A multiple sequence alignment (MSA) of characterized AA11s revealed that three of the four characterized PMOs maintain a three-part architecture, while the other is a single-domain PMO containing only a catalytic domain (Figure 1—figure supplement 1). The three-domain architecture includes an N-terminus AA11 PMO domain Hemsworth et al., 2014; Levasseur et al., 2013 followed by an extended GS-rich linker region that is likely disordered, and a C-terminal X278 domain proposed to be a chitin-binding domain; this domain is present in other predicted AA11 proteins and GH18 chitinases (Gonçalves et al., 2019; Hemsworth et al., 2014; Figure 1A). AA11 proteins, similar to all other PMOs, contain a signal peptide at the N-terminus (predicted by SignalP) that directs the protein to the ER and that is subsequently cleaved during translocation, leaving a histidine residue at the N-terminus.

CWR-1 is a member of the three-region architecture cluster, and many fungi contain both AA11 PMO architectures in their genome. To more fully examine the relationships of predicted AA11 proteins, we performed sequence similarity network (SSN) analysis using the EFI-EST BLAST feature with the CWR-1^HG1 sequence as the search peptide, which yielded ~1700 sequences. The three characterized AA11 proteins and CWR-1 that contained the GS linker and X278 domain were found in a single cluster (blue) (Figure 1—figure supplement 2). The single AA11 domain enzyme (AfAA11A) clustered separately into the red cluster. Since these domain architectures are clustered separately, this suggests a different biological function in fungi (Figure 1—figure supplement 2). This is not solely due to the difference in domain organization, for example, starch PMOs with different architectures cluster together at ~40% sequence ID (Vu et al., 2019). To determine which domain of CWR-1 was involved in allorecognition and the fusion checkpoint, a series of truncations were generated in the HG1 (FGSC2489) cwr-1 allele: full-length cwr-1, removal of the GS linker, and removal of the GS linker and X278 domain (Figure 1A). Full-length and truncated constructs were inserted at the his-3 locus under the native promoter (1111 bp) and transformed into a strain of N. crassa bearing a deletion of cwr-1 (NCU01380) and cwr-2 (NCU01382) (Δcwr-1Δ81Δcwr-2) and expressing cytoplasmic GFP (Gonçalves et al., 2019). These strains were assessed for cell fusion in pairings with isogenic strains containing the cwr-2 allele from each of the HG strains (HG1-6). The cwr-1^HG1 strain gave the strongest cell fusion block in pairings with cwr-2^HG3 germlings (Figure 1—figure supplement 3A and B). This cwr-2^HG3 strain was therefore used to test strains harboring the cwr-1^HG1 truncations that lacked the GS linker or the GS linker and the X278 domain. Strains containing any of the cwr-1 truncations showed a block in cell fusion with the cwr-2^HG3 strain, including the truncation strain with only the catalytic PMO domain (Figure 1B and C). As a control, all of the truncation strains were also paired with a permissive mutant lacking cwr-1 and cwr-2 (Δcwr-1Δ81Δcwr-2); all showed robust cell fusion frequencies (Figure 1—figure supplement 3C and D). These data showed that the PMO catalytic domain was sufficient to cause cell fusion arrest at the CWR allorecognition checkpoint.

Given that the PMO catalytic domain alone was responsible for conferring allorecognition, we performed experiments to probe the activity for this domain. The PMO catalytic domain of the cwr-1^HG1 allele was expressed in Escherichia coli using the periplasmic expression pelB system, which cleaves scarlessly during expression to ensure an N-terminal histidine with a C-terminal cleavable hexahistidine tag (Figure 1D) and subsequently purified using Ni-IMAC. The protein ran slightly higher (~26–27 kDa) on an SDS-PAGE gel than the predicted size of 22.6 kDa (Figure 1—figure supplement 4A); the mass was validated with whole-protein mass spectrometry. The spectrum revealed a deconvoluted mass of 22,624 Da, which corresponds to the exact mass of the protein with three disulfide bonds (22,630 – 3 × (2 Da for each disulfide)) (Figure 1—figure supplement 4B). The CWR-1^HG1 PMO domain bound ~0.7 equivalents of copper after reconstitution and was shown to reduce oxygen to hydrogen peroxide in the presence of ascorbate (Figure 1E). The activity of the CWR-1^HG1 protein was then tested with β-chitin as a substrate. New peaks between 18 and 22 min that corresponded to C1-oxidized products were only observed in the presence of both ascorbate and the CWR-1^HG1 PMO domain (Figure 1E). C1-oxidized standards were generated from chitooligosaccharides (Figure 1—figure supplement 5A) using the AA3, chitin-C1-oxidizing enzyme, ChitO (from Fusarium graminearum) (Figure 1—figure supplement 5B). The ChitO-C1-oxidized oligosaccharides eluted at timepoints similar to the new peaks from CWR-1^HG1 PMO at around 18–22 min (Figure 1—figure supplement 5B, Figure 1E). To further validate the CWR-1^HG1 C1 regioselectivity, tandem MS/MS was performed on the crude reaction from PMO^HG1 action on α-chitin in the presence of ascorbate. The resulting spectra were consistent with C1 oxidation (Figure 1—figure supplement 6). These data were in line with all other characterized AA11 PMOs that exhibit C1-oxidative activity on both α-chitin and β-chitin.

