Article | Published:

Co-delivery of cell-wall-forming enzymes in the same vesicle for coordinated fungal cell wall formation

Nature Microbiology volume 1, Article number: 16149 (2016) | Download Citation


Fungal cells are surrounded by an extracellular cell wall. This complex matrix of proteins and polysaccharides protects against adverse stresses and determines the shape of fungal cells. The polysaccharides of the fungal wall include 1,3-β-glucan and chitin, which are synthesized by membrane-bound synthases at the growing cell tip. A hallmark of filamentous fungi is the class V chitin synthase, which carries a myosin-motor domain. In the corn smut fungus Ustilago maydis, the myosin-chitin synthase Mcs1 moves to the plasma membrane in secretory vesicles, being delivered by kinesin-1 and myosin-5. The myosin domain of Mcs1 enhances polar secretion by tethering vesicles at the site of exocytosis. It remains elusive, however, how other cell-wall-forming enzymes are delivered and how their activity is coordinated post secretion. Here, we show that the U. maydis class VII chitin synthase and 1,3-β-glucan synthase travel in Mcs1-containing vesicles, and that their apical secretion depends on Mcs1. Once in the plasma membrane, anchorage requires enzyme activity, which suggests co-synthesis of chitin and 1,3-β-glucan polysaccharides at sites of exocytosis. Thus, delivery of cell-wall-forming enzymes in Mcs1 vesicles ensures local foci of fungal cell wall formation.

Fungi are one of the oldest and largest clades of microbes1,2. Up to 5 million species are thought to populate all habitats on Earth3. A hallmark of fungi is their cell wall. This extracellular meshwork of polysaccharides and glycoproteins shapes the fungal cell, thereby enabling substrate, soil or host-tissue exploration by polarized growth4. Moreover, the cell wall forms the environmental interface, which withstands abiotic and biotic stresses. As such, it is also crucial for pathogen recognition by mammalian and plant host cells5,6. The fungal cell wall consists mainly of 1,3-β-glucan and chitin, a polymer of N-acetylglucosamine7,8. Both sugars are synthesized at the apical growth region by membrane-bound 1,3-β-glucan synthase (GS) and chitin synthases (CHSs), respectively. Our knowledge about the mechanism of their apical delivery is fragmentary, but it is widely accepted that GS and CHSs travel in secretory vesicles9. Here, the vesicles fuse with the plasma membrane, thereby exposing their cargo, for synthesis of the extracellular cell wall. However, neither the detailed secretion pathways, nor the mechanism by which the activity of GSs and CHSs is coordinated during wall synthesis, is known.

Filamentous fungal genomes usually carry one or two genes encoding GS, whereas CHSs comprise large super-families, with up to 7–10 members10,​11,​12, lying in seven classes13. Among these families, class V and class VII CHSs are characteristic of filamentous asco- and basidiomycetes13 and are required for the virulence of many fungal pathogens11,12,14. Interestingly, class V and several class VII CHSs are chimaeric proteins, consisting of an N-terminal myosin motor domain (MMD), fused to the C-terminal CHS domain11,12,15. Although the MMD can bind F-actin15,16, it does not seem to be involved in directed transport of secretory vesicles17. In the maize smut pathogen Ustilago maydis, vesicles containing the class V myosin-CHS Mcs1 are delivered by myosin-5 and kinesin-1, while the MMD of Mcs1 tethers the arriving vesicles to F-actin, thereby supporting apical exocytosis17. The importance of this myosin–CHS-dependent secretion pathway for fungal cell wall formation is not understood.

Here, we show that U. maydis class V CHS, class VII CHS and GS co-travel in the same secretory vesicle. Co-delivery is followed by co-secretion, which is dependent on Mcs1. Exocytosed enzymes remain closely positioned and stationary within the plasma membrane. This depends on their enzymatic activity and on the cell wall. Thus, co-localization of cell-wall-forming enzymes in Mcs1-containing vesicles ensures the establishment of local foci of coordinated cell wall synthesis.


Class VII CHSs have lost their myosin motor domain

In Aspergillus fumigatus, Aspergillus nidulans, Fusarium oxysporum, Magnaporthe oryzae and Neurospora crassa, class V and class VII carry a putative N-terminal MMD12,15,16,18,​19,​20, whereas in U. maydis only Mcs1 contains an MMD11,21. But how conserved are the CHS MMDs in asco- and basidiomycetes and in the ancient cryptomycete fungus Rozella allomyces? We looked at the conservation of the predicted primary sequence and for the presence of functional motifs, such as the ATP-binding P-loop and A-loop22,23 and a putative actin-binding site24. This revealed that the MMD of R. allomyces myosin–CHS and the basidiomycete class V CHSs are highly conserved (Fig. 1a; note that a set of colour-adjusted main figures, suitable for red-green colour-blind readers, is provided as Supplementary Figs 1–6). In contrast, ascomycete class V CHSs show less sequence conservation (Fig. 1a and Supplementary Fig. 7; domain according to the Protein Family Database (PFAM)). In class VII CHSs, the MMD is either absent or truncated (Fig. 1a). Protein sequence conservation is indicative of functionality25, suggesting that the MMD in class VII CHSs has altered or lost its function during evolution of filamentous fungi.

Figure 1: Domain organization of fungal class V and class VII CHSs.
Figure 1

a, Maximum likelihood tree and domain structure of class V and class VII CHSs from ascomycete and basidiomycete fungi, as well as the ancient cryptomycete fungus R. allomyces. Class V CHSs from asco- and basidiomycetes carry an N-terminal motor domain that shows high sequence conservation (indicated by a high probability of being identified as an MMD in PFAM) and the presence of an ATP-binding P-loop22, a supporting A-loop23 and a residue implicated in actin-binding24. In contrast, the MMD in class VII CHSs is absent or truncated and has lost ATP-binding motifs (Supplementary Fig. 7). The tree is based on the CHS domains and was calculated in MEGA5.2 (ref. 58). Bootstrap values from 1,000 replications are given at nodes. b, Domain structure of Mcs1, Chs6 and the fusion protein MMD–Chs6. c, Localization of GFP–Chs6 in the absence of endogenous Chs6 (Δchs6), Chs6 and Mcs1 (Δchs6Δmcs1), and localization of the chimaeric fusion protein GFP–MMD–Chs6 in the Δchs6Δmcs1 double mutant. Scale bar, 2 µm. d, Fluorescent intensity profiles of GFP–Chs6 in Δchs6 (red curve) and Δchs6Δmcs1 mutants (green curve), and the chimaeric fusion protein GFP–MMD–Chs6 (blue curve). Fusing the MMD to Chs6 partially rescues the polar localization. Each data point is given as mean ± standard error, n = 15 measurements. e, Kymographs showing motility of GFP–Chs6 in a Δchs6Δmcs1 double mutant. Scale bars, 2 s (vertical) and 2 µm (horizontal).

Class VII CHS requires the motor domain of class V CHS for polar localization

Next, we asked why MMD is either absent from or truncated in class VII CHSs. In U. maydis, the MMD of Mcs1 mediates tethering of myosin5- and kinesin1-delivered secretory vesicles to the growth region17. This increases the pausing time of vesicles and thereby supports apical exocytosis17. We hypothesized that Chs6 is delivered in Mcs1-containing and Mcs1-tethered vesicles, which could have made Chs6 MMD redundant. To test this hypothesis, we investigated the localization of a fusion of a green fluorescent protein (GFP)-tag and Chs6 (GFP–Chs6), as well as the fusion protein GFP–MMD–Chs6, where the N-terminal MMD (aa 1–915) was fused to Chs6 (aa 59–1180; Fig. 1b; see Supplementary Table 1 for strain genotypes and Supplementary Table 2 for experimental strain usage). When expressed in the Δchs6 background, GFP–Chs6 localized to the growth region of yeast-like cells of U. maydis (Fig. 1c, left image; Fig. 1d). In the absence of Mcs1 (Δchs6Δmcs1 mutants), GFP–Chs6 lost this apical localization (Fig. 1c, middle image; Fig. 1d), despite its intracellular motility in vesicles (Fig. 1e, kymograph shows motility as diagonal lines). However, fusion of the MMD to Chs6 restored polar exocytosis and accumulation (Fig. 1c, right image; Fig. 1d). These results demonstrate that the MMD of Mcs1 is required for Chs6 positioning. Thus, we speculate that the MMD of class VII CHSs was removed gradually due to functional redundancy.

Class VII and class V CHSs are co-delivered

The role of Mcs1 in Chs6 exocytosis suggested that both could be co-delivered in the same vesicle. Indeed, both fluorescent proteins Mcs1–GFP3 and Chs6–GFP3 covered similar regions of the growing cell tip (Fig. 2a; plasma membrane labelled by mCherry-Sso1, ref. 26). In apical regions of the cell that were photobleached to reduce signal interference, a treatment harmless to the cell17, Chs6–GFP3 moved rapidly in a bidirectional manner, often pausing before turning direction (Supplementary Video 1 and Supplementary Fig. 8a). This pattern was indistinguishable from Mcs1–GFP3 (Supplementary Fig. 8a and Supplementary Video 2). Indeed, the velocity, flux and pausing time of the fluorescent signals for Mcs1–GFP3 and Chs6–GFP3 were not significantly different (Fig. 2b). This is consistent with the hypothesis that Chs6 and Mcs1 travel in the same vesicle. To test this, we co-visualized red-fluorescent mCherry3–Mcs1 and Chs6–GFP3. Indeed, 90% of the mCherry3–Mcs1 signals co-travelled with Chs6–GFP3 (Fig. 2c,d; Supplementary Fig. 8b and Supplementary Video 3). This reinforces the idea that Chs6 and Mcs1 are co-delivered to the growth region in the same carrier.

