Endosidin 5 disruption of the Golgi apparatus and extracellular matrix secretion in the unicellular charophyte Penium margaritaceum

Abstract Background and Aims Endosidins are a group of low-molecular-weight compounds, first identified by ‘chemical biology’ screening assays, that have been used to target specific components of the endomembrane system. In this study, we employed multiple microscopy-based screening techniques to elucidate the effects of endosidin 5 (ES5) on the Golgi apparatus and the secretion of extracellular matrix (ECM) components in Penium margaritaceum. These effects were compared with those caused by treatments with brefeldin A and concanamycin A. Penium margaritaceum’s extensive Golgi apparatus and endomembrane system make it an outstanding model organism for screening changes to the endomembrane system. Here we detail changes to the Golgi apparatus and secretion of ECM material caused by ES5. Methods Changes to extracellular polymeric substance (EPS) secretion and cell wall expansion were screened using fluorescence microscopy. Confocal laser scanning microscopy and transmission electron microscopy were used to assess changes to the Golgi apparatus, the cell wall and the vesicular network. Electron tomography was also performed to detail the changes to the Golgi apparatus. Key Results While other endosidins were able to impact EPS secretion and cell wall expansion, only ES5 completely inhibited EPS secretion and cell wall expansion over 24 h. Short treatments of ES5 resulted in displacement of the Golgi bodies from their typical linear alignment. The number of cisternae decreased per Golgi stack and trans face cisternae in-curled to form distinct elongate circular profiles. Longer treatment resulted in a transformation of the Golgi body to an irregular aggregate of cisternae. These alterations could be reversed by removal of ES5 and returning cells to culture. Conclusions ES5 alters secretion of ECM material in Penium by affecting the Golgi apparatus and does so in a markedly different way from other endomembrane inhibitors such as brefeldin A and concanamycin A.


