Essential amino acids in the Plant-Conserved and Class-Specific Regions of cellulose synthases

Abstract

The Plant-Conserved Region (P-CR) and the Class-Specific Region (CSR) are two plant-unique sequences in the catalytic core of cellulose synthases (CESAs) for which specific functions have not been established. Here, we used site-directed mutagenesis to replace amino acids and motifs within these sequences predicted to be essential for assembly and function of CESAs. We developed an in vivo method to determine the ability of mutated CesA1 transgenes to complement an Arabidopsis (Arabidopsis thaliana) temperature-sensitive root-swelling1 (rsw1) mutant. Replacement of a Cys residue in the CSR, which blocks dimerization in vitro, rendered the AtCesA1 transgene unable to complement the rsw1 mutation. Examination of the CSR sequences from 33 diverse angiosperm species showed domains of high-sequence conservation in a class-specific manner but with variation in the degrees of disorder, indicating a nonredundant role of the CSR structures in different CESA isoform classes. The Cys residue essential for dimerization was not always located in domains of intrinsic disorder. Expression of AtCesA1 transgene constructs, in which Pro₄₁₇ and Arg₄₅₃ were substituted for Ala or Lys in the coiled-coil of the P-CR, were also unable to complement the rsw1 mutation. Despite an expected role for Arg₄₅₇ in trimerization of CESA proteins, AtCesA1 transgenes with Arg₄₅₇Ala mutations were able to fully restore the wild-type phenotype in rsw1. Our data support that Cys₆₆₂ within the CSR and Pro₄₁₇ and Arg₄₅₃ within the P-CR of Arabidopsis CESA1 are essential residues for functional synthase complex formation, but our data do not support a specific role for Arg₄₅₇ in trimerization in native CESA complexes.

An in vivo method employing mutated CesA1 transgenes to complement the Arabidopsis root-swelling1 mutant revealed amino acids essential for function in two plant-unique sequences.

Introduction

Plant cellulose microfibrils are synthesized by protein complexes arranged in hexagonal arrays called “particle rosettes” (Giddings et al., 1980; Mueller and Brown, 1980). These complexes consist of plant cellulose synthase proteins (CESAs), each capable of synthesis of a single (1→4)-β-D-glucan chain that is extruded through membrane channels, where at least 18 chains form the para-crystalline cellulose microfibril at the plasma membrane (Nixon et al., 2016). Plant CESAs are related to ancestral bacterial synthases (Nobles et al., 2001), with catalytic core domains containing four signature sequences containing D, DxD, D, and Q/RxxRW residues essential for substrate binding and catalysis (Saxena et al., 1995; Pear et al., 1996). The catalytic core is sandwiched between seven trans-membrane domains (TMDs), two upstream and five downstream, and an additional membrane-associated Interfacial3 (IF3) helix between TMDs 4 and 5 (Slabaugh et al., 2016; Purushotham et al., 2020).

Plant CESAs contain three unique domains, one forming an extended N-terminus containing a Zn-finger redox-dependent dimerization domain (Kurek et al., 2002), and two sequences located within the catalytic core, the Plant-Conserved Region (P-CR) and the Class-Specific Region (CSR), with little or no similarity to any domain of known function (Berman et al., 2000). The P-CR domain is highly conserved among all CESA isoforms of flowering plants, but the CSR domains, first thought to be hypervariable (Pear et al., 1996; Delmer, 1999), form a multi-clade sub-group structure whose structures are well-conserved across angiosperm species (Holland et al., 2000; Vergara and Carpita, 2001). In angiosperms, at least three isoforms of CESAs are expressed during primary wall formation, and a different set of three is expressed during secondary cell wall synthesis (Tanaka et al., 2003; Appenzeller et al., 2004; Brown et al., 2005). Mutations in each of these isoforms result in cellulose deficiencies, indicating that they possess nonredundant roles within a cellulose synthase complex (Taylor, 2008). The sequence variation of the CSRs defines classes of CesA genes that are expressed specifically during primary or secondary wall synthesis.

The crystal structure of the bacterial cellulose synthase (BcsA) complex from Rhodobacter sphaeroides provided a model for the catalytic domain of plant CESA sequences (Morgan et al., 2013), which is predicted to adopt the same nucleotide-binding fold as those of bacterial sequences (Olek et al., 2014). Locations for the P-CR and CSR domains were proposed from solution structures of recombinant catalytic domains (CatDs) of rice (Oryza sativa) CESA8 obtained by small-angle X-ray scattering (SAXS; Olek et al., 2014). The SAXS-derived molecular envelope showed a two-domain structure, with the smaller extended domain participating in dimerization in vitro. Homology modeling of the SAXS-derived structure placed the catalytic core in the central portion of the envelope, with the CSR filling the extended domain and the P-CR wrapping around the catalytic core domain distal to the CSR. The crystal structure and modeling of a recombinant rice OsCESA8 P-CR domain yielded a coiled-coil domain formed by anti-parallel α-helices and a large connecting loop, with stacks of aromatic residues formed between the helices in the hydrophobic core (Rushton et al., 2017). The P-CR fitted into the catalytic domain of the rice CESA8 (CatD) in a position that was consistent with the solution structure (Olek et al., 2014), with the coiled-coil aligning at the entrance to the catalytic core and the connecting loop filling a hydrophobic pocket (Rushton et al., 2017). However, the structures of full-length, detergent-solubilized, and reconstituted poplar (Populus trichocarpa) PttCESA8 (Purushotham et al., 2020), and cotton (Gossypium hirsutum) GhCESA7 (Zhang et al. 2021), determined by cryo-EM, placed the coiled-coil of the P-CR away from the entrance to the catalytic core and at the distal end of the CatD, and link CESAs into a homotrimer. The crystal structure of a portion of an Arabidopsis (Arabidopsis thaliana) CESA3-CatD also placed the P-CR consistent with a distal location (Qiao et al., 2021).