Figure 2: Class V CHS and class VII CHS co-migrate and are co-secreted.
Figure 2

a, Mcs1–GFP3 (left) and Chs6–GFP3 (right) localization in plasma membrane (mCherry-Sso1-labelled, red). Scale bar, 2 µm. b, Mcs1–GFP3 and Chs6–GFP3 motility parameters. Velocity and pausing time shown as whisker plots, flux shown as mean ± standard error; n = 3 experiments, each data set >45 measurements; NS, non-significant difference at Pvelocity = 0.257, Ppausing = 0.435, PFlux = 0.169. See Supplementary Videos 1 and 2. c, Co-motility kymographs of mCherry3–Mcs1 and Chs6–GFP3. Scale bars, 3 s (vertical) and 2 µm (horizontal). See Supplementary Fig. 8b,c and Supplementary Video 3. d, Degree of co-motility of Mcs1 and Chs6. n = 36 cells; ***P < 0.0001. e, Co-localization of mCherry3–Mcs1 (red) and Chs6–GFP3 (green) after secretion at the growth region (co-localization shown in yellow). Images are deconvoluted, and brightness, contrast and gamma processed. Scale bar, 2 µm. f, Intensity profiles of mCherry3–Mcs1 (red) and Chs6–GFP3 (blue) at the growth region, taken from Supplementary Fig. 9a. Co-localization is indicated by arrowheads. See also Supplementary Fig. 9b–d. g, Chs6–GFP3 secretion. Chs6 concentrates in the apical plasma membrane (mCherry–Sso1, red; Pre). After photobleaching (0 min; mCherry–Sso1 from Pre image), newly secreted signals reappear (5 min, arrowheads). Scale bar, 2 µm. h, Mcs1–GFP3 and Chs6–GFP3 secretion rates at the apical tip and in the mother cell. NS, non-significant difference at PTip = 0.204 and PMother = 0.645; n = 20 cells (Mcs1) and 21 cells (Chs6 at tip, Mcs1 and Chs6 in mother). i, Co-secretion of mCherry3–Mcs1 (red) and Chs6–GFP3 (green), 5 min post-bleaching. Dotted box indicates bleached region; arrowhead indicates secreted signals. Right panels are at higher magnification. Scale bars, 3 µm (left) and 0.5 µm (right). j, Chs6–GFP3 fluorescence at the cell tip in conditional mcs1 mutants. In the presence of mcs1, Chs6–GFP3 locates at the growth region (Mcs1↑), whereas fluorescence is reduced when mcs1 is repressed (Mcs1↓); n = 3 experiments, 56 cells for Mcs1↑ and 79 cells for Mcs1↓; ***P < 0.0001. k, Chs6–GFP3 secretion rate in conditional mcs1 mutants. Repression of mcs1 (Mcs1↓) reduces Chs6–GFP3 secretion. n = 3 experiments, 32 cells for Mcs1↑ and 30 cells for Mcs1↓; ***P < 0.0001. Non-normal distributed data sets were analysed using Mann–Whitney tests and are shown as whisker plots (b,h,k), with 25th and 75th percentiles indicated by blue lines and the median by a red line; whisker ends represent minimum/maximum values. Gaussian-distributed data sets (b,d,j) were analysed by unpaired Student's t-test with Welch's correction and are shown as mean ± standard error.

Class VII and class V CHSs are co-secreted

Mcs1-carrying vesicles fuse with the plasma membrane, thereby allowing chitin synthesis into the cell wall. If Mcs1 and Chs6 are co-delivered, both enzymes should be exocytosed simultaneously. Indeed, mCherry3-Mcs1 co-localized with Chs6–GFP3 in the apical growth region ‘cap’, where cell wall formation occurs (Fig. 2e, co-localization results appear in yellow; Fig. 2f, arrowheads, Supplementary Fig. 9). We gathered quantitative data on CHS secretion rates by monitoring the reappearance of the GFP signal following photobleaching. This recovery was considered to represent the delivery and exocytosis of unbleached fluorescent signals17. We expressed Mcs1–GFP3 and Chs6–GFP3, respectively, in cells that carry the plasma membrane marker mCherry–Sso1. At 5 min after photobleaching the growth region, signals reappeared (Fig. 2g, arrowheads; the mCherry-Sso1 image was taken from the ‘Pre’ image). These fluorescent signals overlapped with the mCherry–Sso1 signal (Supplementary Fig. 10), confirming that fluorescent Chs6–GFP3 has been inserted into the plasma membrane. We determined the number of newly secreted GFP signals in the plasma membrane. We did not find significant differences in secretion rates for Mcs1–GFP3 and Chs6–GFP3 at the apical growth region or in the mother cell (Fig. 2h), where reduced rates of secretion have been reported17. Moreover, in cells co-expressing mCherry3–Mcs1 and Chs6–GFP3, both signals appear together after photobleaching (Fig. 2i). Taken together, these data are consistent with the notion that Mcs1 and Chs6 are exocytosed together.

Next, we asked for further evidence that Mcs1 is required for secretion of Chs6. We generated a conditional mutant, replacing the mcs1 native promoter with the repressible crg promoter27. In the presence of arabinose, mCherry3–Mcs1 was expressed (Supplementary Fig. 11, Mcs1↑, co-localization with Chs6–GFP3 results in yellow signals), whereas mCherry3–Mcs1 was depleted in glucose-containing medium. Consistent with a role of Mcs1 in Chs6 secretion, low Mcs1 levels resulted in reduced Chs6–GFP3 signals at the growth region (Fig. 2j). Instead, Chs6–GFP3 signals accumulated in the cytoplasm (Supplementary Fig. 11, Mcs1↓; arrowheads). The secretion rate of Chs6–GFP3 at buds was consistently and significantly reduced (Fig. 2k). Taken together, our results support a model in which Mcs1 and Chs6 co-travel in the same secretory vesicle, and this results in co-secretion at sites of cell wall formation. Tethering of these vesicles is performed by the MMD of Mcs1, rendering redundant the MMD of Chs6. This could explain the gradual loss of the MMD in class VII CHSs.

GS delivery and secretion requires Mcs1

We further characterized our CHS mutants by ultrastructural analyses. This revealed that Δchs6 mutant cells form significantly thicker cell walls than wild-type cells (Fig. 3a,b). This suggests that Δchs6 cells reinforce their cell wall. Consistently, Δchs6 cells were more resistant to treatment with cell-wall-digesting enzymes than wild-type cells (Fig. 3c,d). Defects in the fungal cell wall are known to activate the cell-wall integrity pathway, which results in reinforcement of the cell wall due to increased activity of CHSs or GS (refs. 28,​29,​30). We labelled chitin in wild-type control and Δchs6, using N-acetylglucosamine-specific rhodamine-conjugated wheat germ agglutinin. Consistent with previous reports11, the fluorescent lectin labelled the growth region in control cells (Fig. 3e). Chitin staining was less pronounced in Δchs6 cells (Fig. 3e,f), suggesting that the cell wall deposit is not chitin but 1,3-β-glucan. Surprisingly, a double mutant, lacking Chs6 and Mcs1 (Δchs6Δmcs1), showed increased sensitivity to cell-wall-lysing enzymes (Fig. 3c,d). Thus, glucan-based reinforcement of the cell wall does not occur in the absence of Mcs1. This raised the possibility that GS is also travelling in Mcs1-containing vesicles. We set out to test this idea by visualizing GS in U. maydis. We identified the single gene encoding a putative GS, gsc1, in the U. maydis genome (NCBI accession no. XP_011387626.1). The predicted amino-acid sequence of Gsc1 shares 52.8/67.9% sequence identity/similarity and a very similar domain organization to Fks1p from Saccharomyces cerevisiae (NCBI accession no. AJV66071.1; Fig. 3g). This includes an Fks1 domain (P = 2.4 × 10−35, predicted by PFAM), a GS domain (P = 0.0; predicted by PFAM), and numerous transmembrane domains (predicted by SMART). To visualize GS in living cells, we fused the endogenous copy of gsc1 to a triple gfp. The resulting strain showed no defects in morphology or growth, suggesting that the fusion protein is biologically active. Consistent with a role in cell wall formation, GFP3–Gsc1 concentrated at the apical growth region, where it co-localized with mCherry3–Mcs1 (Fig. 3h, co-localization results are in yellow; Fig. 3i, closed arrowheads). Moreover, a significant portion of GFP3–Gsc1 localized to regions where mCherry3–Mcs1 was absent (Fig. 3i and Supplementary Fig. 12; open arrowheads). Next, we co-visualized GFP3–Gsc1 and mCherry3–Mcs1 in photobleached buds of U. maydis. We found that 68.3% of the GFP3–Gsc1 signals co-migrated with mCherry3–Mcs1 (Fig. 3j and Supplementary Video 4), while 31.7% travelled alone (n = 50 cells; Fig. 3k). This suggests that a subset of Gsc1 is delivered via an Mcs1-independent secretion pathway, while the majority of GS arrives at the growth region in Mcs1 vesicles. We tested for a role of Mcs1 in secretion of GS by expressing GFP3–Gsc1 in conditional mcs1 mutants. When mcs1 expression is repressed, significantly less GFP3–Gsc1 is found at the growth region (Fig. 3l) and GFP3–Gsc1 accumulates in the cytoplasm (Fig. 3m, arrowheads). Consistently, the secretion rate was significantly reduced when Mcs1 was depleted (Fig. 3n). These results are consistent with the notion that the majority of GS is delivered and exocytosed in Mcs1-containing vesicles.