INTRODUCTION
The extracellular matrix (ECM) of plants includes the cell wall, an assortment of gel-like polysaccharides/proteoglycans and hydrophobic materials that are deposited either onto or in the cell wall. The structural and functional dynamics of the ECM are profoundly important for the life of the plant cell. The cell wall plays a central role in the control of cell expansion and morphogenesis, provides a rigid edifice for physical support, comprises a signalling network as well as a structural and biochemical barrier for resistance to abiotic and biotic stressors, and contributes to absorption and transport of water and solutes (Tenhaken, 2015;Chebli and Geitmann, 2017;Zhang et al., 2021;Cosgrove, 2022). Other ECM components function by providing a hydrated matrix for embryo development during seed germination (Tsai et al., 2021), serving as lubricants for root growth in soils (Driouich et al., 2021) or providing waterproof coatings on various tissues and organs (e.g. cuticles; Domínguez et al., 2011;Niklas et al., 2017;Arya et al., 2021). Production of the ECM requires the coordinated actions of multiple subcellular systems that are synchronized during cell expansion and development and can rapidly modulate in response to environmental stress. The primary machinery for ECM processing (i.e. biosynthesis, packaging and transport to the cell surface, secretion and incorporation into the ECM architecture) consists of the endomembrane system (Kim and Brandizzi, 2014), cytoskeletal network (Gu and Rasmussen, 2022), specific loci of the plasma membrane (vesicle tethering complex or exocyst; Huang et al., 2019a,b;Saeed et al., 2019) and even the ECM itself (Bacete et al., 2018). The endomembrane system is the central subcellular unit in this processing and its main components, the endoplasmic reticulum (ER), the Golgi apparatus, the trans Golgi network (TGN) and various types of vesicles, have been extensively studied in plants (Worden et al., 2012;Drakakaki, 2015;van de Meene et al., 2017;Sinclair et al., 2018;Robinson, 2020;Aniento et al., 2022;Nakano, 2022;Shimizu and Uemura, 2022). The Golgi apparatus is a major 'cog' in the ECM-processing machine. Its distinct architectural design and highly complex biosynthesis machinery are directly involved in the synthesis, packaging and transport of ECM components to the cell exterior.
Analyses of the Golgi apparatus and endomembrane system in plants have greatly benefitted from the application of diverse and novel technologies. These include glycomic analyses of isolated endomembrane components (Okekeogbu et al., 2019;Wilkop et al., 2019), mutation analysis (Kohorn et al, 2021), electron microscopy and tomography (Koga et al., 2016;Otegui and Pennington, 2019;Weiner et al., 2022), 3-D printing (Mai et al., 2017) and the application of a wide array of fluorescent markers (DeVree et al., 2021) including fluorescence-tagged marker proteins in transformed cell lines. Another important tool used in endomembrane system research is chemical genetics/chemical biology (Hicks and Raikhel, 2010;Norambuena and Tejos, 2017;Rodriguez-Furlan et al., 2018;Ma et al., 2022). Here, application of small-molecule inhibitors targeting specific endomembrane components allows for transient manipulation of the membrane trafficking processes in both a notable and reversible manner. Alterations to the endomembrane system and secretory cargoes may then be monitored and even quantified using high-resolution live-cell fluorescence light microscopy (FLM)-based imaging, and detailed 'snapshots' of specific stages of alteration during the treatment may be acquired using various transmission electron microscopy (TEM)based techniques. Inhibition effects may ultimately be checked via recovery experiments and their severity can be controlled by manipulating dosage. Several chemical agents have become powerful tools for dissection of the Golgi apparatus and other endomembrane components (e.g. brefeldin A; Nebenführ et al., 2002;Foissner et al., 2020)).
Plants evolved from a freshwater charophycean green alga (i.e. basal streptophyte) ~500-600 million years ago (Mattox and Stewart, 1984;de Vries and Archibald, 2018;Rensing, 2018). The ECM of these algae was critical to the successful invasion of land and subsequent evolution into the great diversity of modern land plants (Sørensen et al., 2011;Harholt et al., 2016). Phylogenomic analyses have now clearly confirmed that the Zygnematophyceae are the sister group to embryophytes (Delwiche and Cooper, 2015;de Vries and Archibald, 2018;Puttick et al., 2018;Zhou and von Schwartzenberg, 2020;Hess et al., 2022). Like most land plants, most modern-day zygnematophyte taxa possess a complex ECM. This includes a cell wall with constituents that are remarkably similar to those found in many land plants. Many zygnematophytes also secrete large amounts of polysaccharide-based gels called the EPS (extracellular polymeric substance) that function in flotation and gliding, serving as platforms for establishing communication conduits of biofilms and as protection against water loss (Boney, 1981;Domozych and Domozych, 2008;Domozych et al., 2009). The secretory apparatus that is required for the production of this ECM is often very extensive. In some unicellular desmids such as Penium margaritaceum, Closterium acerosum and Micrasterias denticulate (Lütz-Meindl, 2016;Domozych et al., 2020), each cell often contains well over 100 Golgi bodies, and each Golgi body has the capability of processing multiple ECM components and packaging them in various types of vesicles for subsequent transportation to the peripheral cytoplasm. Here, the vesicles enter actin-mediated cytoplasmic streaming channels and stream around the cell periphery. Different vesicles are then selectively removed from the streaming channels and fuse with specific plasma membrane sites where they secrete their constituents to the outside of the cell. Their fast growth rates and ease in experimental manipulation and multiple microscopy-based imaging make these unicellular zygnematophytes valuable for 'dissecting' ECM processing pathways in late divergent charophytes (Lütz-Meindl, 2016;Domozych and Bagdan, 2022) and elucidating subcellular systems that may have been key in successful colonization of land by ancient charophyte algae. Their unicellular phenotype and ease in experimental manipulation and screening (Rydahl et al., 2015) also makes these algae ideal specimens for screening potential chemical agents that target the Golgi apparatus and secretory machinery. In a screening of multiple chemical agents, we noted that endosidin 5 (ES5) results in significant alterations to Golgi apparatus structure and secretion dynamics in Penium margaritaceum. Multiple light-and electron microscopy-based protocols were employed and the acquired data synthesized in order to identify the subcellular targets of ES5 activity. We also examined the effects of two other endomembrane-targeting agents, brefeldin A (BFA) and concanamycin A (ConcA), and compared their effects to that of ES5.

General
Penium margaritaceum was cultured in Woods Hole Medium supplemented with 5 % soil extract (Carolina Biological Supply, USA) using previously described methods (Rydahl et al., 2015). Live cells from 7-10-d-old subcultures were harvested by centrifugation at 700 g for 1 min. The cell pellets were washed three times with fresh Woods Hole Medium + soil extract (WHS) and collected by centrifugation.

Screening protocol: EPS production
For experimental cultures, 12-well uncoated tissue culture Petri dishes were used. One litre of WHS was added to each well of a 12-well uncoated tissue culture Petri dish. Various concentrations of each endosidin, BFA and ConcA were added to each well along with 0.75 µm FITC-fluorescent microspheres (Polysciences, USA). Endosidins 3, 5, 7 and 9 were purchased from ChemBridge Chemicals (USA). Endosidin 2, ConcA and BFA were purchased from Sigma Aldrich (USA). To each well 20 µL of dense cell suspension obtained from the pellet of washed cells (see above) was added. The wells were gently mixed and the dish sealed with Parafilm. Plates were cultured for 24 h under 74 µmol photons m −2 s −1 of cool white fluorescent light with a 16 : 8-h light-dark cycle. The wells were then observed with an IX-83 inverted light microscope (Olympus, USA) with FITC optics. The presence or absence of EPS trails or ensheathments were noted. For those cultures where gliding trails were not observed, the cells were harvested, centrifuged as above, washed three times with fresh WHS, resuspended in WHS containing fluorescent beads and cultured as above. Wells were observed after 24-72 h to determine if the ability to form gliding trails was recovered.