In solution, recombinant CatDs of rice CESA8 dimerize in a redox-dependent manner through the CSR domains and not through the P-CR (Olek et al., 2014), and the lack of oligomerization of recombinant P-CR proteins in solution (Rushton et al., 2017) is inconsistent with the behavior of reconstituted homotrimer structures derived after detergent solubilization. Here, we designed an in vivo approach to resolve these incongruities by determining the impact of substitution of amino acids that might be essential in the P-CR and CSR for assembly of the cellulose synthase complex. We used site-directed mutagenesis of an Arabidopsis cv. Col-0 AtCesA1 to substitute amino acids predicted to function in the CSR and P-CR, and then determined whether the mutated CESA1 could restore the ability of the temperature-sensitive root-swelling1 (rsw1) mutant (Arioli et al., 1998) to grow and synthesize cellulose at restrictive temperatures. From the rice P-CR crystal structure, we predicted several amino acids that might alter the structure of the coiled-coil and connecting loop in Arabidopsis, including the cognate Arg residue proposed to function in trimerization in the poplar CESAs. However, substitution of many other amino acids predicted to be essential in the CSR and P-CR, including the conversion of R₄₅₇ to Ala, the equivalent residue responsible for trimer formation in reconstituted proteins, were able to fully or partially complement the rsw1 mutation in plants grown at the restrictive temperature. Further examination of the 3D structures of the published PttCESA8 and GhCESA7 monomer and trimer revealed other Arg residues in the C-terminus that might function in oligomerization. Our results identify amino acids critical for cellulose synthesis and bring insights to establish the functions of these plant-unique sequences in vivo.

Results

Amino acids of the CSR required for dimerization in vitro

Rice CESA8 contains a 125-amino acid CSR domain between catalytic DxD- and D-motifs 2 and 3 of the catalytic core (Figure 1). The CSR of OsCESA8 is characterized by a single Cys residue and four motifs of three more consecutive Arg or Lys residues (basic motifs) towards the N-terminus of the sequence, and two acid-rich motifs towards the C-terminus (Supplemental Figure S1). Affinity-purified recombinant CESA-CatD behaved as a monomer in 10 mM dithiothreitol (DTT), and dilution of the DTT and concentration of the protein by filtration resulted in formation of a dimer (Supplemental Figure S2). We used site-directed mutagenesis to substitute the C₆₆₂ motif, the four basic motifs, and the first acidic motif with Ser or Ala residues (Supplemental Figure S2A). The mutations did not affect recombinant protein production or stability of the monomer in solution. When the monomer preparations of each were depleted of DTT and concentrated by centrifugal filtration, all recombinant proteins except those with mutated C₆₆₂ residues formed dimers in vitro (Supplemental Figure S2B). Double, triple, or quadruple mutations to eliminate all four basic motifs had no substantial effect on the ability of the recombinant proteins to dimerize. Only in the quintuple mutant, in which all four basic amino acid motifs and the first acidic motif were mutated, was dimerization substantially reduced (Supplemental Figure S2B).

Figure 1.

Comparison of protein models of the bacterial and plant cellulose synthases. All cellulose synthases (CESAs) contain up to eight TMDs (blue boxes) but differ in the position of the CatD. Motifs containing D, DxD, D, and QxxRW amino acids essential for substrate binding and catalysis are indicated. Plant CESAs contain three sequences not found in bacterial synthases and are missing the PilZ regulator domain found in the BscA of R. sphaeroides and CelA of Gluconacetobacter xylinus. The extended N-terminus upstream of the first TMD contains a Zinc-Finger domain, and within the catalytic core sequences are the P-CR and CSR. Of the eight originally predicted TMDs, TMD5 (pink) was found to occupy an interfacial orientation along the membrane, placing both flanking sequences facing the cytoplasm, leaving seven true TMDs (Purushotham et al., 2020; Zhang et al., 2021).

Although these data indicate a specific role of C₆₆₂ in a redox-dependent dimerization of CatDs in vitro, functions of the C₆₆₂ and the basic and acidic motifs characteristic of the CSR in vivo cannot be inferred. For this reason, we designed an in vivo approach employing the temperature-sensitive Arabidopsis root-swelling1 (rsw1; Atcesa1) as a testbed to determine the effect of directed mutations to these motifs, when mutated AtCesA1 transgenes were expressed in the rsw1 background.

Amino acids of the CSR required for cellulose synthesis in vivo

The rsw1 mutant displayed normal root tip formation but slightly reduced root growth compared to Col-0 at the permissive temperature of 22°C. However, an abrupt decrease in root growth rate and a swelling of the tip occurred upon transfer to 31°C, which also resulted in the temperature-sensitive cellulose deficiency (Figure 2, A–C). Root-tip swelling was first evident within 8 h after transfer of plantlets to the restrictive temperature, but we recorded growth phenotypes after 24 h at the restrictive temperature when the phenotypes were more evident. Transformation of the rsw1 mutant with a plasmid containing an AtCesA1 cDNA complemented wild-type root length and normal tip formation at the restrictive temperature (Figure 2, A and B).

Figure 2.

Root growth and cellulose synthesis in plants expressing mutated AtCesA1 transgenes in a rsw1 background. A, Sequence of a portion of the CSR of AtCESA1 with Cys (yellow), basic motifs (blue), or acidic motif (red) indicated and mutated to Ala or Ser residues as shown in light blue. Micrographs of roots grown at 22°C or 31°C for the corresponding wild type (Col-0), rsw1, transgenic rsw1 mutants, and rsw1 complemented with AtCesA1 (rsw1-C). Scale bar = 100 µm for all micrographs. B, Root growth at permissive and restrictive temperatures in plants transformed with mutated AtCesA1 constructs compared to wild type (Col-0), rsw1, and rsw1 lines complemented with AtCesA1. Four-day-old seedlings grown at 22°C were transferred to 31°C and root growth was measured for up to 24 h (red lines) and compared to that of seedlings kept at 22°C (blue lines). Values are the mean ± SD of four independent experiments with 20–25 seedlings each. Letters for 24 h values indicate groups of similarity by t test at P ≤ 0.05. C, The proportion of ¹⁴C-glucose incorporated into acetic–nitric insoluble cellulose in plants transformed with mutated AtCesA1 constructs compared to wild type, rsw1, and rsw1 lines complemented with AtCesA1. Five-day-old seedlings grown at 22°C were transferred to 31°C for 6 h before addition of labeled glucose, and incorporation into cellulose was determined after incubations for 18 h compared to that of seedlings kept at 22°C. Box-whisker plots show the box limits of 1 SD, a black horizontal line for the mean, and a vertical center line for the extreme values of three independent experiments with 20–25 seedlings each. Letters indicate groups of similarity by t test at P ≤ 0.05.

