The biological relevance of the FspTF transcription factor, homologous of Bqt4, in Fusarium sp. associated with the ambrosia beetle Xylosandrus morigerus

Transcription factors in phytopathogenic fungi are key players due to their gene expression regulation leading to fungal growth and pathogenicity. The KilA-N family encompasses transcription factors unique to fungi, and the Bqt4 subfamily is included in it and is poorly understood in filamentous fungi. In this study, we evaluated the role in growth and pathogenesis of the homologous of Bqt4, FspTF, in Fusarium sp. isolated from the ambrosia beetle Xylosandrus morigerus through the characterization of a CRISPR/Cas9 edited strain in Fsptf. The phenotypic analysis revealed that TF65-6, the edited strain, modified its mycelia growth and conidia production, exhibited affectation in mycelia and culture pigmentation, and in the response to certain stress conditions. In addition, the plant infection process was compromised. Untargeted metabolomic and transcriptomic analysis, clearly showed that FspTF may regulate secondary metabolism, transmembrane transport, virulence, and diverse metabolic pathways such as lipid metabolism, and signal transduction. These data highlight for the first time the biological relevance of an orthologue of Bqt4 in Fusarium sp. associated with an ambrosia beetle.


Introduction
Ambrosia beetles are members of the wood-decaying ecosystem, are xylem-borers of dead and stressed trees that perform fungus farming as an ecological strategy, and carry spores of symbiotic fungi in specialized structures of their exoskeletons (the mycetangium). Once the insect is established in the plant host, the fungi colonize the drilled galleries and serve as the only source of food for both the adult beetle and the developing larvae (Six, 2012;Kirkendall et al., 2015;Huang et al., 2020;Dzurenko and Hulcr, 2022). The beetle-fungi associations are variable from the highly specific nutritional symbiosis to less specific association with other fungi acquired from the environment but which are established on the abdomen or external surfaces. Thus, fungal mycobiome is considered complex and includes phytopathogens (Bateman et al., 2016;Miller et al., 2019;Rassati et al., 2019;Hulcr et al., 2020;Morales-Rodríguez et al., 2021).
On the other hand, fungi perform transcriptional reprogramming across their life cycle (Riquelme et al., 2018;Nowrousian, 2022) and during their interaction with plants ( Tan and Oliver, 2017;John et al., 2021), and it is already known that transcription factors (TFs) play a crucial role. In Fusarium spp., the TFs Sge1 (SIX gene expression 1), FTF1, and FTF2 are involved in the regulation of biological processes, such as hyphal differentiation to reproductive structures and the expression of effectors (Jonkers et al., 2012;van der Does et al., 2016;Rocafort et al., 2020;Zhao et al., 2020;Zuriegat et al., 2021). In other phytopathogenic fungi, TFs members of diverse families such as the Zinc (Zn) finger-binding proteins family, bZIP family, Ste12 family, Velvet family, Gti1/Pac2 family, Fork head family, and Asm1, Phd1, Sok2, Efg1, StuA (APSES) family have been characterized, finding roles in the regulation of development and pathogenesis (Tan and Oliver, 2017;John et al., 2021). Specifically, the APSES subfamily belongs to a bigger fungal TF family called the KilA-N.
The KilA-N family comprises Swi4/6, Mbp1, Res1/2, Cdc10 (SMRC), Xbp1, APSES, and Bqt4 TF subfamilies sharing the ancestral KilA-N DNA-binding domain (Medina et al., 2019). The KilA-N domain is considered homologous to the fungal DNA-binding APSES domain. It is reported to adopt the topology of a winged helix-turnhelix (wHTH) domain (Liu et al., 2015;Hu et al., 2019a,b). The KilA-N family members have diverse functions; the APSES subfamily is the most studied, and they are considered key regulators of fungal development, secondary metabolism, and virulence Medina et al., 2019). The SMRC subfamily regulates cell cycle entry and progression (Medina et al., 2016(Medina et al., , 2019, whereas the Xbp1 subfamily is required for morphogenesis, control of cell cycle, and metabolism (Medina et al., 2019;Xin et al., 2020;Zhang et al., 2021). The Bqt4 subfamily is the youngest member of the KilA-N family; few studies are focused on this subfamily, mainly in Schizosaccharomyces pombe in which Bqt4 is a nuclear membrane protein (Chikashige et al., 2009) and regulates the replication of certain chromosomes regions by tethering specific heterochromatic regions ensuring the fidelity of the DNA synthesis and reassembling of the heterochromatin (Ebrahimi et al., 2018) during mitosis. In Aspergillus nidulans, MtgA, the homologous of Bqt4, was reported with a role in nuclear and nucleolar structure and it is required for mitotic-specific tethering to the nuclear envelope of Gle1, an mRNA export factor, but not for attachment of telomers to the nuclear envelope (Chemudupati et al., 2016). So far, the function of Bqt4 in phytopathogenic fungi has yet to be investigated at the molecular level. Therefore, we adopted Fusarium sp. INECOL-BM-06, which was isolated from X. morigerus, as a model, and we performed the characterization of an edited strain, by the CRISPR/Cas9 system, in Fsptf, homologous to Bqt4. The phenotypic analysis of the edited strain suggests an essential role of FspTF in growth and pathogenicity. Moreover, the data obtained through high-throughput approaches suggest that FspTF regulates the expression of genes involved in key biological processes such as secondary metabolism, membrane transport, regulation of gene expression, and lipid metabolism, among others.

Materials and methods
2.1. Fungal strains, media, and growth conditions Fusarium sp. INECOL_BM-06 (Carreras-Villaseñor et al., 2022) was used as the wild-type (WT) strain throughout this work. Fusarium sp. WT and mutant strains were propagated on potato dextrose agar (PDA, DIFCO) at 25°C. For stress responses assays, a minimal medium (MM) was prepared as described by Leslie and Summerell (2006) and supplemented with H 2 O 2 0.08%, SDS 0.025%, Congo Red (CR) 0.02%, Sorbitol 1 M, KCl 0.7 M, or hydroxyurea (HU) 20 mM. The MM was supplemented with bacteriological Agar (SIGMA) 2%. A plug (5 mm in diameter) from the edge of the actively growing colony of the strains grown in PDA was inoculated in the center of the plate. The cultures were incubated at 25°C for 7 days. The experiment was performed with three technical replicates.

