In vitro function, assembly, and interaction of primary cell wall cellulose synthase homotrimers

Reviewer #1 (Public Review):

Cellulose is the major component of the plant cell wall and as such is a major component of all plant biomass on the planet. It is made at the cell surface by a large membrane-bound complex known as the cellular synthase complex. It is the structure of the cellulose synthase complex that determines the structure of the cellulose microfibril, the unit of cellulose found in nature. Consequently, while understanding the molecular structure of individual catalytic subunits that synthesise individual beta 1-4 glucose chains is important, to really understand cellulose synthesis it is necessary to understand the structure of the entire complex.

In higher plants, cellulose is synthesised by a large membrane-bound complex composed of three different CESA proteins. During cellulose synthesis in the primary cell wall, this is composed of members of groups CESA1, CESA3, and CESA6. While the authors have previously presented structural data on CESA8, required for cellulose synthesis in the secondary cell wall, here they provide structural and enzymatic analysis of CESA1, CESA3, and CESA6 from soya beans.

The authors have utilised their established protocol to purify trimers for all three classes of CESA proteins and obtain structural information using electron microscopy. The structures reveal some subtle, but interesting differences between the structures obtained in this study and that previously obtained for CESA8. In particular, they identify a change in the position of transmembrane helices 7 that in previous structures formed part of the transmembrane channel. In the structure of CESA1 TM7 is shifted laterally to a position more towards the periphery of the protomer, where it is stabilised by inter-protomer interactions. This creates a large lipid-exposed channel opening that is likely encountered by the growing cellulose chain. In the discussion, the authors speculate this channel might facilitate lateral movement of cellulose chains in the membrane which would allow them to associate to form the microfibril. There is, however, no explanation for why this might be different for CESA proteins involved in primary and secondary cell wall CESA proteins.

Interactions within the trimer as stabilised by the plant conserved regions (PCR), while in common with previous studies that class-specific regions (CSR) are not resolved, are likely highly disordered as has been suggested in previous studies. As the name suggests these regions are likely to be important for determining how different CESA proteins interact, but it remains to be seen how they achieve this. Similarly, the N-terminal domain (NTD) remains rather intriguing. In the CESA3 structure, the NTD forms a stalk that protrudes into the cytoplasm that was previously observed for CESA8, while it remains unresolved in CESA1 and CESA6. The authors suggest the inability to resolve this region is likely the result of the NTD being able to form multiple conformations. Loss of the NTD does not prevent the formation of trimers and CESA1 and CESA3 are still able to interact. Previous bioinformatic studies suggest that the CSR part of the NTD is also highly class-specific (Carrol et al. 2011 Frontiers in Plant Science 2, 5-5) suggesting it is also likely to participate in interactions between different CESA proteins. This analysis provides little new information on the structure of the NTD or how it functions as part of the cellulose synthase complex.

The other important point regarding cellulose synthesis is how the different CESA trimers function during cellulose synthesis and complex assembly. The authors provide biochemical evidence that mixed complexes of two different CESA proteins are able to synergistically increase the rate of cellulose synthesis. This increase is not dramatic, around 2-fold as it is unclear what brings about this increase and whether it results from the ability to form larger complexes favouring greater rates of cellulose synthesis.

It is clear however from electron microscopy that mixing of CESA proteins can lead to the formation of large aggregates not seen with single CESA proteins. The aggregates observed do not form rosette-type shapes but appear to be much more random aggregates of different CESA trimers. The authors suggest that this is likely a result of the fact that the complexes are not constrained in two dimensions by the membrane, however, if these are biologically relevant interactions that form aggregates it is somewhat surprising that they do not form hexameric structures, particularly since they are essentially forming as a single layer.

Overall the study provides some important data and raises a number of important questions.

Reviewer #2 (Public Review):

Summary:

In their manuscript entitled "In vitro function, assembly and interaction of primary cell wall cellulose synthase homotrimers" Purushotham et al. purify and functionally and structurally characterize the primary cell wall cellulose synthase isoforms from soybeans. Overall, the manuscript is well-written and contributes several important observations.

Strengths:

The structural and functional characterization of all three primary cell wall CesA isoforms contributes significantly to important problems in plant biochemistry.

The demonstration that the isolated CesA monomers and homotrimers are catalytically active in vitro, interact with each other, and show catalytic cooperativity between the homotrimers.

Weaknesses:

The paper could be further strengthened by addressing the following:

Are the interactions between the homotrimers observed via the pull-down assays stable enough to co-elute on the sizing column or are they transient interactions?