Our data showed that the PMO domain of CWR-1^HG1 was the driver for fusion arrest at the cwr allorecognition checkpoint. We therefore compared sequences of the PMO domain bioinformatically from the six different CWR-1 haplogroups. The phylogenetic tree showed identical clades to a tree created using full-length CWR-1 sequences (Gonçalves et al., 2019; Figure 2A). The intra-haplogroup similarity was very high with >95% sequence identity across the PMO domain for alleles within a haplogroup (Supplementary file 1e). All six PMO haplogroups contained the residues essential for PMO activity; two histidines in the histidine brace and residues implicated in the catalytic mechanism, including a tyrosine involved in the secondary coordination sphere and other conserved hydrogen-bonding residues (Span et al., 2017). Of note, the regions that have the most inter-haplogroup PMO domain differences corresponded to the AA9 nomenclature of LC, LS, and L2 loops (Danneels et al., 2018; Liu et al., 2018), which have been shown to play a role in substrate binding and specificity. There were also minor differences in the L8 loop located opposite the substrate-binding surface of the PMO domain (Figure 2B). The differences between the different haplogroup PMO domains were more apparent when a homology model was constructed using the crystal structure of AoAA11 (PDB: 4MAI) as a template (Figure 2C).

It is possible that the product of CWR-1 from each of the haplogroups was either generating a different product profile or recognized a different substrate. Mutations along substrate binding loops can change the regioselectivity of PMOs (C1 vs. C4) partially or entirely, and these residues can determine soluble substrate preferences (Courtade et al., 2018; Danneels et al., 2018; Liu et al., 2018; Vu et al., 2014a). The two regions that exhibited the most prominent differences were the LC and LS loops, which was consistent with the hypothesis that a substrate/product(s) could be the signal for the cwr allorecognition checkpoint (Figure 2C).

Our in vivo data indicated that the PMO domain from cwr-1^HG1 was sufficient to confer the cwr allorecognition checkpoint. We therefore tested whether the full-length CWR-1 protein from each haplogroup was sufficient for fusion arrest and for conferring allelic specificity. Full-length cwr-1 alleles from each haplogroup were expressed in a Δcwr-1Δ81Δcwr-2 strain and confronted against cwr-1^HG1 cwr-2^HG1 germlings. As expected, strains bearing cwr-1 from any of the five different haplogroups showed significantly lower fusion rates as compared to a strain carrying cwr-1^HG1 and paired with cwr-1^HG1 cwr-2^HG1 germlings (Figure 3A and D). As a control to ensure fusion machinery is not impaired, strains bearing the six different cwr-1 haplotypes showed high fusion rates with the Δcwr-1Δ81Δcwr-2 strain (Figure 3—figure supplement 1A and C).

To determine whether the PMO domain from the CWR-1 proteins from all six different haplogroups was essential for allorecognition, chimeric constructs containing the N-terminal PMO domain from each HG (2–5) were fused to the GS linker region and X278 domain of cwr-1^HG1 (Figure 3B) in a Δcwr-1Δ81Δcwr-2 strain (which also expressed cytoplasmic GFP). The resulting germlings were paired with a cwr-1^HG1 cwr-2^HG1 strain, a Δcwr-1Δ81Δcwr-2 mutant, and a Δcwr-1 cwr-2^HG1 strain. The fusion percentages between germlings carrying any of the HG2-5 cwr-1 chimeric constructs were low when paired with a cwr-1^HG1 cwr-2^HG1 strain or with the Δcwr-1 cwr-2^HG1 strain, but were high when paired with the permissive Δcwr-1Δ81Δcwr-2 mutant (Figure 3C–F, Figure 3—figure supplement 1B and C). No significant differences were observed in percentages of fusion when either full-length cwr-1 from the different haplogroups or with chimeric cwr-1 alleles with the PMO domain from the different haplogroups were paired with a cwr-1^HG1 cwr-2^HG1 germlings (Figure 3D). These experiments showed that allorecognition is mediated in trans between a cell containing cwr-1 and a cell containing cwr-2 from the different haplogroups and provide strong evidence that the PMO domain from the six different haplogroups conferred the allorecognition fusion checkpoint.

Our data above indicated that the PMO domain from each CWR-1 haplogroup was required to confer the allorecognition fusion checkpoint. These data suggested that the CWR-1 PMO domain from each haplogroup might generate a unique product distribution that would confer allorecognition. To test this hypothesis, the PMO domain from the remaining five CWR-1 haplogroups was expressed in a heterologous system, and the protein was purified using the same strategy for the expression of the PMO domain from the HG1 strain (FGSC2489) (Figure 1D and E). The CWR-1 proteins from HG2-6 ran slightly high (approximately 1–2 kDa higher than expected) on SDS-PAGE (Figure 1—figure supplement 4A) and were confirmed to be the correct mass (Figure 1—figure supplement 4B), where each deconvoluted mass was consistent with three disulfides present on the protein (±1 Da) similar to the HG1 protein. All six PMOs from the different haplogroups oxidatively degraded both the α- and β- alloforms of chitin (Figure 4A and B), with β-chitin producing more oxidized fragments than α-chitin. Additionally, each of the six CWR-1 PMO proteins from the six different haplogroups generated C1-oxidized products, bound ~1 equivalent of copper and reduced oxygen at similar rates (Figure 4C and D). None of the other substrates tested showed activity (Supplementary file 1f).