Expression analysis by quantitative RT-PCR in fungal and plant tissue
Total RNA from fungal and plant tissue was isolated by phenol extraction and LiAc precipitation. Primers for RT-qPCR were designed to produce amplicons around 150 bp using the IDT PrimerQuest™ tool (Supplementary Table S2). cDNA was synthesized using SuperScript III Reverse Transcriptase (Invitrogen) following the manufacturer's instructions. The quantitative PCR was carried out according to the SYBR Green PCR Master Mix protocol (Applied Biosystems) using the Mx3000P qPCR system (Agilent). EF1-a was used as the standard (housekeeping gene) to normalize the content of cDNA, and the 2 -ΔΔCt method (Livak and Schmittgen, 2001) was employed to determine the relative expression and compare the treated samples. edition by the CRISPR/Cas9 system. Guide 65Fw (ACACCTAAGATGGTCGGCAC) for exon 2 was selected based on its on-target scores and no off-target prediction. The 65Fw protospacer sequence was inserted in plasmid pFC332 by combining two PCR fragments amplified from plasmid pFC334 and the linearized pFC332 plasmid in a NEBuilder reaction (Nødvig et al., 2015;Wenderoth et al., 2017). The primer sequences are listed in Supplementary Table S2. The fragments were amplified from pFC334 with Platinum Taq DNA Polymerase High Fidelity (Invitrogen) following the instruction of the manufacturer with a specific PCR program (First denaturation 94°C for 2 min, 30 cycles of 94°C for 15 s, 63°C for 30 s, 68°C for 30 s, and final elongation 68°C for 10 min). The plasmid pFC332 was linearized with PacI (Invitrogen) and assembled with the purified PCR fragments, following the NEBuilder protocol. DNA sequencing corroborated the plasmid construct. The edited strain was isolated, and the genome edition was confirmed by the amplification of a 1,524 bp fragment with primers that anneal specific sites up-and downstream of the protospacer sequence with posterior sequence analysis (Supplementary Table S2). The deletion construct was generated employing double-joint PCR (Yu et al., 2004) containing the hygromycin phosphotransferase gene (hph) as a selection marker, which confers hygromycin resistance, joined with the 5′ and 3′ flanking regions of the Fusarium sp. tf gene. The hph selection marker was amplified from pCB1004 and the 5′ and 3′ flanking regions from the genomic DNA of Fusarium sp. INECOL_BM-06 strain. The final construct was introduced in Fusarium sp. INECOL_ BM-06 strain. The replacement mutant was isolated and confirmed with primers annealing to the ORF by PCR. The primers used are listed in Supplementary Table S2.

Transformation of Fusarium sp. protoplast
Fusarium sp. was transformed as described previously (Connolly et al., 2013), with some modifications. Briefly, fungal conidia were harvested from a carboxymethylcellulose (CMC) culture (Cappellini and Peterson, 1965) and inoculated in 50 mL of YPD broth for overnight cultivation at 28°C and 200 rpm. The mycelium was harvested by filtering, washed with KCl solution (1.4 M), and digested in a lysing enzyme (Sigma) suspension (40 mg/mL in KCl solution 1.4 M) for 3 h with soft shaking at 70 rpm at 28° C. Protoplast were separated from cell fragments by filtering through to 30 μm membrane and sedimented by centrifugation at 8000 rpm for 10 min at 4°C. The lysing enzyme solution was discarded, and the protoplast was washed once with STC (Sorbitol 1.2 M, Tris-HCl pH 8 10 mM, and CaCl 2 50 mM) and resuspended in 400 μL of STC. Then, 10 μg of DNA and 100 μL of PEG 30% were added to the protoplast followed by 20 min incubation at room temperature. We then added 1 mL of PEG 30%, followed by 5 min incubation at room temperature. Finally, 2 mL of STC was added. Half of the suspension was mixed with warm 30 mL regeneration medium and split into two Petri dishes. The operation was repeated with the rest of the suspension. After overnight incubation at 28°C, the transformation plates were overlayed with a 15 mL warm regeneration medium containing hygromycin (100 μg/ mL). Resistant colonies were picked and purified to obtain homokaryotic strains.
Frontiers in Microbiology 04 frontiersin.org 2.6. Measurement of germination rate and yield of conidia For conidia generation, three mycelium plugs were inoculated in 25 mL of CMC and incubated at 28°C and 200 rpm for 7 days. Then, they were harvested and counted using a hemocytometer. To examine conidia germination levels, conidia of WT and mutants were inoculated into 5 mL of MM at a concentration of 2×10 5 conidia/mL and incubated at 28°C. Then, 2 h post-inoculation (hpi), germination was assessed every 2 h. The percentage of germination was calculated by the number of total conidia and germinated conidia visualized in the hemocytometer.

Pathogenesis assays
Foliar disk of 2 cm in diameter from freshly collected leaves from Populus nigra, Coffea arabica cv. Marsellesa, Citrus sinensis, and Citrus latifolia were injured in the center and inoculated with a mycelium plug from the edge of an actively growing colony of the strains grown in PDA. The foliar disks were placed in wet chambers and incubated at 25°C ± 1 with photoperiod (16 h dark/8 h light) for 7 days. Twelve disks were inoculated with each strain. Symptom development was recorded every day. The lesion area was measured by Image J conducting particle quantification based on image contrast.

Plant cell death and reactive oxygen species detection
Plant cell death was evaluated by staining with trypan blue (Fernández-Bautista et al., 2016). ROS, specifically H 2 O 2 , was determined by DAB staining (Daudi and O'Brien, 2012) after 4 days of infection. Images were acquired with a Zeiss SteREO Discovery. V8 stereomicroscope.

