Cellulose-synthesizing machinery in bacteria

Cellulose is produced by all plants and a number of other organisms, including bacteria. The most representative cellulose-producing bacterial species is Gluconacetobacter xylinus, an acetic acid bacterium. Cellulose produced by G. xylinus, commonly referred to as bacterial cellulose (BC), has exceptional physicochemical properties resulting in its use in a variety of applications. All cellulose-producing organisms that synthesize cellulose microfibrils have membrane-localized protein complexes (also called terminal complexes or TCs) that contain the enzyme cellulose synthase and other proteins. The bacterium G. xylinus is a prolific cellulose producer and a model organism for studies on cellulose biosynthesis. The widths of cellulose fibers produced by Gluconacetobacter are 50–100 nm, suggesting that cellulose-synthesizing complexes are nanomachines spinning a nanofiber. At least four different proteins (BcsA, BcsB, BcsC, and BcsD) are included in TC from Gluconacetobacter, and the proposed function of each is as follows: BcsA, synthesis of a glucan chain through glycosyl transfer from UDP-glucose; BcsB, complexes with BcsA for cellulose synthase activity; BcsC, formation of a pore in the outer membrane through which a glucan chain is extruded; BcsD, regulates aggregation of glucan chains through four tunnel-like structures. In this review, we discuss structures and functions of these four and a few other proteins that have a role in cellulose biosynthesis in bacteria.


Introduction
Cellulose, an assembly of b-1,4-linked linear glucan chains, is the most abundant polysaccharide and biopolymer. It is used widely, without and with further modifications (such as ether and ester derivatives), in a number of industries including clothing, pulp and paper, food, pharmaceuticals, cosmetics, and building materials. Most cellulose is synthesized as a major component of the cell walls in plants, and it is estimated that between 10 10 and 10 11 tons of cellulose is produced each year (Hon 1994). Cellulose is an environmentally friendly and circulating polymer, as plants use carbon dioxide from the atmosphere during photosynthesis to produce the raw materials for cellulose biosynthesis, and it is broken down by a number of organisms.
In addition to plants, other organisms, including bacteria, produce cellulose as an extracellular polysaccharide. Cellulose biosynthesis is reported in a number of bacterial and cyanobacterial species (Napoli et al. 1975; Matthysse et al. 1981;Nobles et al. 2001;Otsuka et al. 2004;Kumagai et al. 2011;Kawano et al. 2011;Castro et al. 2013; Thongsomboon et al. 2018), but it is in the acetic acid bacteria where production of cellulose is best observed and studied. Even among the acetic acid bacteria, it is Gluconacetobacter xylinus/ Gluconacetobacter hansenii (referred to as Acetobacter xylinum in earlier publications, and more recently classified in the genus Komagataeibacter) (Yamada and Yukphan 2008;Yamada et al. 2012) that have been studied extensively for cellulose biosynthesis, and where maximal amount of cellulose production is observed. Both these species of Gluconacetobacter are Gram-negative, rod-shaped, obligate aerobes that are widely used for bacterial cellulose production, and are the representative cellulose-producing bacterial species. Herein, we will refer to both these species as G. xylinus or G. hansenii.
Cellulose synthesis in all organisms is catalyzed by the enzyme cellulose synthase, a transmembrane protein, that uses UDP-glucose as the substrate. In bacteria, cellulose synthase is part of a cellulosesynthesizing machinery that most likely includes other proteins. A linear array of particles associated with cellulose synthesis was observed in G. xylinus by freeze-fracture electron microscopy (Brown et al. 1976), with each particle representing a large transmembrane protein complex and the site of cellulose synthesis. That other proteins, in addition to cellulose synthase, may be part of this complex is suggested by the identification of the gene for cellulose synthase as part of an operon in G. xylinus (Saxena et al. 1990(Saxena et al. , 1994Wong et al. 1990;Nakai et al. 1998;Umeda et al. 1999;Kawano et al. 2002b) and other bacteria (Römling and Galperin 2015), and by proteinlocalization studies (Kimura et al. 2001;Sunagawa et al. 2013;Sun et al. 2017).
Although cellulose synthase is the main component of the cellulose-synthesizing complex, other proteins have a role in cellulose synthesis in vivo. What are these other proteins, what role do they play in cellulose biosynthesis, and how do they interact with each other in the cellulose-synthesizing complex? Some information related to these questions is coming from structural studies, and in this review, we discuss structures of a few proteins, specifically those that are known to have a role in cellulose synthesis in Gluconacetobacter species. Even as the structure and role of individual proteins in cellulose biosynthesis is being worked out, the objective in most cases is to determine the structure of the complete cellulosesynthesizing complex that performs all the events in cellulose synthesis, starting with the addition of glucose from UDP-glucose to a growing glucan chain (polymerization) to extrusion of glucan chains across the cell envelope.