To determine the ability to synthesize cellulose, 5-d-old plantlets were transferred to 31°C for 6 h while controls remained at 22°C. ¹⁴C-Glucose in water was added to the agar medium surrounding the seedlings, and incubations were continued for 18 h. Plantlets (shoots and roots) were flash-frozen in liquid nitrogen, and cell walls were isolated using an SDS-Tris buffer extraction protocol (Mertz et al., 2012). Relative synthesis of cellulose was determined by the proportion of radiolabel in crystalline cellulose remaining after acetic–nitric digestion of noncellulosic polysaccharides (Updegraff, 1969). For wild-type Arabidopsis, about 25% of the total radiolabel in cell walls was found in cellulose, whereas in rsw1 the amount of cellulose synthesized was less than one-half of this proportion at the restrictive temperature (Figure 2, A and C). Complementation of the rsw1 mutation with a wild-type AtCesA1 restored cellulose synthesis to wild-type levels.

As a second assay of complementation, we also found that the rsw1 mutant is unable to elongate a floral stem at 31°C. Plants grown for 5 weeks in a growth room at 22°C were transferred to 31°C for up to 6 d, with a duplicate set kept at 22°C. Prior to transfer, flowering in both wild type and rsw1 was initiated at 22°C. Arabidopsis is sensitive to heat stress at the restrictive temperature, but floral stem development was hastened in wild type, and complementation of rsw1 with a wild-type AtCesA1 restored the ability to form a floral stem (Supplemental Figure S3).

Compared to OsCESA8, the Arabidopsis CESA1 contains two upstream Cys residues, two basic motifs, and two acidic motifs (Supplemental Figures S1 and S2). We next substituted various amino acids through site-directed mutagenesis of AtCesA1 and used Agrobacterium tumefaciens-mediated transformation of Arabidopsis rsw1 mutants to determine the ability of the mutated AtCesA1 to complement the rsw1 mutant phenotypes. Up to twelve homozygous lines were selected for the AtCesA1 transgene and each of the constructs encoding mutated sequence variants. Relative expression of the mutated CesA1 transgenes was variable, and those with diminished expression were less able to restore the wild-type phenotype at the restrictive temperature. For this reason, we identified at least two lines with moderate to strong expression of each mutated AtCesA1 transgene (Supplemental Figure S4). Each of the lines was assayed for root length and swelling phenotype under permissive and restrictive temperatures, and a representative line for each mutant was selected for determination of cellulose synthesis (Supplemental Figure S5).

The C₆₆₂A variant of AtCesA1 failed to complement the rsw1 mutant, exhibiting the severe swelling phenotype and inhibition of growth at the restrictive temperature, whereas elimination of either of the two basic motifs in AtCESA1 or the acidic motif resulted in full complementation (Figure 2, A and B). Consistent with these observations, rsw1 plants expressing the C₆₆₂A variant were unable to form a floral stem, whereas modifications to the AtCesA1 transgene that eliminated either of the basic motifs or the acidic motif were still able to form floral stems (Supplemental Figure S3).

The proportion of ¹⁴C-D-glucose incorporated into acid-insoluble cellulose in the C₆₆₂A variant was consistent with the growth behaviors and the root-swelling phenotypes compared to the rsw1 mutant, with poor accumulation of label into cellulose (Figure 2C). By contrast, both basic-motif and acidic-motif mutants exhibited cellulose synthesis rates that achieved those of wild type at either permissive or restrictive temperatures.

In Arabidopsis, the titers of AtCESA1, AtCESA3, and AtCESA6 are expected to be in a 1:1:1 ratio (Hill et al., 2014). To assess the ability of the mutated AtCesA1 constructs to produce CESA1 proteins in the rsw1 background and to determine if the mutated protein altered the abundance of the three primary wall isoforms, we performed proteomics analysis on the relative titers of AtCESA1, AtCESA3 and AtCESA6 proteins in the Col-0, rsw1, and AtCesA1:rsw1 backgrounds in isolated Golgi membranes from developing rosette leaves. From spectral counts of fragments unique to each CESA isoform, ratios of AtCESA1:AtCESA3 were between 0.7 and 1.2 for all genotypes grown at 22°C, with or without a 6-h treatment at 31°C (Table 1). By comparison, ratios of AtCESA6:AtCESA3 were 0.5–0.7 at both the permissive and restrictive temperatures. Overall, wild type and fully complemented mutants gave average ratios of about 1:1:0.6 (Table 1). These ratios are consistent with expression ratios of 1.1:1.0:0.7 for these primary wall CesAs in Arabidopsis seedlings (Sowinski et al., 2022).

Table 1.

Determination of CesA protein abundance in Golgi membranes of Col-0, rsw1, and ΔAtCesA1:rsw1 lines, at permissive or restrictive temperatures

