amplify the marker gene hygromycin B (hph) from the plasmid Psilent1 [73]. Then, the three PCR fragments (5’—upstream and 3’ downstream regions of the PoAce1 gene and hph) were fused with fusion PCR. The PCR product was then amplified using nested primers Ace1-TAP-CSF/Ace1-TAP-CSR and was transformed into the WT to obtain the Ace1-TAP mutant. The Fbx23-TAP strain was constructed using the same strategy as the Ace1-TAP strain. Briefly, primer pairs Fbx23-TAP-UF/Ace1-TAP-UR and Fbx23-TAP-DF/Ace1-TAP-DR were used to amplify the 5’- upstream and 3’- downstream homologous arms of the Pofbx23 gene. 5’—upstream and 3’ downstream regions of the Pofbx23 gene and hph were fused and amplified using nested primers Fbx23-TAP-CSF/Fbx23-TAP-CSR. The fused product was transformed into the WT to obtain the Fbx23-TAP mutant. The proper fusion of protein PoAce1 or PoFbx23 with the FLAG-HA tag was verified by diagnostic PCR and DNA sequencing.

The construction strategies and verification for the Δfbx23 mutant (the gene Pofbx23 was deleted) and Refbx23 mutant (re-complement of the Pofbx23 gene in the Δfbx23 mutant) are shown in S4D~G Fig. For constructing Δfbx23 strain, primer pairs Δfbx23-UF/Δfbx23-UR and Δfbx23-DF/Δfbx23-DR were used to amplify the 5’- upstream and 3’- downstream homologous arms of the Pofbx23 gene using P. oxalicum WT genomic DNA as the template. Primer pairs hph-F/hph-R were used to amplify the marker gene hph from the plasmid Psilent1. Then, the three PCR fragments (5’—and 3’ -flanking regions of the Poace1 gene and hph) were fused with fusion PCR. The fused product was then amplified using nested primers Δfbx23-CSF/Δfbx23-CSR and was transformed into the WT to obtain the Δfbx23 mutant. For constructing the Refbx23 strain, primer pairs REfbx23-UF/Refbx23-UR were used to amplify the fragment from 5’- upstream of the Pofbx23 gene to Pofbx23 gene terminator using P. oxalicum WT genomic DNA as the template. Primer pairs ptrA-F/ptrA-R were used to amplify the marker gene pyrithiamine hydrobromide (ptrA) from plasmid pME2892 [74]. Then, the two PCR fragments were fused using primers Refbx23-UF/ptrA-R. The fused product was transformed into the Δfbx23 mutant to obtain the Refbx23 mutant. The mutants Δfbx23 and Refbx23 were verified by diagnostic PCR and Southern blot.

The construction strategies for the Fbx23-GFP mutant (the protein PoFbx23 fused with a green fluorescent protein (GFP)) and Ace1-GFP mutant (the protein PoAce1 fused with GFP) are shown in S5A&B Fig. For constructing the Fbx23-GFP strain, primer pairs fbx23-GFP-UF/fbx23-GFP-UR and fbx23-GFP-DF/fbx23-GFP-DR were used to amplify the 5’- upstream and 3’- downstream homologous arms of the Pofbx23 gene using P. oxalicum WT genomic DNA as the template. Primer pairs gfp-F/gfp-R were used to amplify the gfp gene from the plasmid pEGFP. Primer pairs hph-F/hph-R were used to amplify the marker gene hph from the plasmid Psilent1. Then, the four PCR fragments (5’- upstream and 3’- downstream of the Pofbx23 gene, gfp, and hph) were fused with fusion PCR. The PCR product was then amplified using nested primers fbx23-GFP-CSF/fbx23-GFP-CSR and was transformed into the WT to obtain the Fbx23-GFP strain. Similar to the construction of the Fbx23-GFP strain, primer pairs Ace1-TAP-UF/Ace1-GFP-UR and Ace1-TAP-DF/Ace1-TAP-DR were used to amplify the 5’- upstream and 3’- downstream homologous arms of the PoAce1 gene. Then, the four PCR fragments (5’- upstream and 3’- downstream of the PoAce1 gene, gfp, and hph) were fused with fusion PCR. The PCR product was then amplified using nested primers Ace1-TAP-CSF/Ace1-TAP-CSR and was transformed into the WT to obtain the Ace1-GFP strain. For constructing the PoAce1 protein fused the GFP in the Δfbx23 background, primer pairs Δfbx23(Ace1-GFP)-UF/Δfbx23(Ace1-GFP)-UR were used to amplify the fragment from 5’- upstream to terminator of the Pofbx23 gene using P. oxalicum WT genomic DNA as the template. Primer pairs Δfbx23-UF/Δfbx23(Ace1-GFP)-UR and Δfbx23(Ace1-GFP)-DF/Δfbx23-DR were used to amplify the 5’-upstream and 3’-downstream homologous arms of the Pofbx23 gene. Primer pairs ptrA-F/ptrA-R were used to amplify the marker gene ptrA from plasmid pME2892 [74]. Then, the three PCR fragments (5’- and 3’-flanking regions of the Pofbx23 gene and ptrA) were fused with fusion PCR. The fused product was then amplified using nested primers Δfbx23-CSF/Δfbx23-CSR and was transformed into the Ace1-GFP to obtain the ΔF-Ace1-GFP mutant. The proper fusion of protein PoFbx23 or PoAce1 with GFP was verified by diagnostic PCR and DNA sequencing.