Purified cellulose synthase from G. xylinus contained two polypeptides (BcsA and BcsB) (Lin and Brown, Jr. 1989), and BcsA was shown to bind to UDP-Glucose, and therefore, it was identified as the cellulose synthase catalytic subunit ). Following identification of the cellulose synthesis operon (bcs operon) in G. xylinus (Wong et al. 1990;Saxena et al. 1994), other proteins that have a role in cellulose synthesis were identified. From analysis of mutants, it was shown that while BcsC is absolutely required for native cellulose synthesis, BcsD is not absolutely required (Saxena et al. 1994). Genetic and biochemical studies demonstrated that even as synthesis of cellulose (cellulose II to be more specific) could be carried out in vitro by just BcsA and BcsB, other protein(s) were required for synthesis of native cellulose (cellulose I) in vivo. The role of other proteins, in addition to those coded for by genes in the bcs operon, in native cellulose synthesis was suggested through analysis of mutants affected in cellulose production in G. xylinus. Interestingly, though not surprisingly, the genes for these proteins are present adjacent (both upstream and downstream) to the bcs operon in G. xylinus (Standal et al. 1994;Tonouchi et al. 1997;Nakai et al. 1998;Umeda et al. 1999;Tajima et al. 2001;Kawano et al. 2002b). Upstream of the bcs operon are two genes-one encoding endo-b-1,4-glucanase (CMCax) and the other encoding the cellulose complementing protein (Ccp) (Standal et al. 1994); downstream of the bcs operon is a gene encoding b-glucosidase (Bgl) (Tonouchi et al. 1997;Tajima et al. 2001;Kawano et al. 2002b). CMCax exhibits cellulose hydrolyzing activity (Standal et al. 1994;Kawano et al. 2002a;Yasutake et al. 2006) and its expression is induced at a later stage of growth (Kawano et al. 2008). Gene expression analyses of bgl and cmcax using real-time qRT-PCR suggest that increase in cmcax expression after 5 d cultivation is caused by an increase in gentiobiose (a b-1,6-linked disaccharide of D-glucose), possibly synthesized in a reaction catalyzed by Bgl. Earlier work showed that CMCax affects cellulose production in G. xylinus (Tonouchi et al. 1995;Hyun Min et al. 1998;Kawano et al. 2002a); CMCax may play a role in the regulation of cellulose synthesis together with Bgl (Kawano et al. 2008). The function of Ccp in cellulose biosynthesis is not understood, but it is a protein that exists only in some cellulose-synthesizing bacteria such as Gluconacetobacter. Since Ccp interacts with BcsD, it is speculated that it plays a role in cellulose synthesis in collaboration with BcsD ).
The first report of structure determination of a protein coded for by a gene in the bcs operon of G. xylinus was that of BcsD (Hu et al. 2010). Since then, structures are obtained of protein complexes from other organisms capable of in vitro cellulose synthesis, namely the BcsA-BcsB complex from Rhodobacter sphaeroides (Morgan et al. 2013), and the BcsA-BEFRQ (Abidi et al. 2021) and BcsABEFGRQ complexes (Acheson et al. 2021) from Escherichia coli (E. coli). Crystal structures have been obtained of the N-terminal part, containing the TPR-domain, of BcsC from Enterobacter sp. CJF-002 (Nojima et al. 2017), and the C-terminal pore-containing domain of BcsC from E. coli (Acheson et al. 2019). In addition, structures of BcsZ from E. coli (Mazur and Zimmer 2011) and CMCax from G. xylinus (Yasutake et al. 2006) are known. Given the possible function(s) of some of these proteins and their known structures in a few cellulose-producing organisms, the emerging picture is that the cellulose-synthesizing machinery in bacteria may be much more diverse than originally thought of.
Although G. xylinus is the major cellulose-producing bacterium in nature, only a few proteins with a role in cellulose biosynthesis from this bacterium have been characterized structurally. At the same time, a wealth of information at the microscopic level is available on cellulose biosynthesis in this bacterium. One of our objectives is to understand the protein components at the sites of cellulose synthesis in G. xylinus, and how they may be associated and regulated to produce ribbons of cellulose.