Genotype, treatment	CESA1	CESA3	CESA6	CESA1/CESA3	CESA6/CESA3
	Normalized spectral counts			Ratio
Col-0, 22°C	16.4 ± 0.8	13.5 ± 2.1	10.1 ± 1.3	1.2 ± 0.2	0.7 ± 0.2
Col-0, 31°C	15.6 ± 0.2	15.8 ± 1.4	8.6 ± 1.5	1.0 ± 0.1	0.5 ± 0.1
rsw1, 22°C	15.1 ± 0.4	16.2 ± 0.6	8.7 ± 0.2	0.9 ± 0.1	0.5 ± 0.1
rsw1, 31°C	13.6 ± 2.4	15.5 ± 2.1	10.9 ± 0.3	0.9 ± 0.3	0.7 ± 0.1
rsw1-COMP, 22°C	14.3 ± 0.1	16.4 ± 0.6	9.3 ± 0.7	0.9 ± 0.1	0.6 ± 0.1
rsw1-COMP, 31°C	12.6 ± 3.5	18.6 ± 3.6	8.8 ± 0.0	0.7 ± 0.3	0.5 ± 0.1
CSR mutants
C662A, 22°C	6.9 ± 0.2	19.9 ± 1.4	13.2 ± 1.4	0.3 ± 0.1	0.7 ± 0.1
C662A, 31°C	0.7 ± 0.7	20.3 ± 0.3	19.0 ± 1.0	0.1 ± 0.1	0.9 ± 0.1
R666S/K667A/K668S, 22°C	10.9 ± 0.4	21.8 ± 0.3	7.3 ± 0.4	0.5 ± 0.0	0.3 ± 0.1
R666S/K667A/K668S, 31°C	9.0 ± 3.3	16.7 ± 0.4	14.3 ± 2.9	0.5 ± 0.2	0.9 ± 0.2
E696A/D697S-x-D699S/E700A, 22°C	12.1 ± 0.3	20.5 ± 1.0	7.4 ± 1.3	0.6 ± 0.0	0.4 ± 0.1
E696A/D697S-x-D699S/E700A, 31°C	10.3 ± 2.3	21.3 ± 2.7	8.4 ± 0.4	0.5 ± 0.2	0.4 ± 0.1
P-CR mutants
R457A, 22°C	13.3 ± 1.2	17.6 ± 3.1	9.1 ± 1.8	0.8 ± 0.2	0.5 ± 0.2
R457A, 31°C	14.9 ± 3.6	16.5 ± 2.2	8.6 ± 1.4	1.0 ± 0.3	0.5 ± 0.1
F431A, 22°C	11.7 ± 1.7	21.3 ± 2.0	7.0 ± 0.4	0.6 ± 0.2	0.3 ± 0.1
F431A, 31°C	13.0 ± 2.4	18.2 ± 0.3	8.9 ± 2.7	0.7 ± 0.1	0.5 ± 0.2
L812R, 22°C	13.4 ± 0.5	17.1 ± 0.2	9.5 ± 0.3	0.8 ± 0.0	0.6 ± 0.1
L812R, 31°C	11.3 ± 1.2	14.4 ± 0.6	14.3 ± 1.8	0.8 ± 0.1	1.0 ± 0.2
D438A-Y439A, 22°C	9.9 ± 2.4	20.5 ± 2.0	9.6 ± 0.4	0.5 ± 0.2	0.5 ± 0.1
D438A/Y439A, 31°C	12.2 ± 0.5	17.5 ± 0.8	10.4 ± 0.3	0.7 ± 0.1	0.6 ± 0.1
R453K, 22°C	10.0 ± 2.0	16.0 ± 0.0	14.0 ± 2.0	0.6 ± 0.2	0.9 ± 0.1
R453K, 31°C	10.3 ± 1.1	20.1 ± 1.4	9.6 ± 0.4	0.5 ± 0.1	0.5 ± 0.1
P417A, 22°C	10.0 ± 0.0	16.2 ± 0.5	13.8 ± 0.5	0.6 ± 0.0	0.9 ± 0.1
P417A, 31°C	6.3 ± 1.3	20.5 ± 0.5	13.7 ± 2.3	0.3 ± 0.1	0.7 ± 0.2

Ratios of CESA1/CESA3 statistically significant by t test at P = 0.05 are in bold. Values ± variance of normalized spectral counts of diagnostic fragments of two independent Golgi isolations.

Fragments of AtCESA1 that would contain the A₅₄₉V mutation in rsw1 were not recovered, so the relative abundance of CESA1 represents the sum of that in the rsw1 background plus the contribution of the AtCesA1 transgene. The mutated transgenes containing the substitutions R₆₆₆A/K₆₆₇S/K₆₆₈A or E₆₉₆A/D₆₉₇S-x-D₆₉₉S/E₇₀₀A, which fully complemented the rsw1 phenotype (Figure 2), yielded AtCESA1:AtCESA3 ratios of 0.5 and 0.6, respectively, at both permissive and restrictive temperatures (Table 1). By contrast, rsw1 plants expressing the AtCesA1 transgene with a mutated C₆₆₂ residue resulted in ratios of only 0.3 in plants grown at 22°C; upon transfer to 31°C for 6 h, fragments associated with AtCESA1 were barely detectable (Table 1).

Intrinsic disorder and conservation of sequence in the CSR

Because dimerization of the CSR domain is implicated in CESA complex assembly, we explored further the structural features of the domain. The CSRs in moss CESAs are characterized as containing Intrinsically Disordered Regions (IDRs), a feature consistent with a function in protein coupling (Scavuzzo-Duggan et al., 2018). A common Prediction of Naturally Disordered Regions (PONDR) algorithm is the VL-XT, which combines several predictors based on neural-network analyses trained on missing electron densities from X-ray crystallography and disorder from NMR studies (Romero et al., 1997; Li et al., 1999). Long regions of 30 or more residues with PONDR scores of 0.8 or greater are strong predictors of disorder, but short regions of order within long regions of disorder indicate potential binding regions (Garner et al., 1999). Aliphatic residues and enrichment of the basic residues, Lys and Arg, and acidic residues, Asp and Glu, observed in all CSR sequences are strong predictors of disorder. PONDR scores for the CSR of rice CESA8 have two long stretches of sequence with greater than 0.8 scores flanking a short sequence of order between them (Figure 3). This region of predicted order was not associated with the Cys residue 21 (C₆₆₂) implicated in dimerization. Substitution of all four basic amino acid motifs with Ala or Ser residues reduced the PONDR score below the 0.8 threshold (Figure 3), but had no impact on dimerization (Supplemental Figure S2). Mutation of the acidic motif in addition to the four basic motifs increased order substantially in the sequences flanking the motif (Figure 3).

Figure 3.

Substitution of the four basic motifs and the acidic motif of the OsCesA8 CSR reduces the PONDR score. PONDR scores of OsCesA8 were calculated for a sliding window of nine residues across the sequence using the VL-XT method (http://www.pondr.com/). The black box shows the relative position of the Cys₆₆₂ residue. The top panel shows the PONDR scores for the CSR of OsCESA8. The middle panel shows the scores for OsCESA8 with all four basic (K) motifs mutated to substitute these residues with Ser or Ala residues (boxed in the sequence). The lower panel shows scores for OsCESA8 with the acidic motif mutated along with the basic motifs.

Multiple alignment analysis of homologs closest in sequence to primary wall AtCESA1 and AtCESA3 across 33 diverse angiosperm species showed remarkable degrees of identity/similarity within the CSR class, but much greater diversity in AtCESA6 sequences (Supplemental Figure S6, A-C), illustrated using sequence logos (Supplemental Figure S7). The PONDR score profiles of CSRs varied markedly among the three primary wall AtCESA isoforms. Prediction scores of disorder greater than 0.8 were associated with highly divergent sequences enriched in basic- and acid-rich motifs (Figure 4). However, locations of the cognate Cys residues of C₆₆₂ were independent of disorder prediction, where it resided in sequence of predicted low disorder in AtCESA1, high disorder in AtCESA3, but in an ordered sequence between two sequences of high disorder in AtCESA6.

Figure 4.

The patterns of predicted intrinsic disorder vary within CSRs in a class-specific manner. PONDR scores of the CSRs of AtCESA1, AtCESA3, and AtCESA6 were calculated for a sliding window of nine residues across the sequence using the VL-XT method (http://www.pondr.com/). The black box shows the relative position of the common first Cys residue, predicted to function in dimerization.