Untargeted metabolomics assays
The WT and TF65-6 strains were grown in PDA for 7 days in the dark after the mycelium was collected and frozen immediately. For sample preparation, 0.3 g of dried mycelial samples and 30 mL of methanol (HPLC grade, Sigma-Aldrich) were used. Methanolic extracts were obtained by four pulses of 10 min each in an ultrasonic bath (Cole Parmer, United States). The supernatant was recovered and concentrated to dryness in a rotatory evaporator (RII, Büchi, Switzerland). Dried methanolic extracts were redissolved in methanol (50 mg/mL, MS grade with 0.1% of formic acid, Sigma-Aldrich), filtered with 0.2 μm PTFE membranes, and placed in 2 mL Ultra Performance Liquid Chromatography (UPLC) vials. The untargeted metabolomics analyzes were performed in a UPLC (Class I, Waters™, United States) coupled to a quadrupole-time-of-flight Synapt G2-Si mass spectrometer (Waters, United States). The chromatography was carried out on a Waters Acquity BEH column (1.7 μm, 2.1 × 50 mm) with column and sample temperatures of 40 and 15°C, respectively. The mobile phase consisted of (A) water and (B) acetonitrile, both with 0.1% of formic acid (Sigma-Aldrich, United States). The gradient conditions of the mobile phases were 0-20 min, a linear gradient of 1-99% B, 20-24 min 99% B isocratic, and a 24-25 min linear gradient of 90-1% B (total run time 30 min). The flow rate was 0.3 mL/min, and 5 μL of the extract was injected. The mass spectrometric analysis was performed with an electrospray ionization source in positive and negative mode with a capillary, sampling cone, and source offset voltages of 3,000, 40, and 80 V, respectively. The source temperature was 120°C, and the desolvation temperature was 20°C. The desolvation gas flow was 600 L/h, and the nebulizer pressure was 6.5 Bar. Leucineenkephalin was used as the lock mass [556.2771, (M + H) + ; 554.2615, (M-H) − ]. The conditions used for MSe analysis were: mass range 50-1,200 Da, Function 1 CE, 6 V, function 2 CER 10-30 V, and scan time of 0.5 s. The data were acquired and processed with MassLynx (version 4.1) and MarkerLynx (version 4.1) softwares from Waters. The tentative identification of metabolites was performed using the MetaboAnalyst bioinformatics platform. 12 The Statistical Analysis module was used for fold change (FC) analyzes to identify over-and down-accumulated metabolites in the mutant strain (log2FC = ±1).

RNA-seq and differential gene expression analysis
The mycelium of the WT and TF65-6 strains analyzed in the transcriptomic experiment was collected and frozen immediately after growing in PDA for 7 days in the dark. Total RNA was extracted using the TRIzol protocol (Invitrogen). Following the manufacturer's instructions, six libraries (three biological replicates for each strain) were prepared with the TruSeq RNA Sample Preparation Kit (Illumina, San Diego, CA, United States). The six libraries were sequenced using the NextSeq500 platform (Illumina, San Diego, CA, United States) on paired-end reads at 2 × 150 bp format. The raw data obtained from RNA-seq were deposited in the Short Read Archive (SRA) database of the National Center for Biotechnology Information (NCBI), accession PRJNA972669.
Raw paired-end reads were analyzed with Trimmomatic v0.38 (Bolger et al., 2014) to filter out low-quality sequences. Reads mapping to the reference genome (Fusarium sp. isolate INECOL_ BM-06, unpublished assembled genome, access provided by Ibarra-Laclette and Sánchez-Rangel from the INECOL at Xalapa, Veracruz-Mexico) and transcript abundance estimation were performed using Bowtie2 v2.3.5.1 mapper (Langmead and Salzberg, 2012) and Expectation-Maximization (RSEM) v1.3.1 (Li and Dewey, 2011) software, respectively. A transcript abundance matrix was created containing each of Fusarium sp. INECOL_BM-06 genes (rows) and the expected counts (EC). In addition, transcripts per million (TPM) values were calculated for both WT and TF65-6 strains (columns). EC values represent the expression levels calculated by the maximum likelihood estimation approach and posterior mean estimates with 95% credibility intervals. For its part, TPM values can be considered highly consistent when a comparison is performed among different samples (Wagner et al., 2012). These values were used to perform principal component analysis (PCA) and a negative binomial model was used to identify differentially expressed genes (DEGs) that were used to perform pairwise Wald tests. In addition, the 2.11. Annotation of the differentially expressed genes and gene ontology enrichment The annotation of the DEGs was performed by homology-based inference. The annotation process involved BLASTp similarity searches (e-value cutoff of 10 −5 ) against the proteins reported for Fusarium solani, Fusarium oxysporum NRRL 3293, Fusarium fujikuroi IMI 58289, Fusarium verticillioides 7600, and Fusarium graminearum PH-1. Neurospora crassa and Saccharomyces cerevisiae were also added. All protein sets were downloaded from the latest version available on the GenBank database. 13 GO terms were assigned by transferring GO terms from the Fusarium species mentioned above using the software package InterProScan (Jones et al., 2014).
The functional enrichment analysis was performed considering the list of differentially expressed genes using the web tool g:Profiler (Raudvere et al., 2019). 14 The g:SCS method was applied to perform multiple testing corrections using the p-adjusted values with a ≤ 0.05 threshold. The available Fusarium species on web site (Fusarium graminearum str. PH-1, Fusarium fujikuroi, Fusarium oxysporum f. sp. lycopersici 4,287, Fusarium pseudograminearum CS3096, Fusarium vanettenii 77-13-4 (F. solani), and Fusarium verticillioides 7600) were used as references.

Statistical analyzes
All the experimental data collected were statistically evaluated using the GraphPad Prism 9.5.1 software (GraphPad Software Inc., San Diego, United States), with p < 0.05 considered statistically significant in all the cases.  Table S3). The phylogenetic tree grouped the fungal sequences into four clades ( Figure 1A). The bigger clade comprised 13 https://www.ncbi.nlm.nih.gov/ 14 http://biit.cs.ut.ee/gprofiler/ members of the subfamily SMRC such as Swi4 and Mbp1 of S. cerevisiae, and Res1 and Res2 of S. pombe. In this clade, two sequences of Fusarium sp. were identified, named SMRC1 and SMRC2, which are closer to Swi4 and Swi6 of N. crassa, respectively. Another clade grouped the proteins of the APSES subfamily members, homologous to StuA of A. nidulans; this clade included one sequence from Fusarium sp. denominated as StuA. A third clade enclosed the sequences homologous to Xbp1 of S. cerevisiae, in which a sequence of Fusarium sp. named Xbp1 was included. The last clade grouped the proteins homologous to Bqt4 of S. pombe and MtgA of A. nidulans, which have been previously characterized (Chikashige et al., 2009;Chemudupati et al., 2016). The Fusarium sp. sequence, denominated as FspTF, was more related to the N. crassa, F. graminearum, and M. oryzae proteins whose function has not yet been determined. Since the family Bqt4 is the youngest and least distributed and, therefore, least studied member of the family KilA-N/APSES in filamentous fungi, we characterized this gene in Fusarium sp. associated with X. morigerus. FspTF is a protein with 431 aa, Mw of 47.5 kDa, and pI 5.4. The analysis by SMART (Letunic and Bork, 2018) identified a KilA-N domain (IPR018004) in the N-terminal region (residues 79-165), a coiled-coil region (residues 313-340), and a single-pass C-terminal transmembrane helix region (residues 411-430) ( Figure 1B).

