Different constructs of MoHTR2 without signal peptide (MoHTR2Δsp) and MGG_13063Δsp were PCR amplified from the cDNA of the wild-type strain (KJ201) of M. oryzae using the following primers (Supplementary Table 1) and Platinum SuperFi II Green PCR Master Mix (Thermo Fisher Scientific, Waltham, MA, USA). The amplified constructs were purified using the MEGAquick-spin Plus Total Fragment DNA Purification Kit (iNtRON Biotechnology, Seongnam, Korea). For transient expression in rice protoplasts, the purified constructs were first cloned into the pENTR/D-TOPO gateway vector (Thermo Fisher Scientific) and then reintroduced into the final eGFP expression vector p2GWF7 (Karimi et al., 2002) using the Gateway LR Clonase II Enzyme Mix (Invitrogen, Carlsbad, CA, USA). Before the protoplast transfection assay, the final plasmids were prepared using the PureLink HiPure Plasmid Midiprep Kit (Invitrogen).

Peeled rice seeds (Oryza sativa cv. Nakdongbyeo) were surface-sterilized using 70% ethanol and 2% NaOCl. We planted 25-30 sterilized seeds in 100 mL 1/2 MS medium and incubated them in a dark chamber at 25°C for 10-14 days. We then isolated rice protoplasts using a previously described method (Lim et al., 2024; Zhang et al., 2011). We finely chopped the 10-14-day-old rice tissues and incubated with lysing enzyme solution (containing 1.5% cellulase and 0.75% Macerozyme R-10) for 3 h to extract protoplast of 2 × 10⁶ cells/mL. For a transient expression of plasmid, we used the PEG-mediated transfection method by adding 5 μg of eGFP tagged MoHTR2 variants and 5 μg of nucleus marker, OsABF1::mRFP in 100 μL of rice protoplast. After 16 h, we observed at least 120 rice protoplasts per experiment under a fluorescence microscope (Leica DM6000B, Leica Microsystems, Wetzlar, Germany).

To find the NLS and the core motif in NLS of MoHTR2, we performed site-directed mutagenesis using Golden mutagenesis (Püllmann et al., 2019), QuickChange Site-Directed Mutagenesis Kit (Agilent, Santa Clara, CA, USA), and GeneArt Site-Directed Mutagenesis PLUS System (Thermo Fisher Scientific), following the manufacturer’s instruction. The primers used for the individual site-directed mutagenesis are provided in the supplementary data (Supplementary Table 1).

For the bait, pENTR/D-TOPO::MoHTR2Δsp was cloned to pDEST32, respectively. For the prey, three rice importin αs and three rice importin βs, selected from the Rice Genome Annotation Project Database (https://rice.uga.edu/) (Peng et al., 2022), were used. OsImpα (Os01g24060), OsImpα1a (Os01g14950), OsImpα1b (Os05g06350), OsImpβ_1 (Os03g18350), OsImpβ_2 (Os12g38110), and OsImpβ1 (Os05g28510) were PCR amplified from the rice cDNA, and were cloned to pENTR/D-TOPO, then finally LR reacted to pDEST22. The gateway recombination reactions were performed using Gateway LR Clonase II Enzyme Mix (Invitrogen). The bait and prey plasmids were co-transformed into the yeast strain MaV203, following the manual provided by ProQuest Two-Hybrid System (Invitrogen). The transformation was confirmed by growing the transformed yeast on SC-Leu-Trp (SC-LT) agar media at 30°C for 3 days. A single colony was selected and regrown in YPAD broth in the shaking incubator at 30°C, 150 rpm, overnight, then washed twice with 0.9% NaCl. The protein-protein interaction was tested on SC-Leu-Trp-His + 25 mM 3-amino-1,2,4-triazole (SC-LTH + 3-AT) agar plate via serial drop inoculation.

Out of the 1,899 proteins known to be secreted by M. oryzae, according to the Fungal Secretome Database, we previously identified 440 proteins with predicted NLS using three NLS prediction tools: WoLF PSORT, cNLS mapper, and NLStradamus (Lim et al., 2024). Among these, we identified five proteins with an ‘HH’ motif in their predicted NLS. Of these five proteins, we successfully amplified three protein encoding genes from KJ201 cDNA and performed quantitative reverse transcription polymerase chain reaction (qRT-PCR) to examine effector characteristics.