An overview of cellulose synthesis in bacteria
Since the first report of cellulose synthesis in bacteria (Brown 1886), the list of bacteria where cellulose production has been demonstrated, and the number of bacteria where genes for cellulose biosynthesis have been identified, has grown longer. In addition to the acetic acid bacteria that includes species of Gluconacetobacter, production of cellulose is demonstrated in bacterial species such as Agrobacterium tumefaciens (Matthysse et al. 1981), E. coli (Thongsomboon et al. 2018), Rhizobium (Napoli et al. 1975), Enterobacter (Otsuka et al. 2004) and others. Interestingly, analysis of genome sequences of a large number of bacteria led to identification of genes for cellulose biosynthesis in many more bacterial species than are known to produce cellulose. In most cases, the genes for cellulose biosynthesis are organized in an operon. Based on the organization of genes in the operons, they have been classified into four major groups (I to IV) as shown in Fig. 1 (Römling and Galperin 2015). Considerable variation is observed with respect to the morphology and amount of cellulose synthesized by different bacteria, and although the function of cellulose in bacteria is not fully understood, in many cases it is part of the biofilm.
Cellulose biosynthesis is most extensively studied in the acetic acid bacteria, mainly because of the amount and nature of the cellulose produced by Gluconacetobacter. In general, these bacteria are found on the surface of fruits, and other locations where there is an abundant amount of sugars. In fact, the acetic acid bacterium (G. intermedius NEDO-01), which we currently use for the production of cellulose nanofibers, was isolated from a fruit grown in Hokkaido (Kose et al. 2013). In many cases, cellulose synthesized by Gluconacetobacter is called bacterial cellulose (BC), and it has a dense three-dimensional network structure consisting of nanofibers (Fig. 2). The main characteristics of BC synthesized by acetic acid bacteria are as follows: (1) lignin-free, hemicellulose-free, (2) hierarchical assembly of nanofiber (Penttilä et al. 2018), (3) developed three-dimensional network structure, (4) high mechanical strength (high Young's modulus and tensile strength), (5) high biodegradability, (6) high biocompatibility, and (7) high water retention. Because of these characteristics, and the ability to produce it in large amounts, BC is used in making a variety of products, including food items such as Nata de coco, speaker acoustic diaphragms (Nishi et al. 1990), artificial blood vessels (Klemm et al. 2001;Zang et al. 2015), wound dressings (Fontana et al. 1990;Portela et al. 2019), high-strength transparent materials (Yano et al. 2005;Nogi et al. 2006a,b;Ifuku et al. 2007), display devices (Shah and Brown 2005), and capacitors (Wang et al. 2016).
Cellulose is an extracellular polysaccharide and, in general, its biosynthesis involves three steps: (1) polymerization, (2) translocation, and (3) assembly or aggregation of glucan chains/crystallization (Ross et al. 1991). In Gram-negative bacteria, polymerization occurs in the cytosolic domain of the inner membrane-localized cellulose synthase, and nascent glucan chains are translocated through the inner membrane, periplasm, and outer membrane before they assemble or aggregate into higher-order structures. In Gluconacetobacter, the glucan chains assemble in a stepwise manner just outside the cell surface to form ribbons of crystalline cellulose I. Brown et al. visualized a linear array of particles and pores in the cell envelope of G. xylinus following freeze-fracture electron microscopy, and identified these as sites of cellulose synthesis (Brown et al. 1976). Zaar also observed in this bacterium a linear array of pores on the lipopolysaccharide membrane beneath the cellulose ribbons by freeze-etching, and correlated these pores with cellulose production (Zaar 1979). The pores have a rim diameter of 12-15 nm and a central hole or deepening of * 3.5 nm, and these are hypothesized to be export sites of sub-elementary fibrils (Zaar 1979). In G. xylinus, the average  frequency of the pores is 3.8 per 100 nm, and so 50-100 pores are believed to exist along the long axis of a cell (Brown et al. 1976;Zaar 1979).
Fine structure of the cellulose ribbon was observed by transmission electron microscope (TEM) following incubation of Gluconacetobacter in medium with Calcofluor White, a fluorescent brightener Haigler et al. 1980;Haigler and Chanzy 1988). Fibers with a width of 1.5 nm were frequently observed, and it was suggested that fibers (subelementary fibrils) having a width of 1.5 nm were the first form of cellulose to be secreted out. Low molecular weight compounds such as fluorescent brighteners and cellulose derivatives such as carboxymethylcellulose (CMC) bind to the surface of cellulose fibers and prevent the formation of large bundles of microfibrils. Using these compounds, it was hypothesized that in Gluconacetobacter the assembly of cellulose microfibrils and ribbon occurs in a stepwise and hierarchical manner in the extracellular space (Fig. 3), and yet this process is described as celldirected because the mutual orientation and association of glucan chains, aggregates, microfibrils, bundles, and ribbons is clearly governed by the original pattern of extrusion sites (Ross et al. 1991). The linear assembly of pores on the cell surface (sites from which cellulose is secreted out, and corresponding to cellulose-synthesizing sites) that gives rise to a microfibril is referred to as a terminal complex (TC), and in the case of Gluconacetobacter, three pores are suggested to form a terminal complex (Brown 1996;Mehta et al. 2015). Furthermore, the cellulose ribbon produced by Gluconacetobacter exhibits a twist, suggesting that the cells undergo rotation during cellulose synthesis ( Fig. 2A).
