(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Autophagy and cell wall integrity pathways coordinately regulate the development and pathogenicity through MoAtg4 phosphorylation in Magnaporthe oryzae [1]

['Pusheng Guo', 'Department Of Plant Pathology', 'College Of Plant Protection', 'Nanjing Agricultural University', 'Key Laboratory Of Integrated Management Of Crop Diseases', 'Pests', 'Ministry Of Education', 'Nanjing', 'The Key Laboratory Of Plant Immunity', 'Yurong Wang']

Date: 2024-02

Autophagy and Cell wall integrity (CWI) signaling are critical stress-responsive processes during fungal infection of host plants. In the rice blast fungus Magnaporthe oryzae, autophagy-related (ATG) proteins phosphorylate CWI kinases to regulate virulence; however, how autophagy interplays with CWI signaling to coordinate such regulation remains unknown. Here, we have identified the phosphorylation of ATG protein MoAtg4 as an important process in the coordination between autophagy and CWI in M. oryzae. The ATG kinase MoAtg1 phosphorylates MoAtg4 to inhibit the deconjugation and recycling of the key ATG protein MoAtg8. At the same time, MoMkk1, a core kinase of CWI, also phosphorylates MoAtg4 to attenuate the C-terminal cleavage of MoAtg8. Significantly, these two phosphorylation events maintain proper autophagy levels to coordinate the development and pathogenicity of the rice blast fungus.

Autophagy is a conserved cellular degradation and recycling process important in maintaining proper cellular homeostasis whereas the cell wall integrity (CWI) signaling pathway is also a conserved regulatory pathway in stress response. CWI signaling not only operates alone but also functions through crosstalk with autophagy during fungal infection of host plants. Despite the importance, little is known about how autophagy and CWI signaling coordinate their regulatory roles. Here, we have found that autophagy-related (ATG) protein MoAtg4 phosphorylation mediates the coordination between autophagy and CWI signaling, which underlies the development and pathogenicity of Magnaporthe oryzae. Both ATG protein MoAtg1 and CWI kinase MoMkk1 phosphorylate MoAtg4, which inhibits the deconjugation and C-terminal cleavage of MoAtg8, respectively, to maintain proper autophagy levels. Our study revealed a novel coordination between autophagy and CWI signaling and how the two pathways coordinate to regulate the development and pathogenicity of the blast fungus.

Funding: This research was supported by the key program of Natural Science Foundation of China (Grant No: 32030091 to ZZ) and the program of NSFC (Grant No: 32293245 to ZZ). PW received support from National Institutes of Health (US) (award numbers: AI156254 and AI168867). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

As the critical processes in response to environmental stresses, connections between autophagy and CWI signaling have been previously examined in S. cerevisiae and M. oryzae [ 16 , 26 – 28 ]. In S. cerevisiae, Slt2 promotes mitophagy by affecting mitochondrial recruitment to the PAS, and also regulates pexophagy without affecting nonselective autophagy [ 26 – 28 ]. In M. oryzae, endoplasmic reticulum (ER) stress induced by abnormal protein synthesis activates MoAtg1-dependent MoMkk1 phosphorylation to enhance CWI signaling, thus promoting infection [ 16 ]. At the same time, CWI signaling addresses cell wall stress by targeting protein production, implying a possible important relationship between the CWI signaling pathway and nonselective autophagy (hereafter “autophagy”). However, the underlying mechanism of how autophagy coordinats with CWI signaling to regulate development and pathogenicity is unknown. Here, we revealed that two phosphorylation events of MoAtg4, MoMkk1-mediated phosphorylation in the cytoplasm and MoAtg1-mediated phosphorylation at PAS, collectively maintain proper autophagy and cellular homeostasis.