Screening protocol: cell wall expansion
To examine cell wall expansion following treatment with the perturbation agents, cell walls were labelled with an anti-pectin antibody, JIM5, followed by a TRITC-labelled secondary antibody as previously described (Domozych et al., 2014). Twenty microlitres of cell suspension of JIM5-TRITC-labelled cells was added to the wells. Cells were cultured as above for 24 h and observed with an Olympus IX-83 inverted light microscope equipped with TRITC optics. Cell wall expansion was determined by the presence of unlabelled zones at the cell isthmus, surrounded by JIM5-labelled zones (Domozych et al., 2014). If no wall expansion was noted, cells were harvested from the well, centrifuged as above, washed three times with WHS and then resuspended in 1 mL of fresh WHS. The cells were cultured as above and monitored for cell expansion within 24-48 h.
An estimate of the percentage of new cell wall material produced was done by assuming the cell to be rectangular in shape. The percentage of new cell wall material was calculated by: where L t is the total length of the cell and L o is the zone(s) composed of the old cell wall material that was labelled with JIM5-TRITC. The average percentage was calculated from all of the cells found in three images taken on an Olympus IX83 using the 10× objective (typically ~10 cells per image). No difference in the average/s.d. of new cell wall production was seen in control cells measured from using three images or when the number of images was increased to 10. For time-lapse imaging of cell wall expansion, Penium cells walls were labelled as described above. Labelled cells were kept in the dark until use. Cells were immobilized in 2 % agarose solution within the Petri dish, and covered with fresh WHS. Timelapse images were taken every hour with an Olympus IX83 inverted microscope using the 10× objective. Between images the cells were left on the microscope under white light with the power of the light set to 5 V.

Live cell labelling
Live cells from 7-10-d-old subcultures or from wells of treated cultures were harvested by centrifugation at 800 g for 1 min were collected in 1.5-mL microcentrifuge tubes and cells were concentrated by centrifuging at 2500 g for 1 min. The cell pellet was washed three times with distilled water and resuspended in a final of volume of 1 mL. The cells were then incubated with either 1.25 μm Yeast Vacuole Membrane marker MDY-64 (Invitrogen, USA) or 2 μm LysoTracker™ Red (Invitrogen, USA), for 30 min in the dark with constant rotation, and washed three times with distilled water. Cells were then imaged on a Fluoview 1200 confocal laser scanning microscope (Olympus, USA) using either the FITC filter to image the MDY-64 label or TRITC filter to image the LysoTracker TM Red label. The TOTO filter was used to view chloroplast autofluorescence.

TEM preparation
Cells were harvested from 7-14-d-old cultures and washed three times with WHS as described above. The cell pellet was resuspended in 0.5 mL of WHS. Cells were then spray frozen using a commercial artist's airbrush into 10 mL of liquid propane cooled to −185° C in a dewar of liquid nitrogen. The frozen cells were poured into precooled (−80° C) glass scintillation vials containing 0.5 % glutaraldehyde/0.2 % tannic acid in acetone (for osmicated ultrastructural analysis) The vials were placed in a −80° C freezer for 24 h. For cells prepared for osmicated ultrastructural analysis, 0.1 g of osmium tetroxide was added to the scintillation vial and the vial was placed back in the −80° C freezer for 24 h. After this time, the vial was slowly warmed to room temperature over 16 h. The cells were then collected into a pellet by centrifugation at 700 g for 1 min. The supernatant was discarded and the pellet was washed with acetone and recentrifuged. This was repeated twice more. The cells/pellet was then infiltrated for 3 h each in cocktails of 25 % Spurrs Low Viscosity Plastic (SLVP; EMS, USA)/75 % acetone, 50 % SLVP/50 % acetone and 75 % SLVP/25 % acetone at room temperature (RT). The cells were then placed in 100 % SLVP for 2 h at RT. The cells were then pelleted into Beem capsules (EMS, USA) and polymerized at 55° C for 8 h.
For ultrastructural analysis, 70-100-nm sections were cut using a Diatome diamond knife with a LeicaUltracut microtome, collected on Formvar-coated copper grids (EMS, USA) and stained with conventional uranyless/lead citrate (EMS, USA). Cells were viewed with an Hitachi 5800 TEM or a Zeiss Libra TEM at 120 kV.

