ATP prevents Woronin bodies from sealing septal pores in unwounded cells of the fungus Zymoseptoria tritici

Abstract Septa of filamentous ascomycetes are perforated by septal pores that allow communication between individual hyphal compartments. Upon injury, septal pores are plugged rapidly by Woronin bodies (WBs), thereby preventing extensive cytoplasmic bleeding. The mechanism by which WBs translocate into the pore is not known, but it has been suggested that wound‐induced cytoplasmic bleeding “flushes” WBs into the septal opening. Alternatively, contraction of septum‐associated tethering proteins may pull WBs into the septal pore. Here, we investigate WB dynamics in the wheat pathogen Zymoseptoria tritici. Ultrastructural studies showed that 3.4 ± 0.2 WBs reside on each side of a septum and that single WBs of 128.5 ± 3.6 nm in diameter seal the septal pore (41 ± 1.5 nm). Live cell imaging of green fluorescent ZtHex1, a major protein in WBs, and the integral plasma membrane protein ZtSso1 confirms WB translocation into the septal pore. This was associated with the occasional formation of a plasma membrane “balloon,” extruding into the dead cell, suggesting that the plasma membrane rapidly seals the wounded septal pore wound. Minor amounts of fluorescent ZtHex1‐enhanced green fluorescent protein (eGFP) appeared associated with the “ballooning” plasma membrane, indicating that cytoplasmic ZtHex1‐eGFP is recruited to the extending plasma membrane. Surprisingly, in ~15% of all cases, WBs moved from the ruptured cell into the septal pore. This translocation against the cytoplasmic flow suggests that an active mechanism drives WB plugging. Indeed, treatment of unwounded and intact cells with the respiration inhibitor carbonyl cyanide m‐chlorophenyl hydrazone induced WB translocation into the pores. Moreover, carbonyl cyanide m‐chlorophenyl hydrazone treatment recruited cytoplasmic ZtHex1‐eGFP to the lateral plasma membrane of the cells. Thus, keeping the WBs out of the septal pores, in Z. tritici, is an ATP‐dependent process.

The mechanism by which WBs plug the septal pore is not understood. The most widely accepted hypothesis is that wound-induced bulk flow of cytoplasm "flushes" septum-associated WBs from the unwounded cell into the septal pore (Jedd & Chua, 2000;Maruyama et al., 2005). However, quantitative electron microscopy studies revealed that a single WB closes the septum after wounding, whereas other WBs remain largely unaffected. This was taken as an argument against a pressure-driven mechanism of pore sealing by WBs . Alternatively, a contractile tether may pull WBs into the septal pore . This hypothesis is supported by optical laser trapping experiments in Nectria haematococca, which revealed "elastic" tethering of WBs to the septal pore (Berns, Aist, Wright, & Liang, 1992). Indeed, studies in N. crassa identified the protein LAH1 as being such tether (Ng, Liu, Lai, Low, & Jedd, 2009), and its homologue in A. fumigatus and Aspergillus oryzae was shown to anchor WBs at the septal pore Han et al., 2014;Leonhardt, Carina Kakoschke, Wagener, & Ebel, 2017). Lah-homologues share sequence similarity to motifs in the muscle protein titin (Ng et al., 2009), which confer calcium-dependent elasticity to titin (Labeit et al., 2003). With this finding, controlled contraction of Lah was suggested to mediate WB plugging (Han et al., 2014). However, no experimental evidence for such a mechanism exists. Interestingly, mutant studies in N. crassa strongly suggest a role of the septum-associated protein SPA9 in preventing Woronin-based septal pore plugging (Lai et al., 2012).
The molecular mechanism behind this is not known, but this finding adds strong support to the notion that WB-based pore plugging is an active process.