Fsptf is responsive to different stress conditions
To infer the biological function of FspTF, the expression of Fsptf was evaluated during different growth conditions. Fsptf expression was induced during oxidative stress, and plant infection process with 4.7 and 1.6-fold, respectively. Conversely, it was repressed in 30% in response to cell wall (CR 0.02%) stress, and in osmotic stress exerted by KCl 0.7 M, sorbitol 1 M, and DNA stress (HU 20 mM), the repression was 91, 39, and 28%, respectively. Fsptf was not responsive to the growth in SDS 0.025% ( Figure 1C). These data suggest a role of Fsptf during stress responses and pathogenesis.

Fsptf gene edition by the CRISPR/Cas9 system in Fusarium sp. results in a DNA region deletion
Once the homokaryotic strain TF65-6 was obtained after the protoplast transformation of Fusarium sp., sequence analysis of its genomic DNA revealed a genomic deletion of 179 bp in exon 2 ( Figure 1D; Supplementary Figure S1, S2). Fsptf gene structure and protein domain organization in the edited strain TF65-6 were predicted using FGENESH (Solovyev et al., 2006) and SMART (Letunic and Bork, 2018), respectively. In TF65-6, the deletion of 179 bp in exon 2 produced a frameshift that generated a longer intron 1 and smaller exon 2. This edited gene was predicted to encode a protein with 349 amino acids lacking the KilA-N DNA-binding domain but keeping the coiled-coil and transmembrane helix regions ( Figure 1E). The expression analysis in growth in PDA, MM, oxidative stress, and in interaction with plant tissue indicated that Fsptf was downregulated by 74, 60, 70, and 67%, respectively, in TF65-6 in comparison with the expression in the WT (Supplementary Figure S3). The colony morphology of the edited strain, TF65-6, was contrasting to the WT. The WT strain presented pigmented mycelium with a concentric ring of aerial mycelium and secreted a compound that pigmented the culture medium; in contrast, TF65-6 developed white mycelium without aerial mycelium and lost pigmentation in both mycelium and medium. At 14 dpi, the contrasting phenotypes between both strains were maintained ( Figure 2A). Additionally, the colony area of the TF65-6 strain was 2.4 and 1.5-fold bigger than that of the WT at 7 and 14 dpi, respectively ( Figure 2B). Conidia production was affected in TF65-6, with a 58% lower quantity  Figure 2C); however, the germination rate increased, specifically, in the early stages, with 11 and 30% higher rates at 2 and 4 hpi, respectively ( Figure 2D). To investigate whether FspTF has a role in stress responses, Fusarium sp. WT and edited strain were grown in MM amended with compounds known to cause different types of stress, such as oxidative and osmotic, and cell wall, membrane, and DNA damage ( Figures 3A,B). In the presence of H 2 O 2 (0.08%), the growth of the WT was reduced by 56%, and in TF65-6 it was reduced by 62%. In osmotic stress induced by KCl 0.7 M, TF65-6 was more sensitive with 57% of inhibition, and WT had 50% of inhibition. Interestingly, TF65-6 was more tolerant to DNA damage triggered by HU 20 mM, with 7% inhibition compared with 63% of inhibition in the WT ( Figure 3B).

Pathogenesis of Fusarium sp. is regulated by FspTF
Since Fusarium sp. was reported as a potential phytopathogen vectored by X. morigerus (Carreras-Villaseñor et al., 2022), we investigated the role of FspTF in pathogenicity against the foliar disk of Populus nigra (Figure 4), Coffea arabica, Citrus sinensis, and Citrus latifolia (Supplementary Figure S4). The lesion in the plant tissue by the WT strain was characterized by the increasing necrosis and chlorosis area from the inoculation site along the time course tested. Interestingly, the mutant strain showed a significant negative impact on their infection capacity. Nevertheless, the impact was different among the plant species. In P. nigra, the edited strain provoked necrosis around the inoculation site, however the percentage of plant tissue damage at the end of the tested period was 54 and 9.8% by WT and TF65-6, respectively, denoting a clear difference in the infection progression ( Figures 4A,B). The differences in the infection capacity were also noted during cell death and ROS detection in the plant tissue since the blue staining and the brown precipitate of DAB reacting with ROS were of lesser magnitude in foliar disks infected with TF65-6 ( Figures 4C,D). In C. arabica, as in P. nigra, WT and TF65-6 established the infection process, but the progression was slower in the tissue infected by the mutant, denoted by the significant difference in the percentage of tissue damage along the tested period, being 17.7 and 14.3% at 7 dpi, respectively (Supplementary Figure S4). The progression of the infection by the mutant in the citrus species was slower until a certain time point; in C. sinensis the percentage of damage of the plant tissue was minor up to 2 dpi by TF65-6 (6.6%) in comparison by WT (9.3%), then the damaged area was comparable among the strains. In C. latifolia the damage in plant tissue by all the strains was comparable from 5 dpi, before that, the foliar disk    Figure S4). Later, we evaluated the defense system of P. nigra in response to infection with Fusarium sp. WT and TF65-6. Interestingly, the expression of PR1 and PR2 as marker genes of the salicylic acid (SA) pathway and JAR1 and COI1 involved in the jasmonic acid (JA) pathway in response to TF65-6 was different in comparison to WT, mainly showing a delay and less intensity in the response ( Figure 5). Specifically, the expression profile of PR1 in response to the WT implied an induction from 30 min (2.2 fold), with a major increase in the expression level, 6.5 fold, at six hpi; then, repression was observed reaching a basal level at 12 hpi and the repression increased at 24 hpi (0.4 fold), with a tendency toward induction from 48 hpi with a 3.4 fold increase. In contrast, in the PR1 expression profile in response to TF65-6, an induction was observed at 30 min (2.1fold, similar in response to the WT), followed by a tendency to repress the expression up to 12 hpi (1.2 fold), and, at 24 hpi (1.8 fold), a trend toward induction was observed until reaching a similar expression in response to both strains at 72 hpi (2.7 fold). The expression profile of PR2 in response to the WT strain implies "transcriptional waves" of induction-repression of expression up to 24 hpi, with the higher induction at 30 min (1.7 fold), and at 6 hpi the expression was the lowest (0.2 fold). From 48 hpi, the trend was toward induction of expression, reaching the highest level at 72 hpi (2.7 fold); however, the plant infected with TF65-6 did not seem to respond until 12 hpi, where induction of PR2 expression was observed (1.8 fold) with subsequent repression at 24 hpi (0.7 fold), followed by a tendency toward induction from 48 hpi onwards, reaching an induction of 2.3 fold at 72 hpi, as well as the response against the WT strain ( Figure 5). Regarding the expression of marker genes for the JA pathway, JAR1 was repressed in the interaction with the WT, with greater repression being observed at 72 hpi (0.4 fold). On the other hand, there seemed to be no response against the TF65-6 mutant strain or with a slight tendency to repression, to a lesser extent than against the WT strain, especially at 72 hpi (0.7fold). In relation to the expression of the COI1 gene, it was observed that for both strains, the response was the repression of the expression, presenting practically the same levels in response to both strains, except at 24 hpi, where the repression was more significant in the strain TF65-6 (0.4 fold and 0.6 for WT) ( Figure 5).