Tomography
For tomography, serial sections of osmicated cells embedded in SLVP were collected on copper slot grids coated with Formvar. Serial sections (150 nm) were viewed on a Libra 120 plus TEM (Zeiss, Germany) or HT7800 TEM (Hitatchi, Japan) at 120 kV from 8000 to 12 000× magnification. A tilt series of images was collected for each of the -three to six serial sections from −55° to 55° or −40° to +40° in 1° increments. Images were captured on a Cantega G2 digital camera (Olympus, USA) for images taken on the Zeiss Libra 120 plus TEM or BioSprint16 High-Definition CCD Camera (AMT, USA) for images taken on the Hitachi 7800 TEM. To obtain data for a dual axis tomogram, an orthogonal tilt series was taken after manually rotating the sample grid 90° in the sample rod. Image focus, tilt, alignment and capture were controlled by the WinTEM software (Zeiss, Germany) on the Zeiss Libra 120 plus TEM and Hitachi TEM system software (Hitachi, Japan) on the Hitachi HT7800 TEM. This process was repeated for all serial sections for each sample.
Dual axis tomograms were reconstructed using the Etomo software interface in the IMOD software package (https:// bio3d.colorado.edu/imod/). The two orthogonal single-axis tomograms were merged into one with a warping procedure (Mastronarde, 1997). The above process was repeated for each section. The sections were then joined by manual alignment using Midas in the IMOD software. The alignment was then refined using fiducial markers through the adjacent tomograms (Mastronarde et al., 1992).
After completing the tomographic reconstruction, the 3D model was built using the 3dmod software interface. Each cisterna was modelled as a separate object, by tracing the structures as contours in consecutive tomogram slices. After the contours were drawn on each of the tomogram slices, the IMOD mesh tool was used to create a 3D surface for each object. Vesicles around the Golgi stack were modelled by a single sphere at that position. Due to the electron beam causing some degree of section collapse, the z-dimension of the models was stretched by a scaling factor to accurately display the model, based on the know thickness of the sections (150 nm).

RESULTS
Penium margaritaceum is a unicellular, freshwater zygnematophyte with a cylindrical phenotype measuring ~17 µm in width and 125-225 µm in length . Each cell is composed of two semi-cells surrounding a central, isthmus zone, which is the focal zone for cell wall expansion and cell division. The nucleus resides within the isthmus zone, flanked by one or two chloroplasts in each semicell (Fig. 1A). The ECM is composed of a cell wall and a gellike component known as EPS that is secreted beyond the cell wall.
The cell wall of Penium consists of an inner layer composed of cellulose microfibrils which act as a scaffold for the outer lattice composed of pectic polymers -primarily homogalacturonan (HG), cross-linked by calcium (Ca 2+ ) ions (Domozych et al., 2014). The monoclonal antibody, JIM5, labels the HG lattice that is located on the outer layer of the cell wall ( Fig. 1B) and has been used as a live cell label for monitoring expansion over time (Rydahl et al., 2015). During cell expansion of JIM5labelled cells, new HG that is unlabelled (i.e. a non-fluorescent zone) is secreted at the isthmus and displaces the older labelled cell wall to the polar zones (Fig. 1C).
Secretion of the gel-like EPS can be monitored with 0.75µm fluorescent microspheres (Polysciences, USA) which bind to these polysaccharide-based gels (Jiao et al., 2020). Regions where these beads are concentrated can be detected by FLM. EPS secretion was monitored by placing washed cells (i.e. washed free of existing EPS) into a well of a 12-well uncoated culture plate containing 1 mL of growth medium with a particular agent and the fluorescent microspheres. Large numbers of treatments and concentrations of the chemical agent were then monitored using FLM. In control cells after 1 h, the fluorescent beads adhered to a thin layer of EPS secreted around the cell periphery (Fig. 1D). After 2 h, the cells secrete EPS from localized points on the cell surface ( Fig. 1E) to form trails for gliding. Penium employs localized EPS secretion and subsequent hygroscopic swelling of the EPS as the motive force to move cells for the gliding mechanism (Domozych et al., 2005). After 24 h, the trails of these gliding cells elongate substantially (Fig. 1F).

EPS screen
The first level of screening using the fluorescent beads assessed changes to EPS secretion in treated cells and in cells recovering from chemical treatment. In the latter, treated cells that exhibited no or notably altered EPS secretion were extensively washed, placed into fresh growth medium with the fluorescent beads, cultured and observed after a 24 h period using FLM.
An agent and specific treatment application were considered reversible if washed cells subsequently produced EPS trails. We selected those treatments that (1) altered EPS production, (2) did not kill cells and (3) whose effects were readily reversed by removal and recovery of cells in growth medium.
ES5 clearly inhibited the secretion of EPS. EPS is seen to encapsulate the cells, but no EPS trails are formed ( Fig. 2A). EPS production could be restored by washing the cells with fresh media and allowing the cells to recover for 1-2 d (Fig. 2D). Similar suppression of EPS secretion (Fig. 2B, C) and recovery (Fig. 2E, F) could be seen for BFA and ConcA respectively. The final concentrations of each perturbation agent used were those which produced an effect on EPS production without killing the cells, and allowed for complete recovery of EPS production upon removal of the perturbation agent, indicating reversibility of the treatment. Concentrations were 10 µm for ES5, 1.5 µm for BFA and 1 µm for ConcA.
Endosidin 2 (ES2, 20 µm) and endosidin 3 (ES3, 15 µm) also inhibited the secretion of EPS, whereas endosidin 7 (ES7) and endosidin 9 (ES9) did not have any effect on EPS production. The concentration threshold was analysed based on the solubility of each agent and the solvent used for stock cultures (i.e. DMSO). Analysis of cells treated with DMSO up to 30 µL mL -1 caused no changes to EPS production (Supplementary Data Fig. S1).