Secondary wall AtCESA isoforms were equally diverse among the 33 angiosperm species, with the strongest conservation among homologs closest in sequence to AtCESA8, slightly less so with AtCESA7 (Supplemental Figures S8, A–C; Supplemental Figure S9), with weakest conservation of sequence observed among homologs of AtCESA4. Only CESA4 exhibited highly predicted high disorder associated with the basic and acidic motifs (Supplemental Figure S10). A short sequence of disorder is predicted for AtCESA7, but AtCESA8 exhibits high disorder only in the first 20 amino acids of the N-terminus (Supplemental Figure S10). The cognate Cys residues of C₆₆₂ were in regions of predicted moderate disorder for both AtCESA7 and AtCESA8, but in domains of predicted high order in AtCESA4 (Supplemental Figure S10). Taken together, our data indicate that Cys residues predicted to be essential for dimerization are independent of domains of intrinsic disorder. Further, mutations of the basic motifs associated with domains of intrinsic disorder do not impact the ability of AtCESA1 proteins to form a cellulose synthase complex.

Amino acids of the P-CR required for cellulose synthesis in vivo

With 86% sequence identity and 96% similarity (Supplemental Figure S1), the crystal structure of the OsCESA8 P-CR was expected to closely match that of AtCESA1. Of several structure prediction models, RaptorX (Kallberg et al., 2012) produced the lowest RMSD of 1.09Å between the coiled-coils of P-CRs of OsCESA8 and AtCESA1 (Supplemental Figure S11). A homology model of the AtCESA1 P-CR preserves the OsCESA8 three-dimensional structure, but insertions and deletions (indels) upstream of the P-CR in both sequences results in the AtCESA1 sequence being offset by one residue in its alignment with the OsCESA8 sequence (Supplemental Figure S1). Mutations were introduced into the AtCESA1 P-CR of P₄₁₇A in HX1, of E₄₃₀A, F₄₃₁A, and a double mutant D₄₃₈A/Y₄₃₉A in the connecting loop, and of R₄₅₂A, R₄₅₇A, and V₄₇₀A of HX2. Another mutation was introduced into the IF2 helix of (L₈₁₂R) in the catalytic core near the potential interaction site with the P-CR F₄₃₁ (W₄₃₂ in OsCESA8). An additional mutation in R₄₅₃K was made to recreate the identical mutation found in the fragile fiber6 (fra6; Atcesa8) mutant (Zhong et al., 2003). Rationales for selection of mutations in the P-CR are summarized in Supplemental Table S1.

At least two and typically three lines with substantial expression of each transgene were selected for evaluation of root tip phenotype and root length (Supplemental Figure S12), and assayed for root growth and swelling phenotype under permissive and restrictive temperatures (Supplemental Figure S13). A representative line for each mutant was selected for determination of cellulose synthesis. Substitution of Ala for R₄₅₇ or V₄₇₀ of HX2 was able to fully complement the rsw1 mutation with respect to cellulose synthesis and root tip swelling but with slight reductions in root length at permissive and restrictive temperatures (Figure 5, B–D). Substitution of Ala for R₄₅₂ in HX2, or for E₄₃₀ or F₄₃₁ in the connecting loop, resulted in some loss of root growth at permissive and restrictive temperatures, but these mutants lacked the swelling of the root tip. Similarly, mutation of L₈₁₂ of the catalytic core at the predicted site of interactions with F₄₃₁ caused some loss of growth at permissive and restrictive temperatures, but root-tip morphology was normal (Figure 5, B and C). Substitution of P₄₁₇ with Ala in HX1 or R₄₅₃ with Lys in HX2 resulted in failure to complement the rsw1 mutation with respect to both root growth and root-tip morphology. The double exchange of Ala for D₄₃₈ and Y₄₃₉ resulted in failure to fully complement root-tip morphology and showed substantial loss of root growth at the restrictive temperature (Figure 5, B and C).

Figure 5.

Root growth and cellulose synthesis in plants expressing mutated AtCesA1 transgenes. A, Best homology model of the AtCESA1 P-CR domain using RaptorX (Kallberg et al., 2012) showing the Atcesa1 mutants (yellow, stick models) constructed for the A. thaliana rsw1-1 complementation experiments. Rationale for mutant construction is shown in Supplemental Table S1. Sequence of the P-CR portion of AtCESA1 with amino acids mutated to Ala, Ser, Lys, or Arg (blue) as indicated. B, Micrographs of roots grown at 22°C or 31°C for corresponding wild type (Col-0), rsw1, and transgenic rsw1 mutants. Col-0 and rsw1 images iterated from Figure 2 in samples run in parallel. Scale bar = 100 µm for all micrographs. C, Root growth at permissive and restrictive temperatures in plants transformed with mutated AtCesA1 constructs compared to wild type (Col-0) and rsw1. Four-day-old seedlings grown at 22°C were transferred to 31°C, and root growth was measured for up to 24 h and compared to that of seedlings kept at 22°C. Values are the mean ± SD combined from four independent experiments with 20–25 seedlings each. Letters for 24 h values indicate groups of similarity by t test at P ≤ 0.05. D, Proportion of ¹⁴C-glucose incorporated into acetic–nitric insoluble cellulose in plants transformed with mutated AtCesA1 constructs compared to wild type (Col-0) and rsw1. Col-0 and rsw1 data were iterated from Figure 2 in samples run in parallel. Five-day-old seedlings grown at 22°C were transferred to 31°C for 6 h before addition of labeled glucose, and incorporation into cellulose was determined after incubations for 18 h compared to that of seedlings kept at 22°C. Box-whisker plots show the box limits of 1 SD, a black horizontal line for the mean, and a vertical center line for the extreme values of three independent experiments with 20–25 seedlings each. Letters indicate groups of similarity by t test at P ≤ 0.05.

The proportion of label in cellulose during synthesis under permissive and restrictive temperatures was generally proportional to root length rather than strictly with root-tip swelling morphology. Substitution of Ala for E₄₃₀ or F₄₃₁ resulted in some loss in the capacity for cellulose synthesis without root-tip swelling. Mutation of L₈₁₂, P₄₁₇, R₄₅₃, or D₃₂₈/Y₄₃₉ resulted in failure to fully complement the cellulose-deficient phenotype (Figure 5D). The ability to form a floral stem was generally associated with the root growth, root-tip morphology, and capacity for cellulose synthesis in the wild type vs. rsw1 grown under restrictive conditions (Supplemental Figure S14).