Our in vitro data showed no selectivity differences between the CWR-1 PMOs from the six different haplogroups on α- and β-chitin alloforms. We therefore tested whether these HG1-6 PMOs had differences in activity on a more physiologically relevant substrate, the cell wall from an HG1 strain (FGSC2489). This cell wall was purified and assessed as a substrate for the PMO domains from the six different CWR-1 haplogroups. New peaks that corresponded to C1-oxidized chitooligosaccharides and soluble chitooligosaccharides appeared only in the presence of both reductant and the PMO. The products from the CWR-1 PMOs from haplogroups 1,4 and 5 showed a similar pattern of four peaks in the C1-oxidized chitin fragments region (Figure 4E), whereas the CWR-1 PMO from haplogroups 2, 3, and 6 showed only two prominent C1-oxidized peaks (Figure 4E). Since this pattern did not correlate with phenotypic aspects of cell fusion incompatibility for the six CWR-1 haplogroups, we attributed this result to small rate differences between the CWR-1 PMO domains from the different haplogroups and heterogeneity in the purified cell wall suspension. As a control, cell wall components from the Δcwr-1Δ81Δcwr-2 strain were purified and used as a substrate. Although all six CWR-1 PMO haplogroups exhibited oxidized C1-oxidized products, there appeared to be two main groupings of peak patterns: CWR-1 from HG1 and 5 contained a third oxidized chitin fragment, while CWR-1 from HGs 2, 3, 4, and 6 showed only two prominent peaks (Figure 4F). These data indicated that cell walls from both HG1 and the Δcwr-1Δ81Δcwr-2 strains were similar and that differences could be attributed to small rate differences or cell wall sample heterogeneity. Efforts to identify low-abundance carbohydrates or other potential products through ESI-MS were unsuccessful.

PMO activity was also determined on purified cell walls from a self-incompatible strain (cwr-1^HG1cwr-2^HG1; cwr-1^HG2); this strain shows reduced growth, asexual spore production, and lack of self-fusion (Gonçalves et al., 2019). However, the pattern of products on this substrate for all six PMO domain haplogroups had very similar product profiles (Figure 4G). The PMO reaction products from CWR-1^HG1 and CWR-1^HG4 on chitin and purified cell walls from FGSC2489 (HG1) and JW242 (HG4) strains were also used in the germling fusion assays. However, neither the addition of the PMO reaction products nor the cell walls from the incompatible strains blocked or reduced cell fusion frequencies between compatible strains (Figure 4—figure supplement 1).

Since there were no striking differences between CWR-1 PMO haplogroups for either products or substrates, the requirement for catalytic activity in fusion blockage was tested. In Serratia marcescens and Pseudomonas aeruginosa, alanine variants in the second histidine (corresponding to His78 for CWR-1^HG1) of the histidine brace in AA10 PMOs (CBP21 and CbpD, respectively) disrupt PMO activity with no oxidized products observed (Askarian et al., 2021; Vaaje-Kolstad et al., 2010). Therefore, we constructed cwr-1^HG1 and cwr-1^HG6 variants containing His→Ala substitution of the first histidine (His20), a second set of variants containing a His→Ala substitution at H78 and double His→Ala substitutions (H20A; H78A). Similar to the WT PMOs, the variant proteins’ deconvoluted mass was consistent with three disulfides and confirmed the construct was correctly expressed (Figure 5—figure supplement 1). These CWR-1 PMO His→Ala variants were assessed for catalytic activity, copper-binding, and/or subsequent hydrogen peroxide generation. All single and double histidine variants of PMO^HG1, PMO^HG6, and the Tyr→Ala variant PMO^HG1 were inactive on β-chitin, where no oxidative activity could be detected (Figure 5A and B). PMO^HG1 H20A and H78A variants and PMO^HG6 H79A variants retained copper-binding after purification (Figure 5C). These variants were functional in an oxygen reduction assay but exhibited slightly lower rates as compared to WT CWR-1 PMO (Figure 4C). Both PMO^HG1 and PMO^HG6 His→Ala double variants greatly reduced copper-binding and had lower oxygen reduction rates (Figure 5D). X-band continuous-wave EPR spectra were acquired for the WT PMO^HG1 and PMO^H78A proteins. Both PMO^HG1 and its H78A variants bound copper; this active site copper center is in the Cu(II) oxidation state under aerobic conditions. Cu(II) has an S = 1/2 spin state, and EPR is a sensitive probe for the primary coordination sphere at the active site. The EPR spectra of the WT and H78A PMO^HG1 variant proteins were compared to identify any differences in Cu(II) binding between WT and H78A mutant PMO proteins (Figure 5—figure supplement 2). The WT PMO^HG1 spectrum is consistent with other reported PMOs (Frandsen et al., 2016) that contain axial Cu (II) sites with observed g values of g1 = 2.245, g2 = 2.065, and g3 = 2.005 with visible N–hyperfine splitting. The spectrum from the PMO^HG1 H78A variant is distinct from WT PMO^HG1, displaying more splitting g2 and g3 values, indicative of a more rhombic coordination environment: g1 = 2.245, g2 = 2.08, and g3 = 2.05. The EPR differences between WT and the PMO H78A variant suggests a significant change to the primary coordination sphere of the Cu(II) ion and is presumably because the rigid histidine brace has been changed and replaced by water molecules. Since the PMO^HG6 variant proteins were not substantially different from the PMO^HG1 variant proteins, these results likely represent all the PMO proteins in the six haplogroups.

Previously, we tested whether a tyrosine residue in the PMO^HG1 domain that is predicted to be important for catalysis was essential for allorecognition in pairings with a wild-type strain (JW199, HG5); a statistically significant level of fusion was observed in these pairings, suggesting that Tyr159 was important for allorecognition (Gonçalves et al., 2019). Here, we further assessed these findings using an HG1 strain bearing a mutation in cwr-1 (cwr-1 ^{HG1 Y159A}) in a Δcwr-1Δ81Δcwr-2 deletion strain and paired with isogenic germlings bearing a cwr-2^HG3 allele in a Δcwr-1Δ81Δcwr-2 background. In these pairings, a significant block in cell fusion was observed (Figure 5E and F). These data suggest that other genetic factors in the JW199 strain affected cell fusion frequencies when paired with the HG1 cwr-1^Y159A mutant in an otherwise FGSC2489 background. However, using isogenic strains here showed that allorecognition and a significant cell fusion block occurred between a strain bearing only CWR-1^Y159A and a strain bearing only CWR-2^HG3.