Cell wall expansion screen
The second level of screening entailed labelling live cells using JIM5 and a secondary antibody conjugated with TRITC and then placing the cells back into growth medium containing the various test chemical agents (Domozych et al., 2014). Timelapse imaging can be used to follow the bi-directional growth of the cell wall at the isthmus of each cell. Initially, a single dark zone forms at the centre of the cell which expands over time as new cell wall material is secreted at the isthmus. If ES5-, BFA-and ConcA-treated cells were washed with media to remove the perturbation agent and allowed to grow for an additional 1-2 d, recovery of cell wall expansion can be seen by the appearance of new unlabelled zones in the cell wall. This indicates that the action of these chemicals was reversible.
Treatment of the cells with ES2, ES3, ES7 or ES9 did not completely inhibit cell wall expansion over 24 h. When the percentage of new cell wall material was compared between control and treated cells, no difference was seen for ES2, ES7 and ES9. Treatment with ES3 (15 µm) led to a small but significant reduction in the percentage of new wall material produced in 24 h from 46 ± 6 to 34 ± 6 % (Supplementary Data Fig. S2K-P).

Alterations in endomembrane components
The endomembrane system and structural alterations resulting from treatment with chemical agents were assessed by both LM and TEM methodologies. The lobed chloroplasts are the largest organelles in the cell (Fig. 3A). The lobes surround valleys of cytoplasm that extend into the interior of the cell (Fig.  3B). The cytoplasmic valleys hold mitochondria, ER, linear arrays of Golgi bodies and numerous vacuoles ( Fig. 3C; see also Domozych et al., 2020). Our main focus of chemical screening was the Golgi apparatus. The fluorescent label, MDY-64, was used for identification of the Golgi bodies and their orientation in the cytoplasm (Fig. 3D). In control cells, Golgi bodies are arranged in a linear orientation in the cytoplasmic valleys, appear as curved bodies and typically number 125-150 per cell. Unlike plant cells whose secretory networks have been studied in detail (e.g. Arabidopsis; Robinson, 2020), Penium's Golgi bodies are non-mobile. Each Golgi body consists of 12-15 cisternae that are tightly stacked (Fig. 3E). Numerous vesicles are produced by the Golgi body. During EPS production, large EPS-containing vesicles form at the swollen peripheries of the medial-trans cisternae (Fig. 3F)   are also apparent (Fig. 3G). Smaller vesicles carry cell wall cargo from the Golgi body to the cell surface (see . The EPS vesicles are carried to the peripheral cytoplasm ( Fig. 3H-J) where they are transported around the cell via cytoplasmic streaming.