Figure 3: Null mutants of mcs1 and chs6 exhibit different cell wall phenotypes.
Figure 3

a, Electron micrographs of control and Δchs6 cell walls (dotted red). Scale bar, 0.1 µm. b, Cell wall thickness, measured in electron micrographs. n = 53 cells, two experiments; ***P < 0.0001. c, Protoplast formation in control and mutants at 10 min of exposure to cell-wall-degrading enzymes. Scale bar, 10 µm. d, Protoplast formation in control and mutants at 10 min of exposure to cell-wall-degrading enzymes. Δchs6 forms fewer protoplasts (P = 0.0179), while Δchs6Δmcs1 forms more (P < 0.0001) protoplasts than Control; n = 5 experiments, 510–593 cells. e, Chitin staining in control cells and Δchs6. Insets: growth regions; arrowheads indicate bud scars. Scale bar, 2 µm. f, Chitin staining in control cells and Δchs6. n = 90 (control) and 71 (Δchs6) cells, two experiments; ***P < 0.0001. g, Domain organization of U. maydis Gsc1 and S. cerevisiae Fks1p. h, Co-localization of mCherry3–Mcs1 (red) and GFP3–Gsc1 (green) at the growth region (co-localization shown in yellow). Scale bar, 2 µm. i, Intensity profile of mCherry3–Mcs1 (red) and GFP3–Gsc1 (blue) at the growth region. Small arrowheads indicate co-localization and large arrowheads indicate Gsc1 alone. See also Supplementary Fig. 11. j, Kymographs of mCherry3–Mcs1 and GFP3–Gsc1 co-motility. Scale bars, 2 s (vertical) and 1 µm (horizontal). See Supplementary Video 4. k, Degree of co-motility of Mcs1 and Gsc1. n = 593 (control), 542 (Δchs6) and 561 (Δchs6Δmcs1) from five experiments; ***P < 0.0001. l, Secreted GFP3–Gsc1 at growth region in conditional Δmcs1. n= 81 (Mcs1↑) and 90 (Mcs1↓) cells from three experiments; ***P < 0.0001. m, GFP3–Gsc1 at the growth region in conditional Δmcs1. Repression of mcs1 results in accumulation of internal signals (Mcs1↓, arrowheads). Scale bar, 2 µm. n, GFP3–Gsc1 secretion rate in conditional mcs1 mutants. Repression of mcs1 (Mcs1↓) reduces GFP3–Gsc1 secretion. n = 30 cells for Mcs1↑ and Mcs1↓, three experiments; **P = 0.002. Non-normal distributed data sets were analysed using Mann–Whitney tests and are shown as whisker plots (b,f,l), with 25th and 75th percentiles indicated by blue lines and the median by a red line; whisker ends represent minimum/maximum values. Gaussian-distributed data sets (k,n) were analysed by unpaired Student's t-test with Welch's correction and are shown as bars of mean ± standard error.

Secretory vesicles contain various numbers of CHSs and GS

Our live-cell imaging experiments suggested that Mcs1, Chs6 and Gsc1 are delivered in the same vesicle. We gathered further evidence for this by ultrastructural investigation of secretory vesicles, purified from strains AB33Ch3Mcs1_Chs6G3 and AB33Ch3Mcs1_G3Gsc1, following protocols published for S. cerevisiae31. In epifluorescent microscopy, these vesicles combined red and green fluorescence, confirming that they contain all three enzymes (Fig. 4a). We then performed immuno-gold labelling, using commercial antibodies raised against mCherry and GFP. The chemically fixed vesicle preparation was homogeneous, and antibodies exclusively labelled the membranes (Supplementary Fig. 13). These experiments confirmed that Mcs1, Chs6 and Gsc1 co-localized in the same vesicles (Fig. 4b). These vesicles had a diameter of 35 nm (Fig. 4c; mean ± standard deviation: 36.81 ± 4.27 nm, n = 16, for vesicles containing Mcs1 and Chs6; 34.04 ± 4.14 nm, n = 21, for vesicles containing Mcs1 and Gsc1). This size is concordant with the dimensions of neuronal secretory vesicles (30–60 nm; refs 32,33). We found variation in the number of Mcs1, Chs6 and Gsc1 per vesicle (1–4 signals; Fig. 4b). To address this quantitatively, we measured the fluorescence intensity of moving Mcs1–GFP3, Chs6–GFP3 and GFP3–Gsc1 in living cells and compared these intensities with fluorescent nuclear pores. Each nuclear pore contains sixteen Nup107–GFP nucleoporin fusion proteins and thus has a uniform fluorescent signal intensity (Fig. 4d; arrowheads). This allows them to be used as an internal calibration standard for measuring protein numbers in living cells of U. maydis34,35. This approach revealed that secretory vesicles usually carry one or two Mcs1 and Chs6 molecules, but more than two Gsc1 molecules co-travel in the vesicles (Fig. 4e; number of Gsc1 significantly different from Mcs1 and Chs6 at P < 0.0001, Mann–Whitney test).

Figure 4: Mcs1, Chs6 and Gsc1 locate in the same vesicle.
Figure 4

a, Merged fluorescent images of secretory vesicle preparations, showing co-localization of mCherry3–Mcs1 (red) and Chs6–GFP3 (red, left panel), as well as mCherry3–Mcs1 and GFP3–Gsc1 (right panel). Note that both channels were shifted slightly to better visualize co-localization. Scale bar, 2 µm. b, Electron micrographs showing immuno-gold-based co-localization of mCherry3–Mcs1 and Chs6–GFP3 (top row; Chs6–GFP3 indicated by large gold particles (large filled circles), mCherry3–Mcs1 by smaller gold particles (small filled circles)), as well as mCherry3–Mcs1 and GFP3–Gsc1 in vesicle preparations that were chemically fixed (bottom row; GFP3–Gsc1 indicated by large gold particles (large filled circles), mCherry3-Mcs1 by smaller gold particles (small filled circles)). Scale bars, 50 nm. c, Diameter of mCherry3–Mcs1 and Chs6–GFP3-, and mCherry3–Mcs1 and GFP3–Gsc1-carrying vesicles. Means ± standard error are shown, n = 16 (Chs6+Mcs1) and 21 (Gsc1+Mcs1), two experiments; NS, non-significant difference at P = 0.057 from Student's t-test with Welch's correction. d, False-coloured image of the fluorescent nucleoporin GFP–Nup107. Individual nuclear pores (arrowheads) contain a known number of GFP–Nup107 molecules and their uniform fluorescent intensity was used as internal calibration standard to determine the number of Mcs1–GFP3, Chs6–GFP3 and GFP3–Gsc1 molecules per moving vesicles (e). Scale bar, 1 µm. e, Estimated numbers of Mcs1–GFP3, Chs6–GFP3 and GFP3–Gsc1 molecules per moving vesicle. Gsc1 is often present in numerous copies. Median is indicated by the red dotted line; n = 100 signals (Mcs1, Chs6, Gsc1), three experiments.

Mcs1, Gsc1 and Chs6 concentrate in stationary focal points of cell wall formation