Only P-CR mutations in R₄₅₃ and P₄₁₇ were unable to even partially complement the rsw1 mutation (Figure 5). AtCESA1:AtCESA3 ratios in P₄₁₇ and R₄₅₃ plants were both 0.6, when grown at 22°C; upon transfer to 31°C, the ratio in R₄₅₃ plants was relatively unchanged, but that in P₄₁₇ plants fell to 0.3. Mutations in F₄₃₁, L₈₁₂, and D₄₃₈/Y₄₃₉ in the AtCesA1 transgenes only partly restored the Col-0 phenotype even though they exhibited AtCESA1:AtCESA3 ratios between 0.5 and 0.8 at the permissive or restrictive temperatures (Table 1). Mutation of R₄₅₇ in the transgene, which was able to fully restore the wild-type phenotype (Figure 5), exhibited AtCESA1:AtCESA3 ratios of 0.8 to 1.0 at the permissive or restrictive temperatures (Table 1).

Reconstruction of the CESA-CatD and oligomerization models

Reconstruction of models of the full-length PttCESA8 (PDB: 6WLB; Purushotham et al., 2020) and AtCESA3-CatD (PDB: 7CK1 Qiao et al., 2021) CatDs placed the P-CR coiled-coil at the distal end of the CatD (Figure 6A). We had predicted that the W₄₃₂ residue (F₄₃₁ in AtCESA1) in the connecting loop of OsCESA8 interacted with an L₈₁₂ in the catalytic core (Rushton et al., 2017), but this interaction was not observed in either PttCESA8 or AtCESA3. The CSR in the recombinant AtCESA3-CatD was too disordered to obtain structural information (Qiao et al., 2021), but the CSR observed in the full-length PttCESA8 structure formed distended α-helices at the opposite end of the CatD (Figure 6A). A major distinction between the crystal structure of AtCESA3-CatD, and the CatD within the full-length PttCESA8, is the position of the α-helix containing the QxxRW motif, which aligns along the plasma membrane in the full-length protein but is rotated nearly 90° upward in the absence of the TMDs in the AtCESA3-CatD structure (Figure 6A). A 21-amino acid cytosolic sequence between the IF3 helix and TMD5 containing the FxVTxK motif had insufficient densities to be modeled for the PttCESA8 (Purushotham et al., 2020). However, when this sequence was modeled within the PttCESA8 sequence by AlphaFold (Jumper et al., 2021), the FxVTxK-containing sequence was predicted to extend downward over the entrance to the active site (Figure 6B).

Figure 6.

Reconstruction of the AtCESA3-CatD, PttCESA8 monomer, and PttCESA8 trimer structures. A, Comparison of the structures of AtCESA3-CatD (Qiao et al. 2021) with the CatD within the full-length PttCesA8 (Purushotham et al., 2020). The P-CR is in blue, and the CSR (partial) is in green. Catalytic motifs of DxD and QxxRW are in orange. For the PttCESA8, amino acids flanking the FxVTxK motif are in red; R₃₅₆ and R₉₃₄ residues of the PttCESA8 that potentially participate in trimerization are in magenta. B, The location of a 21-amino acid sequence containing the FxVTxK motif, missing from the PttCESA8 structure, was predicted using AlphaFold (Jumper et al., 2021) and is colored yellow, with the FxVTxK motif in red. C, Top view of the proposed trimer structure stabilized by Arg₃₅₆, compared to the bottom view of the C-terminus loops with Arg₉₃₄ residues shown as alternative points of potential trimerization. Monomer structures were reconstructed from the published coordinates (PDB: 6WLB; Purushotham et al., 2020; PDB: 7CK3; Qiao et al., 2021).

The R₃₅₆ in the P-CR coiled-coil HX2 of PttCESA8 forms an identical interaction with a phosphate ion to produce the trimer structure defined by cryo-EM (Purushotham et al., 2020), which was predicted to serve as a linchpin at the distal end of the CatD (Figure 6C). Mutation of the AtCesA1 transgene to convert the cognate R₄₅₇ to Ala resulted in full complementation of the wild-type phenotype and cellulose synthesis in the rsw1 mutant (Figure 5, A–D), indicating that the cognate residue R₃₅₆ is not essential for trimerization of the PttCESA8. When the structures were reconstructed using the published coordinates of the PttCESA8 trimer, we observed an R₉₃₄ (R₁₀₃₈ in AtCESA1) in the C-terminus domain in close proximity within the trimer structure that might be an alternative coupling site adjacent to the plasma membrane (Figure 6, A and C).

Discussion

The nature of the cellulose inhibition in the rsw1 mutant and its complementation

Although the three-dimensional structure of plant CESAs is becoming resolved, how different isoforms of CESA assemble into complexes is still not known. Mutations that alter this higher-order assembly could possibly impact the formation of the cellulose microfibril, and not specifically the catalytic activity to produce a single glucan chain. In the rsw1 testbed, our complementation assay only considered recovery of the ability to synthesize acid-resistant crystalline cellulose in the mutant. In the original report of the rsw1 mutation (Arioli et al., 1998), the authors indicated that a noncrystalline glucan (amorphous cellulose) forms instead of the crystalline form. Similar soluble glucans were observed in plantlets grown in the presence of the cellulose synthesis inhibitor CGA 325’615 (Peng et al., 2001). However, the soluble glucans induced by the inhibitor precipitated upon protease treatment, indicating that short crystalline chains might have been kept in solution by the CESA proteins themselves. Other mutations in CESAs and responses to other cellulose synthesis inhibitors also resulted in loss of cellulose crystallinity, as detected by X-ray diffraction (Harris et al., 2012; Fujita et al., 2013). However, it is not known how loss of an X-ray diffraction signal might correspond to reductions in resistance of these altered microfibrils to hydrolysis in acetic–nitric acid.

Loss of protein titer observed with the C₆₆₂A and P₄₁₇A mutants would most certainly cause structural abnormalities in formation of a robust cellulose synthase complex. The association of R₄₅₃ with R₄₅₇ to disrupt possible oligomerization (Rushton et al., 2017), disruption of the complex by the R₄₅₃K mutation in AtCesA1 and the equivalent mutation fra6 in AtCesA7 (Zhong et al., 2003) would also be expected to cause defects in assembly. By contrast, mutations L₈₁₂R and D₄₃₈A/Y₄₃₉A, which only partly complement the rsw1 mutation yet maintain CESA protein titer, might be deficient for other reasons, such as defects in microfibril formation that result from improper CESA complex formation, or unrelated defects in regulation of catalytic activity or in trafficking of the complexes to the plasma membrane unrelated to protein structure.