Secondary metabolism is controlled by FspTF
We performed a comparative untargeted metabolomic analysis between the mycelium of the WT and TF65-6 obtained after 7 days of growth in PDA ( Figure 6). We detected 1,694 spectrometric features (retention time-mass/charge ratios; rt-m/z) in both ESI + and ESI − modes in the dataset. A differential accumulation analysis made it evident that fewer rt-m/z signals were accumulated in the TF65-6 ( Figure 6A), with 351 rt-m/z signals over-accumulated, and 833 rt-m/z signals down-accumulated in TF65-6 compared to WT, and 465 rt-m/z signals were not statistically different ( Figure 6B). Among the rt-m/z signals detected and over-accumulated in TF65-6, fusaproliferin, classified as an emerging Fusarium mycotoxin (Gruber-Dorninger et al., 2017; Ćeranić et al., 2021), and some lipids, such as linolenic and palmitic acids ( Figure 6B; Table 1) were putatively identified. Among those that were down-accumulated were fusarubin, 6-O-demethyl-5-deoxyanhydrofusarubin, and javanicin, which are pigmented polyketides produced by different Fusarium species (Studt et al., 2012;Kristensen et al., 2021) and fumonisin B4, the less common fumonisin of the B series (Awuchi et al., 2021;Ekwomadu et al., 2021) tentatively identified ( Figure 6B; Table 1).

Overview of the transcriptome analysis in TF65-6
We used high-throughput sequencing to examine the genomewide transcript profiles of the WT and TF65-6 mycelium samples collected from colonies grown on PDA for 7 days. The raw sequencing data consist of 13 to 18 million reads for each library, suggesting good coverage and sequencing depth. After the quality analysis, a total of 90,642,783 high-quality (HQ) paired-end reads were obtained (Supplementary Table S4). These HQ reads were mapped against the Fusarium sp. INECOL_BM-06 genome and quantified (Supplementary Table S5). To carry out the differential expression analysis (Figure 7), a principal component analysis (PCA), using the estimated TPM values, was performed to detect the major sources of variance underlying the biological replicates (R1-R3) of each strain. The two-dimensional PC plot with the first two principal components (PC1 and PC2) best illustrated the variance among the expression profiles, with a proportion of explained variance of 80 and 6%, respectively ( Figure 7A). The plot, together with a hierarchical clustering tree of the six libraries ( Figure 7B), indicated that the biological replicates have high reproducibility. Based on these results, a pairwise comparison (TF65-6 vs. WT) was performed to identify differentially expressed genes (DEG). In total, 4,326 genes were encountered with differential expression of at least two-fold (Log2FC = ±1) and an adjusted significant value of p of ≤0.05 (Figures 7C-E; Supplementary Tables S6-S7), and as expected, the expression profile among the biological replicates was comparable ( Figure 7C). From the differential genes, 1,787 genes were induced, and 2,539 genes were repressed in TF65-6 relative to WT ( Figure 7D), most of them with Log2FC = ±5 ( Figure 7E; Supplementary Table S7).

Regulation of secondary metabolism: pigments and mycotoxins
Since the evident differences in pigment production between mutant and WT strains and the metabolomic results (Figures 2, 6; Table 1), we delved into the analysis of genes related to secondary metabolism. In this sense, genes predicted to encode polyketides synthases (PKS) and non-ribosomal peptide synthases (NRPS) are downregulated in mutants, such as PGL1 (ID: g05609 FC: -6.7), BIK1 (ID: g8257 FC: −5.1), and FUB1 (ID: g07760, FC: −3.89), that are required for perithecium pigment, bikaverin and fusaric acid production in F. verticilloides, respectively (Table 2). In addition, homologous genes of some FSR cluster genes, required for fusarubin synthesis in F. fujikuroi (Studt et al., 2012), were encountered as downregulated, and homologous genes of FUS cluster genes, required for fusarin production in F. verticilloides (Brown et al., 2012) were identified as repressed. In addition, FUM 13 (ID: g01406, FC: −1.8), a short-chain dehydrogenase/reductase required for the fumonisin biosynthesis in F. verticilloides (Butchko et al., 2003), was also repressed ( Table 2).  Figure S4). In this context, we searched for genes with possible roles in pathogenicity, with emphasis on those that encode secreted hydrolytic enzymes and virulence factors or effectors. We focused on the enzymes that target cuticle and plant cell walls, the first barriers of pathogen invasion, and proteases (Supplementary Table S9).
Moreover, in TF65-6, some DEGs encode glycoside hydrolases (GH) from different families that degrade the cellulose, hemicellulose, and pectin, components of the primary plant cell wall, as well as oxidases that degrade lignin, the secondary cell wall (Supplementary Table S9). There are 26 DEGs predicted to be involved in cellulose degradation. The highest induced gene encodes weak endoglucanase (ID: g08682, FC: 5.31) belonging to the lytic polysaccharide monooxygenase (LPMO) auxiliary superfamily (cl41755), and the most repressed gene was predicted to be a β-glucosidase (ID: g06562, FC: −5.72) of the BglX superfamily (cl34276). Most of the induced genes belonged to the GH6 (cl38036), GH7 (cl21662), GH12 (cl03302), GH45 (cl03405), and GH61 Enrichment of the Biological Process Gene Ontology (GO) terms. Significantly enriched GO terms are shown; highlighted are the GO terms to which belong most of the DEGs in TF65-6.