Alterations in the Golgi apparatus
ES5 produced the most striking alterations to the endomembrane system. After overnight treatment, the positioning and structure of the Golgi bodies in the cell are altered. MDY-64 labelling shows that the Golgi bodies take on a circular profile and are unevenly distributed in the cytoplasm (Fig. 4A). After thorough washing to remove ES5 and allowing the cells to recover for 24 h, the shape and distribution of the Golgi bodies return to normal (Fig. 4B). TEM imaging further highlights the structural changes to the Golgi body and its distribution in the cytoplasm (Fig. 4C). Four hours of treatment with ES5 reveal that the cisternae at the medial-to-trans loci of the Golgi curve inward to form circular profiles. The number of cisternae is also reduced (Fig. 4D). Aggregates of altered Golgi bodies appear in the cytoplasm and are interspersed with cytoplasm zones that are Golgi-free (Fig. 4C, E). After 24 h of treatment the Golgi body is reduced to a few scattered cisternae dispersed in the cytoplasm. After 24 h of recovery, Golgi bodies return to their original state (Fig. 5G).
Treatment with BFA produced significant changes to the endomembrane system. MDY-64 labelling of cells after overnight treatment with BFA shows that the positioning and architecture of the Golgi bodies are altered. The Golgi adopt a curved or disc-like shape and are no longer arranged in linear arrays along the cytoplasmic valleys (Fig. 5A). After washing to remove BFA and allowing the cells to recover for 24 h, the shape and distribution of Golgi bodies can be seen to return to normal (Fig. 5B). TEM imaging shows the progression of Golgi disruption over time. After 2 h of treatment, there is a reduction in the number of cisternae and an elongation of the remaining cisternae (Fig. 5). Secretory vesicles can still be seen at the trans face. After 4 h of treatment, production of vesicles can still be seen, and the elongate cisternae begin to curl inward at the trans face (Fig. 5D). After 24 h treatment, the Golgi bodies have curled inward even further, surrounding the secretory vesicles (Fig. 5E).
ConcA treatment also resulted in substantial changes to the endomembrane system. After overnight treatment with concanamycin, MDY-64 labelling shows a large number of small, round Golgi bodies that are often clustered together (Fig. 6A). After recovery for 24 h, the shape, number and distribution of Golgi bodies returns to normal (Fig. 6B). TEM imaging of cells treated for 4 h with ConcA shows a reduction in the number of cisternae in Golgi bodies. The Golgi bodies form aggregates in the cell. In some Golgi bodies, the trans face cisternae curve inward. Swelling can often be seen at the ends of the cisternae, as well as an accumulation of vesicles near these Golgi bodies (Fig. 6C). After 24 h of treatment, the Golgi bodies further transform into irregular cisternal aggregates, including the formation of multivesicular bodies (MVBs) (Fig. 6D). Golgi bodies return to their normal state after 24 h of recovery (Fig. 6E).
ES7 and ES9 did not alter EPS production or cell wall expansion and so were not analysed for structural changes to the Golgi apparatus. ES2 and ES3 were assessed for changes to the Golgi apparatus. No structural changes were seen for ES2treated cells, where ES3 treatment led to a mixture of normal and altered Golgi (Supplementary Data Fig. S3).

Alterations in the vesicle network
In ES5-treated cells, labelling with lysotracker shows vesicles in the peripheral cytoplasm in similar numbers to that seen in untreated cells (Fig. 4H), in agreement with the number of vesicles seen in the TEM micrographs of ES5-treated cells (Fig. 4C). Cells treated with ConcA also show a similar number of vesicles in the peripheral cytoplasm to that observed in untreated cells using lysotracker fluorescence (Fig. 6F). BFAtreated cells, on the other hand, display a large increase in the number of vesicles seen in the cytoplasm using lysotracker (Fig. 5H) and in TEM images (Fig. 5C).

Alterations in the pectin lattice
We used JIM5 labelling and freeze shattering/SEM protocols to monitor changes to the outer wall, pectin-based lattice in treated cells. ES5 and ConcA treatments resulted in a decrease or absence of the HG lattice while no change was observed under BFA treatment. Based on the cell expansion experiment, we posit that BFA stopped the secretion and organization of the cell wall. ES5 and ConcA allowed some cell expansion but the HG lattice was incompletely formed. In all cases, the production of the lattice returned to normal upon recovery.

Tomography models of Golgi bodies
To acquire additional structural information about the Golgi bodies, dual-axis tomography of serial EM sections was performed. The serial tomograms were then joined to achieve a total sample thickness of 450-900 nm. Three-dimensional profiles could be reconstructed from the tomography data using the 3dmod software (Kremer et al., 1996). The differences between the Golgi of treated and untreated cells becomes even more ob- secretory vesicles are present near the trans face but are reduced in number compared to the control. The Golgi of BFA-treated bodies also show interesting features in the 3D reconstruction (Fig. 7E, and Supplementary Data Video S3). The elongated cisternae are curved at the centre but the edges of the cisternae remain flat, creating a 'U'-like shape. While substantial curvature is seen, no cup-like structures or fusion between cisternae was seen. Lastly, tomographic reconstruction was performed on Golgi of ConcA-treated cells (Fig. 7F, and Video S4). Several Golgi bodies can be seen in close proximity. The cisternae can be seen to have a variety of different curved shapes including hollow tubes and cup-like structures. Some of the internal-most membranes have fused completely at both the top and bottom to form spheres surrounding smaller vesicles, forming MVBs. Secretory vesicles are still seen surrounding the Golgi bodies. A tomographic reconstruction of a curved Golgi body found in an ES3-treated cell was also ( Fig. S4 and Video S5). Tomography data confirm structural information about the Golgi bodies seen in the micrographs obtained from thin TEM sections and also provide unique insights into the Golgi structure, particularly for treated samples. The 3D curling of the cisternae in ES5-, BFA-and ConcA-treated cells can only be elucidated from tomographic data (Fig. 8).