In this study, we use electron microscopy and live cell imaging to elucidate WB dynamics after laser-based hyphal wounding in Z. tritici. This fungus causes Septoria wheat blotch and poses a serious challenge to wheat producing agricultural industry (Fones & Gurr, 2015). However, despite its economic importance, its cell biology is poorly understood (Steinberg, 2015). We show that cell injury creates a pressure gradient, which is consistent with cytoplasmic flow-driven movement of WBs into the septal pore. However, a subpopulation of the WBs moves against the flow from the ruptured cell into the septal pore, suggesting an active mechanism of WB-based pore plugging. In agreement with this notion, we report that reduced cellular ATP levels trigger movement of WBs into the septal pore in intact hyphae.
2 | RESULTS 2.1 | A large number of Woronin bodies "guard" the septal pore As a first step in our study, we set out to analyse WB localisation, number, and dimension in the Z. tritici wild-type strain IPO323, using electron microscopy techniques in chemically fixed cells. Consistent with reports in other fungi, spherical WBs were closely associated with the septal pore (Figure 1a, 1b). These rounded organelles were surrounded by a single membrane and displayed a fine-granular homogeneous matrix. They had a diameter of~129 nm, whereas the septal pore opened only~41 nm and were located at average~300 nm away from the pore (Table 1). To determine the number of septum-associated WBs, we generated image stacks, derived from 24 to 26 serial sections per septum. Using this 3D information, we determined that three to four WBs "guard" each side of the septal pore in Z. tritici (Table 1; Figure 1c, Movie S1). Next, we treated cells of Z. tritici with quartz sand crystals and visualised septal pores in these wounded cells.
We found that septa were always plugged by a single WB (n = 20; Figure 1d). The remaining two to three WBs in the intact cell only slightly changed their position relative to the plugged septal pore (average distance to septal pore: 284.02 ± 21.22 nm, n = 45; mean ± standard error of the mean; not significantly different from control, p = 0.6978).
2.2 | Woronin bodies plug the pore against high pressure Cell injury induces cytoplasmic bleeding. This, in turn, may "flush" WBs from neighbouring intact cells into the septal pore (Jedd & Chua, 2000;Maruyama et al., 2005). We investigated cytoplasmic bleeding by performing controlled cell wounding experiments, using a 405-nm laser pulse to rupture cells. We did this in cells that express cytoplasmic green fluorescent protein (GFP), or the plasma membrane marker enhanced GFP (eGFP)-ZtSso1, and observed the effect of wounding on cytoplasmic bleeding and on the septa, using live cell imaging. We found that wounding induced bleeding of the GFP-containing cytoplasm from the wounded cell into the extracellular space (Movie S2). Rapid sealing by the pore This indicates that rapid plugging prevents further damage in adjacent cells.
Hex1 encodes the major protein in the hexagonal crystals in WBs in filamentous ascomycetes. It was used to visualise WBs in living fungal cells (overview in Steinberg et al., 2017). We used the predicted amino acid sequence of Hex1 from N. crassa and identified a putative homologue, ZtHex1, in the published genomic sequence of Z. tritici  Table 1). In the remaining 7.6%, a single ZtHex1-eGFP "dot" was located in the septal pore region, indicating that the septum was closed by WBs in a small number of cells. It is worth mentioning that the strong ZtHex1-eGFP signals most likely represent numerous septa-associated WBs, which cannot be separated spatially by light microscopy.
