Structural and functional analysis of the cerato-platanin-like protein Cpl1 suggests diverging functions in smut fungi

Abstract

Plant-pathogenic fungi are causative agents of the majority of plant diseases and can lead to severe crop loss in infected populations. Fungal colonization is achieved by combining different strategies, such as avoiding and counteracting the plant immune system and manipulating the host metabolome. Of major importance are virulence factors secreted by fungi, which fulfil diverse functions to support the infection process. Most of these proteins are highly specialized, with structural and biochemical information often absent. Here, we present the atomic structures of the cerato-platanin-like protein Cpl1 from Ustilago maydis and its homologue Uvi2 from Ustilago hordei. Both proteins adopt a double-Ψβ-barrel architecture reminiscent of cerato-platanin proteins, a class so far not described in smut fungi. Our structure-function analysis shows that Cpl1 binds to soluble chitin fragments via two extended grooves at the dimer interface of the two monomer molecules. This carbohydrate-binding mode has not been observed previously and expands the repertoire of chitin-binding proteins. Cpl1 localizes to the cell wall of U. maydis and might synergize with cell wall-degrading and decorating proteins during maize infection. The architecture of Cpl1 harbouring four surface-exposed loop regions supports the idea that it might play a role in the spatial coordination of these proteins. While deletion of cpl1 has only mild effects on the virulence of U. maydis, a recent study showed that deletion of uvi2 strongly impairs U. hordei virulence. Our structural comparison between Cpl1 and Uvi2 reveals sequence variations in the loop regions that might explain a diverging function.

Keywords: Ustilago maydis; biotrophy; cerato-platanin-like; chitin.

FIGURE 1

Crystal structure of Cpl1. (a) Secondary structure of Cpl1. The boundaries of the N‐terminal domain (NTD) and the double‐ψβ‐barrel (DPBB) are shown above. (b) Cartoon model of Cpl1 coloured in deep teal (NTD) and blue (DPBB). Secondary structure elements, the two disulphide bridges (DBs), and the four loop regions (LRs) are labelled above the respective elements. (c) Cartoon model of the Cpl1 dimer with the two monomers coloured in green and blue, respectively. (d) Size‐exclusion chromatography‐multi‐angle light scattering (SEC‐MALS) analysis of Cpl1 at pH 7.5 and pH 5.0 shows the presence of a 50‐kDa species (theoretical mass of Cpl1 dimer = 52.4 kDa). The inset shows a Coomassie‐stained SDS‐PAGE of the peak fraction.

FIGURE 2

Cpl1 has high structural homology to cerato‐platanin‐like proteins and binds to chitin oligomers. (a) Superposition of a Cpl1 monomer (blue) with MpCP5 (yellow; PDB code: 3SUM). The loop regions of Cpl1 have been truncated for better visibility of the superposed regions with MpCP5 in the right panel. N‐terminal domain (NTD), double‐ψβ‐barrel (DPBB). (b) Microscale thermophoresis (MST) experiments with Cpl1. Shown are normalized fluorescence values (F‐norm, black dots) of Cpl1 titrated against cellobiose, chitobiose, and chitotetraose (concentration in nM). Where possible, a curve was fitted (red) used to calculate the dissociation constant (K_D). Data points not included in the fitting are shown as residual data (curve identity, blue). The K_D values of Cpl1 binding chitobiose and chitotetraose were 7.4 μM (± 1.4) and 68.8 μM (±28.8), respectively. (c) Areas of Cpl1 that exhibited reduced deuterium incorporation within at least two time points in the presence of chitotetraose (compare to Figure S5) are coloured in blue or dark teal for the α‐ and β‐subunits, respectively, and areas not covered by peptides are shown in black. The side chains of the disulphide bridge‐forming cysteine residues 77/96 and 124/157 are shown as red sticks. (d) The chitin‐binding interfaces of MpCP5 (yellow) and MpCP3 (orange) superposed with Cpl1 (blue) show no structural similarity between the proteins in these regions. LR, loop region.

FIGURE 3

A groove in the dimer interface at Cpl1 is important for binding of soluble chitin oligomers. (a) Cpl1 dimer (blue) with a chitotetraose (yellow) docked into the dimer groove using AutoDock Vina. The inset shows a closeup of the chitotetraose fitted in between the α3–β4–loop (LR1) structures of both protomers. (b) Amino acids that were varied to alanines and the results of microscale thermophoresis (MST) experiments employing Cpl1_N151A/R155A and Cpl1_D69A/E72A. (c) Closeup of differently docked chitotetraose molecules. N151 and R155 are highlighted and have been mutated to alanines.