Mutations in the histidine brace resulted in chitin-inactive variants, although the purified PMO^HG1 H20A, PMO^HG1 H78A, and PMO^HG6 H79A variants could still bind copper. With the retention of copper, hydrogen peroxide could still be generated from these single His→Ala variants. We, therefore, tested whether the single histidine brace variants (H20A and H78A) and the double His→Ala variant (H20A; H78A) were affected in allorecognition and cell fusion. Isogenic fungal strains were constructed bearing HG1 cwr-1^H20A, cwr-1^H78A, and cwr-1 ^{H20A; H78A} alleles. Germlings from these strains were paired with germlings expressing cwr-2^HG3 in the cwr-1 cwr-2 deletion strain (cwr-2^HG3; Δcwr-1Δ81Δcwr-2) that also expressed cytoplasmic GFP. Contrary to expectations, the single cwr-1 histidine variants (H20A and H78A), including the inactive double mutant (H20A; H78A), still retained the ability to trigger allorecognition and block cell fusion when paired with germlings that expressed an incompatible cwr-2^HG3 allele (Figure 5E and F). As a control, all the His→Ala cwr-1^HG1 variants were paired with cwr-1^HG1 cwr-2^HG1 (FGSC2489) germlings, and all underwent robust cell fusion (Figure 1—figure supplement 3C and D). These data rule out hydrogen peroxide generation as a contributing factor in blocking cell fusion as the copper-deficient cwr-1^{H20A; H78A} variant was not different from those that retained some copper-binding. Taken together, these data suggest that the cwr fusion block checkpoint is not dependent on enzymatic activity, and the CWR-1 PMO domain is not generating a signal to block cell fusion via a substrate preference or generated product.

Our data showed that the PMO domain is essential for allorecognition, but that catalytic activity of the PMO domain was not required. An examination of a homology model of the six different CWR-1 PMO haplogroups based on the structure of AoAA11 revealed that differences appeared to be most significant in the region of the LS loop (Figure 2C). To assess whether this region of CWR-1 was important for conferring allorecognition specificity, we constructed a loop-swap chimeric construct: the first where the entire LS loop and surrounding amino acids from PMO^HG1 (I85 to G129) was replaced with the sequence from PMO^HG6 (V86 to T130) (Figure 6A). If this region of the CWR-1 protein was necessary for allelic specificity, the chimeric strains would have a higher fusion frequency with a cwr-2^HG6 (in a Δcwr-1Δ81Δcwr-2 background) strain and a lower fusion frequency with the cwr-1^HG1 cwr-2^HG1 strain or cwr-2^HG1 in the Δcwr-1Δ81Δcwr-2 background. However, the cwr-1^HG1LS ^HG6 chimera maintained the same specificity as HG1 strain (Figure 6B and C). These studies indicate that this region, which is different in all six CWR-1 PMO haplogroups, was insufficient to switch allelic specificity. A second chimeric construct was generated; this new chimera includes LS loop, L8 loop, and the first half of the LC loop. The L8 and LC regions showed differences between haplogroups in the MSA and homology models, but more so in the case of the LC region. We hypothesized that a combination of these three regions could impart specificity. The region I85 to D201 of CWR-1^HG1 was replaced with V86-D202 from CWR-1^HG6 (Figure 6A). This cwr-1^HG1/HG6 chimera showed increased fusion percentages (~55%) with cwr-2^HG6 cells, but a decreased fusion percentage (~35%) with both the HG1 strain (FGSC2489) and a cwr-2^HG1 strain in the Δcwr-1Δ81Δcwr-2 background (Figure 6D and E). These results suggest that the V86-D202 region in CWR-1 contains some but not all required residues to elicit the allorecognition response. To determine whether we had created an allele with a new cwr-1 specificity, we assessed fusion percentages of this chimera with strains bearing the other cwr-2 haplogroup alleles in a Δcwr-1Δ81Δcwr-2 background. The cwr-1^HG1/HG6 chimeric strain showed fusion percentages of ~21, ~20, and~27% with strains bearing cwr-2^HG2, cwr-2^HG3, and cwr-2^HG5, but a higher fusion percentage with a cwr-2^HG4 strain (~50%) (Figure 6E). Consistent with alterations in allelic specificity, these results were different from the cwr-1^HG1 germlings paired with cwr-2 germlings from the different haplogroups (Figure 6E). Thus, strains with the cwr-1^HG1/HG6 chimera showed the highest cell fusion percentages with the cwr-2^HG4 and cwr-2^HG6 strains but were not fully compatible with any of the strains containing cwr-2 from the different haplogroups. These data and further studies into chimeric constructs of cwr-1 in concert with chimeric constructs of cwr-2 will help to define specific residues/regions required for haplogroup specificity and allorecognition that will enable the development of molecular models for somatic cell fusion arrest via the cell wall remodeling checkpoint.

Plasmids and oligonucleotides used in this work are listed in Supplementary file 1a–c. All plasmids used to transform N. crassa were derived from plasmid pMF272 (Freitag et al., 2004). PCR reactions were performed in MiniAmp Plus Thermal Cycler with Q5 High-Fidelity DNA polymerase (2000 U/mL) (NEB, Ipswich, MA) according to the manufacturer’s instructions. Genes were amplified using genomic DNA of N. crassa strains (FGSC2489 or P4471, JW228, JW242, JW258, P4476, D111) as a template. The fragments amplified were purified using Monarch DNA Gel extraction Kit (T1020S, NEB). The cwr-1 chimeras were made by fusing two fragments, a fragment that encodes the signal peptide and the PMO domain corresponding to alleles from each of the six haplogroups of cwr-1 (JW228, JW242, JW258, P4471, P4476, D111) and a second fragment that included the region corresponding to the GS linker and X278 domain of cwr-1^FGSC2489 (haplogroup 1). The fragments were fused using polymerase Phusion high-fidelity DNA Polymerase (M0530S, NEB) following the manufacturer’s instructions. We designed primers flanking the ORF of the cwr-1 (NCU01380), cwr-2 (NCU01382), and chimeras of cwr-1 with XbaI and PacI restriction sites for the 5′ and 3′ ends, respectively. Amplified and gel-purified PCR ORFs were XbaI (R0145S, NEB) and PacI (R0547S, NEB) digested and inserted into XbaI/PacI digested pMF272::his-3-Ptef-1-V5-Tccg-1. The fragments were gel purified and ligated using T4 DNA ligase (15224-017, Invitrogen Life Technologies, Waltham, MA). The resulting constructs were used to transform N. crassa (△NCU01380△NCU01381△NCU01382::hph; his-3^-).