KJ201 mycelia were cultured in CM broth at 25°C in a shaking incubator set at 150 rpm for 3 days. The collected mycelia were freeze-dried and ground into a fine powder using the FastPrep-24TM 5G homogenizer (MP Biomedicals, Irvine, CA, USA). The powdered sample was treated with genomic DNA (gDNA) extraction buffer (100 mM EDTA, 100 mM Tris-HCl, 250 mM NaCl, 100 μg/mL RNaseA, and 100 μg/mL proteinase K). The mixture was incubated at 50°C for 1 h, followed by adding 10% N-lauroylsarcosin and a second incubation at 55°C for 1 h. The lysate was centrifuged, and the supernatant was treated with 1:1 phenol:chloroform solution, followed by incubation on a rocking shaker at 15 rpm for 15 min. After a second centrifugation, the supernatant was treated with a 24:1 chloroform: isoamyl alcohol solution. Following a third centrifugation, the pellet was washed sequentially with isopropanol and ethanol and then eluted with water.

To generate ΔMohtr2:: MoHTR2^ΔIHH, we first generated ΔMohtr2 protoplast. The knockout strain, ΔMohtr2, was obtained from the previous study (Kim et al., 2020). We grew ΔMohtr2 in a complete medium broth for 3 days in the shaking incubator at 25°C, 200 rpm. Afterwards, we treated the mycelia with CW-lytic F1 (Amicogen, Jinju, Korea) and washed it with 2× STC buffer using the previously described method (Lee et al., 2023). To make the MoHTR2^ΔIHH construct, we extracted the gDNA from KJ201 using the protocol mentioned above. By using KJ201 gDNA as a template, we amplified the upstream and downstream of MoHTR2 ORF, which excludes the protein sequence ‘IHH’ in the middle, using Platinum SuperFi II Green PCR Master Mix (Thermo Fisher Scientific) and Phusion Hi-Fi DNA polymerase (Invitrogen). The geneticin cassette from PII99 (Namiki et al., 2001) was used as a selection marker. Before use, all constructs were purified with MEGAquick-spin Plus Total Fragment DNA Purification Kit (iNtRON Biotechnology). The prepared construct and geneticin cassette were introduced into ΔMohtr2 protoplast using the PEG-mediated transformation method. The complemented strains were selected from a TB3 agar plate containing 800 ppm geneticin.

For sheath inoculation, KJ201, ΔMohtr2, and ΔMohtr2::MoHTR2^ΔIHH were each grown on V8 agar for 6 days. 2 × 10⁴/mL conidial suspension was applied to the sheath isolated from 6-week-old Nakdongbyeo. At 48 hour post-inoculation (hpi), we observed invasive hyphal growth at least 120 infection sites per experiment under a microscope (Leica DM6000B). For spray inoculation, KJ201, ΔMohtr2 and ΔMohtr2::MoHTR2^ΔIHH were each grown on oatmeal agar for 10 days. 10⁵/mL conidial suspension with 250 ppm Tween 20 (Bio-Rad, Hercules, CA, USA) was sprayed onto 4-week-old Nakdongbyeo and kept in a dark dew chamber with 100% humidity overnight. The inoculated Nakdongbyeo was further grown in a plant growth chamber under 25°C, 70% humidity. At 6 days post-inoculation (dpi), we collected inoculated leaves and the diseased area was calculated using ImageJ (Schneider et al., 2012).

We extracted RNA from mycelia and from rice infected with wild type, ΔMohtr2, and ΔMohtr2::MoHTR2^ΔIHH using the easy-spin Total RNA Extraction Kit (iNtRON Biotechnology), following the manufacturer's instructions. After confirming the purity of extracted RNA using Nanodrop 1000 (Thermo Fisher Scientific), we synthesized cDNA using the ImProm-II Reverse Transcription system (Promega, Madison, WI, USA). Each qRT-PCR reaction containing 2× Rotor-Gene SYBR Green PCR master mix (Qiagen, Hilden, Germany), 25 ng cDNA and 15 pmol of forward and reverse primer (Supplementary Table 1) was run in Rotor-Gene Q real-time PCR cycler (Qiagen). The gene expression level was quantified using the Delta-Delta Ct method (Livak and Schmittgen, 2001).