FIGURE 4

Structural comparison between Cpl1 and Uvi2. (a) Cartoon model of Uvi2 (yellow) alone and superposed with Cpl1 (blue). The loop regions (LRs) and disulphide bonds (DBs) are labelled on the respective elements. N‐terminal domain (NTD), double‐ψβ‐barrel (DPBB). (b) Dimer of Uvi2 as observed in the crystal structure with the two protomers coloured in yellow and salmon. (c) Size‐exclusion chromatography‐multi‐angle light scattering (SEC‐MALS) analysis of Uvi2 at pH 7.5 (green graph) and pH 5.0 (red graph) shows the presence of a 50‐kDa species. The inset shows a Coomassie‐stained SDS‐PAGE gel of the peak fraction. (d) Amino acid sequence alignment of Uvi2 and Cpl1 coloured according to the CLUSTAL colouring scheme (Sievers et al., 2011). The regions involved in dimer formation are underlined in black and residues that were mutated to diminish binding of soluble chitin oligomers are indicated with black arrows. The four LRs showing the highest degree of sequence deviation are indicated with grey boxes.

FIGURE 5

Cpl1 is dispensable for virulence, suppresses chitin‐induced medium alkalization, and binds to chitin, but does not protect hyphal growth against chitinase hydrolysis. (a) Maize infection assays using deletion strains in different genetic backgrounds of Ustilago maydis. cpl1 has been deleted from the genomes of FB1, FB2, and SG200 (from left to right). (b) HA‐tagged Cpl1 was incubated with magnetic chitin beads, shrimp shell chitin, chitosan, cellulose, and xylan for 6 h, and after centrifugation, pellets and supernatants were analysed by protein gel electrophoresis. (c) Microscopic pictures of Trichoderma viride grown in vitro in the presence of Cladosporium fulvum Avr4 or Ecp6 or U. maydis Cpl1 for 2 h, followed by addition of chitinase or water as control. Pictures were taken at 6 h after chitinase addition. (d) The pH of Nicotiana tabacum ‘Bright Yellow‐2’ cell suspensions treated with 100 nM chitin hexamer, 1 μM U. maydis Cpl1, 1 μM C. fulvum Ecp6 (positive control), or mixtures of chitin hexamer and effector protein was measured for 34 min. The assay was performed twice with similar results.

FIGURE 6

Cpl1 decorates the fungal cell wall of Ustilago maydis hyphae and is copurified with cell wall‐degrading and cell wall‐decorating virulence factors during infection. (a–d) Cpl1‐HA is secreted and binds to the fungal cell wall of U. maydis filaments grown on a Parafilm surface. Confocal microscopy of strains immunostained with an anti‐HA primary antibody and an AF488‐conjugated secondary antibody. Shown are merge, green fluorescent protein (GFP) channel, and differential interference contrast (DIC) images. Panels (a) and (b) show several hyphae and a closeup of SG200Δcpl1 complemented with cpl1‐HA under the control of the constitutive otef promoter, respectively. Panels (c) and (d) show two representative images of SG200 hyphae. (e) Maize infection assay with the indicated U. maydis strains. (f) Western blot analysis of Cpl1‐HA derived from infected maize leaves 3 days post‐infection. For the control bait mCherry‐HA in the ip locus fused to a sequence encoding the signal peptide of Cmu1 under the control of its own promoter was used. (1) Whole lysate of infected maize leaves, (2) insoluble fraction, (3) cleared lysate, (4) wash fraction, and (5) elution fraction. (g) Volcano plot depicting the results of the third of the three co‐immunoprecipitation experiments conducted with FB1 and FB2 containing Cpl1‐HA as bait and mCherry‐HA as the control sample. Shown are normalized label‐free quantification (LFQ) intensities as either the −log(p) value (Student's t test; y axis) or the LFQ intensity difference. Data were acquired by liquid chromatography coupled with mass spectrometry combined with the detection and quantification of peptide intensities. A high −log(p) value (y axis) indicates that many peptides were found. A negative difference (x axis) means that peptides were specific towards the bait sample containing Cpl1‐HA (green and bold) and positive values mean that peptides were specific towards the mCherry‐HA control (blue and bold).