The plasmids to transform N. crassa with the truncated version of cwr-1 (cwr-1△GS^HG1, cwr-1△GS△X278^HG1), histidine mutants (cwr-1H20A^HG1, cwr-1H78A^HG1, cwr-1H20AH78A^HG1), and the loop chimera (cwr-1^HG1LS^{HG6(V86-T130))} were constructed using the plasmid pMF272::his-3-Pcwr-1-cwr-1(NCU01380)-Tcwr-1^FGSC2489HG1 as a backbone. According to the manufacturer’s instructions, the truncated versions of cwr-1 and the histidine mutant plasmids were prepared using the Q5 Site-Directed Mutagenesis Kit (E0552S, NEB). Gibson assembly (E2611S, NEB) cloning was used for both the loop chimera constructs and the pET22B PMO domain protein-expression constructs (for all alleles) according to the manufacturer’s instructions.

The generated plasmids were transformed into NEB 5-alpha Competent E. coli (NEB #C2987) for propagation and storage. Plasmid DNA was extracted with QIAprep Spin Miniprep Kit (QIAGEN, Redwood City, CA; Cat# 27104) and linearized with NdeI (NEB, R0111S), SspI-HF (R3132S), or PciI (NEB, R0655S) to transform N. crassa.

The standard protocols for N. crassa growth, crosses, and maintenance are available at the Fungal Genetics Stock Center (FGSC) website (http://www.fgsc.net/Neurospora/NeurosporaProtocolGuide.htm; accessed 1/23/2019). All the strains used in this work are listed in Supplementary file 1d and are available from the FGSC. The strains were grown in Vogel’s minimal medium (VMM) (Vogel, 1956) solidified with 1.5% agar with supplements as required. For crosses, Westergaard’s synthetic cross-medium was used (Westergaard and Mitchell, 1947). The FGSC2489 laboratory strain was used as a parental strain and as a WT control for all experiments. The wild N. crassa strains used in this study were isolated from Louisiana, USA (Gonçalves et al., 2019; Heller et al., 2016; McCluskey et al., 2010; Palma-Guerrero et al., 2013).

Conidia of the auxotrophic histidine strain in the triple delete background (ΔNCU01381ΔNCU01381NCU01382::hph; his-3^-) (Gonçalves et al., 2019) of N. crassa were transformed with linearized plasmid or PCR products by electroporation on a Bio-Rad Pulse controller plus and Bio-Rad gene pulser II. The conidia were electroporated using 1 mm gap cuvettes (Bio-Rad Gene Pulser/MicroPulser Cuvette; Bio-Rad, Hercules, CA) at 1.5 kV, 600 ohm, 25 μF as previously described (Margolin et al., 1997). For each transformation, around 30 (His⁺) prototroph transformants were selected and transferred to tubes containing VMM with Hygromycin B. The integration of the constructs in the selected transformants was confirmed by using F-130WH Phire Plant Direct PCR Kit (Thermo Fisher Scientific, Waltham, MA) and Sanger sequencing. Hygromycin B (10687010, 50 mg/mL; Thermo Fisher Scientific) was used at a final 200 µg/mL concentration for selection of transformants. Cyclosporin A (30024, Sigma, St. Louis, MO) was used at a final concentration of 5 µg/mL for selection. The L-histidine (L-histidine hydrochloride monohydrate, 98%, Acros Organics, Waltham, MA) was used at a final concentration of 0.5 mg/mL to support the growth of his-auxotrophic strains.

Heterokaryotic transformants were used as a female parent strain and were grown on Westergaard’s synthetic cross-medium at 25°C with constant light until protoperithecia were observed. The protoperithecia were fertilized with a conidial suspension from a male parental strain, FGSC9716/his^- or FGSC2489::GFP/his^- (FGSC2489; csr-1::Pccg-1-gfp; his-3^-), and grown for an additional 10 days or until ascospores had been shot onto the lid of the Petri dish. To select homokaryons, the resulting ascospores were heat-shocked at 60°C for 40 min and inoculated onto bottom agar plates (VMM with 1.5% agar and a mixture of 20% sorbose, 0.5% fructose, 0.5% glucose) and incubated at 30°C overnight. The germinated ascospores were dissected using a stereomicroscope Stemi SV6 Zeiss and transferred to slants with VMM. The selection of the homokaryons was made depending on their (His⁺) prototrophy, Hygromycin B (Hyg⁺), and/or Cyclosporin A (Cyclosporin⁺) resistance. Additionally, PCR was done to confirm the presence of the construct using F-130WH Phire Plant Direct PCR Kit (Thermo Fisher Scientific) and Sanger sequencing.