The cell wall integrity (CWI) signaling pathway is critical for maintaining strong but elastic cell wall during growth and development and also in responses to the external stress [ 14 – 16 ]. In S. cerevisiae, CWI signals stimulate the small G protein Rho1 that binds to the protein kinase C (Pkc1) to activate the conserved mitogen-activated protein kinase (MAPK) cascade, composed of Bck1, Mkk1/2, and Slt2/Mpk1, which then targets the transcription factors Rlm1 and the SBF complex (Swi4 and Swi6) to govern the expression of cell wall biosynthesis and cell cycle genes, respectively [ 17 – 19 ]. Many plant pathogenic fungi, including M. oryzae, Fusarium graminearum, and Claviceps purpurea, utilize this conserved signaling pathway to balance the cell wall homeostasis, which is critical for development and pathogenicity [ 20 – 23 ]. In M. oryzae, deletions of the CWI MAPK cascade components MoMCK1, MoMKK1, or MoMPS1 lead to abnormal chitin distribution and virulence attenuation [ 22 , 24 , 25 ]. Recent studies have also shown that deletions of these kinases results in aberrant autophagy by unknown mechanisms [ 16 ].

Autophagy is a highly conserved catabolic process that mediates intracellular component degradation through the autophagosomes to maintain cellular homeostasis in response to cellular and environmental stresses [ 1 , 2 ], which can be categorized into nonselective and selective autophagy based on its selectivity for substrates [ 3 ]. Nonselective autophagy is critical for pathogenicity, whereas selective autophagy (e.g., pexophagy and mitophagy) is dispensable for appressorium-mediated host infection in the rice blast fungus Magnaporthe oryzae [ 4 – 6 ]. In the Baker’s yeast Saccharomyces cerevisiae, the autophagosome formation correlates well with the amount of autophagy-related (ATG) protein 8 (Atg8)-phosphatidylethanolamine (PE), the lipidation form of Atg8 [ 7 – 9 ]. At the same time, the cysteine protease Atg4 cleaves the C-terminus of the newly synthesized cytoplasmic Atg8 to produce an active variant that undergoes lipidation modification and further conjugates PE of the phagophore assembly site (PAS) [ 10 , 11 ]. Once autophagosome forms at PAS, Atg4 deconjugates Atg8 from the Atg8-PE anchor on the autophagosome membrane, a critical process for Atg8 recycling and subsequent autophagosome fusion with vacuoles [ 12 , 13 ]. Studies in S. cerevisiae also found that Atg4-mediated deconjugation of Atg8-PE is inhibited by ULK/ULK protein kinase Atg1-dependent phosphorylation, ensuring Atg8 anchoring until autophagosome formation [ 1 ].

Results

Cell wall stress-induced autophagy is dependent on CWI kinases in M. oryzae To explore the coordination between autophagy and the CWI signaling pathway, we treated the wild-type strain Guy11 and CWI kinase mutant stains ΔMomck1, ΔMomkk1, and ΔMomps1 with cell wall stressor Calcofluor White (CFW) and performed autophagic body (AB) straining using monodansylcadaverine (MDC) [29–31]. Vacuoles with AB were significantly increased after 5 h of treatment in Guy11 (~65%), but not in the mutants (Fig 1A and 1B). In addition, the expression of the autophagic marker protein RFP-MoAtg8 in these strains revealed that vacuole-localized RFP was significantly increased in Guy11, but not in the mutants (Fig 1C). This result was confirmed by Western blotting analysis that showed an elevated ratio of free RFP in Guy11 in comparison to the mutants (Fig 1D). We also used sodium dodecyl sulfate (SDS) as a second stressor and observed a similar result (S1A–S1C Fig). In addition, the degradation of MoSec63-GFP, an ER-phagy marker of M. oryzae [32], was detected under CFW treatment. These results showed that cell wall stress did not induce ER-phagy (S1D Fig). Together, these results indicate that cell wall stress-induced autophagy is dependent on CWI kinases in M. oryzae. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Cell wall stress-induced autophagy is dependent on CWI kinases. (A) Hyphae of Guy11, ΔMomck1, ΔMomkk1, and ΔMomps1 strains cultured in CM were treated without or with CFW for 5 h (both adding 2 mM PMSF), and autophagic bodies (AB) were observed by confocal microscopy after MDC staining. C: cytoplasm. V: vacuole. (B) Quantification of vacuoles with AB as shown in (A). (C) Quantification of cells with vacuole-localized RFP-MoAtg8 after CFW treatment. (D) Total mycelial proteins after CFW treatment were extracted and analyzed by Western blot analysis with anti-RFP and anti-Actin antibodies. The amount of free RFP was compared with the total amount of RFP-MoAtg8 and free RFP to quantify autophagic levels. CFW treatment: treated with 1 mg/ml CFW. Data (n = 100) from three independent experiments were used for statistical analysis by two-way ANOVA with Tukey’s HSD. Scale bar: 5 μm. https://doi.org/10.1371/journal.ppat.1011988.g001