Moreover, the cytoplasm contained additional weaker ZtHex1-GFP signals, which may represent a population of cytoplasmic WBs (arrowheads Figure 2d, maximum projection of a z-axis image stack;  Note. All values given as mean ± standard error of the mean (sample size).  We considered it likely that WBs are responsible for the wound-induced rapid plugging of the septal pore. To test this, we observed ZtHex1-eGFP and mCherry-ZtSso1 in laser-induced Immediately after injury, the cellular pressure drops in the wounded cell (lower half of images, indicated by "Dead"). The WB of the intact cell has plugged the septal pore, whereas WBs in the wounded cell remain stationary or move slightly away from the septum, whereas the cytoplasm bleeds out. Time after wounding is given in seconds. Scale bar represents 1 μm. See also Movies S6 and S7. (g) Bar chart showing the behaviour of WBs in laser-wounded cells. In most cases, the WBs in the ruptured cell stay associated with the septum. Mean ± standard error of the mean is shown; sample size n is 3 data sets, 48 experiments. (h) WB "ballooning" in a laser-wounded cell of Z. tritici. After injury of the cell (left half of images, indicated by "Dead"), the WB of the intact cell plugs the pore. Within a few seconds, the WB balloons out, whereas the cytoplasm bleeds out of the ruptured cell. Time after wounding is given in seconds; scale bar represents 1 μm. See also Movie S8. (i) Image series showing "ballooning" of ZtHex1-eGFP and mCherry-ZtSso1 after injury of a cell. The "balloon" contains the integral syntaxin ZtSso1, suggesting that the plasma membrane in the unwounded cell (indicated by "intact") sealed after wounding and extends due to the pressure gradient into the wounded cell (indicated by "injured"). Time in seconds given in the upper right corner; scale bar represents 1 μm. See also Movie S9. (j) Electron micrograph of the septal pore of a wild-type cell after wounding with quartz sand. A WB has sealed the septal pore and formed a "balloon" into the injured cell (dead cell). Note that the membrane of the WB extends into the "bubble," suggesting that the pressure gradient between the injured cell (dead cell) and the living cell (live cell) causes shape change of the organelle. Electron microscopy confirmed that the "balloons" were surrounded by a double membrane and contain peripheral granular material ( Figure 2j). This may represent ZtHex1-eGFP, visible as a GFP lining of the expanding plasma membrane "balloon" (Movies S8 and S9).
The origin of this ZtHex1-eGFP lining is not known, but it appears likely that it is recruited from a cytoplasmic pool of this protein.
The physiological relevance of such cytoplasmic ZtHex1-eGFP, however, is not known, but "ballooning" of mCherry-ZtSso1 suggests rapid sealing of the plasma membrane after rupture of the cell, which is too weak to resist the pressure gradient between the intact and bleeding cells.

| An active mechanism participates in Woronin body closure of septal pores
Our ultrastructural analysis also revealed that some pores were sealed by WBs from the ruptured cell side (Figure 3a; Figure S1; seen in 3 out of 20 cases, equals~15%). Indeed, live cell imaging of ZtHex1-eGFP in cell wounding experiments confirmed that WBs are able to move from the ruptured cell into the septal pore (seen in 12.3% of all experiments, n = 3 data sets; Figure 3b, "Move to pore"; Movie S10). This movement is best seen in fluorescent intensity scans over the pair of ZtHex1-eGFP signals. Here, the two fluorescent intensity peaks, representing the WBs in the intact and ruptured cells, form one intensity peak within 1-2 s after wounding (Figure 3c). Considering the woundinduced drop in pressure, this WB motion occurs against the cytoplasmic bleeding. This suggests the existence of an active mechanism of WB movement into the septal pores.

| Woronin body plugging of pores requires cellular ATP
The cytoskeleton has been implied in septal pore plugging by WBs . To test for a role of the cytoskeleton in pore closure, we observed ZtHex1-eGFP in laser-injured cells that were treated with benomyl and latrunculin A. These inhibitors have been shown to disassemble microtubules and F-actin in Z. tritici . Upon cell wounding, WBs moved into the septal pore in control cells, as well as in cells treated with benomyl and latrunculin A (Figure 3d). This result suggests that the cytoskeleton is not involved in WB-mediated pore plugging. To test if ATP-dependent enzymatic activity is required to close the septal pore, we treated the cells with the carbonyl cyanide m-chlorophenyl hydrazone (CCCP). This chemical inhibitor of oxidative phosphorylation reversibly depletes cellular ATP in fungal cells (Lin et al., 2016). We found that CCCP-treated cells were not impaired in that drags WBs towards the septal pore .
We discuss our results in the light of these proposed mechanisms.