Using a combination of direct-staining of reducing ends of cellulose chains and microdiffraction-tilting electron crystallographic analysis, Koyama et al. demonstrated that the reducing ends of growing glucan chains in the cellulose ribbon of Gluconacetobacter point away from the bacterium, suggesting that polymerization of glucose residues takes place at the nonreducing end of the growing glucan chains (Koyama et al. 1997). Thus, all the glucan chains point in the same direction for the ''parallel-up'' packing in unit cells of cellulose I. This packing arrangement is also consistent with the mechanism of cellulose synthesis suggested by the structure of cellulose synthase from R. sphaeroides (Morgan et al. 2013). Considering that TCs are present along the longitudinal axis of Gluconacetobacter cells and they participate in synthesis of the cellulose ribbon, it has been possible to localize a few proteins to these cellulosesynthesizing complexes using electron and fluorescence microscopy. BcsB was the first protein that was localized to the TCs in Gluconacetobacter by immunolabeling of freeze-fracture replicas followed by electron microscopy (Fig. 4) (Kimura et al. 2001;Sun et al. 2017). The localization of BcsA (Fig. 5), BcsD (Figs. 5 and 6), and Ccp (Fig. 7) to the longitudinal axis of bacterial cells was also demonstrated by fluorescence microscopy, suggesting that these proteins are present in the TCs of Gluconacetobacter Sun et al. 2017). Figures 4, 5, 6, 7 show the presence of BcsA, BcsB, BcsD, and Ccp along the longitudinal axis of a cell of Gluconacetobacter, whereby the linear arrangement  Ross et al. 1991, and is used with permission of American Society for Microbiology of these proteins matches with the linear arrangement of TCs in these bacteria, suggesting that these proteins are part of the cellulose-synthesizing machinery present along the same axis. However, so far it was unclear as to how the TCs line up linearly along the longitudinal axis in these bacteria. Recently, Nicolas et al. studied cellulose-synthesizing Gluconacetobacter in a frozen-hydrated, near-native state by cryoelectron tomography and focused-ion-beam milling, and observed a new cytoskeletal structure called the cortical belt, adjacent to the inner membrane and below the site where cellulose is seen emerging from cells ( Fig. 8) (Nicolas et al. 2021). The cortical belt was not observed in other cellulose-producing bacteria, such as A. tumefaciens and E. coli 1094. Since these other bacterial species do not produce organized cellulose ribbons, Nicolas et al. proposed that the cortical belt holds the cellulose-synthesizing complex in a row to form higher-order cellulose structures such as sheets and ribbons (Nicolas et al. 2021).

Cellulose synthase (BcsA-BcsB)
Cellulose synthase is the key enzyme in cellulose biosynthesis. The first purified form of this enzyme from Gluconacetobacter showed that it is made of two different-sized polypeptides that were later identified as BcsA and BcsB (Lin and Brown, Jr. 1989). BcsA was identified as the UDP-glucose-binding and catalytic subunit ) but the function of BcsB was not fully understood until the structure of cellulose synthase (BcsA-BcsB) from R. sphaeroides was determined (Morgan et al. 2013). BcsA is a transmembrane protein that is present in the inner membrane, and it contains a glycosyltransferase family 2 (GT-2) domain, a PilZ domain, and eight transmembrane helices (TMHs).
The first structure of cellulose synthase was determined following expression of the R. sphaeroides bcsA and bcsB genes in E. coli, and purification and crystallization of the catalytically active BcsA-BcsB complex (Morgan et al. 2013). BcsA and BcsB from R. sphaeroides form a 1:1 stoichiometric complex (Morgan et al. 2013), similar to what is observed in purified cellulose synthase from Gluconacetobacter (Chen and Brown 1996). The BcsA-BcsB complex of R. sphaeroides has a membrane-spanning domain, a large cytoplasmic loop of BcsA that contains the GT domain, and a periplasmic domain made mostly of BcsB (Morgan et al. 2013). A helix of BcsB spans the inner membrane; at the same time, BcsB also interacts in some manner with the outer membrane (Kimura et al. 2001;Sun et al. 2017).