CWI kinase MoMkk1 interacts with and phosphorylates MoAtg4 in the cytoplasm Previous studies identified 22 ATG proteins that are essential for autophagy in M. oryzae [6], among which several were found to interact with MoMkk1 via yeast-two-hybrid (Y2H) screening, including MoAtg3, 4, 5, and 16 that are directly involved in Atg8 lipidation [16]. In addition, MoMkk1 mediates the crosstalk between the CWI signaling pathway and autophagy under ER stress [16]. We, therefore, hypothesized that autophagy coordinates CWI signaling through direct phosphorylation of ATG proteins by MoMkk1. To test this hypothesis, in vitro phosphorylation assays were performed using phosphoprotein gel staining [16,33,34], and the results showed significantly increased phospho-fluorescence only between MoMkk1 and MoAtg4, but not MoAtg3, 5, or 16 (Fig 2A). This result suggested that MoMkk1 can phosphorylate MoAtg4. Consequently, we focused on the interaction between MoMkk1 and MoAtg4. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. The CWI kinase MoMkk1 interacts with and phosphorylates MoAtg4 in the cytoplasm. (A) In vitro phosphorylation reaction of GST-MoMkk1 with His-MoAtg4, His-MoAtg3, GST-MoAtg5, or His-MoAtg16 fusion proteins in the presence of 50 μM ATP, and phosphorylation proteins were stained using Pro-Q Diamond Phosphorylation Gel Stain. Cytation3 microplate reader was then used to measure the Phospho-Fluorescence signal at 590 nm (excited at 530 nm). (B) GST-MoMkk1, empty GST, and His-MoAtg4 proteins were obtained for GST pull-down analysis. (C) MoMkk1-GFP and MoAtg4-Flag were co-expressed in the Guy11 strain, and proteins were extracted for co-IP assay by Western blot analysis using anti-GFP and anti-Flag antibodies. Proteins from Guy11 co-expressing GFP and MoAtg4-Flag were used as the control. (D) BiFC observation in the strain co-expressing MoMkk1-nYFP and MoAtg4-cYFP during various developmental stages. The strains co-expressing empty nYFP with MoAtg4-cYFP and MoMkk1-nYFP with empty cYFP were used as controls. Scale bar: 5 μm. (E) Peptides of MoMkk1-dependent MoAtg4 phosphorylation were identified by LC-MS/MS analysis under CFW treatment. (F) GST-MoMkk1, His-MoAtg4, His-MoAtg4S97A, His-MoAtg4S276A, His-MoAtg4T278A, and His-MoAtg43A fusion proteins were obtained for in vitro phosphorylation analysis. 3A: replaced S97, S276, and T278 with A to mimic nonphosphorylated form of MoAtg4. Data from three independent experiments were used for statistical analysis by one-way ANOVA with Tukey’s HSD. https://doi.org/10.1371/journal.ppat.1011988.g002 We next examined the localizations of MoMkk1 and MoAtg4 in M. oryzae. Both were localized to the cytoplasm with or without cell wall stress, though MoAtg4 was also accumulated at PAS that was marked with MoApe1-RFP (S2A and S2B Fig). CFW treatment did not affect the numbers of cells with PAS-localized MoAtg4 (S2C Fig). These results implied a direct interaction between MoMkk1 and MoAtg4. To test the physical interaction between the two, GST pull-down and co-immunoprecipitation (co-IP) assays were performed that showed positive interaction (Fig 2B and 2C). Furthermore, a bimolecular fluorescence complementation (BiFC) assay confirmed that the interaction occurs in the cytoplasm, and there was no obvious co-localization with the MDC dye (Figs 2D and S2D). This result indicated that MoMkk1 interacts with MoAtg4 in the cytoplasm. To test whether MoAtg4 is also phosphorylated in vivo, proteins were extracted from ΔMoatg4/MoATG4-GFP and ΔMomkk1/MoATG4-GFP strains. Mn2+-Phos-tag SDS-PAGE showed that the phosphorylated MoAtg4 band was separated from the unphosphorylated one in the ΔMoatg4/MoATG4-GFP strain, while a similar pattern was observed in the ΔMomkk1/MoATG4-GFP strain (S2E Fig). Cell wall, ER, or nitrogen starvation stress all affect the function of MoMkk1 or MoAtg4 [5,16,22]. Thus, MoAtg4 phosphorylation under CFW, dithiothreitol (DTT), and nitrogen starvation minimal medium (MM-N) treatment was initially analyzed by phos-tag electrophoresis, and no difference was found under these conditions (S2F Fig). To further validate MoMkk1-dependent MoAtg4 phosphorylation, we purified the MoAtg4 protein from ΔMoatg4/MoATG4-GFP and ΔMomkk1/MoATG4-GFP strains and performed liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis with or without CFW. Data showed that serine 97 (S97), S276, and threonine 278 (T278) of MoAtg4 in ΔMoatg4/MoATG4-GFP treated with CFW, but not ΔMomkk1/MoATG4-GFP with CFW and the two strains without CFW treatment, were phosphorylated (Figs 2E and S2G). However, T101 was phosphorylated in both ΔMoatg4/MoATG4-GFP and ΔMomkk1/MoATG4-GFP strains, which was consistent with phos-tag electrophoresis results, suggesting that a complex MoAtg4 phosphorylation pattern (S2G Fig). To validate S97, S276, and T278 are MoMkk1-dependent phosphorylation sites, we performed in vitro phosphorylation analysis using the nonphosphorylated protein His-MoAtg43A in which these residues were replaced with alanine (S97A, S276A, and T278A) and single-site phosphorylation mutations His-MoAtg4S97A, His-MoAtg4S276A, and His-MoAtg4T278A. Results showed that phosphorylation levels of MoAtg4S97A, MoAtg4S276A, and MoAtg4T278A were significantly reduced compared with MoAtg4, while MoAtg43A did not react with MoMkk1 (Fig 2F), indicating that these residues of MoAtg4 are important for MoMkk1-dependent phosphorylation under cell wall stress, and kinases other than MoMkk1 might also phosphorylate MoAtg4 in M. oryzae.