DISCUSSION
A multifunctional ECM that dynamically modulates in response to environmental pressures was critical to the evolution of plants and serves as a foundation for a plant cell's life functions. To process the large numbers of diverse ECM components, plant cells employ a complex endomembrane system that synthesizes, packages and transports ECM cargoes to the cell surface. Central to this secretory processing is the Golgi apparatus. While recent research has greatly improved our understanding of Golgi-based membrane trafficking in plants, many basic questions remain to be answered. For example, how do Golgi dynamics change in response to particular abiotic and biotic stress agents in order to produce necessary modulations in the ECM? Likewise, how has the endomembrane system evolved in ancestral taxa of plants and how might information derived from studies of charophytes help provide insight into plant evolution and ecophysiology? Penium is an outstanding alga for elucidating endomembrane structure/function and ECM secretion in zygnematophytes. It can also provide insights into those subcellular systems that are critical for life at the aquatic/terrestrial interface which may have been key in the successful colonization of land by an ancient alga. Some of the efficiacious attributes of Penium include: (1) it has a wellcharacterized ECM that includes a pectin-rich cell wall and extensive EPS, (2) it is both easy to maintain in the laboratory and screen using large arrays of chemical agents, (3) it lends itself well for live cell fluorescence microscopy with a variety of probes (e.g. antibodies, subcellular labelling agents) and electron microscopy (e.g. cryofixation by spray freezing), and (4) it has an extensive endomembrane system that processes the ECM . In this study, we exploit these attributes to analyse the effects of ES5 and other endomembrane system-disrupting agents. Chemical biology and specifically the identification/application of chemicals that target and perturb endomembrane structure and function have become one of the valuable tools in the dissection of the endomembrane system. In this study, we describe multiple screening techniques that can be used to analyse the effects of endomembrane-targeting chemicals in the unicellular charophyte Penium margaritaceum. Following the initial screens, we focused our analyses on the effects of ES5, one of the low-molecular-weight, bioactive molecules that have been shown to interact with endomembrane trafficking and affect the architecture of endomembrane compartments (Robert et al., 2008;Drakakaki et al., 2011;Dejonghe and Russinova, 2017;Davis et al., 2020;De Caroli et al., 2020). We also compared these effects with BFA, a widely used endomembrane systemdisrupting agent that has been used with Penium , other zygnematophytes and green algae (Dairman et al., 1995;Salomon and Meindl, 1996;Domozych, 1999;Hummel et al., 2007;Lütz-Meindl, 2016), and ConcA), a widely used V-ATPase inhibitor that has been shown to disrupt the endomembrane system in plants (Seidl, 2022).
ES5 is the endosidin that causes the most notable perturbations to ECM production and the endomembrane system. During the first 4 h of treatment there were substantial alterations to Golgi body positioning in the cytoplasmic valleys and the structure of the Golgi apparatus. Golgi bodies no longer aligned in the typical linear arrays and the cisternae of the medial-to-trans loci of the Golgi body curl inward at the trans face. Secretory vesicle number decreases near the Golgi bodies but they are still present in the peripheral cytoplasm. After longer treatments (e.g. 24 h), the Golgi bodies are reduced to a few scattered cisternae. During ES5 treatment, EPS and cell wall secretion are inhibited or curtailed substantially. These effects indicate that ES5 directly targets the structural integrity of the Golgi apparatus. TEM imaging revealed that the trans-face cisternae are the first to be disrupted, with the cis-face cisternae remaining relatively intact in early stages of treatment We also noted that typical TGN is not found in ES5-treated cells. It is possible that the movement of cisternae towards the trans face and their ultimate transformation into the TGN are inhibited, which in turn leads to curling of the cisternae. At present, little information is available concerning factors that maintain the infrastructure of the Golgi body of Penium. For example, this would include specific proteins, such as Golgins and Golgi Reassembly and Stacking Proteins (GRASPs; Rui et al., 2022), that maintain Golgi structure, including cisternal stacking. However, the recent sequencing of the Penium genome has revealed many Golgin-like genes (Jiao et al., 2020). Further work will be required at both the molecular and cellular levels to identify those Golgi-specific maintenance components and the exact roles of ES5 in their alteration.
ES7 and ES9 had no discernible impact on the secretion of ECM components and were not considered further for this study. Both ES2 and ES3 were found to inhibit EPS production. ES2 has been shown to inhibit exocytosis in plants and human cells and to specifically target the EXO70 subunit of the exocyst complex (Zhang et al., 2016). Little is known about the exocyst of Penium but screening of the Penium genome reveals many candidate genes for the exocyst in this alga (Jiao et al. 2020). ES3, interestingly, showed a slight inhibition of cell wall expansion and a mixture of normal and curved Golgi structures. The curved Golgi bodies in ES3-treated cells display similarities to those following treatment with ES5 ( Supplementary Data Fig.  S3). ES3 has been shown to affect membrane trafficking but the specific target is unknown (Drakakaki et al., 2011).