We modeled the structure of the BcsA-BcsB complex from Gluconacetobacter using Phyre2 (http://www.sbg.bio.ic.ac.uk/*phyre2/html/page. cgi?id=index) with the structure of BcsA-BcsB from R. sphaeroides (PDB: 5EJ1) as a template. Figure 9 shows that the modeled structure of the G. hansenii ATCC 23769 BcsA-BcsB complex is quite similar to the BcsA-BcsB complex of R. sphaeroides. Alignment of the amino acid sequences and secondary structures of BcsA from R. sphaeroides and G. hansenii ATCC 23769 shows that the active site in BcsA of G. hansenii exists between amphipathic interface (IF) helices 1 and 2, similar to what is observed in BcsA from R. sphaeroides (Fig. 10). The three aspartic acid residues in the conserved D,D,D,Q-X-X-R-W motif identified in cellulose synthases and other b-glycosyltransferases (Saxena et al. 1995) correspond to D171, D221, and D316 of the G. hansenii ATCC 23769 BcsA. The amino acids residues in the Q-X-X-R-W motif of G. hansenii ATCC 23769 BcsA correspond to Q352, R353, V354, R355, and W356, and these residues are present in IF2 (Fig. 10).
The structure of the BcsA-BcsB complex from G. hansenii ATCC 23769 is determined at 23.4 Å resolution using negative stain images (Fig. 11) (Du et al. 2016). The structure reveals that BcsA-BcsB of Glucose is transferred from UDP-glucose to the non-reducing end of the glucan chain, which is the growing end (Koyama et al. 1997). Based on the twofold helical symmetry of the glucan chain in cellulose, it would appear that UDP-glucose alternately approaches from two opposite sides of the non-  Sun et al. 2017, and is used with permission of Springer Nature reducing terminal residue of the extending glucan chain. Instead, a mechanism is proposed in which the glucose residue, immediately after the rearrangement reaction, rotates with respect to the glycosidic bond and is inverted with respect to the adjacent residue on the reducing end side (Morgan et al. 2016). Knott et al. confirmed this proposed mechanism by molecular simulation (Fig. 12) (Knott et al. 2016). A multi-scale simulation applying quantum chemistry analyzed the three-dimensional structural changes, including the activation energy of the rearrangement reaction and the proton transfer along the reaction coordinates. The proposed synthesis cycle is as follows: Glycosyl transfer can add the glucose moiety from UDP-glucose in the same orientation as the acceptor glucose ('same side', with the hydroxymethyl of both rings on the same side) or in the opposite orientation ('opposite side').
Basically, the same process occurs for both the 'opposite side' (top six panels) and the 'same side' (bottom six panels) portions of the cycle: (upper left) gating loop is open and active site is empty; (upper middle) UDP-glucose binds in the active site; (upper right) the gating loop inserts into the active site; (lower right) glycosyl transfer produces an elongated cellulose chain and UDP product; (lower middle) cellulose translocation moves the chain into the transmembrane tunnel; (lower left) gating loop opens facilitating UDP product expulsion (Fig. 12). In addition, it is suggested that the rotation of glucose residues immediately after the rearrangement reaction proceeds thermodynamically due to interaction with the surrounding aromatic side chains, and the up-and-down movement of finger helix that guides the cellulose molecular chain in the transmembrane (TM) direction is thermodynamically possible (Knott et al. 2016).
The structure of the PilZ domain of BcsA from G. hansenii ATCC 23769 ) is similar to the PilZ domain of BcsA from R. sphaeroides (Morgan et al. 2013) (Fig. 13). The arginine (R580) within the TM8-b-barrel linker and glutamine (E371) residues in BcsA from R. sphaeroides, which are responsible for the regulation of cellulose synthesis, are conserved and correspond to R557 and E344 in BcsA from G. hansenii ATCC 23769 (Fig. 10). The similarities in the PilZ domain of BcsAs from R. sphaeroides and G. hansenii suggests that cellulose synthase is regulated by a similar mechanism in these two bacteria.
The structure of BcsB in the R. sphaeroides BcsA-BcsB complex shows the presence of a periplasmic BcsB domain (function unknown) and one transmembrane domain on the C-terminal side (Morgan et al. 2013). The BcsB domain consists of two cellulosebinding domains (CBDs 1 and 2) and two flavodoxinlike domains (FDs 1 and 2). The CBDs are positioned at a 45°angle relative to one another, and stabilized by disulfide bonds. CBD1 has a conserved amino acid sequence (H159, R160, I161, L171, and W172)  Sunagawa et al. (2013), and is used with permission of Elsevier B.V similar to that of a bacterial CBD which forms a complex with glucuronic acid dimers, suggesting that this region could be involved in translocation of a glucan chain (Morgan et al. 2013). BcsB is absolutely required for BcsA catalytic activity. Studies in R. sphaeroides identified the BcsB C-terminal transmembrane helix and the preceding amphipathic helix as being necessary for the BcsA catalytic activity, possibly through stabilization of the transmembrane region of BcsA and completing its inner membrane transport domain (Omadjela et al. 2013). Although the stoichiometry of BcsA to BcsB is 1:1 in Gluconacetobacter and R. sphaeroides, cryo-EM studies of E. coli cellulose synthesizing complexes show a hexamer of BcsBs associating with a single BcsA (Abidi et al. 2021;Acheson et al. 2021).