MoMkk1-dependent MoAtg4 phosphorylation plays a role in the development and virulence of M. oryzae To investigate the biological function of MoMkk1-dependent MoAtg4 phosphorylation, we replaced S97, S276, and T278 residues with aspartic acid (D) to mimic phosphorylation, and introduced nonphosphorylated MoATG43A-GFP and phosphorylation-mimic MoATG43D-GFP vectors into the ΔMoatg4 mutant, respectively. We found that ΔMoatg4/MoATG43A and ΔMoatg4/MoATG43D strains were defective in conidiation and appressorium formation. In addition, the phosphorylation-mimic strain showed more severe defects than nonphosphorylated ones (Figs 3A, S3A and S3B). The ΔMoatg4/MoATG43A strain showed a similar virulence as Guy11, in contrast to the attenuated one in the ΔMoatg4/MoATG43D strain (Fig 3B–3D). The invasive hyphae (IH) growth assay showed that about 30% of sites were type 4 in infection by Guy11, ΔMoatg4/MoATG4, and ΔMoatg4/MoATG43A, whereas only approximately 5% were type 4 in infection by the ΔMoatg4/MoATG43D strain, and none by ΔMoatg4 (Fig 3E). Collectively, these results indicated that MoMkk1-dependent MoAtg4 phosphorylation plays a role in the development and pathogenicity of M. oryzae. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. MoMkk1-dependent MoAtg4 phosphorylation functions in the development and virulence of M. oryzae. (A) Strains cultured on SDC (straw decoction and corn) medium at 28°C for 7 d in the dark, followed by 3 d of continuous illumination under fluorescent light for conidiation assay. (B) Conidial suspensions of Guy11, ΔMoatg4, ΔMoatg4/3A, ΔMoatg4/3D, and ΔMoatg4/MoATG4 strains were sprayed onto 2-week-old rice seedlings (cultivar CO39) for virulence analysis. Diseased rice leaves were photographed at 7 days post inoculation (dpi). (C) Quantification of diseased leaf area as shown in (B) by Image J. (D) Severity of rice blasts was evaluated by quantitative PCR of M. oryzae genomic 28S rDNA relative to rice genomic RUBQ1 DNA as shown in (B). (E) Close observation and statistical analysis of the invasive hyphae (IH) growth in rice leaf sheath at 36 hours post inoculation (hpi). Statistical analysis of IH at 100 appressorium penetration sites by rating from type I to type IV (type 1, no penetration; type 2, a single IH with no branch; type 3, IH extended but was limited in one plant cell; type 4, IH extended to neighboring cells). The arrow points to appressorium, and the asterisk indicates IH. Scale bar: 5 μm. 3D: replaced S97, S276, and T278 with D to mimic phosphorylated form of MoAtg4. Different letters indicate statistically significant differences (Duncan’s new multiple range test, p < 0.05 or p < 0.01). https://doi.org/10.1371/journal.ppat.1011988.g003 To analyze the impact of cell wall stress on ΔMoatg4/MoATG43A and ΔMoatg4/MoATG43D strains, we investigated the growth, conidiation, and pathogenicity of the strains under CFW treatment. Although CFW inhibited 25% of the growth rate of ΔMoatg4/MoATG43A and 21% of ΔMoatg4/MoATG43D strains, it showed no difference compared to Guy11, and inhibitions of the strains’ virulence by CFW treatment were also similar. In addition, CFW treatment did not impact conidiation (S4A–S4C Fig). Since the CWI signaling pathway requires proper MAPK cascade phosphorylation and signal transduction [22], we examined the effect of MoMkk1-dependent MoAtg4 phosphorylation on CWI signaling through CFW staining. Chitin distribution of ΔMoatg4/MoATG43A and ΔMoatg4/MoATG43D strains is even and concentrated on the growing apices, and they also did not exhibit hyphal autolysis (S4D–S4F Fig). MoMps1 phosphorylation in ΔMoatg4/MoATG43A and ΔMoatg4/MoATG43D under CFW treatment was unaffected (S4G Fig), suggesting that MoMkk1-dependent MoAtg4 phosphorylation might not play a role in CWI signaling.