The co-delivery of Gsc1, Chs6 and Mcs1 in the same vesicle raises the possibility that these enzymes, once exocytosed into plasma membrane, cooperate in wall formation. To investigate this, we observed the ‘behaviour’ of mCherry3–Mcs1, Chs6–GFP3 and GFP3–Gsc1 after insertion into the apical plasma membrane of U. maydis cells. We found that exocytosed Mcs1 and Chs6 or Mcs1 and Gsc1 continued to co-localize and remain stationary in the plasma membrane (Fig. 5a). Secreted mCherry3–Mcs1, Chs6–GFP3 or GFP3–Gsc1 usually co-localized for several minutes (Fig. 5b), suggesting that diffusion in the plasma membrane is restricted. This finding is surprising, as the fluorescent syntaxin GFP–Sso1 diffuses into photobleached regions (Supplementary Fig. 14). Thus, cell-wall-forming enzymes are actively immobilized after exocytosis. We tested if the cytoskeleton restricts diffusion of secreted Mcs1–GFP3 and GFP3–Gsc1 by counting newly secreted fluorescent signals in photobleached growth regions in control cells and mutants that were truncated in the MMD of Mcs1 (ref. 26). In addition, we depolymerized the cytoskeleton with latrunculin A and benomyl, which disassemble F-actin and microtubules in U. maydis, respectively36. In control cells, 80% of all Mcs1 signals are stationary and non-diffusive (Fig. 5c, Control; Supplementary Video 5). This number did not change significantly when the MMD was truncated or the cytoskeleton was disrupted (Fig. 5c, LatA + Ben; no significant difference, P = 0.422, Kruskal–Wallis test). Thus, we conclude that the cytoskeleton has no obvious role in anchoring secreted cell-wall-forming enzymes. Next, we asked whether the enzymes are physically linked to the wall and therefore restricted in diffusion within the plasma membrane. We removed the wall with cell-wall-degrading enzymes and investigated Mcs1–GFP3 in protoplasts (Fig. 5d). In the absence of the wall, most Mcs1–GFP3 showed random motility, and the number of non-diffusive Mcs1–GFP3 signals was reduced significantly (Fig. 5c,e, ‘No cell wall’; P < 0.0001, Student's t-test; Supplementary Video 5). This raised the possibility that synthetic enzymes are anchored via their newly formed polysaccharide chains. We tested this by observing newly secreted Mcs1–GFP3 and GFP3–Gsc1 in the presence of the CHS inhibitor nikkomycin Z (ref. 37) and the GS inhibitor caspofungin38. Indeed, simultaneous treatment with both inhibitors increased the diffusive motion of secreted Mcs1–GFP3 (Fig. 5c,f, ‘NikkoZ+CSG’; Supplementary Video 5) and GFP3–Gsc1 (Fig. 5g, ‘NikkoZ+CSG’). Interestingly, treatment with either nikkomycin Z or caspofungin alone had an intermediate effect on GFP3–Mcs1 mobility (Fig. 5c, ‘NikkoZ’ and ‘CSG’). On the other hand, blocking CHS activity with nikkomycin Z increased the diffusion of Gsc1 (Fig. 5g, ‘NikkoZ’). These results suggest that immobilization of one enzyme depends, in part, on the activity of the other enzymes, suggesting their cooperation during cell wall synthesis. Taken together, these results support a model in which CHSs and GS are co-delivered in the same secretory vesicle, to ensure co-exocytosis and the formation of local foci of cell wall synthesis activity.

Figure 5: Mcs1, Chs6 and Gsc1 remain co-located in stationary wall-forming foci.
Figure 5

a, Kymographs of secreted mCherry3–Mcs1/Chs6–GFP3 and mCherry3–Mcs1/GFP3–Gsc1 at photobleached growth regions. Scale bar, 2 µm. b, Co-localization of newly secreted mCherry3–Mcs1 and Chs6–GFP3 (Mcs1+Chs6), and mCherry3–Mcs1 and GFP3–Gsc1 (Mcs1+Gsc1) 5, 7 and 10 min after photobleaching. n = 3 experiments, 58 cells (Mcs1+Chs) and 56 cells (Mcs1+Gsc1). c, Motility behaviour of secreted Mcs1–GFP3. Signals are stationary in control cells, in cells expressing Mcs1ΔMM, deleted in the MMD of Mcs1 (ΔMM; ref. 26), and after disruption of the cytoskeleton (LatA+Ben). Diffusion increases when the cell wall is digested (‘No cell wall’) or when enzyme activity is inhibited by CHS inhibitor nikkomycin Z (NikkoZ) or β-GS inhibitor caspofungin (CSG). Simultaneous treatment had a stronger effect, suggesting that enzymes are anchored by nascent polysaccharides in the plasma membrane. n = 664 (control), 277 (ΔMM), 419 (LatA+Ben), 288 (No cell wall), 293 (NikkoZ+CSG), 372 (NikkoZ), 387 (CSG), two experiments; ***P < 0.0001; NS, non-significant differences (Control, ΔMM, LatA+Ben: P = 0.422; No cell wall, NikkoZ+CSG: P = 0.646; NikkoZ, CSG: P = 0.965). See Supplementary Video 5. d, Localization of Mcs1–GFP3 in cell-wall-less protoplasts. Scale bar, 2 µm. e, Motility behaviour of secreted Mcs1–GFP3 in control cells and protoplasts (No cell wall). Diffusion increases when the cell wall is removed. Scale bars, 5 s (vertical) and 1 µm (horizontal). f, Motility behaviour of secreted Mcs1–GFP3 when enzyme activity was inhibited by simultaneous treatment with the CHS inhibitor nikkomycin Z (NikkoZ) and the β-GS inhibitor caspofungin (CSG). Scale bars, 5 s (vertical) and 1 µm (horizontal). g, Motility behaviour of secreted GFP3–Gsc1 in the presence of the specific CHS inhibitor nikkomycin Z (NikkoZ) or β-GS inhibitor caspofungin (CSG). Inhibiting enzyme activity increases diffusive motility of GS in the plasma membrane. n = 330 (control), 290 (NikkoZ+CSG), 342 (NikkoZ), 361 (CSG), two experiments; ***P < 0.0001; NS, non-significant differences (NikkoZ+CSG and NikkoZ: P = 0.459; NikkoZ+CSG, NikkoZ, CSG: P = 0.060). Non-normal distributed data sets were analysed using Mann–Whitney tests (pairwise comparison in c and g) and Kruskal–Wallis tests (multiple comparison in c and g), and are shown as whisker plots (c,g), with 25th and 75th percentiles indicated by blue lines and the median by a red line; whisker ends represent minimum/maximum values. Gaussian-distributed data sets (b) are shown as means ± standard error.

Cell wall synthases co-travel in small secretory vesicles in hyphae

U. maydis is a dimorphic fungus: it grows as yeast-like cells and also as tip-growing elongate hyphae. The results described so far were derived from an investigation of the yeast-like growth form, and we set out to extend our study to hyphal cells. Consistent with the extension of the hyphal tip, fluorescent Mcs1, Chs6 and Gsc1 concentrated at the growing hyphal apex (Fig. 6a). After photobleaching, rapidly moving vesicles carrying mCherry3–Mcs1/Chs6–GFP3 and mCherry3–Mcs1/GFP3–Gsc1 were observed (Fig. 6b; Supplementary Fig. 15 and Supplementary Video 6). A small portion of Chs6 and Gsc1 travelled independently of Mcs1 (Fig. 6c), which confirms the results described for yeast-like cells and suggests additional Mcs1-independent secretion pathways. Next, we purified vesicles from hyphal cells of mCherry3–Mcs1/Chs6–GFP3 and mCherry3–Mcs1/GFP3–Gsc1 co-expressing strains. In contrast to the preparation from yeast-like cells, we did not chemically ‘fix’ these vesicles. This improved the quality of the electron micrographs and revealed a largely homogeneous population of small vesicles (diameter 23.40 ± 3.15 nm, n = 30). Immuno-gold labelling indicated that only a small number of these vesicles contain Mcs1. A portion of these Mcs1-positive vesicles was either paired with Chs6 or with Gsc1 (Fig. 6d). Thus, in U. maydis, co-delivery of cell wall synthases in the same vesicle is characteristic of hyphae as well as yeast-like cells.

Figure 6: Cell wall synthases co-travel in the same vesicle in hyphae.
Figure 6

a, Localization of GFP3–Gsc1, Mcs1–GFP3 and Chs6–GFP3 in hyphal cells. Scale bar, 5 µm. b, Kymographs showing anterograde co-motility of mCherry3–Mcs1 and GFP3–Gsc1. Scale bars, 2 s (vertical) and 2 µm (horizontal). c, Degree of co-motility of mCherry3–Mcs1/Chs6–GFP3 and mCherry3–Mcs1 and GFP3–Gsc1. n = 50 for each column, two experiments. Results shown as mean ± standard error. d, Electron micrographs showing immuno-gold-based co-localization of mCherry3–Mcs1/Chs6–GFP3 and mCherry3–Mcs1/GFP3–Gsc1 using antibodies against mCherry and GFP. Note that vesicles were not chemically fixed. Red arrowhead: double-labelled vesicle; yellow arrowhead: only Chs6-GFP3; blue arrowhead: only mCherry3-Mcs1. Scale bars: 30 nm (left) and 100 nm (right). e, Model of secretion of cell-wall-forming enzymes. Kinesin-1 and myosin-5 deliver Mcs1-containing vesicles to the growth region17. These carriers contain other cell-wall-forming enzymes, including Chs6 and Gsc1. Near the plasma membrane, Mcs1 tethers the vesicles, so increasing the likelihood of exocytosis. After insertion into the plasma membrane, the enzymes begin cell wall synthesis. The nascent polysaccharide anchors the enzymes at their location for several minutes. This ensures coordinated formation of the complex fungal cell wall.


This study provides evidence for a central role of fungal myosin–CHSs in focal point secretion of cell-wall-forming enzymes. We integrated fluorescent tags into the native loci of two CHSs and a GS in U. maydis, allowing us to visualize native levels of these cell-wall-forming enzymes. We used these strains to show, both by live-cell imaging and ultrastructural studies, that the Mcs1, Chs6 and Gsc1 co-travel in secretory vesicles. Consistent with a central role of the MMD of Mcs1, we demonstrate that secretion of Chs6 depends largely on Mcs1 activity. On the other hand, exocytosis of Gsc1 is less dependent on Mcs1. A significant proportion of Gsc1 neither co-localizes with Mcs1 nor requires Mcs1 for exocytosis. This suggests the existence of an Mcs1-independent secretion pathway for GS. However, this putative pathway is not sufficient to reinforce the cell wall when chs6 and mcs1 are deleted. This suggests that the Mcs1-dependent delivery of cell-wall-forming enzymes is the dominant pathway. U. maydis contains four polar-localized CHSs (ref. 11), and we speculate that these enzymes are delivered in Mcs1 vesicles, but experimental evidence to support this is missing. However, it is important to note that our ultrastructural results, as well as low rates of anterograde delivery of Mcs1 vesicles17 suggest that cell wall synthase-containing carriers represent a minor population of all secretory vesicles.