The construction strategy for the Ace1-YFP-Fbx23 mutant (the strain for bimolecular fluorescence complementation (BiFC)) is shown in S5C&D Fig. BiFC strains were constructed through a previously described method [36]. The plasmids pMD18-T-NYFP carrying the coding sequence of the N-terminus (1–155 aa) of yellow fluorescent protein (YFP) and pUC19-CYFP carrying the coding sequence of the C-terminus (156–238 aa) of YFP were used as the expression vectors for BiFC. The Poace1 gene was amplified using primers Ace1-NF/Ace1-NR. The PCR product was inserted into the multiple cloning sites (MCS) of plasmid pMD18-T-NYFP to construct the recombinant vector pMD18-T-NYFP-Ace1, which contains protein PoAce1 fused with the N-terminus of YFP (S5C Fig, left). The Pofbx23 gene was amplified using primers fbx23-NF/fbx23-NR. The PCR product was inserted into the MCS of plasmid pUC19-CYFP to construct the recombinant vector pUC19-CYFP-Fbx23, which contains protein PoFbx23 fused with the C-terminus of YFP (S5C Fig, right). The recombinant vectors pMD18-T-NYFP-Ace1 and pUC19-CYFP-Fbx23 were linearized and simultaneously transformed into the WT to obtain the Ace1-YFP-Fbx23 mutant. The proper fusion of protein PoAce1 with the N-terminus of YFP and protein Pofbx23 with the C-terminus of YFP was verified by diagnostic PCR and DNA sequencing.

Phenotypic analysis of P. oxalicum strains

For observation of colony morphology, 1 μL of fresh spore suspension (10⁷ conidia ml^-1) was point-inoculated onto VMMG agar and then cultivated at 30°C. The colony diameters were measured every day for 9 days. To count the spore yield of different mutants, we coated 150 μL of fresh spore suspension (10⁷ conidia ml^-1) evenly onto VMMG agar. After cultivated at 30°C for 5 days, 10-mm diameter agars were taken from each plate, and spores were washed with 2 mL of physiological saline containing 0.2% tween 80 by vortexing. Then, the spores were counted using a hemocytometer. The biomass was determined by the dry weight method described previously [75].

Phylogenetic analysis and domain architecture analysis

The amino acid sequences of Fbx23 homologous proteins from different species were obtained from the UniProt database (https://www.uniprot.org/). The Clustal X and MEGA 7.0 software were used to construct multiple sequence alignments and construct a phylogenetic tree with the neighbor-joining method. The SMART database (https://smart.embl.de/) was used for domain analysis of proteins. The domain architecture patterns were constructed in proportion to the corresponding protein sequences.