Potential roles for the CSR in CESA structure and interaction in CESA complexes

Based on the modeling of the conserved catalytic core into the SAXS-derived structure, the predicted redox-dependent dimerization of the recombinant OsCESA8-CatD occurs through the CSR (Olek et al., 2014). Although the P-CR is implicated in the trimerization of the recombinant PttCESA8, two roles for the CSR in dimerization of the CESA trimer units into the CESA complex have been proposed (Purushotham et al., 2020). Of the characteristic motifs of the CSR, only the mutation C₆₆₂A in the AtCesA1 transgene failed to complement the rsw1 mutant phenotype (Figure 2). Whereas rice OsCESA8 has a single cognate Cys residue, Arabidopsis AtCESA1 has two. The failure of mutated C₆₆₂ to complement the rsw1 phenotype indicates that C₆₆₃ cannot substitute for the first Cys residue in AtCESA1. Transgenes of all mutations were moderately to highly expressed in the transformed rsw1 backgrounds, but proteomics analysis of enriched Golgi fractions, the site of assembly of cellulose synthase complexes, demonstrated a reduced titer of AtCESA1 protein relative to AtCESA3. Unlike any other mutation, mutation of C₆₆₂ caused a substantial reduction in the AtCESA1:AtCESA3 ratio at the permissive temperature, and upon transfer to 31°C, the AtCESA1 was almost undetectable (Table 1). These data indicate that the mutation causes an interference of assembly and targeted removal of AtCESA1 specifically, not only of the transgene products, but of the endogenous rsw1 product as well. One possible explanation could be if AtCESA1 forms a homodimer, where the mutated form recognized for removal would result in loss of the dimer.

Strong sequence conservation was found within the CSR among homologs of AtCESA1, AtCESA3, and AtCESA6 for primary wall expressed isoforms, and AtCESA4 and AtCESA8 of secondary wall isoforms, with more limited conservation among homologs of AtCESA7 (Carroll and Specht, 2011). Consistent with our analyses of 33 angiosperm species, the CSR domains of CESAs of moss (Physcomitrium patens) were classified as IDRs (Scavuzzo-Duggan et al., 2018). IDRs are defined as entire proteins or regions of proteins that lack a fixed tertiary structure of the backbone, yielding extended structures with high degrees of flexibility (Dunker et al., 2002). Among the many functions ascribed for IDRs are protein–protein interactions, where the IDR might undergo disorder-to-order transition upon binding to their protein partners (Dunker et al., 2001, Dyson and Wright, 2002). Defining the CSRs of CESAs as IDRs implicates them in assembly of the CESA proteins into the synthase complex (Scavuzzo-Duggan et al., 2018). However, the IDRs do not extend over the entire CSR. We also found that predicted intrinsic disorder in the N-terminal sequences of Arabidopsis and rice CSRs is associated with the basic motif-rich regions, but the high-sequence conservation across the angiosperm classes of CESAs in regions of predicted high order indicates that other structural features downstream are also important (Figures 3 and 4, Supplemental Figures S7 and S9). From binding studies in vitro with the recombinant OsCESA8-CatD, only the Cys residue is essential for binding, and elimination of all four basic motifs reduced the predicted intrinsic disorder yet had little impact on dimerization (Figure S2). Mutation of the Cys residue in an AtCesA1 construct failed to restore the wild-type phenotype of the rsw1 mutant, but elimination of either basic motif had no effect on the ability to restore wild-type growth and cellulose synthesis (Figure 2).

How might the diverse isoforms of the CSR optimize CESA assembly of the multiple isoforms into the cellulose synthase complex? Moss CESAs are classified into two noninterchangeable groups by size and structure (Roberts and Bushoven, 2007). A knock-out of a member of clade A, PpCesA5, is unable to form gametophores, but constructs of a PpCesA5 modified to contain the CSR sequences from clade B genes were able to restore gametophore formation, indicating that CSRs were not involved in specific isoform interactions (Scavuzzo-Duggan et al., 2018). Similar reciprocal domain-swap experiments with several domains of Arabidopsis secondary wall AtCESA4, AtCESA6, and AtCESA7 were tested for their ability to rescue floral stem growth and cellulose synthesis when expressed in their respective T-DNA insertional null mutants (Kumar et al., 2017). AtCESA7 is a better universal donor, restoring stem development, and cellulose synthesis when several of its domains replaced those of AtCESA4 or AtCESA8. In particular, swaps of the CSR differed in their ability to restore phenotype. Insertion of the CSR from AtCESA4 into AtCESA7 restored about 50% of the cellulose content, but the CSR from AtCESA8 was a poor donor (Kumar et al., 2017). In reciprocal swaps, CSR from AtCESA7 was able to restore about 60% of cellulose synthesis in AtCESA4 but only about 25% in AtCESA8.

Although IDRs are inferred to constitute the hypervariable domains of the CSR and are interchangeable (Scavuzzo-Duggan et al., 2018), the IDRs of the CSR are not only confined to poorly conserved domains enriched in basic motifs but also extend into the highly conserved acid-rich, class-specific C-terminal domains (Figure 4, Supplemental Figures S7 and S9). Other variable domains in the N-terminus and C-terminus contributed greatly to class specificity, but strong conservation of sequence was observed across most of the angiosperm CSR sequences. Based on molecular modeling of the CatDs of the major isoforms of the moss CESAs, the impact of the IDRs on flexibility was greatly dependent on isoform (Scavuzzo-Duggan et al., 2018). If the IDRs of the CSR represent domains of protein–protein interaction in a lock-and-key mechanism, then the more ordered CSRs might be analogous to the key, and the more flexible CSRs might be the tumblers of the lock. Consistent with this hypothesis, the AtCESA6 isoform for primary wall cellulose synthesis, and the AtCESA4 isoform secondary wall cellulose synthesis, exhibit much greater species diversity in protein structure compared to the other isoforms that could lead to more flexibility in a species-dependent manner (Supplemental Figures S7 and S9).