Regulation of lipid metabolism: cell membrane
We recognized repressed genes involved in the biosynthesis of the lipid components of the cell membrane, such as glycerophospholipids, sphingolipids, and sterols (Table 2). For example, the homologous of CDS1 (ID: g04816, FC: −1.09), which catalyzes the critical enzymatic reaction in the biosynthesis of all major phospholipids for producing the liponucleotide precursor, CDP-diacylglycerol (CDP-DAG; Sant et al., 2016). In addition, the most repressed gene involved in the biosynthesis of sphingolipids was the homologous of LCB2 (ID: g11362, FC: −2.83), a serine palmitoyltransferase responsible for the first committed step of their biosynthetic pathway (Fabri et al., 2020). Worth mentioning is the ergosterol biosynthetic pathway. We identified 10 DEGs that were homologous to those involved in biosynthesis. Homologous to ERG10, the acetyl-CoA C-acyltransferase, which mediates the first step for the synthesis of the intermediate mevalonate, was the most repressed (ID: g08093, FC: −7.23). Genes involved in the late biosynthetic steps were deregulated; considering the two ratelimiting steps, ERG1 and ERG11 were induced, and genes that encode last enzymes of the pathway, suggested to display low substrate affinities, such as ERG6 and ERG4, were induced, conversely, ERG5 was repressed (Table 2; Supplementary Figure S5) (Jordá and Puig, 2020). Together these data suggest that the  composition of the cell membrane may be altered, and because the changes in membrane composition might be necessary for cold stress response (Aguilera et al., 2007), we evaluated the growth phenotype of WT and TF65-6 by subjecting both strains to cold stress (4°C) during 72 h, and then resumed their growth at 25°C. We found that the mutant showed 6% more growth inhibition in comparison with the WT data, a discrete but statically significant difference ( Figures 9C,D) denoting that TF65-6 was less tolerant to this stress than WT.

Discussion
Fungi modify gene expression during their life cycle and in response to environmental stimuli through the action of a variety of transcription factors (TFs). In fungal genomes, hundreds of TFs have been predicted and classified into approximately 80 TF families based on their DNA-binding domains. Interestingly, three TF families are fungal-specific: KilA-N/APSES, Mating-type MAT α1, and Copper fist (Shelest, 2017;Yu et al., 2023). However, few members of these families have been characterized. In phytopathogenic fungi, members of 10 families have been studied; specifically in F. oxysporum, it was recently reported that from nearly 800 TFs predicted, only 26 TFs had been functionally characterized (John et al., 2021;Zuriegat et al., 2021), highlighting the universe of TFs that remains to be characterized. The KilA-N/APSES family is described as exclusive in fungi; however, this domain is found in DNA viruses of prokaryotes and eukaryotes (Iyer et al., 2002). Some TFs of the KilA-N family have a solid connection to cell cycle regulation but also diversified their function in other aspects of fungal cell biology Medina et al., 2019). In this study, we identified five KilA-N TFs in Fusarium sp. INECOL_ BM-06 associated with Xylosandrus morigerus ( Figure 1A; Supplementary Table S3) in accordance with the distribution and conservation encountered in fungal species of the subphylum Pezizomycotina (Medina et al., 2019). The Bqt4 is the youngest and least distributed member of the KilA-N family and, therefore, less studied. Here, FspTF, a homologous of Bqt4 of S. pombe (Chikashige et al., 2009), and MtgA of A. nidulans (Chemudupati et al., 2016) were functionally characterized in Fusarium sp. associated with X. morigerus. Bqt4 and MtgA are proteins embedded in the inner nuclear membrane (INM), and the KilA-N domain in Bqt4 of S. pombe is reported to have moderate and unspecific DNA-binding activity and to mediate protein-protein interactions (Chikashige et al., 2009(Chikashige et al., , 2010Hu et al., 2019a,b). In this study, we generated an edited strain, TF65-6, by the CRISPR/Cas9 system, in which Fsptf was downregulated (Supplementary Figure S3) and the protein was predicted to lack the KilA-N domain ( Figure 1E) and, in consequence, it was suggested to fail in both binding DNA and protein interaction through this domain.
The phylogenetic analysis indicated that FspTF is less related to Bqt4 and MtgA, suggesting that functional differences exist between them; even Bqt4 and MtgA were not highly related, denoting their differences regarding telomere tethering, among others. Instead, FspTF formed a cluster with the orthologous of Fusarium species that were more related to the orthologous of N. crassa, which have not been characterized, meaning that this is the first study exploring this TF in Fusarium spp.
The Fsptf expression analysis that we conducted in different conditions ( Figure 1C) indicated a possible role in different biological processes of Fusarium sp., including growth and pathogenesis. In congruence, the edition of Fsptf in Fusarium sp. resulted in various phenotypical changes, such as increased mycelial growth, reduction of conidia production (but accelerated conidia germination rate), decreased virulence, and secondary metabolites production, and it also modified the response to diverse stresses (Figures 2-4). In addition, Δtf, a replacement mutant, showed the same phenotypical changes as TF65-6, regarding mycelial growth, conidia germination rate, virulence, secondary metabolites production, and responses to certain stresses (Supplementary Figure S6-S8). The high-throughput data provided evidence of gene candidates that may be regulated by the action of FspTF to modulate pigment production, growth, and virulence.