Tandem affinity purification

Fresh conidial suspensions of the P. oxalicum WT and the TAP strains were inoculated in 2 L of VMMG liquid at 30°C and 180 rpm for 48 h. The mycelia were filtered through miracloth and washed twice with 0.96% NaCl (w/v) containing 1% DMSO and 1 mM PMSF. Afterward, the mycelia were ground in liquid nitrogen. We used about 40 g of ground mycelia per tandem affinity purification and mass spectrometry (TAP-MS) system. Each treatment for TAP-MS was performed in three biological replicates. The TAP-MS procedure was performed as previously described [36]. Briefly, Protein lysis buffer (0.9 g NaCl, 1 M Tris-HCl, pH 7.5, 10 mL glycerol, 0.1 mL NP40, and 0.05 mL protease inhibitor per 100 mL) was used to extract the total protein. Ezview TM Red ANTI-FLAG M2 Affinity Gel (SIGMA, USA) was used for the first affinity purification step. 500 μL 3 × FLAG peptide (150 ng/μL) was used to compete with the ANTI-FLAG M2 affinity resin to obtain the first protein eluate. ANTI-HA resin (SIGMA, USA) was used for the second affinity purification step. Finally, 80 μL of 8 M urea was used to elute ANTI-HA resin to obtain the final protein eluate. The final eluate was separated into three sections following two affinity purification steps. One part of the eluate was assayed with 12.5% SDS-PAGE, followed by silver staining. One part of the eluate was assayed with Western blot using rabbit anti-HA as an antibody (ABclonal, China). The other part of the eluate was assayed through LC-MS/MS for protein identification (APT, Shanghai, China). Rabbit anti-K6-, K11-, K27-, K29-, K33-, K48-, and K-63 linkage specific polyubiquitin antibodies (ABclonal, China) were used to detect the ubiquitin modification of the protein.

Proteasome inhibitor MG132 treatment

Strain culture and mycelia collection are performed as described in the “Tandem affinity purification” section above. Then, the mycelia are vacuum-infiltrated with 50 μM MG132 (dissolved in DMSO) or DMSO alone for ten minutes and immersed in the same solution for two hours. The target protein was detected as described in the “Tandem affinity purification” section above.

Microscopy of GFP and BiFC strains

Hyphae of the GFP strains Ace1-GFP and Fbx23-GFP and the BiFC strain Ace1-YFP-Fbx23 were observed under a high sensitivity laser scanning confocal microscope (ZEISS LSM900) (Carl Zeiss, Germany). The nuclei were stained in the dark for 15 minutes by Hoechst 33342 (Sigma, United States). The blue nuclei stained with Hoechst 33342 were observed with excitation light at 405 nm. The green fluorescence of the Ace1-GFP and Fbx23-GFP strains was observed by excitation light at 488 nm. The yellow fluorescence of the Ace1-YFP-Fbx23 strain was observed by excitation light at 488 nm.

Yeast two-hybrid assay

For yeast two-hybrid assay (Y2H) analysis, the coding sequence (CDS) of two truncated PoAce1, Ace1- (1 to 396 aa) and Ace1- (484 to 816 aa) which avoid DNA binding domain of PoAce1, were amplified by primers A396-ADF/A396-ADR and A816-ADF/A816-ADR, respectively. Then, two CDS were cloned into the plasmid pGADT7 and transformed into S. cerevisiae Y187. The CDS of PoFbx23 were amplified by primers Fbx23-BDF/Fbx23-BDR, cloned into the plasmid pGBKT7, and transformed into S. cerevisiae Y2H Gold. Y2H strains were grown on QDO (quadruple-dropout, SD-Ade/-His/-Leu/-Trp) and QDO/x-α-gal/Aba (QDO supplemented with X-α-gal and aureobasidin A) agar to test possible interactions.

Transcriptome analysis and GO analysis

Fresh spores (10⁷ conidia ml^-1) were precultured in VMMG liquid at 30°C for 24 h. Equal mycelia was collected through vacuum filtration and was then transferred (0.3 g mycelia/50 mL medium) to VMMG or VMMC. Fresh mycelia of the WT and the Δfbx23 after 24-h of cultivation (30°C, 200 rpm) in VMMG or VMMC were harvested and fully ground in liquid nitrogen. Total RNA was extracted with RNAiso Plus reagent (Takara, Japan) and was incubated with 10 U DNase I (Takara, Japan) at 37°C for 30 min to remove the genomic DNA. Transcriptome analysis based on BGISEQ-500 RNA-Seq was performed by the Beijing Genomics Institute (BGI, Wuhan, China). Saturation analysis of each sample indicated their availability for omics analysis (6S Fig). Sequenced reads were mapped against the reference genome of P. oxalicum. The filtered clean reads for each gene were normalized to fragments per kilobase transcriptome per million mapped reads (FPKM) for differential expression analysis (S5 Data). Significantly different expression between samples were identified through a significance test with combined thresholds (fold change ≥ 2, Q-Value < 0.05). OmicsBox was used for GO annotation and function enrichment analysis with a threshold at FDR ≤ 0.05 [76]. Genesis software was used for Cluster analysis [56]. The raw transcriptome profiling data have been deposited in NCBI’s Gene Expression Omnibus database under the accession number GSE228483.