MoAtg1-mediated MoAtg4 phosphorylation is involved in the development and pathogenicity of M. oryzae Our previous analysis showed that MoAtg4 remains phosphorylated in the ΔMomkk1 mutant (S2E and S2F Fig), suggesting a possibility of phosphorylation by other protein kinases. Since the yeast and human Atg1/Ulk1 phosphorylate Atg4/Atg4B [1,37], we tested whether MoAtg1 also phosphorylates MoAtg4. The interaction between MoAtg1 and MoAtg4 was first observed at PAS (S6A and S6B Fig). The in vitro phosphorylation assay showed MoAtg1 phosphorylates MoAtg4 (S6C Fig). Further LC-MS/MS data revealed that the phosphorylation of MoAtg4 at S67 was detected in Guy11, but not in the ΔMoatg1 mutant, under nitrogen starvation, Guy11 and ΔMoatg1 without starvation (S6D Fig). Although not covered in repeated LC-MS/MS data, S364, a conserved residue previously reported to be phosphorylated by Atg1 in S. cerevisiae, was mutated to evaluate the function in MoAtg1-mediated phosphorylation reaction (S6E Fig). In vitro phosphorylation analysis showed that phospho-fluorescence was partially reduced in MoAtg4S67A and MoAtg4S364A, while MoAtg42A containing mutations of both S67A and S364A showed more significantly reduced phospho-fluorescence (S6C Fig), indicating that S67 and S364 residues are important for MoAtg4 phosphorylation by MoAtg1. To investigate the role of MoAtg1-mediated MoAtg4 phosphorylation in M. oryzae, we generated nonphosphorylated ΔMoatg4/MoATG42A and phosphorylation-mimic ΔMoatg4/MoATG42D strains, and single-site phosphorylation mutations strains, including ΔMoatg4/MoATG4S67A, ΔMoatg4/MoATG4S364A, ΔMoatg4/MoATG4S67D, and ΔMoatg4/MoATG4S364D. Compared with Guy11, ΔMoatg4/MoATG42A and ΔMoatg4/MoATG42D strains showed defects in conidiation and appressorium formation. In addition, MoAtg42D caused more severe defects than MoAtg42A, and MoAtg42A resulted in more severe defects than MoAtg4S67A or MoAtg4S364A, and similarly, MoAtg42D produced more severe defects than MoAtg4S67D or MoAtg4S364D (S3C, S3D and S6F Figs). Moreover, ΔMoatg4/MoATG42D exhibited reduced pathogenicity and more severe than ΔMoatg4/MoATG4S67D and ΔMoatg4/MoATG4S364D, while the ΔMoatg4/MoATG4S67A, ΔMoatg4/MoATG4S364A, and ΔMoatg4/MoATG42A strains were as virulent as Guy11 (S6G and S6H Fig). We therefore concluded that MoAtg1-mediated MoAtg4 phosphorylation is also involved in appressorium formation and pathogenicity of M. oryzae.