FspTF control pigment production
TF65-6 showed a lack of mycelium and medium pigmentation (Figure 2A), down-accumulation of the pigments javanicin and fusarubin (Table 1), and those genes belonging to the FSR cluster were repressed (Table 2). Altogether, these results can explain the unpigmented phenotype and give clues about the possible direct or indirect targets of FspTF. In addition, we informed for the first time that Fusarium sp. INECOL_BM-06 has the capacity to produce javanicin and fusarubin, which is relevant given the ecological niche of this fungus and that these pigments are reported with antifungal and antibacterial activity  and, in the ambrosia F. ambrosium, are suggested to prevent the invasion of endophytic fungi in galleries of Euwallacea formicatus, protecting its habitat (Kehelpannala et al., 2018). This is in agreement with the succession of metabolic roles that occur between bacteria and fungi that inhabit the galleries to guarantee the optimal development of ambrosia beetles (Ibarra-Juarez et al., 2020).

Growth is controlled by the expression of transmembranal nutrient transporters, regulated by FspTF
In TF65-6, most DEG-encoding transporters were induced (Supplementary Table S8), denoting the active income of nutrients that contribute to the increased colony growth. Fungi use a variety of sugars such as hexoses as primary carbon source; in yeast, the Frontiers in Microbiology 18 frontiersin.org transport of hexoses is mediated by hexose transporters (Hxt, Hxt1 to Hxt17) ( Barata-Antunes et al., 2021). TF65-6 overexpresses homologous of HXT1, HXT9, HXT13, HXT15, and HXT17. HXT1 is a low-affinity glucose transporter (Bisson et al., 2016), HXT9 is a functional glucose carrier (Wieczorke et al., 1999), and HXT13, HXT15, and HXT17 are polyol transporters (Jordan et al., 2016); the expression of HXT13 and HXT15 was induced in the presence of non-fermentable carbon sources (Greatrix and van Vuuren, 2006). However, the homologous of CAT8, a TF required for the growth in non-fermentative conditions (Hedges et al., 1995;Haurie et al., 2001), was repressed. Furthermore, the homologous of HXT2 and HXT6, high-affinity transporters, were repressed. The same situation was observed with regard to the uptake of maltose, with genes encoding maltose permeases being induced; conversely, the TFs that regulate maltose utilization, MAL13 and MAL33 (Orikasa et al., 2018), were repressed. Taken together, these results suggest that sufficiency of glucose is sensed by the mutant but can incorporate alternative C sources, although it cannot use them efficiently. This hypothesis must be proved in future research work.
On the other hand, TF65-6 repressed specific genes encoding TFs related to the biosynthesis of certain amino acids, ARG81, Lys14, and ARO80, and overexpressed transporters for different amino acids (Supplementary Table S8), denoting the necessity of taking them from the environment, which, in addition, can be used as carbon and nitrogen sources and as substrates for the biosynthesis of other biomolecules, such as NAD + , folate, glutathione, nucleotides, polyamines, and phospholipids (Bianchi et al., 2019). In addition, the repression of nit-4 (Yuan et al., 1991;Huberman et al., 2021), the induction of UGA3 (Cardillo et al., 2011;Sylvain et al., 2011), and the induction of transporters of non-preferred nitrogen sources as urea, purines, oligopeptides, and allantoate (Supplementary Table S8) indicates the possibility of assimilating alternative nitrogen sources. In this context, TF65-6 seems to respond to its intracellular state by regulating the transport systems across biological membranes to uptake essential nutrients to sustain active growth, a common compensation mechanism in fungi (Ramos et al., 2016).

FspTF is relevant for controlling the pathogenic process
Our results suggest that TF65-6 does not entirely lose its pathogenicity but dramatically changes the progression velocity of the infection process. The hypothesis above is sustained by the characterization of several pathosystems using histochemical tests and analyzing some plant defense marker genes (Figures 4-5). However, the changes depend totally on the host tested (Supplementary Figure S4). Based on these observations, we encountered genes that can be related to the phenotype in pathogenesis. In this sense, our data suggest that TF65-6 can transgress the cuticle and degrade the plant cell wall in order to establish within the host, denoted by the induced genes related to lipid degradation, hydrolytic enzymes, and oxidases (Supplementary Table S9) and the evident establishment of the infection ( Figure  4; Supplementary Figure S4). Nevertheless, inside the host, TF65-6 cannot take nutrients to support the progression of the infection, given that most of the proteases differentially expressed were repressed (Supplementary Table S9).
On the other hand, the genes related to effectors were induced (Table 2). Therefore they could have a minor role in pathogenesis, could participate in the establishment of the infection, but not in its progression, and could allow the induction of defense responses in P. nigra ( Figure 5; Zeng et al., 2023). This is consistent with previous studies in which it was proven that phytopathogenic species of Fusarium genus associated with ambrosia beetles strongly induced genes related to defense in their host (Pérez-Torres et al., 2021). In addition, TF65-6 downregulated different PKS, including that involved in the synthesis of fusaric acid, recognized as a virulence factor in Fusarium (López-Díaz et al., 2018), and diverse NRPS (Table 2), which can participate in the synthesis of biologically active cyclic depsipeptides and other phytotoxic nitrogen-containing metabolites, which are important players for the development of plant disease symptoms (Xu et al., 2021). Moreover, fumonisins are reported as potential virulence factors, with FB1 being the most relevant in this regard (Sánchez-Rangel et al., 2012;Xie et al., 2021). TF65-6 downaccumulated fumonisin B4 (Table 1) and repressed FUM13 (Table 2), which can contribute to the observed phenotype related to virulence. Fusaproliferin was over-accumulated, and genes of the FUS cluster involved in the biosynthesis of fusarins were downregulated in TF65-6. The role of these mycotoxins in the ecology of Fusarium is not clear, and their function as virulence factors, in development, in protection against competitors and predators, and as an adaptation mechanism cannot be discarded (Geisen et al., 2017). Additionally, in F. graminearum, ABC-transporters have a role in virulence and azole tolerance, specifically, FgABC3, suggesting encoding a transporter protecting the fungus from host-derived antifungal molecules (Abou Ammar et al., 2013). TF65-6 downregulated genes encoding for ABC transporters (Supplementary Table S8) and the edited strain had enhanced sensitivity to azole fungicides ( Figures 8A,B). Thus, these genes could also have an active role in virulence in Fusarium sp. However, we cannot discard a specific transcriptional profile during the interaction with the host, which makes it possible to identify genes more closely related to the infection process.