BcsC
BcsC is the largest protein (* 130 kDa) encoded by genes present in the bcs operon. Along with bcsA and bcsB, bcsC is part of the bcs operon in almost all bacteria. In most cases, transcription of bcsC follows that of bcsA and bcsB. Cellulose synthase activity is not affected in Gluconacetobacter mutants in which the bcsC gene is disrupted, but these mutants do not produce cellulose, suggesting that BcsC is an essential  (Saxena et al. 1994). Proteins similar to BcsC play a role in secretion of exopolysaccharides in other bacteria (Whitney and Howell 2013). BcsC has an N-terminal tetratricopeptide repeat (TPR) domain and a C-terminal poreforming b-barrel domain. Acheson et al. determined the crystal structure of a truncated BcsC protein from E. coli that contains a TPR repeat, a linker region and the complete b-barrel pore (Acheson et al. 2019). Modeling of structure of the C-terminal part of BcsC from G. hansenii ATCC 23769 using Phyre2 shows that the higher-order structure of G. hansenii BcsC is similar to the BcsC from E. coli (Fig. 14). BcsC is localized in the outer membrane and periplasmic space, and in E. coli it is estimated that one phosphoethanolamine-modified glucan chain (pEtN-cellulose) is excreted from the cells through the b-barrel channel (Acheson et al. 2019). It is not known if BcsC in Gluconacetobacter transports a single glucan chain or a bundle of glucan chains (sub-elementary fibril).
The TPRs towards the C-terminal side of BcsC are composed of two antiparallel a-helices connected by a turn. These TPRs are stacked on each other to form a superhelical structure, something observed in other TPR-containing proteins (D'Andrea and Lynne 2003). Figure 15 shows the TPR domain of BcsC from Enterobacter sp. CJF-002 (Nojima et al. 2017). The hinge (turn) between a5 and a6 allows the C-terminal part to move, suggesting that this region may be involved in extrusion of the glucan chain.

BcsD
In Gluconacetobacter, bcsD follows bcsC, and is the last gene in the bcs operon. For many years, bcsD was identified only in Gluconacetobacter, but as more bacterial genome sequences became available it was observed in bcs operons in a few other bacteria. Analysis of Gluconacetobacter mutants, where bcsD is disrupted, provided evidence that while BcsD is not required for cellulose synthase activity, this protein has a non-essential role in cellulose production (Saxena et al. 1994;Mehta et al. 2015). bcsD mutants Fig. 9 Model structure of the BcsA-BcsB complex from G. hansenii ATCC 23769 (red) predicted by Phyre2 (http://www.sbg.bio.ic.ac. uk/*phyre2/html/page.cgi?id=index) using BcsA-BcsB complex (PDB: 5EJ1, blue) from R. sphaeroides as a template Fig. 10 Modeling of BcsA from G. hansenii ATCC 23769 by Phyre2 (http:// www.sbg.bio.ic.ac.uk/ *phyre2/html/page.cgi?id= index) using BcsA from R. sphaeriides as a template of Gluconacetobacter produce reduced amounts of native cellulose I; at the same time, they produce detectable amount of cellulose II, suggesting that BcsD probably influences assembly of glucan chains during formation of the cellulose ribbon. Interestingly, no homolog of BcsD has been identified in the protein databases and it maybe a unique protein. Although no signal sequence for translocation of this protein is identified in its amino acid sequence, biochemical analysis suggests that BcsD is present in the periplasmic space in Gluconacetobacter (Iyer et al. 2011).
Interestingly, the crystal structure of BcsD is the first one to be determined of a protein coded in the bcs operon of Gluconacetobacter (Hu et al. 2010). BcsD assembles into a cyclic octamer, with each subunit folded into a globular structure and exhibiting five ahelices and a four-stranded b-sheet (Fig. 16). In the BcsD octamer, two monomers interact with each other using the two N-terminal helices a1 and a2, which have a hook-like arrangement, resulting in a two-fold symmetrical assembly of a stable homodimer (Fig. 16A). The octamer, a tetramer of the dimers AB, CD, EF, and GH, exhibits a cylindrical structure along the non-crystallographic four-fold axis of symmetry, with a height of 65 Å , an outer diameter of 90 Å , and an inner diameter of * 25 Å (Fig. 16B).