The strains were grown in VMM in slant tubes for 4 days at 30°C; then, the tubes were placed at room temperature (RT) with constant light for 2 days. The conidia were resuspended in 2 mL of sterile ddH₂O and filtered using cheesecloth. The styryl dye N-(3-triethylammoniumpropyl)–4-(6-(4-(diethylamino) phenyl) hexatrienyl) pyridinium dibromide (FM4-64, 514/670 nm absorption/emission Invitrogen, Waltham, MA) was used at a final concentration of 16.5 μM in ddH₂O from a 16.5 mM stock solution in DMSO, to stain germlings of N. crassa. An aliquot of 200 µL of the filtered conidia was stained with 40 µL of FM4-64 (16.5 µM) and incubated at RT for 15 min in the dark. The stained conidia were centrifuged in a microcentrifuge at 5000 rpm for 2 min, the supernatant was removed, and the conidia were washed twice with 1 mL of sterile ddH₂O. The conidia were resuspended in 100 µL of sterile ddH₂O and counted using the hemocytometer. The conidial suspension was adjusted to 3 × 10⁷ conidia/mL. An aliquot of 45 µL of the conidial suspension from the strain stained with FM4-64 was mixed with 45 µL of a conidial suspension from the strain that expresses cytoplasmic GFP. Then, 80 µL of the final mixed suspension was spread on VMM agar plates (60 mm × 15 mm). Plates were incubated for 3.5 hr at 30°C. Agar rectangles of 3 cm × 2 cm approximately were excised and observed with a microscope ZEISS Axioskop 2 MOT. The microscope is equipped with an Illuminator Cool LED microscope, simply Better Control pE-300 white. The images were captured with a Q IMAGING FAST 1394 COOLED MONO 12 BIT microscope camera RETIGA 2000R SN:Q31594 (01-RET-2000R-F-M-12-C) viewed through a Ph3 ×40/1.30 ∞/0.17 Plan-Neofluar oil immersion objective and then processed using iVision-Mac Scientific Image Processing Bio Vision Technologies (iVision 4.5.6r4). The fusion percentage of germlings was evaluated through observation of cytoplasmic mixing of GFP into germlings stained with FM4-64. Pictures for at least 15 fields were taken (for each filter DIC, Blue, GreenRed), depicting a total of at least 100 germling-contact events, with three biological replicates. Quantification is shown with error bars as SD. All the images were further processed using Fiji (ImageJ2, version 2.3.0/1.53f; Schneider et al., 2012), Adobe Photoshop 2022 version 23.1.0, and Adobe Illustrator 2022 version 26.0.2. ANOVA statistical analyses were performed using GraphPad Prism 9. ANOVA statistical analysis was selected to assess differences in data between different strains/genotypes. One-way ANOVA for the analysis that involves one independent variable and two-way ANOVA for analysis with two independent variables was followed by Tukey’s test to perform multiple comparisons to assess significant differences. All p-values and 95% CIs for each analysis are reported in p-values source data.

Each cwr-1 allele was amplified from genomic DNA and cloned into the pet22B plasmid containing the pelB leader sequence. Each construct contained the PMO catalytic domain of the cwr-1 gene with the signal peptide removed, followed by a C-terminal hexa-histidine tag with a thrombin cleavage site. The constructs were transformed into E. coli BL21 Ros2pLysS cells (QB3, Macrolab, Berkeley, CA). Dense overnight cultures were grown in 100 µg/mL ampicillin (RPI), 20 µg/mL chloramphenicol (RPI) in LB, and then 12 mL were added to 1 L of TB that contained 100 µg/mL ampicillin and 20 µg/mL chloramphenicol. At A600 of ~0.6–0.8, IPTG was added to a final concentration of 100 µM, and cells were grown overnight at 18°C. Each pellet was then resuspended in 1:4 weight/volume ratio of 50 mM Tris, 30% sucrose, pH 8.0, 1 mM AEBSF, 1 mM benzamidine, 0.1 mg/mL DNAase, and 0.1 mg/mL lysozyme. This suspension was nutated for 30 min at 4°C and then centrifuged at 5000 × g for 55 min. The supernatant was decanted, and then the pellet was resuspended with 1:4 g/mL of ice-cold 5 mM Tris 5 mM MgCl₂ pH 8.0. This suspension was nutated for 30 min at 4°C and then centrifuged at 8000 × g for 10 min. The sucrose and Mg supernatants were decanted and further clarified if needed by centrifuging at 4000 × g for 30 min and then combined. The combined supernatant was then run over a gravity column of His60 (Takara, Kusatsu, Shiga, Japan) resin. The column was washed with 20 column volumes (CV) of Buffer A (10 mM imidazole, 50 mM NaPO₄, 300 mM NaCl, pH 8.0), then eluted with 5 CV of a stepwise gradient of 40 mM, 70 mM, 125 mM, and 250 mM imidazole. The buffer with 250 mM imidazole is herein referred to as Buffer B. The pure fractions determined by SDS-PAGE Stain-Free gel (Bio-Rad) were combined and concentrated to ~1 mL and dialyzed using 7 KMWCO Snake Skin dialysis tubing (Thermo Scientific) into 20 mM Tris, 150 mM NaCl, pH 8.4. Thrombin agarose beads (15 µL) (Sigma) were washed 3× with 500 µL of Buffer C (20 mM Tris, 150 mM NaCl, 2.5 mM CaCl₂, pH 8.4), where the last wash of the resin was incubated for 30 min to activate the thrombin. The dialyzed protein was added to the beads and was gently rocked overnight at RT. The solution was centrifuged for 1 min at 1000 × g and then washed with 2 × 1 mL of Buffer B. The washes and supernatant were combined and run over a His60 bed. The resin was washed with 10 CV of Buffer A and 10 CV of Buffer B. The flowthrough and the Buffer A fractions were combined, concentrated, and dialyzed into 50 mM NaOAc, 50 mM NaCl, pH 5.5, then into 50 mM NaOAc, 50 mM NaCl, 5 µM CuSO₄, pH 5.5, and then into 50 mM MOPS, 50 mM NaCl, pH 7.0. The protein was then stored at 4°C and showed negligible loss of chitin-degrading activity over several months. A protein gel of each of the haplogroup 1–6 purified PMO domains is provided in Figure 1—figure supplement 4—source data 1.