Hyphal cells build up internal turgor pressure (Lew, 2011), which, upon wounding, causes bleeding of the cytoplasm into the extracellular space. It is widely assumed that such bulk flow of bleeding cytoplasm takes WBs from the intact cell into the septal pore (Jedd & Chua, 2000;Maruyama et al., 2005), thereby restricting the loss of cytoplasm largely to the injured cell . Here, we show that wound-induced cytoplasmic bleeding in Z. tritici results in a drastic drop of pressure, indicated by bending of the adjacent septa towards the collapsed wounded cell and the occasional "ballooning" of the plasma membrane, which rapidly resealed over the wounded septal pore. Thus, we consider it likely that cytoplasmic bleeding from the intact cell could sweep WBs into the septal pore. It was reported that septa are sealed by a single WB , this study) while the others are barely changing their position relative to the pore. Indeed, our 3D reconstruction of serial sections reveals only a small and insignificant shift of nonplugging WBs towards the septal pore (Student's t test; p = 0.6978). One may argue that a cytoplasmic bulk flow mechanism should reposition all WBs. The fact that this is not found was taken as an argument against a passive, bulk flow-driven movement of WBs . Unless a passive WB sealing mechanism is highly efficient, these results argue for a more active mechanism of WB translocation.
Although the majority of septal pores is sealed off from intact cells, we also find that WBs move from the ruptured cell into the septum. This raises more doubt about a passive cytoplasmic bulk flow-driven mechanism, as WBs move against the cytoplasmic bleeding. Live cell The two septum-associated WB signals fuse into one signal, which is located in the septal pore. In addition, small amounts of ZtHex1-eGFP were found at the plasma membrane (arrowhead). Scale bar represents 3 μm. (f) Examples showing septum-associated WBs in untreated control cells (upper gallery) and in CCCP-treated cells (lower gallery). Depletion of cellular ATP concentrates WBs in the septal pore. Scale bar represents 1 μm.
(g) Graph showing intensity curves over WBs at septa of control cells and after treatment with 100 μM CCCP for 15 min. In the presence of the respiration inhibitor, the bimodal distribution of ZtHex1-GFP turns into a unimodal distribution, indicating that the WBs have moved into the septal pore. Note that the cells were not injured. Also note that CCCP treatment reduces the overall ZtHex1-GFP fluorescent intensity in the cell. (h) Bar chart showing the number of septa that are sealed after treatment of 10-15 min with 100 μM CCCP or 0.1% sodium azide in unwounded multicellular structures. Mean ± standard error of the mean is shown; sample size n is 3-6 data sets, with 156-423 septa analysed per bar. (i) Electron micrograph showing WBs at septal pores in CCCP-treated cells (+CCCP). Reducing the ATP levels results in WBs plugging on both sides of the septum. Scale bar represents 0.2 μm observation of WBs shows that this movement occurs within~1 s after cell injury, suggesting that it is mediated by force-generating mechanisms. Active transport processes along the cytoskeleton spatially to organise the fungal cell (Lin et al., 2016), and it was suggested that microtubules are involved in WB motility into the septal pore . To test this possibility, we performed laser-rupture experiments in the presence of inhibitors that prevent formation of F-actin or microtubules. However, we found no evidence for an involvement of the cytoskeleton in WB-based closure of septal pores.
We found that WB movement into the septal pore of unwounded cells is induced when cellular ATP levels are depleted. This raises the possibility that chemical energy, or at least the presence of ATP, is required to keep the septal pore open. At present, the exact way by which ATP prevents WB activation is unclear. WBs are tethered to septa via Lah proteins Han et al., 2014;Leonhardt et al., 2017), which show sequence similarity with the contractile muscle protein titin (Ng et al., 2009). Upon conformational change, titin can generate force (Martonfalvi, Bianco, Naftz, Ferenczy, & Kellermayer, 2017); this activity involves ATP binding to a kinase domain within titin (Puchner et al., 2008). Lah proteins lack such kinase domain, but our finding that cellular ATP is required to keep septal pores open suggests that as yet unknown kinases control WB movement into the pore.