The N-terminus of each subunit is present inside the octameric ring, forming four tunnel-like structures (Hu et al. 2010). To determine if glucan chains pass through these tunnel-like structures, BcsD was cocrystallized with cellopentaose (CPT), an oligosaccharide representing the glucan chain of cellulose. Results from these experiments showed that CPT was present in each of the four tunnel-like structures inside the ring (Hu et al. 2010). Whether the tunnel-like structures, formed by N-terminals of the BcsD subunits, have any role in cellulose synthesis was studied by analyzing BcsD with different lengths of N-terminal deletions (Fig. 17). These studies show that cellulose yield is reduced significantly in strains that express BcsD with N-terminal deletions, and where no tunnel-like structures are observed, suggesting that the tunnel-like structures play an important role in optimal cellulose synthesis (Hu et al. 2010).
Construction of a computerized model and molecular dynamics (MD) simulation of BcsD complexed with four cellulose chains of DP = 12 (Glc12) revealed an unexpected S-shaped pathway with a flexion region between BcsD dimers (Fig. 18) in the octamer (Uto et al. 2021). Glucose residues located in the region where the pathway bends show reversible changes in ring conformation, and this phenomenon might be associated with the role BcsD plays in cellulose microfibril production (Uto et al. 2021). Furthermore, molecular modeling and molecular simulation shows that the sugar chain pathway of BcsD is much narrower than that of BcsB and BcsC, and that the cellulose chain is strongly compressed at the BcsD surface, especially in the region where the pathway bends. Since BcsD has a homogeneous octamer structure that is overall symmetrical, the cellulose chains are compressed at the BcsD surface at the same time in the four sugar chain pathways (Uto et al. 2021). The movement of nascent cellulose chains extending from four sets of BcsA-BcsB complexes is synchronized as they pass through the BcsD octamer, such that

Cellulose complementing factor (Ccp)
Cellulose complementing factor (Ccp) is an essential protein required for synthesis of cellulose in acetic acid bacteria. Mutation in the ccp gene results in loss of cellulose synthesis in Gluconacetobacter (Standal et al. 1994). A fusion protein of Ccp and a fluorescent protein (EGFP) was used to determine the requirement of Ccp in cellulose synthesis and its localization . Expression of the fusion protein in a Ccp-deficient strain (Cel-1) of Gluconacetobacter resulted in gain of cellulose synthesis in the mutant strain, and the fusion protein was found to localize along the longitudinal axis of the cell (Fig. 7) ). More interestingly, pull-down assays revealed that Ccp directly interacts with BcsD ( Fig. 19), suggesting that Ccp is part of the cellulosesynthesizing complex . Presently, the function and structure of Ccp is not known.

Endo-glucanase (CMCax)
CMCax is an endo-type glucanase that hydrolyzes glucan chains to reduce the molecular weight, and is classified in Glycosyl hydrolase Family 8 (GH-8). In Gluconacetobacter, the cmcax gene is present upstream of the bcs operon, together with the ccp gene ( Fig. 1: Ia, bcsZ = cmcax, bcsH = ccp) (Standal et al. 1994). Based on immunostaining and TEM observations, CMCax is shown to be present not only in the growth medium but also on the cell surface of Gluconacetobacter cells (Yasutake et al. 2006). In CMCax-deficient Gluconacetobacter mutants, cellulose yield is significantly reduced and the cellulose ribbon is extremely twisted compared to that produced by the wild-type strain, suggesting that CMCax has a function in eliminating strain and facilitating ribbon formation (Nakai et al. 2013). Fig. 15 The structure of BcsC-TPR(N6) from Enterobacter sp. CJF-002 A BcsC-TPR(N6) is composed of six TPR motifs (colored blue, light blue, green, yellow, orange, and red) and two unpaired a-helices (gray). B Schematic diagram of the secondary structure of BcsC-TPR(N6). The boxes indicate a-helices and the lines indicate turns. The color scheme is the same as in (A). (Nojima et al. 2017: S. Nojima, A. Fujishima, K. Kato, K. Ouchi, N. Shimizu, K. Yonezawa, K. Tajima, and M. Yao. Sci Rep 2017, 7, 13018 https://doi.org/10.1038. Published by Springer Nature GH-8 enzymes hydrolyze glycosidic linkages using general acid catalysis, and a proton donor and a nucleophile or general base are required (Alzari et al. 1996). From sequence alignment of GH-8 enzymes, Glu57 in CMCax is suggested as a putative proton donor (Yasutake et al. 2006). Endo-glucanase (CtCelA) from Clostridium thermocellum, which belongs to GH-8, has five aromatic residues for sugar recognition by stacking interactions. These residues correspond to Trp132, Trp205, Tyr372, Tyr277, and Tyr369, and form five sugar-recognition subsites, -3, -2, ? 1, ? 2, and ? 3, respectively (Fig. 20, left) (Schmidt et al. 2002). However, residue corresponding to Tyr369 of CtCelA is not conserved in CMCax, suggesting that CMCax lacks subsite ? 3 (Fig. 20, right). Other endoglucanases involved in cellulose synthesis, BcsZEc from E. coli (Mazur and Zimmer 2011) and BcsZEn from Enterobacter (Sunagawa et al. 2012), are also deficient in the ? 3 subsite, suggesting that this structural change may have a role in cellulose synthesis.