Protein was diluted to 20 µM and digested in 2% nitric acid overnight at RT in 1.7 mL tubes (Starstedt, Nümbrecht, Germany). The samples were then centrifuged to remove insoluble material at 20,000 × g for 10 min, and the supernatant was transferred to new tubes. Samples were analyzed on a Thermo Fisher iCAP Qc ICP mass spectrometer in kinetic energy discrimination (KED) mode against a standard curve of known copper concentrations (CMS-5, Inorganic Ventures, Christiansburg, VA), with Ga (20 μg/L, Inorganic Ventures) as an internal standard. Each experiment was carried out in technical triplicate. Data are plotted with SEM error bars as calculated by GraphPad Prism 9.2.0. ICP data is provided in Figure 1—source data 2, Figure 4—source data 3, and Figure 5—source data 3.

Cell wall material was purified as previously described (Maddi et al., 2009) with slight modifications: three 1 L flasks of 200 mL, 1.5% agar, 2% sucrose, 1× Vogel’s salts were inoculated with conidia from FGSC2489, JW242, Δcwr-1Δ81Δcwr-2, and FGSC2489; cwr-1^{D111 HG5} strains and allowed to grow at RT for 7 days with shaking (200 rpm). The conidia were separated from the mycelia with ~200 mL of sterile Milli-Q water and filtered through four layers of cheesecloth. The solution was then centrifuged at 1200 × g for 5 min, washed by resuspending the conidia in 15 mL of PBS, and then centrifuged at 1200 × g for 5 min. The conidia were then resuspended in PBS to a 1.0 × 10⁸ conidia/mL titer. The conidia were lysed in a bead beater (BioSpec, Bartlesville, OK) with 0.1 mm glass beads for 5 × 1 min while resting on ice for 2 min in between cycles. Then the mixture was filtered using cheesecloth to remove the beads, and the disrupted conidia were resuspended in 15 mL of PBS. The solution was then centrifuged at 4000 × g for 5 min, decanted, and resuspended in 15 mL of PBS with 10% SDS and centrifuged at 4000 × g for 5 min. The supernatant was decanted and conidia resuspended in fresh 15 mL of PBS with 10% SDS. The suspension was boiled for 15 min and cooled at RT. The solution was then centrifuged at 10,000 × g for 5 min, washed two times with 15 mL of PBS, and then two times with 15 mL Milli-Q water. After gently pipetting away excess water, the disrupted conidia were frozen in liquid nitrogen and lyophilized overnight on a Labconco Freezone 2.5 plus Lyophilizer (Labconco, Kansas City, MO). The dried conidial cell wall material was brought to RT before use.

Chitin (80 mg; α-chitin from powdered shrimp shells; Sigma) and β-chitin (Mahtani Chitosan Pvt. Ltd., Veraval, India) were dispersed and sonicated in 100 mM EDTA for 30 min and washed with Milli-Q water by first resuspending with 1 mL of water and then centrifuging at 1000 × g for 10 min. The supernatant was then discarded. The wash procedure was performed five times; then, the resulting chitin was resuspended to a stock concentration of 80 mg/mL. Other substrates examined were tested at 20 mg/mL including PASC that was prepared as previously described (Wood, 1988). Purified cell walls were resuspended in 50 mM MOPS, pH 7.0. Then, 10 µM PMO was incubated with 20 mg/mL of the substrate in 50 mM MOPS, pH 7.0. The reactions were initiated with 1 mM ascorbate and incubated for 1 hr, shaking at 37°C. The samples were centrifuged at 20,000 × g at 4°C for 10 min, and the supernatant was transferred into new tubes to stop the reaction by removing from substrate. The products were analyzed with HPAEC-PAD using a Dionex ICS-5000 system (Thermo Fisher) with a gradient of three buffers: (A) 10 mM NaOH, (B) 100 mM NaOH, and (C) 100 mM NaOH 500 mM NaOAc. Buffers were prepared as dictated by the Thermo Fisher manual, and the linear gradient (5 curve) was as follows: 0.4 mL/min flow rate: 15–50% B for 15 min; 50–40% B, 5–60% C to 40% B, 60% C for 20 min; isocratic 100% C for 5 min; then 15% B for 15 min. A gold electrode in carbohydrate-quadrupole mode was used as the detector. The autosampler was kept at 4°C to preserve sample integrity. Each protein and substrate combination were analyzed at least in biological triplicate. The HPAEC-PAD source data are provided in Figure 1—source data 1, Figure 4—source data 1, Figure 4—source data 2, Figure 4—source data 5, Figure 4—source data 6, Figure 4—source data 7, Figure 5—source data 1, Figure 5—source data 2, and Figure 4—figure supplement 1—source data 1.

The standards were generated as previously described (Loose et al., 2014) with slight modifications. 3 mM of chitooligosaccharides (DP 3–6) (Megazyme, Wicklow, Ireland) were treated with 0.12 mg/mL of the AA3 from F. graminearum, ChitO (Gecco Biotech, Groningen, The Netherlands) in 50 mM MOPS, pH 7.0, overnight at RT and then aliquoted and stored at –20°C until use. Source data are provided in Figure 5—figure supplement 1—source data 1 and Figure 1—figure supplement 5—source data 2.