While co-delivery of CHSs and GS in the same vesicle is an efficient way of ensuring focused co-exocytosis and coordinated activity of cell-wall-forming enzymes, it may not be the only way of constructing a fungal cell wall. Studies of the localization of cell wall synthases in the Spitzenkörper of N. crassa suggest that several CHSs and GS travel in distinct subpopulations of secretory vesicles20,39,​40,​41. However, N. crassa is a very fast-growing fungus, with estimates of 38,000 vesicles fusing with the expanding tip per minute42. Thus, this fungus may have developed a different strategy for ensuring the rapid formation of the cell wall.

How common is this myosin–CHS dependent secretion pathway? Studies in U. maydis suggest that the class V myosin–CHS supports polar exocytosis by tethering vesicles to apical F-actin17. We demonstrate here that this pathway also delivers other cell-wall-forming enzymes, including class VII CHS and GS. We propose that this co-delivery renders the MMD of class CHS VII redundant in U. maydis. If so, the loss of this MMD in other fungi suggests co-delivery of cell-wall-forming enzymes is of general importance in filamentous fungi. In support of this, we found a single myosin–CHS with a highly conserved MMD in ascomycete and basidiomycete genomes. Even the ancient cryptomycete fungus R. allomyces carries a highly conserved myosin–CHS (ref. 43,44). Thus, we can speculate that this is, indeed, an ancestral pathway. However, variations on this theme are likely, as the genomes of the zygomycete Mucor circinelloides and chytridiomycete Spizellomyces punctatus carry numerous myosin–CHSs, suggesting multiple secretion pathways. More studies are needed to investigate this possibility. On the other hand, ascomycete fungi appear to use the largely reduced MMD of class VII CHSs to bind F-actin16,45, possibly involving a core set of conserved amino acids in their truncated MMD (Supplementary Fig. 16; 53% identical; PFAM). Although this reduced MMD cannot replace its counterpart in class V CHSs (ref. 46), it seems that both CHSs cooperate in the secretion of wall-forming enzymes in A. nidulans47. In essence, fungal cell wall formation involves an ancestral myosin–CHS-dependent secretion pathway. Cell wall synthases are co-delivered in myosin–CHS-containing transport vesicles (Fig. 6e) and, after co-secretion, cooperate in cell wall formation. We consider it likely that the delivery of such ‘cell wall factories’, consisting of multiple cell wall synthases, is a widespread mechanism found in all classes of fungi, including filamentous ascomycetes. As such, this pathway is pivotal to polarized fungal growth, and essential for colonizing the environment and invading animal and plant hosts.


Strains, plasmids and expression constructs

The U. maydis strains AB33, SG200, FB1ΔChs6, FB1ΔChs6ΔMcs1, SG200ΔChs6, AB33_Mcs1G3, AB33mCh3Mcs1, SG200G3Mcs1ΔMM, SG200G3Mcs1, FB1Chs6G3, AB33GSso1 and FB2N107G have been described previously11,17,21,26,48,​49,​50,​51. The genotypes of all strains used in this study are summarized in Supplementary Table 1, and the experimental usage of each strain is summarized in Supplementary Table 2. All plasmids were generated using standard techniques or in vivo recombination in S. cerevisiae, following published protocols52. Cloning primers are summarized in Supplementary Table 3.

Plasmid construction

The pnChs6G3 plasmid contains the U. maydis class VII CHS gene, fused to three copies of the egfp gene. This plasmid was constructed by replacing the phleomycin resistance cassette from plasmid pChs6G3 (ref. 17) with the hygromycin phospho-transferase gene resistance cassette (hygR), by standard procedures.

The pnG3Gsc1 plasmid contains a triple egfp tag fused to the gene for the catalytic subunit of GS. Fragments of 1,011 bp from the gsc1 5′ upstream region, 1,032 bp of the gsc1 promoter, and 1,006 bp open reading frame (ORF) of gsc1 were amplified by PCR from genomic DNA of U. maydis strain 521. A 717 bp fragment from the egfp ORF and 2,798 bp of the hygR cassette were amplified from the plasmid pHrpl25G (ref. 53). This was done using primers YH166, YH269, YH270, YH273, YH421, YH422, YH423, YH424, YH425 and YH426 (Supplementary Table 3). Ligation of all fragments was carried out by in vivo recombination in S. cerevisiae, following published protocols52. Two additional copies of egfp were introduced into a BsrGI restriction site of the resulting plasmid pnGGsc1 using standard ligation techniques.

The pnGChs6 plasmid contains the U. maydis class VII CHS gene, fused to a single copy of the egfp gene. The fusion construct is expressed under the chs6 native promoter. It was constructed from a DNA fragment of 2,220 bp, covering the chs6 promoter, 717 bp from the egfp ORF, 3,543 bp from the chs6 ORF and 1,201 bp from the chs6 terminator amplified by PCR either with genomic DNA of U. maydis strain 521 or plasmid pn3GMcs1 (ref. 17) as templates, and using primers MU213, MU214, MU215, SK179, MU216, MU217, MU211 and MU218 (Supplementary Table 3). The products were recombined into the BamHI and HindIII digested plasmid pNEBcbx-yeast34 in S. cerevisiae, resulting in pnGChs6.

The pnGM-Chs6 plasmid contains the MMD of the U. maydis class V CHS gene, fused to the class VII CHS gene and to a single copy of the egfp gene. The fusion construct is expressed under the chs6 native promoter. It was constructed from a DNA fragment of 2,220 bp, covering the chs6 promoter, 717 bp from the egfp ORF, 2,745 bp from the mcs1 N-terminal ORF, 3,369 bp from the chs6 C-terminal ORF and 1,201 bp of the chs6 terminator. The fragments were amplified by PCR either with genomic DNA of U. maydis strain 521 or plasmid pn3GMcs1 (ref. 17) as templates, and using primers SK50, SK434, SK179, MU209, MU210 and MU211 (Supplementary Table 3). The products were recombined into the BamHI and HindIII digested plasmid pNEBcbx-yeast34 in S. cerevisiae, resulting in pnGM-Chs6.

The pPcmCh3Mcs1 plasmid allows expression of the fusion gene mcherry3-mcs1 under the conditional crg-promoter27. It was generated by replacing the 958 bp fragment of the mcs1 promoter in plasmid pmCh3Mcs1 (ref. 17) with a 3,525 bp fragment containing the crg1 promoter, which was obtained from pCcrgKin1rigor (ref. 17) using primers YH461 and YH462. The plasmid was generated using in vivo recombination in S. cerevisiae, following published protocols52.

Strain generation

To generate strains AB33Chs6G3 and AB33mCh3Mcs1_Chs6G3, plasmid pChs6G3 was digested with BglI and integrated into the chs6 locus of strains AB33 (ref. 48) and AB33mCh3Mcs1 (ref. 17), respectively. Strain AB33Mcs1G3 was generated by homologous integration of plasmid pMcs1G3 (ref. 17) into strain AB33. Strains AB33G3Gsc1 and AB33Ch3Mcs1_G3Gsc1 were generated by linearization and integration of plasmid pG3Gsc1 into the gsc1 locus of strains AB33 or AB33mCh3Mcs1 (ref. 17). AB33Mcs1G3_ChSso1 and AB33Chs6G3_ChSso1 were generated by ectopic integration of plasmid pomChSso1 (ref. 26), linearized with PciI, into the genome of strains AB33Mcs1G3 and AB33Chs6G3. Strain FB1ΔChs6GChs6 was generated by ectopic integration of plasmid pnGChs6, linearized with HpaI, into the genome of FB1ΔChs6 (ref. 21). FB1Chs6G3_rCh3Mcs1 and AB33G3Gsc1_rCh3Mcs1contain plasmid pPcmCh3Mcs1, digested with SapI and integrated into the mcs1 locus of strains FB1Chs6G3 (ref. 17) and AB33G3Gsc1. Strains FB1ΔChs6ΔMcs1GChs6 and FB1ΔChs6ΔMcs1GChs6MMD were generated by ectopic integration of plasmids pnGChs6 or pnGM-Chs6 linearized with HpaI into the genome of FB1ΔChs6ΔMcs1 (ref. 11).

Growth conditions

U. maydis liquid cultures were grown for 8–12 h in complete medium (CM, ref. 54) containing 1% (wt/vol) glucose (CMglucose) with shaking at 200 revolutions per minute (r.p.m.) at 28 °C. To induce hyphal growth, cells were washed and transferred to nitrate minimal medium (NM, ref. 48) supplemented with 1% (wt/vol) glucose (NMglucose). Cells were grown under these conditions for 5–10 h at 200 r.p.m. at 28 °C. Strains FB1Chs6G3_rCh3Mcs1 and AB33G3Gsc1_rCh3Mcs1 were grown in complete medium containing 1% (wt/vol) arabinose, which induces gene expression from the crg1 promoter27. To repress expression of mcs1, cells were transferred into CMglucose for 12 h at 28 °C, with shaking at 200 r.p.m.