Real-time quantitative PCR

Strains were cultivated as previously described in the “Transcriptome Analysis and GO analysis” section. Fresh mycelia after 4-, 24-h of cultivation (30°C, 200 rpm) in VMMG or VMMC were harvested and fully ground in liquid nitrogen. Total RNA of different strains was extracted using RNAiso Plus reagent (Takara, Japan). Genomic DNA removal and cDNA synthesis were performed using RT Reagent Kit With gDNA Eraser (Takara, Japan). The obtained cDNA was used as the template for quantitative PCR. The primers used for RT-qPCR are listed in S2 Table. For each gene, biological triplicates were conducted. Gene expression copy numbers were calculated using the standard curves constructed for each gene, and the data were then normalized with the expression levels of the actin gene.

Cellulase and hemicellulase activity assay

Fresh spores (10⁷ conidia ml^-1) were precultured in VMMG liquid at 30°C for 24 h. Equal mycelia was collected through vacuum filtration and transferred to VMMC liquid (0.3 g mycelia/50 mL) at 30°C. The culture supernatants from different time points were collected via centrifugation and diluted with 0.2 M NaAc buffer solution (pH 4.8). The activities of filter paper activity (FPA), endo-β-1,4-glucanase (indicated by CMCase activity), and xylanase were assayed following the methods described by Li et al. [27]. Whatman No. 1 filter paper, carboxymethylcellulose sodium salt (CMC-Na) (Sigma-Aldrich, St. Louis, USA), and birch xylan (Ryon, Shanghai, China) were used as substrates. Briefly, 0.5 mL of suitably diluted culture supernatants and 50 ± 1 mg Whatman No. 1 filter paper, or 1.5 mL of 1% (m/v %) CMC-Na or 1% (m/v %) xylan, which was dissolved in NaAc buffer (pH 4.8), were added into 25-mL colorimetric tubes. The mixture was gently stirred and incubated in a 50°C water bath for 30 min. 3 mL 3,5-dinitro salicylic acid (DNS) reagent (10 g 3, 5-dinitro salicylic acid, 20 g sodium hydroxide, 200 g sodium potassium tartrate, 2.0 g redistilled phenol, and 0.5 g sodium sulfite anhydrous per 1000 mL DNS reagent) was then added to stop enzymatic reaction. The tubes were placed in boiling water for 10 min. Finally, 20 mL of distilled water was added and mixed, and the absorbance of the reaction mixture was measured at 540 nm to determine enzymatic activities. One enzyme activity unit was defined as the amount of enzyme required for producing 1 μmol glucose or 1 μmol xylose per minute under the assayed conditions.

Chromatin immunoprecipitation-quantitative PCR

Strains were cultivated as previously described in the section “Transcriptome Analysis and GO analysis” section. ChIP assays were performed by a previously described protocol [43]. Briefly, the hyphae cultivated in VMMG or VMMC were fixed by adding 1% formaldehyde at 30°C for 10 min with shaking. Then, the glycine was added to obtain a final concentration of 125 mM. The mycelia were filtered by a vacuum pump filter, ground in liquid nitrogen, and mixed in the ChIP-lysis buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.5% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 1 mM PMSF (phenylmethanesulfonyl fluoride), 0.1% protease inhibitor cocktail). Then, the chromatin was sheared into 200~500 bp fragments through sonication. Immunoprecipitation (IP) was performed with an anti-HA antibody (Proteintech, USA) with equal amounts of extracted chromatin (1 mg). The obtained IP products and 0.1 mg input chromatin DNA (without IP) of each sample were subjected to RNase digestion to remove RNA, as well as high heat and proteinase K digestion to reverse crosslinks, followed by phenol-chloroform extraction and ethanol precipitation to get purified IP DNA and input DNA. Then, quantitative PCR was performed. The relative enrichment of IP DNA was defined as the ChIP efficiency and calculated by the input% method as follows: ChIP efficiency = 2−^ΔCt × 100%, ΔCt = Ct_IP − (Ct_Input − log₂10), where Ct = threshold cycle of PCR. Three biological replicates were established for each sample. Primers used in the RT-qPCR are listed in S2 Table.