In a previous study , treatment of Penium with BFA resulted in notable changes to both the endomembrane system and cell wall expansion at 1 μg mL -1 (3.57 µm) after 2 h. In this study we tested lower concentrations of BFA which would not kill the cells over 24-48 h of treatment. Penium's endomembrane system was shown to be very sensitive to BFA even at a very low concentration (1.0 µm). As in the previous study, BFA altered the positioning of the Golgi bodies in the cytoplasm and disrupted Golgi architecture. The number of cisternae decreases from 12-15 to fewer than six and cisternae become greatly expanded. Both EPS secretion and cell wall expansion are inhibited. Secretory vesicle production in treated cells for up to 48 h still occurs at the Golgi body and the peripheral cytoplasm is filled with secretory vesicles. None of the large vacuole-like structures found in the previous study  were seen here, probably due to the reduction in the concentration of BFA. These results indicate that in addition to BFA's disruption of the Golgi architecture, this agent also significantly reduces or inhibits secretory vesicle fusion with the plasma membrane and secretion. This corresponds with other studies whereby BFA caused a near complete blockage of constitutive secretion in both animal cells (Rosa et al., 1992;De Lisle and Bansai, 1996) and plant cells (Takác et al., 2011;Drdová et al, 2013;Rounds et al., 2014;Scali et al., 2021). BFA has also been to shown to cause disassembly of the Golgi apparatus and the absorbance of most Golgi components into the ER in plant cells (Nebenführ et al., 2002;Lam et al., 2009;Foissner et al., 2020). It was shown that the majority of Golgi cisternae fuse directly with the ER, leading to the formation of an ER-Golgi hybrid compartment. Additionally, BFA treatment often causes both the Golgi apparatus and TGN to form distinct aggregates or 'BFA bodies' (Lam et al., 2009). In Penium, neither ER-Golgi compartments nor BFA bodies were observed and in fact the altered Golgi stacks were devoid of any notable connections of the Golgi body to the ER and TGN-like structures. This leads to three conclusions: (1) BFA targets Golgi body architecture in Penium in a notably different way than in higher plant cells and causes effects at much lower concentrations; (2) different concentrations of BFA cause different effects on the Golgi architecture but all stop the secretion of cell wall and EPS; and (3) the production of secretory vesicles in the Golgi body probably does not require transport through the TGN. Future analyses of the TGN in Penium and other zygnematophytes and its role as a hub for exo-and endocytosis activities will be needed to provide a more complete understanding of endomembrane dynamics in these algae. The differences seen in BFA effects on Penium as compared to land plants is perhaps unsurprising due to known differences in the Golgi apparatus between Penium and other green organisms. These differences may limit the number of parallels that can be drawn between Penium and land plants, although similarities do exist. For example, BFA's suppression of cell wall expansion as observed in Penium has also been observed in a diverse array of plant cells (Rutten and Knuiman, 1993;Yasuhara and Shibaoka, 2000;Grebnev et al., 2020;Kim et al., 2021;Yan et al., 2022).
ConcA is a known inhibitor of V-ATPase, an important proton pump found throughout the endomembrane system of both plants and algae (Grunow et al., 1999;Seidel, 2022). In Chlamydomonas, ConcA treatment caused changes to autophagy/vacuoles (Couso et al., 2018). Our TEM analyses of Penium's vacuolar/vesicle network did not reveal major changes but future work will be required to specifically examine both autophagy and experimental manipulation of the autophagic mechanism in this alga. In Penium, ConcA altered the number, positioning and architecture of the Golgi bodies. The Golgi bodies become clustered together, each only containing a few, highly curved cisternae. Swelling at the ends of the cisternae and the formation of MVBs can be seen, similar to results found in Arabidopsis (Dettmer et al., 2006). ConcA treatment in tobacco BY-2 cells also leads to a curving of the cisternae and a reduction in cisternae number but also leads to a prominent 'vacuolation' of the Golgi apparatus that was not seen in Penium (Robinson et al., 2004). Secretory vesicles are still found in the cytoplasm but both EPS secretion and cell wall expansion are inhibited. This is perhaps unsurprising as the inhibition of V-ATPase in plants has been shown to inhibit exocytosis, including delivery of cell wall material (Brüx et al., 2008;Luo et al., 2015;De Caroli et al., 2020).
Endosidins have become important tools in 'dissecting' the dynamics of exo-and endocytosis in plant cells (Drakakaki et al., 2011). In this study we show that ES5 is a potent perturbation agent of Golgi architecture and the processing of the cell wall and EPS of Penium. The exact target and action of ES5 are not known but the effects clearly differ from other endomembrane inhibitors such as BFA and ConcA. Further detailed studies will be required to elucidate the mechanism of action, but this chemical offers a new tool in the arsenal of agents available to dissect Golgi architecture and membrane trafficking in late-divergent charophytes. Such inhibitors could play an important role in our understanding of the role played by the endomembrane system in the transition of an ancient charophyte ancestor on to land.