Determination of subsite affinities show that Glc2 (cellobiose) cannot bind rigidly at the active site of 1,4-b-glucosidase to release cellobiose (Tahara et al. 1998b). Modeling of Bgl from G. hansenii ATCC 23769 by Phyre2 using BGL (PDB: 3AC0) from Kluyveromyces marxianus as a template suggests that the active site of Bgl is located near the center and is bag-shaped (Fig. 21). This suggestion corresponds with results of experiments on determination of the subsite of Bgl (Tahara et al. 1998b).
It is also known that Bgl catalyzes glycosylation reactions depending on the reaction conditions, and produces various b-glycosidic disaccharides (Kawano et al. 2008). Amongst the different disaccharides produced by Bgl, gentiobiose (b1,6-linked disaccharide of glucose) is known to induce expression of CMCax. From temporal changes in the expression of cmcax and bgl, it is suggested that Bgl and CMCax work together to modulate cellulose in the late stages of culture of Gluconacetobacter (Kawano et al. 2008).

Conclusions
Cellulose is produced mainly by plants, and is a major sink of atmospheric carbon dioxide. In addition to plants, a number of bacteria produce cellulose, many   Sunagawa et al. (2013), and is used with permission of Elsevier B.V as part of the biofilm. The acetic acid bacteria, mainly Gluconacetobacter, are unique in producing large amounts of cellulose with many desirable properties. Strains of G. xylinus and G. hansenii are useful for understanding the basic mechanisms of cellulose biosynthesis, as well as large-scale production of bacterial cellulose. Even as new information is becoming available on the various proteins and their assembly into a cellulose-synthesizing complex in bacteria such as E. coli, limited understanding exists on the composition of the cellulose-synthesizing machinery in Gluconacetobacter and how it is assembled. In E. coli, phosphoethanolamine-modified glucan chains form part of the biofilm matrix, and these are synthesized and secreted by a cellulosesynthesizing complex with at least 9 subunits  (BcsRQABZCEFG). Using cryo-EM, three-dimensional reconstructions have been obtained of E. coli cellulose-synthesing complexes with many of the subunits, and that show an unexpected subunit stoichiometry of BcsA and BcsB (Abidi et al. 2021;Acheson et al. 2021). The cellulose-synthesizing complex in Gluconacetobacter produces unmodified glucan chains that assemble into microfibrils and ribbons of cellulose. Based on genes present in and adjacent to the bcs operon in Gluconacetobacter, and the known structure and function of proteins coded by these genes, we believe that the structure of the cellulose-synthesizing complex in Gluconacetobacter will be considerably different from the structure of the E. coli cellulose-synthesizing complex. Currently, we are proceeding with structural and functional analyses of proteins associated with cellulose production in Gluconacetobacter and hope to look at assemblies of proteins in the near future.
Determining the overall structure of the cellulosesynthesizing complex and elucidating the entire mechanism of cellulose synthesis in Gluconacetobacter will be meaningful in improving productivity of bacterial cellulose, controlling features of higherorder structures such as fiber width and fiber length, and creating new cellulose-based polysaccharides by modifying substrate specificity. At the same time, research is continuing on the preparation of composites and copolymers by addition of various substances to the medium (Luo et al. 2008;Perotti et al. 2011;Kose et al. 2013;Orelma et al. 2014;Tajima et al. 2017;Gao et al. 2019) and by genetic engineering (Lee et al. 2001;Yadav et al. 2010;Fang et al. 2015;Florea et al. 2016;Teh et al. 2019) to produce cellulose-based environmentally-friendly materials using Gluconacetobacter.
Currently, drastic changes in the environment, caused by elevated levels of greenhouse gases, has become a major problem. Consequently, there is a strong demand for environmentally-friendly manufacturing processes using renewable resources as raw materials. Gluconacetobacter has the unique ability to synthesize large amounts of nanocellulose fibers from low molecular weight biomass compounds via natural processes. It is expected that in the near future various high-strength and high-performance materials will be produced by utilizing the natural cellulose-synthesizing ability of Gluconacetobacter.