Stress responses coordinated by FspTF
Coordinated stress responses are essential for fungi to survive in hostile conditions encountered in the environment. In the mutant, most of the key enzymes of metabolic systems that remove celldamaging ROS, such as superoxide dismutase, catalase, and glutathione/glutaredoxin/thioredoxin (Pócsi et al., 2004), were downregulated (Supplementary Table S10), which explains its sensibility to oxidative stress. Besides, osmotic stress is known to generate intracellular oxidative stress, and the deregulation of ROS scavenging pathways can participate in the low tolerance to this stress ( Figure 3).
On the other hand, the less sensitive phenotype of the edited strain during HU treatment was entirely unexpected. However, our subsequent analysis identified some candidate DEGs that may promote this phenotype. HU is an inhibitor of the ribonucleotide reductase (RNR) that catalyzes the synthesis of deoxyribonucleotides from ribonucleotides, slowing replication forks and arresting the cell cycle in the S phase; in consequence, the DNA replication checkpoint (DRC) pathway is activated to upregulate RNR and increase dNTPs production, which promotes fork progression (Xu et al., 2016). In TF65-6, some components of the DRC were induced, such as the homologous of MEC1 (ID: g07221, FC: 1.13) and CHK1 (ID: g14022, FC: 1.08) (Dörter and Momany, 2016). Furthermore, RNR1 (ID: g00986, FC: 1.02), homologous of the large subunit of the RNR, was induced. These genes can contribute to the observed tolerance of TF65-6 to HU. In addition, there is evidence that the ergosterol pathway is essential to counter this stress since the mutant of S. pombe in erg11, with sterol deficiency, is more sensitive to HU treatment (Xu et al., 2016). Our data suggest that ergosterol biosynthesis was altered in the TF65-6 edited strain since ERG genes were differentially expressed (Table  2; Supplementary Figure S5), and mainly ERG11 was upregulated, which can improve the tolerance to this genotoxic. Finally, HU also rescues the conditional lethality of mutants in S. cerevisiae, defective in diverse cellular pathways, including chromosome segregation, redox homeostasis, and sterol biosynthesis by an unknown mechanism (McCulley et al., 2014). Given the alterations of diverse cellular pathways in TF65-6, a chemical suppression exerted by HU can function in TF65-6, evidenced by the alleviated effect of the compound during growth.

Signal transduction and gene expression regulated by FspTF
In fungi, intra-and extracellular stimuli are translated into specific physiologic responses, through transcription regulation, by the precise coordination of signaling pathways that are reported to play an essential role in growth, development, pathogenesis, and mycotoxin biosynthesis (Guo et al., 2016;Barman et al., 2018;Yang et al., 2022). In TF65-6, we encountered DEGs that encoded components of G protein, MAPK, PKA, and PKC signaling pathways. Reversible protein phosphorylation by protein kinases (PK) regulates various fungal biological processes. In TF65-6, different PKs are differentially expressed, and some, such as CKA1, are important for various aspects of growth, developmental and infection processes in F. graminearum (Wang et al., 2011). Small Ras GTPases, monomeric G proteins, also participate in signal transduction as molecular switches with multiple functions; in TF65-6, RAS2 and RHO1 were induced, which are reported with a role in gene expression regulation, polar growth, membrane traffic, and morphogenesis in fungi (Dautt-Castro et al., 2021). Furthermore, we find specific TFs involved in the regulation of metabolic pathways (Supplementary Table S8) and other diverse TFs belonging to different families and chromatin remodelers (Supplementary Table S11). Thus, given that in the edited strain, multiple components of signaling pathways were deregulated, and the signaling pathways were interconnected and functionally pleiotropic, it was most likely that the phenotypes that we identified related to growth, pathogenesis, and stress were also the consequence of the "new" complex network which was functional in the edited strain.
In summary, this study revealed that FspTF plays a relevant role in overall growth, development, metabolism, stress responses, and pathogenesis by regulating gene expression. However, the open question that arises is which is/are the mechanism(s) and which is/are the direct target(s) of FspTF in Fusarium sp. associated with X. morigerus. Recently, it was encountered that Bqt4 interacts directly or indirectly with proteins with diverse cellular functions such as nucleoplasmic transport, protein degradation by 19S proteasome, cytoplasmic translation, cell cycle, and lipid metabolism, among others (Hirano et al., 2023a,b). It was proposed that Bqt4 interacts with proteins related to histone modification, ubiquitination, and DNA repair (Hu et al., 2019a), and participates actively with the function of Lem2, an INM protein (Ebrahimi et al., 2018;Hirano et al., 2018) with roles in an array of genome-regulatory functions including epigenetic silencing, DNA replication, and post-transcriptional regulation (Banday et al., 2016;Barrales et al., 2016;Tange et al., 2016;Ebrahimi et al., 2018;Hirano et al., 2018;Martín Caballero et al., 2022). Protein-protein interaction in the nuclear envelope can be the main molecular mechanism involved in how FspTF can function as a Frontiers in Microbiology 20 frontiersin.org major regulator in Fusarium sp; however, this hypothesis remains to be investigated in future work.

Conclusion
Despite transcription factors being key players in the regulation of biological processes in all organisms, a universe of fungal transcription factors remains uncharacterized, highlighting many aspects of fungal biology expected to be studied. In this work, we identified that FspTF, a member of the fungal-specific family of transcription factor KilA-N/APSES, of Fusarium sp. associated with the ambrosia beetle Xylosandrus morigerus, is involved in growth, conidia production and germination, responses to stress, and pathogenesis, revealing new aspects of the TFs grouped in the KilA-N/ APSES family in Fusarium genus, with particular emphasis in those species that are symbiotes of ambrosia beetles. These findings open the possibility of identifying targets for the prevention and mitigation of plant diseases caused by the ambrosia complex.

Data availability statement
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary material.

Author contributions
NC-V and DS-R: contributed to the conceptual design of the study, formal analysis, and wrote the manuscript. NC-V, LM-R, EI-L, JM-V, BR-H, and JG-A: methodology. EI-L: transcriptome assembly, annotation, and statistical analysis. JM-V and JG-A: untargeted metabolomics analysis. DS-R funding acquisition and project administration. All authors read and approved the final manuscript.