Laser-based epifluorescence microscopy

Cells were placed onto a 2% (wt/vol) agar cushion and observed using an IX81 motorized inverted microscope (Olympus), equipped with a PlanApo ×100/1.45 oil TIRF objective (Olympus) and a VS-LMS4 Laser-Merge-System with solid-state lasers (488 nm/70 mW and 561 nm/70 mW, Visitron System). For photobleaching experiments, a 405 nm/60 mW diode laser was used, which was attenuated by an ND 0.6 filter, resulting in 15 mW output power, coupled into the light path by an OSI-IX 71 adaptor (Visitron System) and controlled by a UGA-40 controller (Rapp OptoElectronic) and VisiFRAP 2D FRAP control software for Meta Series 7.5.x (Visitron System). Simultaneous observation of red and green fluorescent protein fluorescence was achieved with a Dual-View Microimager (Photometrics) equipped with a dual line beamsplitter (z491/561, Chroma), an emission beamsplitter (565 DCXR, Chroma), an ET-Bandpass 525/50 (Chroma) and a BrightLine HC 617/73 (Samrock). Images were acquired using a Photometrics CoolSNAP HQ2 camera (Roper Scientific). All parts of the system were under the control of the software package MetaMorph (MDS Analytical Technologies). Samples were observed for no longer than 10 min, to prevent oxygen depletion. All image processing was performed in MetaMorph.

Statistical analyses

All statistical analyses were performed using the software Prism5 (GraphPad). Before comparison, data sets were tested for normal distribution using a Shapiro–Wilk normality test. When normally distributed (P < 0.05), they were further analysed using unpaired two-tailed Student's t-testing, with Welch's correction, to account for potential differences in the variances of the data sets. The means ± standard error of such normally distributed data are shown as bar charts in the figures.

In cases where at least one data set was non-normally distributed (P > 0.05 in Shapiro–Wilk tests), we used non-parametric Mann–Whitney testing. Multiple data sets, where at least one data set was non-normal, were compared using non-parametric Kruskal–Wallis testing. Analysis of non-normal data sets is represented by whisker plots (= box plots), which show 25th/75th percentiles as blue lines, the median as a red line, and the minimum and maximum values by the ends of the whiskers. All values that are incorporated into the main body of the manuscript are given as means ± standard deviation unless stated otherwise.

Quantitative analysis of fluorescent intensities and motility

All measurements were carried out in 14-bit images using the software MetaMorph. Motility measurements of Mcs1G3 and Chs6G3 were determined in strains AB33Mcs1G3 and AB33Chs6G3. For velocity, frequency and pausing time measurements of Mcs1G3 and Chs6G3, an image series of 150 frames at 200 ms of the 488 nm observation laser were taken after photobleaching of the whole bud by a 75 ms light pulse using a 405 nm laser (60 mW) at 100% laser power (beam diameter 30 pixels). Velocities, frequencies and pausing times of Mcs1G3 and Chs6G3 signals were measured in kymographs using MetaMorph. Intensity measurements of Chs6G3 and G3Gsc1 were determined in strains FB1Chs6G3_rCh3Mcs1 and AB33G3Gsc1_rCh3Mcs1 under mCh3Mcs1 induced or repressed conditions. Pictures were taken at an exposure time of 200 ms of the 488 and 561 nm observation lasers. Analysis of the signal intensities of Chs6G3 and G3Gsc1 at the growth region was performed by defining an area of interest in the apical growth regions, followed by measuring the average intensities at the plasma membrane using MetaMorph. All measured values were corrected for background and the mean average values were calculated.

Quantitative analysis of co-motility

Co-motility of Ch3Mcs1 with Chs6G3 or G3Gsc1 was analysed in the yeast-like and hyphal growth forms of strains AB33Ch3Mcs1_Chs6G3 and AB33Ch3Mcs1_G3Gsc1. Cells were placed onto 2% (wt/vol) agar cushions and the entire bud or 20 µm of the hyphal tip was photobleached with a 75 ms light pulse from a 405 nm laser (60 mW) at 100% laser power (beam diameter 30 pixels). After a 5 s pause, image series of 75 frames at 150 ms and binning 2 were taken using both the 488 nm and the 562 nm observation lasers at 50% output power. Co-motility was analysed in kymographs using MetaMorph.

FRAP-based secretion assays

Mcs1G3, Chs6G3 and G3Gsc1 secretion rates were determined essentially as described in ref. 17. In brief, cells of strains AB33Mcs1G3_ChSso1, AB33Chs6G3_ChSso1, FB1Chs6G3_rCh3Mcs1 and AB33G3Gsc1_rCh3Mcs1 were observed by live-cell microscopy. Reference images were taken at 70% output power of the 488 nm and 100% output power of the 561 nm observation lasers at an exposure time of 250 ms followed by a 100 ms light pulse using a 405 nm laser (60 mW) at 100% laser power (beam diameter 30 pixels) to bleach the growth region or one flank of the mother cell. An image was acquired directly after laser treatment to confirm successful photobleaching. Subsequently, an image series of 75 frames at 5 min after photobleaching was acquired, using 70% of the 488 nm and 100% of the 561 nm observation lasers at an exposure time of 250 ms. Stable insertions of fluorescent signals at the cell periphery, confirmed in kymographs, were considered exocytosed enzymes. The number of inserted signals per 1 µm plasma membrane and per 5 min after photobleaching was determined. To determine the position of secreted Chs6G3 relative to the mChSso1-labelled plasma membrane, intensity profiles of inserted signals were taken and overlaid with mChSso1 images, taken prebleaching. Intensity values of 14 secreted signals were measured and averaged.

Co-localization before and after secretion

mCh3Mcs1 and Chs6G3 or G3Gsc1 co-localizations were determined by photobleaching the growth region in cells of strains AB33Ch3Mcs1_Chs6G3 and AB33Ch3Mcs1_G3Gsc1, placed on a 2% (wt/vol) agar cushion, using a 100 ms 405 nm laser light pulse (60 mW; beam diameter 30 pixels). For co-localization during delivery to the plasma membrane, an image series of 75 frames was taken 2 s after bleaching using 70% of the 488 nm and 80% of the 561 nm observation lasers at an exposure time of 150 ms. For co-localization, after secretion to the plasma membrane, an image series of 75 frames was taken 5, 7 or 10 min after bleaching using 70% of the 488 nm and 80% of the 561 nm observation lasers at an exposure time of 200 ms. Stable insertion of individual signals was confirmed in kymographs using MetaMorph. The percentage of inserted Chs6G3 or G3Gsc1 signals that co-localized with mCh3Mcs1 after 5, 7 or 10 min was determined.

Protoplast formation

Protoplast formation was carried out following previously described procedures55. Strains SG200, SG200ΔChs6 and FB1ΔMcs1ΔChs6 were grown overnight in YEPSlight liquid media (0.4% (wt/vol) yeast extract, 0.4% (wt/vol) peptone and 2% (wt/vol) sucrose), reaching an optical density at 600 nm of 0.6–0.8. Cells were centrifuged in a Heraeus Biofuge Stratos benchtop centrifuge (Kendro Laboratory Products) at 3,000 r.p.m. for 10 min. The cell pellet was washed by resuspending in sodium citrate sorbitol buffer (SCS, 20 mM sodium citrate, pH 5.81, 1 M sorbitol) and subsequent centrifugation. The resulting sediment was resuspended in SCS containing 12.5 mg ml−1 novozyme lytic enzymes (Novo Nordisk) and kept at room temperature for 10–15 min. Subsequently, the degree of protoplast formation was assessed by light microscopy. Cells were considered ‘protoplasts’ when the cells lost their ‘cigar-shaped’ appearance and became partially or fully rounded.

Secretory vesicle purification

Secretory vesicles were purified from the yeast and hyphal growth forms of strains AB33mCh3Mcs1_Chs6G3 and AB33mCh3Mcs1_G3Gsc1, following published procedures31. For purification from yeasts, cells were grown overnight in YEPSlight liquid medium. For hyphal cells the strains were grown over night in CM (ref. 12) containing 1% (wt/vol) glucose (CMglucose), with shaking at 200 revolutions per minute (r.p.m.) at 28 °C. To induce hyphal growth, cells were washed and transferred to nitrate minimal medium (NM, ref. 48) supplemented with 1% (wt/vol) glucose (NMglucose). Cells were grown under these conditions for 6 h at 200 r.p.m. at 28 °C. Protoplasts were prepared as previously described55. Protoplasts were sedimented by centrifugation and resuspended in 100 mM PIPES buffer (pH 7.2), containing 0.8 M sorbitol, 1 mM EDTA adjusted to pH 7.2 and protease inhibitor cocktail (Sigma-Aldrich). The protoplasts were disrupted in a Dounce homogenizer on ice and the cell extract was centrifuged at 10,000g for 10 min, using a Heraeus Biofuge Stratos benchtop centrifuge. The supernatant was collected and ultracentrifuged at 100,000g for 1 h, using a TLA 120.1 rotor and Optima MAX Ultracentrifuge (Beckman Coulter). After removal of the supernatant, the clear microsomal pellet was resuspended in 40 µl PIPES buffer.

Determining protein numbers in secretory vesicles

Numbers of G3Gsc1, Chs6G3 and Mcs1G3 molecules in moving fluorescent dots were estimated in cells of strains AB33Mcs1G3, AB33Chs6G3 and AB33G3Gsc1. To determine the number of proteins in a moving signal, short image series were captured. Moving signals were identified and their fluorescent signal intensity, corrected for the adjacent background, was determined in the first frame of the image sequence. This fluorescent intensity was compared to the fluorescent intensity of individual fluorescent nuclear pores in strain FB2N107G (ref. 56), following published protocols34,35.