Proteins with cNLSs are known to bind with importin α for nuclear import. The importin-mediated pathway is the most extensively studied mechanism for delivering NLS-carrying proteins into the nucleus (King et al., 2020). A previous study demonstrated that MoHTR1 interacts with rice importin αs for nuclear localization (Lim et al., 2024). To investigate whether MoHTR2 also utilizes the same nuclear localization transport mechanism, we performed a yeast two-hybrid assay to test its interactions with rice importin αs. We used MoHTR2 as bait and importin subunits as prey. The results showed that serially diluted yeast transformants failed to grow on SC-LTH + 25 mM 3-AT medium, indicating that MoHTR2 does not interact with the importin α subunits (OsImpα, OsImpα1a, and OsImpα1b) in vitro (Fig. 1A). Some NLS-carrying proteins are known to directly interact with importin β, without the presence of importin α, for nuclear localization (Cingolani et al., 2002). Therefore, we further examined the interaction between MoHTR2 and three rice importin βs (OsImpβ_1, OsImpβ_2, and OsImpβ1) in vitro. None of importin β subunits interacted with MoHTR2 (Fig. 1B). These findings suggest that MoHTR2 does not rely on classical importin α-mediated pathways for nuclear localization, unlike MoHTR1.

MoHTR2 is a 110-amino-acid (a.a.) protein comprising an N-terminal signal peptide and a C₂H₂ zinc finger domain (residues 68-101). A predicted NLS, LKPSKG, is located within the zinc finger domain, as identified by the NLS prediction tool, cNLS mapper (Fig. 2A). To understand whether this predicted sequence functions as a true NLS, we truncated MoHTR2Δsp into N-terminal (19-63 a.a.) and C-terminal (64-110 a.a.) segments, tagged each with eGFP, and compared their localization in rice protoplasts. The transfection assay reveals that MoHTR2_N localizes in the nucleus in 59 % of transfected rice protoplasts, comparable to the nuclear localization of full-length MoHTR2, which is also 59%. In contrast, MoHTR2_C, which contains the predicted NLS, displayed a much reduced nuclear localization of 35% (Fig. 2B and C). Having confirmed that the predicted NLS of MoHTR2 does not function as a true NLS, we performed serial truncations by adding fragments of MoHTR2_N to MoHTR2_C to identify the location of a potential NLS. After adding the last 15 amino acids of MoHTR2_N (49-63 a.a.), shown as MoHTR2_C_2, we observed 60% nuclear localization, an increased nuclear localization compared to MoHTR2_C (35%). To further confirm that no additional amino acids are required for nuclear transport, we added the last 30 amino acids of MoHTR2_N and examined the localization. MoHTR2_C_1 (34-110 a.a.) has 58% nuclear localization. Taken together, our data indicate that the predicted NLS of MoHTR2 does not function as a true NLS. Instead, the functional NLS is likely located near the N-terminus, between 49-63 a.a.

Basic amino acids such as lysine (K), arginine (R), and histidine (H) are known to play essential roles in NLS (Chaudhary et al., 2014). To investigate their role in the potential NLS of MoHTR2, we substituted three arginine residues (R56, R60, and R63) and two histidine residues (H53 and H54) with alanine (A) and observed changes in nuclear localization. The mutation of the three arginines, MoHTR2_R-A, did not affect the nuclear localization of MoHTR2. However, the additional mutation of the two histidine residues, MoHTR2_H, R-A, reduced nuclear localization to 32% (Fig. 3A and B). To further confirm the roles of the double histidine in the NLS of MoHTR2, we performed serial site-directed mutagenesis of 49-63 a.a. in MoHTR2 by sequentially replacing three amino acids with three alanine residues. The mutation of the first three amino acids, MoHTR2_3aa-A, caused no significant change in nuclear localization compared to MoHTR2 by having 66% nuclear localization. However, with the mutation of three additional amino acids containing the double histidine, MoHTR2_6aa-A, nuclear localization decreased to 43%, similar to the level of MoHTR2_C (33%) (Fig. 4A and B). To consolidate the core NLS, we further observed the nuclear localization of MoHTR2_9aa-A. Three extra mutations had similar levels of nuclear localization to MoHTR2_3aa-A. Therefore, these findings suggest that ⁵³HH⁵⁴ is the core NLS of MoHTR2.