The authors show that the monomeric CesA isoforms can interact with each other using pull-down assays (Figure Supplement 4e). Are these interactions stable or transient? Have the authors tried running the mixed monomers over a sizing column? If you mix all three isoform monomers can you form heterotrimers?

The authors demonstrate via truncation that the N-terminus of the CesA is not involved in the interactions between the isoforms and propose that the CSR hook-like extensions are the primary mediator of trimer-trimer interactions. This argument would be strengthened by equivalent truncation experiments in which the CSR region is removed.

The statement on page 6 that "All CesA isoforms show greatest catalytic activity at neutral pH" seems to contradict the data in Figure 1e and the subsequent statements.

Reviewer #3 (Public Review):

Summary:

Cellulose is a major component of the primary cell wall of growing cells and it is made by cellulose synthases (CESAs) organized into multi-subunit complexes in the plasma membrane. Previous results have resolved the structure of secondary cell wall CESAs, which are only active in a subset of cells. Here, the authors evaluate the structure of CESAs from soybeans (Glycine max, Gm) via cryo-EM and compare these structures to secondary cell wall CESAs. First, they expressed GmCESA1, GmCESA3, or GmCESA6 in insect cells, purified these proteins as both monomers and homotrimers and demonstrated their capacity to incorporate 3H-labelled glucose into the cellulase-sensitive product in a pH and divalent cation (e.g., Mg2+) -dependant fashion (Figure 1). Although CESA1, CESA3, and a CESA6-like isoform are essential for cellulose synthesis in Arabidopsis, in this study, monomers and homotrimers both showed catalytic activity, and there was more variation between individual isoforms than between their oligomerization states (i.e., CESA3 monomers and trimers showed similar activities, which were substantially different from CESA1 monomers or trimers).

They next use cryo-EM to solve the structure of each homotrimer to ~3.0 to 3.3 A (Figure 2). They compare this with PttCESA8 and find important similarities, such as the unidentified density at a positively-charged region near Arg449, Lys452, and Arg453, and differences, such as the position and relatively low resolution (suggesting higher flexibility) of TM7, which presumably creates a large lateral lipid-exposed channel opening, rather than the transmembrane pore in PttCESA8. Like PttCESA8, an oligosaccharide in the translocation channel was co-resolved with the protein structure. Neither the N-terminal domains nor the CSRs (a plant-specific insert into the cytosolic loop between TM2 and TM3) are resolved well.

Several previous models have proposed that the cellulose synthase complexes may be composed of multiple heterotrimers, but since the authors were able to isolate beta-glucan-synthesizing homotrimers, their results challenge this model. Using the purified trimers, the authors investigated how the CESA homotrimers might assemble into higher-order complexes. They detected interactions between each pair of CESA homotrimers via pull-down assays (Figure 3), although these same interactions were also detected among monomers (Supplemental Figure 4). Neither catalytic activity nor these inter-homotrimer interactions required the N-terminal domain (Figure 4). When populations of homotrimers were mixed, they formed larger aggregations in vitro (Figure 4) and displayed increased activity, compared to the predicted additive activity of each enzyme alone (Figure 5). Intriguingly, this synergistic behavior is observed even when one trimer is chemically inactivated before mixing (Supplemental Figure 6), suggesting that the synergistic effects are due to structural interactions.

Strengths:

The main strength of this manuscript is its detailed characterization of the structure of multiple CESAs, which complements previous studies of secondary cell wall CESAs. They provide a comprehensive comparison of these new structures with previously resolved CESA structures and discuss several intriguing similarities and differences. The synergistic activity observed when different homotrimers are mixed is a particularly interesting result. These results provide fundamental in vitro support for a cellulose synthase complex comprised of a hexamer of CESA homotrimers.

Weaknesses:

There are several weaknesses in the manuscript. The authors do not present any data to indicate that GmCESA1, GmCESA3, and GmCESA6 are primary cell wall CESAs (e.g. expression patterns, phylogenetic evidence). Furthermore, their evidence that these proteins make cellulose in vitro is limited to the beta-glucanase-sensitive digestion of the product. Previous reports characterizing CESA structures have used multiple independent methods: sensitivity and resistance of the product to various enzymes, linkage analysis, and importantly, TEM of the product to ensure that it makes genuine cellulose microfibrils, rather than amorphous beta-glucan. Without demonstrating that GmCESA1, GmCESA3, and GmCESA6 are genuinely synthesizing cellulose microfibrils (via TEM) and that they are primary cell wall CESAs (via expression patterns & phylogenetic evidence), it is difficult to place the results into context. Finally, the authors indicate that they were unable to isolate heterotrimers in vitro, but they do not present any evidence of these experiments, which is essential to evaluate their conclusion that these CESAs operate as homotrimers in vitro.