Ultrastructural studies

Liquid cultures of SG200 and SG200ΔChs6 were grown overnight, cells were collected by centrifugation at 4,000g for 5 min in a 15 ml Falcon tube, and the pellet was fixed in 2% (vol/vol) glutaraldehyde and 2% (vol/vol) formaldehyde in 0.1 M PIPES buffer pH 7.2, overnight. Samples were post-fixed in 2% (wt/vol) potassium permanganate in dH2O, dehydrated through increasing concentrations of ethanol (50–100%) and embedded in Durcupan resin (Sigma Aldrich). Ultrathin sections (80 nm) were collected on pioloform-coated EM copper grids (Agar Scientific), contrasted with lead citrate and examined using a JEOL JEM 1400 transmission electron microscope operated at 120 kV. Images were taken with a digital camera (ES 100 W charge coupled device, Gatan). To quantify the cell wall thickness, sections were sampled using systematic uniform random procedures and images were analysed using MetaMorph. Measurements were made from profiles showing a visible plasma membrane, to ensure that a medium section was chosen.

For immuno-gold labelling, purified secretory vesicles from strains AB33Ch3Mcs1_Chs6G3 and AB33Ch3Mcs1_G3Gsc1 were fixed in 1% formaldehyde (in 0.1 M PIPES, pH 7.2) and adhered to pioloform-coated 100-mesh copper EM grids (Agar Scientific). Alternatively, vesicles from hyphal cells were adhered to EM grids without prior chemical fixation. After washing with PBS (pH 7.2), grids were incubated with 0.5% (wt/vol) fish skin gelatine (Sigma Aldrich) in PBS for 10 min, followed by incubation with a rabbit anti-mCherry antibody (MBL, PM005) for 30 min. After washing with PBS (3 × 5 min), the grids were incubated with 5 nm protein A gold (BBI Solutions) for 20 min. This was followed by fixation in 1% (vol/vol) glutaraldehyde/PBS for 5 min and additional incubation in 0.05% (wt/vol) glycine/PBS for 5 min. After washing with PBS, the primary rabbit anti-GFP antibody (Abcam, ab6556) was applied for 30 min followed by subsequent washes in PBS (3 × 5 min) before incubating with 10 nm protein A gold (BBI Solutions). All antibodies and protein A gold were diluted with 0.5% (wt/vol) fish skin gelatine in PBS (pH 7.2). Finally, the samples were washed in PBS (6 × 5 min) and distilled water (10 × 1 min) before contrasting in a 1:9 mixture of 2% (wt/vol) uranyl acetate: 2% (wt/vol) methylcellulose for 10 min. The grids were then air-dried and analysed using a JEOL JEM 1400 electron microscope. The specificity of the gold signal was verified by applying the protein A gold alone. The average diameter of labelled vesicles was estimated from digital images.

Wheat germ agglutinin labelling of chitin

Fluorescein isothiocyanate (FITC)-conjugated wheat germ agglutinin stain was performed as described previously11,57. Stained cells of strains SG200 and SG200ΔChs6 were placed onto 2% (wt/vol) agar cushions, and images of cells were taken using 40% of the 488 nm observation laser at an exposure time of 50 ms. Measurement of wheat germ agglutinin stain intensity at the growth region was performed by determining an area of interest at the apical growth regions, using digital images. This was followed by measuring the average intensities at the cell wall using MetaMorph. All measured values were corrected for background and the mean average values were calculated.

Inhibitor experiments

For all inhibitor experiments in liquid culture, logarithmically growing cells of strains SG200G3Mcs1 and SG200G3Mcs1ΔMM were incubated for 30 min with benomyl at 30 µM (stock: 30 mM in DMSO; Fluka), latrunculin A at 10 µM (stock: 20 mM in DMSO; provided by K. Tenney, University of California) or nikkomycin Z at 5 µM (stock: 5 mg in mQH2O; Sigma-Aldrich) and caspofungin 10 µgml−1 (stock: 10 mgml−1 in mQH2O; provided by N. J. Talbot, University of Exeter). Control cells were treated with respective amounts of the solvent DMSO used in the inhibitor assays. Cells were placed onto a 2% (wt/vol) agar cushion containing the same concentration of their respective inhibitors and observed immediately. For measurements of plasma membrane insertion of individual signals of G3Mcs1 and G3Mcs1ΔMM, image series of 100 frames were taken 5 min after photobleaching, using 70% of the 488 nm observation laser at an exposure time of 200 ms. Stable insertion of individual signals was confirmed in kymographs using MetaMorph. The number of stationary and non-stationary signals per 1 µm plasma membrane and per 5 min was determined.


Orthologues of class V and class VII CHSs were identified using a BLAST search at NCBI (, using the predicted amino-acid sequence of U. maydis Mcs1 (NCBI accession no. XP_011389642.1) and Chs6 (NCBI accession no. XP_011389509.1) as bait. Predicted CHS sequences for S. punctatus were obtained from the Broad Institute ( All accession numbers for the predicted protein sequences are provided in Fig. 1a, Supplementary Fig. 1a and Supplementary Fig. 16. The degree of sequence identity and similarity between proteins was determined by using EMBOSS needle ( The protein domain predictions were done in PFAM ( and SMART ( Sequence alignments were done using ClustalOmega ( Maximum likelihood phylogenetic trees were calculated in MEGA 5.2 (; ref. 58), using Bootstrap testing and 1,000 replications.


  1. 1.

    et al. Reconstructing the early evolution of fungi using a six-gene phylogeny. Nature 443, 818–822 (2006).

  2. 2.

    & Dating the molecular clock in fungi—how close are we? Fungal Biol. Rev. 24, 1–16 (2010).

  3. 3.

    The fungi: 1, 2, 3 … 5.1 million species? Am. J. Bot. 98, 426–438 (2011).

  4. 4.

    Hyphal growth: a tale of motors, lipids, and the Spitzenkorper. Eukaryot. Cell 6, 351–360 (2007).

  5. 5.

    & Mannosylation in Candida albicans: role in cell wall function and immune recognition. Mol. Microbiol. 90, 1147–1161 (2013).

  6. 6.

    et al. Effector-mediated suppression of chitin-triggered immunity by Magnaporthe oryzae is necessary for rice blast disease. Plant Cell 24, 322–335 (2012).

  7. 7.

    The cell wall: a carbohydrate armour for the fungal cell. Mol. Microbiol. 66, 279–290 (2007).

  8. 8.

    Tasting the fungal cell wall. Cell. Microbiol. 12, 863–872 (2010).

  9. 9.

    & The Spitzenkorper: a choreographer of fungal growth and morphogenesis. Curr. Opin. Microbiol. 20, 27–33 (2014).

  10. 10.

    & Multiple genes for chitin synthase in the zygomycete fungus Phycomyces blakesleeanus. J. Gen. Appl. Microbiol. 43, 333–340 (1997).

  11. 11.

    , , & Polar localizing class V myosin chitin synthases are essential during early plant infection in the plant pathogenic fungus Ustilago maydis. Plant Cell 18, 225–242 (2006).

  12. 12.

    et al. Different chitin synthase genes are required for various developmental and plant infection processes in the rice blast fungus Magnaporthe oryzae. PLoS Pathog. 8, e1002526 (2012).

  13. 13.

    & Large-scale phylogenetic classification of fungal chitin synthases and identification of a putative cell-wall metabolism gene cluster in Aspergillus genomes. PloS One 9, e104920 (2014).

  14. 14.

    , & The myosin motor domain-containing chitin synthase PdChsVII is required for development, cell wall integrity and virulence in the citrus postharvest pathogen Penicillium digitatum. Fungal Genet. Biol. 67, 58–70 (2014).

  15. 15.

    , & CsmA, a class V chitin synthase with a myosin motor-like domain, is localized through direct interaction with the actin cytoskeleton in Aspergillus nidulans. Mol. Biol. Cell 16, 01961–01970 (2005).

  16. 16.

    , , & Myosin motor-like domain of the class VI chitin synthase CsmB is essential to its functions in Aspergillus nidulans. Biosci. Biotechnol. Biochem. 73, 1163–1167 (2009).

  17. 17.

    et al. Myosin-5, kinesin-1 and myosin-17 cooperate in secretion of fungal chitin synthase. EMBO J. 31, 214–227 (2012).

  18. 18.

    , , & ChsVb, a class VII chitin synthase involved in septation, is critical for pathogenicity in Fusarium oxysporum. Eukaryot. Cell 7, 112–121 (2008).

  19. 19.

    et al. Chitin synthases with a myosin motor-like domain control the resistance of Aspergillus fumigatus to echinocandins. Antimicrob. Agents Chemother. 56, 6121–6131 (2012).

  20. 20.

    et al. Traffic of chitin synthase 1 (CHS-1) to the Spitzenkörper and developing septa in hyphae of Neurospora crassa: actin dependence and evidence of distinct microvesicle populations. Eukaryot. Cell 10, 683–695 (2011).

  21. 21.

    et al. Loss of virulence in Ustilago maydis by Umchs6 gene disruption. Res. Microbiol. 155, 87–97 (2004).

  22. 22.