Samples of oligosaccharides were analyzed using a 1200 series liquid chromatography (LC) system (Agilent Technologies, Santa Clara, CA) that was connected in line with an LTQ-Orbitrap-XL mass spectrometer equipped with an electrospray ionization (ESI) source (Thermo Fisher Scientific). The LC system was equipped with a reversed-phase analytical column (length: 150 mm; inner diameter: 1.0 mm; particle size: 5 µm; Viva C18, Restek, Bellefonte, PA). Acetonitrile, formic acid (Optima LC-MS grade, 99.5+%, Fisher, Pittsburgh, PA), and water purified to a resistivity of 18.2 MΩ·cm (at 25°C) using a Milli-Q Gradient ultrapure water purification system (Millipore, Billerica, MA) were used to prepare LC mobile phase solvents. Solvent A was 99.9% water/0.1% formic acid, and solvent B was 99.9% acetonitrile/0.1% formic acid (volume/volume). The elution program consisted of isocratic flow at 1% B for 5 min, a linear gradient to 95% B over 1 min, isocratic flow at 95% B for 4 min, a linear gradient to 1% B over 0.5 min, and isocratic flow at 1% B for 19.5 min, at a flow rate of 150 µL/min. The column compartment was maintained at 25°C, and the sample injection volume was 10 µL. Full-scan, high-resolution mass spectra were acquired in the positive ion mode over the range of mass-to-charge ratio (m/z) = 150–2000 using the Orbitrap mass analyzer, in profile format, with a mass resolution setting of 100,000 (at m/z = 400, measured at full width at half-maximum peak height [FWHM]). For tandem mass spectrometry (MS/MS or MS²) analysis, selected precursor ions were fragmented using collision-induced dissociation (CID) under the following conditions: MS/MS spectra acquired using the Orbitrap mass analyzer, in centroid format, with a mass resolution setting of 15,000 (at m/z = 400, FWHM), normalized collision energy 35%, activation time 30 ms, and activation Q 0.25. Mass spectrometry data acquisition and analysis were performed using Xcalibur software (version 2.0.7, Thermo Fisher Scientific). Source data are provided in Figure 1—figure supplement 6—source data 1.

Acetonitrile (Optima grade, 99.9%, Fisher), formic acid (1 mL ampules, 99+%, Pierce, Rockford, IL), and water purified to a resistivity of 18.2 MΩ·cm (at 25°C) using a Milli-Q Gradient ultrapure water purification system (Millipore) were used to prepare mobile phase solvents S3 for LC-MS. ESI-MS of proteins was performed using an Agilent 1260 series liquid chromatograph outfitted with an Agilent 6224 time-of-flight (TOF) LC-MS system (Santa Clara, CA). The LC was equipped with a Proswift RP-4H (monolithic phenyl, 1.0 mm × 50 mm, Dionex) analytical column. Solvent A was 99.9% water/0.1% formic acid (v/v) and solvent B was 99.9% acetonitrile/0.1% formic acid (v/v). PMOs were buffer exchanged into 25 mM ammonium bicarbonate buffer pH 7.5 using Biospin 6 (Bio-Rad) columns according to the manufacturer’s protocol and then spinning through a 0.22 µm cellulose acetate centrifugal spin filter. Samples of 1–5 µL were injected onto the column, corresponding to 10 pmol of protein. Following sample injection, a 5–100% B elution gradient was run at a 0.30 mL/min flow rate over 8 min. Data were collected and analyzed by deconvolution of the entire elution profile (using Agilent Mass Hunter Qualitative Analysis B.05.00) to provide reconstructed mass spectra. Spectra were analyzed with open-source Chartograph software (http://chartograph.com/; accessed 3/21/21) and compared to the mass predicted by Benchling (http://chartograph.com/; accessed 3/21/21). Source data are provided in Figure 1—figure supplement 4—source data 2 and Figure 5—figure supplement 1—source data 1.

The formation of H₂O₂ from PMO activity in the absence of polysaccharide substrate was measured using a horseradish peroxidase (HRP)-coupled assay that oxidized Amplex red to produce a colorimetric readout. A PMO sample (2 µM) was incubated at RT with 100 µM Amplex red 1.3 µM HRP (Sigma), 2 mM ascorbate in 50 mM MOPS, pH 7.0. The absorbance at 540 nm was measured at 30 s intervals for 10 min on SpectraMax340 spectrophotometer (Molecular Devices, San Jose, CA). Each reaction was done in at least biological triplicate. Data are plotted with SEM error bars as calculated by GraphPad Prism 9.2.0. Data for HRP assays are provided in Figure 1—source data 3, Figure 4—source data 4, and Figure 5—source data 4.

The sequence of each haplogroup was used without the signal peptide as predicted by SignalP 5.0 (http://www.cbs.dtu.dk/services/SignalP/; accessed 4/24/2021) and modeled in Swiss Prot using the copper-bound structure of AoAA11 (PDB 4MAI) as the template. The resulting models were overlaid and visualized in ChimeraX 1.1.

The protein sequence of cwr-1^HG1 (NCU01380) was retrieved from the UniProtKB database and used as a search sequence in the EFI-EST website (Gerlt et al., 2015; accessed 4/25/2021) using the blast option with a maximum of 10,000 hits and an e value of 5. These retrieved sequences (~1700) were used to generate an SSN. An alignment score of 63 was used to generate and visualize the SSN in Cytoscape. The SSN data set is available in Supplementary file 1g. Source raw data for the SSN are available in Figure 1—figure supplement 2—source data 1. ClustalW (accessed 4/25/2021) was used to generate an MSA of the cwr-1 haplogroup sequences and determine intra-haplogroup sequence ID%.

EPR spectra were acquired of 225 µM of CWR-1^HG1 and 540 µM of CWR-1^H78A in 50 mM MOPS, 10% glycerol, pH 7.0 at 40 K. X-band continuous-wave (CW) EPR spectra were recorded on a Bruker ELEXSYS E500 spectrometer equipped with a cylindrical TE011-mode resonator (SHQE-W), an ESR 900-liquid helium cryostat, and an ITC-5 temperature controller (Oxford Instruments, Abingdon, UK). Spectra were acquired under slow-passage nonsaturating conditions.