The roles of lysine (K) and arginine (R) in NLS have been studied in depth, as they are common basic amino acids found in cNLS motifs (Lu et al., 2021). However, the role of histidine (H) in NLS function has not been thoroughly investigated. To address this gap and investigate the possible role of double histidine in NLS of other nuclear effectors, we identified M. oryzae nuclear effector candidates that contain a double histidine within its predicted NLS from the Fungal Secretome Database using NLS prediction tools (Supplementary Table 2). Among the HH-containing effector candidates, we revealed that MGG_13063 exhibits a significant increase in expression during the infection stage by performing qRT-PCR (Fig. 5A). MGG_13063 has two predicted NLSs with double histidine within the NLS predicted by NLStradamus and WoLF PSORT (Fig. 5B). Furthermore, it predominantly localizes in the rice nucleus, showing 63% nuclear localization (Fig. 5C and D). To determine whether the double histidine contributes to the nuclear trafficking of MGG_13063, we performed site-directed mutagenesis. Mutation of the two histidines (H277 and H278) reduced nuclear localization to 39% (Fig. 5C and D). These results established that double histidine is required for the nuclear localization of MGG_13063, implying that double histidine also plays an important role in the NLS of other fungal nuclear effectors.

We conducted pathogenicity assays on rice to further investigate the role of MoHTR2’s core NLS. First, we inoculated 6-week-old rice sheath cells with 2 × 10⁴/mL conidial suspension of the wild type, ΔMohtr2, and two independent ΔMohtr2::MoHTR2^ΔIHH strains. After 48 hpi, we classified the infected cells into four disease types based on the degree of invasive hyphal growth originating from the initially penetrated cells. A previous study highlighted that ΔMohtr2 exhibits decreased invasive hyphal growth compared to the wild type (Kim et al., 2020). Similarly, we observed reduced pathogenicity of ΔMohtr2 compared to the wild type, as indicated by 38% of Type I and 52% of Type II cells for the wild type and 60% of Type I and 34% of Type II cells for the ΔMohtr2. Upon inoculation with the two independent ΔMohtr2::MoHTR2^ΔIHH strains, we found 59% of Type I cells and 33-35% of Type II cells, indicating fewer Type II cells and more Type I cells compared to the wild type that is similar to ΔMohtr2 (Fig. 6A). To further examine pathogenicity, we performed spray inoculation by applying 10⁵/mL conidial suspension onto 4-week-old rice seedlings and measured the diseased area 6 dpi. The wild type exhibited a diseased area of 27%, and ΔMohtr2 displayed a reduced diseased area of 14%, showing a trend consistent with previous findings (Kim et al., 2020). In our work, similar to ΔMohtr2, ΔMohtr2::MoHTR2^ΔIHH demonstrated a reduced percentage of diseased area (11%) compared to the wild type (Fig. 6B). Since the ΔMohtr2::MoHTR2^ΔIHH strain exhibited markedly reduced pathogenicity compared to the wild type, we assessed the expression levels of pathogenesis-related (PR) genes in rice, PR1a, PR1b, PR2, and PR10a, previously reported to be upregulated in ΔMohtr2-infected rice (Kim et al., 2020). Expression of these PR genes was significantly increased not only in ΔMohtr2-infected rice but also in ΔMohtr2::MoHTR2^ΔIHH-infected rice compared to the infected rice with wild type (Fig. 6C). These results underscore that the double histidine motif, the core NLS of MoHTR2, affects the pathogenicity of M. oryzae.