    , , & Distantly related sequences in the alpha- and beta-subunits of ATP synthase, myosin, kinases and other ATP-requiring enzymes and a common nucleotide binding fold. EMBO J. 1, 945–951 (1982).

  23. 23.

    , , & The A-loop, a novel conserved aromatic acid subdomain upstream of the Walker A motif in ABC transporters, is critical for ATP binding. FEBS Lett. 580, 1049–1055 (2006).

  24. 24.

    , , , & Deletion of the myopathy loop of Dictyostelium myosin II and its impact on motor functions. J. Biol. Chem. 274, 37840–37844 (1999).

  25. 25.

    & Conserved residue clustering and protein structure prediction. Proteins 52, 225–235 (2003).

  26. 26.

    , , & The myosin motor domain of fungal chitin synthase V is dispensable for vesicle motility but required for virulence of the maize pathogen Ustilago maydis. Plant Cell 22, 2476–2494 (2010).

  27. 27.

    , & Isolation of a carbon source-regulated gene from Ustilago maydis. Mol. Gen. Genet. 253, 342–352 (1996).

  28. 28.

    , & Cell wall integrity signaling in human pathogenic fungi. Cell Microbiol. (2016).

  29. 29.

    Regulation of cell wall biogenesis in Saccharomyces cerevisiae: the cell wall integrity signaling pathway. Genetics 189, 1145–1175 (2011).

  30. 30.

    Chitin and glucan, the yin and yang of the fungal cell wall, implications for antifungal drug discovery and therapy. Adv. Appl. Microbiol. 83, 145–172 (2013).

  31. 31.

    & Purification and characterization of constitutive secretory vesicles from yeast. J. Cell Biol. 105, 163–174 (1987).

  32. 32.

    Intracellular organelle dynamics at nm resolution. Methods Cell. Biol. 90, 19–37 (2008).

  33. 33.

    et al. Molecular anatomy of a trafficking organelle. Cell 127, 831–846 (2006).

  34. 34.

    , , , & Transient binding of dynein controls bidirectional long-range motility of early endosomes. Proc. Natl Acad. Sci. USA 108, 3618–3623 (2011).

  35. 35.

    et al. Controlled and stochastic retention concentrates dynein at microtubule ends to keep endosomes on track. EMBO J. 30, 652–664 (2011).

  36. 36.

    , & Microtubules are dispensable for the initial pathogenic development but required for long-distance hyphal growth in the corn smut fungus Ustilago maydis. Mol. Biol. Cell 16, 2746–2758 (2005).

  37. 37.

    , , & Nikkomycin Z is a specific inhibitor of Saccharomyces cerevisiae chitin synthase isozyme Chs3 in vitro and in vivo. J. Bacteriol. 176, 5857–5860 (1994).

  38. 38.

    et al. The antifungal echinocandin caspofungin acetate kills growing cells of Aspergillus fumigatus in vitro. Antimicrob. Agents Chemother. 46, 3001–3012 (2002).

  39. 39.

    , & Functional stratification of the Spitzenkorper of Neurospora crassa. Mol. Microbiol. 74, 1044–1053 (2009).

  40. 40.

    & Live imaging of β-1,3-glucan synthase FKS-1 in Neurospora crassa hyphae. Fungal Genet. Biol. 82, 104–107 (2015).

  41. 41.

    et al. Dissecting the function of the different chitin synthases in vegetative growth and sexual development in Neurospora crassa. Fungal Genet. Biol. 75, 30–45 (2015).

  42. 42.

    & Hyphal tips of wild-type and spreading colonial mutants of Neurospora crassa. Arch. Microbiol. 99, 353–368 (1974).

  43. 43.

    & No jacket required—new fungal lineage defies dress code: recently described zoosporic fungi lack a cell wall during trophic phase. BioEssays 34, 94–102 (2012).

  44. 44.

    et al. Shared signatures of parasitism and phylogenomics unite cryptomycota and microsporidia. Curr. Biol. 23, 1548–1553 (2013).

  45. 45.

    , , & Aspergillus nidulans class V and VI chitin synthases CsmA and CsmB, each with a myosin motor-like domain, perform compensatory functions that are essential for hyphal tip growth. Mol. Microbiol. 59, 1380–1394 (2006).

  46. 46.

    , & Myosin motor-like domain of class VI chitin synthase CsmB of Aspergillus nidulans is not functionally equivalent to that of class V chitin synthase CsmA. Biosci. Biotechnol. Biochem. 77, 369–374 (2013).

  47. 47.

    et al. Transportation of Aspergillus nidulans class III and V chitin synthases to the hyphal tips depends on conventional kinesin. PloS One 10, e0125937 (2015).

  48. 48.

    , , & Identification of genes in the bW/bE regulatory cascade in Ustilago maydis. Mol. Microbiol. 42, 1047–1063 (2001).

  49. 49.

    , & Dynamic rearrangement of nucleoporins during fungal ‘open’ mitosis. Mol. Biol. Cell 19, 1230–1240 (2008).

  50. 50.

    , , & Genetic regulation of mating and dimorphism in Ustilago maydis. Can. J. Bot. 73, 320–325 (1995).

  51. 51.

    & The dynamic fungal cell. Fungal Biol. Rev. 25, 14–37 (2011).

  52. 52.

    & Yeast recombination-based cloning as an efficient way of constructing vectors for Zymoseptoria tritici. Fungal Genet. Biol. 79, 76–83 (2015).

  53. 53.

    , , & Early endosome motility spatially organizes polysome distribution. J. Cell Biol. 204, 343–357 (2014).

  54. 54.

    in Ustilago maydis (ed. King, R. C.) 575–595 (Plenum Press, 1974).

  55. 55.

    et al. The b alleles of U. maydis, whose combinations program pathogenic development, code for polypeptides containing a homeodomain-related motif. Cell 60, 295–306 (1990).

  56. 56.

    et al. Motor-driven motility of fungal nuclear pores organizes chromosomes and fosters nucleocytoplasmic transport. J. Cell Biol. 198, 343–355 (2012).

  57. 57.

    , , & A putative endosomal t-SNARE links exo- and endocytosis in the phytopathogenic fungus Ustilago maydis. EMBO J. 19, 1974–1986 (2000).

  58. 58.

    et al. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol. Biol. Evol. 28, 2731–2739 (2011).

Download references


The authors thank U. Fuchs, S. Milne and P. Splatt for technical support. M.M.-U. thanks N.J. Talbot for financial support. G.S. acknowledges J. Stajich for fruitful discussions. This work was supported by the Biotechnology & Biological Sciences Research Council (grants BB/H019774/1 and BB/I020667/1 to G.S.).

Author information

Author notes

    • Martin Schuster
    •  & Magdalena Martin-Urdiroz

    Present address: Faculty of Agriculture, Kyushu University, Hakozaki 6-10-1, Fukuoka 812-8581, Japan

    • Yujiro Higuchi

    These authors contributed equally to this work.


  1. Biosciences, University of Exeter, Stocker Road, Exeter EX4 4QD, UK

    • Martin Schuster
    • , Magdalena Martin-Urdiroz
    • , Yujiro Higuchi
    • , Christian Hacker
    • , Sreedhar Kilaru
    • , Sarah J. Gurr
    •  & Gero Steinberg


  1. Search for Martin Schuster in:

  2. Search for Magdalena Martin-Urdiroz in:

  3. Search for Yujiro Higuchi in:

  4. Search for Christian Hacker in:

  5. Search for Sreedhar Kilaru in:

  6. Search for Sarah J. Gurr in:

  7. Search for Gero Steinberg in:


G.S. developed the research plan and the experimental strategy, directed the project, analysed data, assembled all figures and videos and wrote the manuscript (with the exception of the Methods, which was written mainly by M.M.-U., M.S. and C.H.). M.M.-U. generated strains and plasmids, performed experiments and analysed data. M.S. performed experiments, generated strains and analysed data. Y.H. and S.K. generated strains and plasmids. C.H. performed electron microscopy. S.J.G. discussed the project and data. All authors corrected the manuscript and discussed data.

Competing interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to Gero Steinberg.

Supplementary information

PDF files

  1. 1.

    Supplementary Information

    Supplementary Figures 1-6 (Colour-adjusted main figures for readers with red-green colour blindness), Supplementary Figures 7-16, Supplementary Tables 1-3, Supplementary References


  1. 1.

    Supplementary Video 1

    Motility of Chs6-GFP3 in U. Maydis.

  2. 2.

    Supplementary Video 2

    Motility of Mcs1-GFP3 and Chs6-GFP3 in a yeast-like cell of U. maydis.

  3. 3.

    Supplementary Video 3

    Co-motility of mCherry3-Mcs1 and Chs6-GFP3 in a yeast-like cell of U. maydis.

  4. 4.

    Supplementary Video 4

    Co-motility of mCherry3-Mcs1 and GFP3-Gsc1 in a yeast-like cell of U. maydis.

  5. 5.

    Supplementary Video 5

    Diffusive motion of Mcs1-GFP3 in control cells, cell wall-less protoplasts and in the presence of the CHS inhibitor nikkomycin Z and the GS inhibitor caspofungin.

  6. 6.

    Supplementary Video 6

    Co-motility of mCherry3-Mcs1 and GFP3-Gsc1 in a hyphal cell of U. maydis.

About this article

Publication history





Further reading