MoAtg1-mediated MoAtg4 phosphorylation negatively regulates autophagy by inhibiting deconjugation of MoAtg8-PE Since both MoAtg1 and MoAtg4 are essential for autophagy, we next investigated the effect of MoAtg1-mediated MoAtg4 phosphorylation on autophagy in M. oryzae. MDC staining showed that the number of vacuoles with AB was nearly 100% in Guy11, but reduced to 77% in ΔMoatg4/MoATG42A and 28% in ΔMoatg4/MoATG42D strains, respectively (S7A and S7B Fig). Consistently, TEM showed defective autophagy in the aberrantly phosphorylated strains (S7C and S7D Fig). These results indicated that MoAtg1-mediated MoAtg4 phosphorylation is critical for autophagy. In yeast, Atg4 is responsible for Atg8-PE deconjugation and Atg8 recycling to promote the fusion of autophagosomes with vacuoles and flowing autophagosomes formation, which is inhibited by Atg1-mediated Atg4 phosphorylation [1,38]. To assay whether the phosphorylation of MoAtg4 controls the deconjugation of MoAtg8-PE, we expressed MoAtg8G116 with an N-terminal RFP tag (RFP-MoAtg8G116), which could conjugate to PE without the initial step of MoAtg4-mediated cleavage [1,38], in the ΔMoatg4/MoATG42A and ΔMoatg4/MoATG42D strains. As indicated by vacuolar fluorescent dye CMAC [1,39], RFP signals were localized in the vacuoles of Guy11 and ΔMoatg4/MoATG42A strains but were restricted in the cytoplasm and the edge of vacuoles in ΔMoatg4 and ΔMoatg4/MoATG42D strains (S7E Fig). Blocked deconjugation of MoAtg8-PE was also observed at the appressorium maturing stage in ΔMoatg4 and ΔMoatg4/MoATG42D strains. Western blotting analysis showed that protein levels of free RFP were lower in ΔMoatg4 and ΔMoatg4/MoATG42D (S7F Fig). These findings suggest that the phosphorylation of MoAtg4 by MoAtg1 inhibits deconjugation of MoAtg8-PE. To understand the underlying mechanism, we examined whether MoAtg4 phosphorylation by MoAtg1 affects its interaction with MoAtg8. Y2H results showed that the phosphorylation-mimic MoAtg42D, but not the nonphosphorylated variant MoAtg42A, significantly reduced the affinity with MoAtg8 (S7G Fig). Co-IP analysis was also performed with proteins extracted from strains expressing RFP-MoAtg8G116, and a weak interaction was detected between MoAtg42D and MoAtg8 (S7H Fig). Interestingly, vacuole-localized RFP signals were observed in both MoMkk1-dependent nonphosphorylated ΔMoatg4/MoATG43A and phosphorylation-mimic ΔMoatg4/MoATG43D strains, which expressed RFP-MoAtg8G116 (S8A Fig), while this localization was not observed in ΔMomck1, ΔMomkk1, and ΔMomps1 (S8B Fig), suggesting that MoMck1, MoMkk1, and MoMps1, but not MoMkk1-dependent MoAtg4 phosphorylation, is involved in the deconjugation of MoAtg8-PE. Overall, MoAtg1-mediated MoAtg4 phosphorylation negatively regulates autophagy by inhibiting deconjugation of MoAtg8-PE at PAS.

[END]
---
[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1011988

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.