Indeed, studies in A. nidulans have shown that the NIMA kinase is involved in selective closure of the septal pore, yet this level of control occurs independently of WBs (Shen, Osmani, Govindaraghavan, & Osmani, 2014). Alternatively, ATP may bind directly to Lah or to interacting proteins, thereby affecting their activity. Such mechanisms have been well documented for a wide variety of membrane proteins and molecular chaperones (Suzuki & Yura, 2016;Wellhauser, Luna-Chavez, D'Antonio, Tainer, & Bear, 2011). Although the detailed mechanism of WB translocation at low ATP levels is not known, it appears to be an efficient mechanism to ensure that cell rupture, and associated ATP depletion, results in rapid septal pore closure by WBs.
In conclusion, our results support a combinatorial mechanism for pore sealing by WBs. Although cytoplasmic bulk flow may be the primary way to close a pore, active recruitment of WBs for the ruptured cell may provide an alternative mechanism. The latter process appears to be ATP sensitive, and we speculate that it involves conformational changes in the Lah protein. Such an ATP-dependent mechanism may also facilitate WB-based closure of septal pores in intact cells (Bleichrodt et al., 2012;Markham, Collinge, Head, & Poole, 1987).
Reversible closing of pores by WBs was suggested to underpin hyphal heterogeneity and cell specialisation (Bleichrodt et al., 2012).

| Fungal strains and growth conditions
The Z. tritici wild-type isolate IPO323 (Goodwin et al., 2011) was used to generate the strains IPO323_CHex1eGFP and IPO323_ CHex1eGFP_HmCherrySso1, as well as strains IPO323_CeGFP  and IPO323_ GFPSso1 .
All strains were grown in 20 ml YG media (yeast extract, 10 g/L; glucose, 30 g/L) at 18°C with 200 rpm for 48 hr.

| Molecular cloning
Vector pCHex1eGFP was designed for integration into the succinate dehydrogenase locus . It contains the gene for GFP, egfp, fused to geneZthex1, placed under the control of constitutive Zttub2 promoter and limited by the Zttub2 terminator . In detail, plasmid pCHex1eGFP carries a 12,530-bp fragment of pCeGFPTub2  digested with BsrGI), a 1149-bp Z. tritici α-tubulin promoter (amplified with SK-Sep-14 and SK-Sep-47;   Table 2). The vector was generated by in vivo ligation of these DNA fragments in the yeast Saccharomyces cerevisiae . Subsequent transformation into IPO323 was done as previously described , resulting in strain IPO323_CHex1eGFP. To covisualise WBs and the plasma membrane, vector pHmCherrySso1  was ectopically integrated into IPO323_CHex1eGFP, resulting in strain IPO323_CHex1eGFP_HmCherrySso1.

| Laser-based epifluorescence microscopy
Fluorescence microscopy was performed as previously described (Kilaru, Schuster, Studholme, et al., 2015). In brief, the cells were inoculated in YG media (yeast extract, 10 g/L; glucose, 30 g/L) and grown Sigma-Aldrich, Gillingham, UK) and mixed for 15 min in a 2-ml reaction tube, using IKA Vibrax shaker (IKA, Staufen, Germany). Glass beads were removed by centrifugation at 1000 rpm for 30 s. The cell-containing supernatant was prepared for electron microscopy as described below.

| Ultrastructural studies
For ultrastructural studies, liquid cultures were fixed and embedded as previously described . For serial section analysis, sections of 70 nm were placed on pioloform-coated copper slot grids (Agar Scientific, Stansted, UK) and contrasted with lead citrate. Sections were examined using a JEOL JEM 1400 transmission electron microscope operated at 120 kV, and images taken with a digital camera (ES 100W CCD, Gatan, Abingdon, UK). Estimation of WB numbers at septal pores was done by the physical disector method (Sterio, 1984). To this end, 25 ± 1 sections were acquired at a magnification of 60,000 × g. WBs were identified by their characteristic appearance, excluding those that were still in contact with peroxisomes. The average distance of WBs to the septal pore was measured by defining the centre of the pore in Photoshop CS6 and the centre of WBs, taking z-axis information into account. Three-dimensional models of septum-associated WBs were reconstructed from the serial micrographs using IMOD software and ETomo (http://bio3d.colorado.edu/imod/), and video files created with ImageJ (https://imagej.nih.gov/ij/).