Potential roles for the P-CR in CESA function

A completely different picture of the interaction domains is presented in modeling detergent extracted and reconstituted Arabidopsis AtCESA1-CatDs (Vandavasi et al., 2016) or full-length recombinant poplar PttCESA8 (Purushotham et al., 2020) and cotton GhCESA7 (Zhang et al., 2021) homotrimers. Reconstituted AtCESA1-CatD monomers gave no discernable SAXS-derived structures but spontaneously formed trimers did (Vandavasi et al., 2016). By contrast, the native recombinant OsCESA8-CatD provided a SAXS-defined two-domain structure that dimerized but did not trimerize (Olek et al., 2014). The best fit of the monomer into the reconstituted OsCESA8-CatD trimer placed the P-CR in the junction interface and the CSRs on the periphery (Rushton et al., 2017). Although the recombinant rice OsCESA8-P-CR concentrated as a monomer, trimers of the coiled-coil domain were observed only in the crystal structure. In this structure, the Arg₄₅₈ residues in the rice OsCESA8 are coordinated by a phosphate residue (Rushton et al., 2017). This coordination around three Arg residues of the P-CR was identical to that observed in the poplar PttCESA8 homotrimer structure derived by cryo-EM (Purushotham et al., 2020). Replacement of the Arabidopsis R₄₅₇ residue with Ala in AtCesA1 complemented the rsw1 mutation, indicating that this residue is unlikely to be essential to stabilize a CESA trimer in the native CESA complex (Figure 5, A–D). Although a nearly identical trimer formed by the OsCESA8-P-CR was observed in the cotton GhCESA7 structure, its cognate R₄₃₇ was not coordinated by phosphate directly but rather to a proposed basic amino acid cluster including R₄₃₁, R₄₃₂, and K₄₃₆ (Zhang et al., 2021). Purushotham et al. (2020) pointed out that, although the Arg residues could serve to stabilize the trimer structure, the P-CR tight interactions are primarily hydrophobic contacts among other residues of the helices. The formation of glucagon trimers during fiber formation represents an example of coupling through ends of α-helices alone (Košmrlj et al., 2015), which might explain the defect in the fra6-equivalent R₄₅₃K mutation through disruption of this interaction, a residue closer to the end of HX2 (Figure 5A; Supplemental Table S1). Replacement of R₄₅₃ with Lys to create the equivalent fra6 mutation renders the mutated AtCesA1 unable to complement the phenotype. The cognate R₄₅₃ in rice interacts with R₄₅₇ in the P-CR crystal structure (Rushton et al., 2017), but it is unlikely to be able to coordinate the phosphate residue in the absence of R₄₅₇. The R₄₅₃K substitution would not be expected to alter the charge of the basic amino acid cluster (Zhang et al., 2021), indicating a more precise structural disruption by the Lys residue. A potential way to test this is to determine if the poplar PttCESA8 is able to trimerize if the cognate R₃₅₆ residues are mutated in the recombinant protein. If such a trimer represents a structure in a native CESA complex, other residues might be essential in stabilizing the structure. Alternatively, a potential equivalent interaction as the Arg₄₅₇ residues would be cognate Arg₁₀₃₈ residues in the C-terminus (Arg₉₃₄ in PttCESA8) that would be in close proximity within a trimer (Figure 6, A and B). Zhang et al. (2021) also noted that the cognate R₁₀₀₀ might be involved in trimerization in GhCESA7.

The only two amino acid substitutions in the P-CR that rendered the AtCesA1 transgene unable to complement the rsw1 mutation were P₄₁₇A and R₄₅₃K within the coiled-coil domains. Mutations in R₄₅₃ of the P-CR maintained the AtCESA1 titer, indicating that the mutation might affect assembly or structure of the complex but does not target the protein to proteolysis. On the other hand, the P₄₁₇A substitution resulted in substantial loss of titer at the restrictive temperature (Table 1). This finding indicates that the bend in HX1 introduced by P₄₁₇ is essential for maintenance of the synthase structure.

Three mutations that partly reduced the ability of the transgene to complement the rsw1 mutation were the D₄₃₈A/Y₄₃₉A substitution in the connecting loop, and the L₈₁₂R substitution in the catalytic core. Our modeling of the P-CR of OsCESA8 would, in AtCESA1, place the connecting loop into the catalytic core, where an interaction between F₄₃₁ and L₈₁₂ would be predicted to stabilize the loop, embedding the D₄₃₈A/Y₄₃₉A, which would otherwise be solvent exposed (Rushton et al., 2017). These results are consistent with the modeling of the P-CR in the AtCESA3-CatD (Qiao et al., 2021) crystal structure, and the PttCESA8 and GhCESA7 cryo-EM structures, which place these residues of the connecting loop at the distal end of the protein. In this configuration, the L₈₁₂ is unlikely to form an interaction with F₄₃₁, as originally proposed (Rushton et al., 2017), but the L₈₁₂ might be part of a hydrophobic cluster that would introduce disorder in the structure if replaced by Lys (Figure 6A). Placement of the P-CR coiled-coil at the distal end of the CatD indicates a more likely role in protein–protein interaction with other CESAs or accessory proteins in regulation of catalytic activity or in trafficking the complexes to the plasma membrane.

In bacterial synthases, the PilZ domain interacts with these sequences to regulate entry of the UDP-Glc (Morgan et al., 2014). Within the CatD of the R. sphaeroides BscA cellulose synthase, a “gating loop” containing a central FxVTxK motif overlies the entrance to the active site and interacts with the PilZ domain upon binding of cyclic-di-GMP activator (Omadjela et al., 2013; Morgan et al., 2014). Plant CESAs are missing the PilZ domain; the FxVTxK motif is conserved but is displaced downstream of IF3 (formerly TMD5; Slabaugh et al., 2016). In genetic complementation studies like those here, mutations F₉₅₄L, V₉₅₆D, and K₉₅₉E in AtCesA1 each failed to rescue the rsw1 root phenotype at restrictive temperatures, indicating a critical role for the FxVTxK motif in cellulose synthesis and cell growth (Slabaugh et al., 2014, 2016). By contrast, the cognate mutation F₉₀₁L, in AtCesA7 was able to fully complement the collapsed xylem phenotype and restore cellulose synthesis in the irregular xylem3-1 (irx3-1) mutant (Burris et al., 2021). Mutations V₉₀₃D and K₉₀₆E partially restored the wild-type phenotype. A 21-amino acid sequence containing the FxVTxK motif lacked sufficient density to be modeled in the PttCESA8 (Purushotham et al., 2020), but amino acids flanking these sequences closely align with the α-helix containing the QxxRW catalytic motif (Figure 6B). When the PttCESA8 was modeled by AlphaFold (Jumper et al., 2021), the 21-amino acid sequence containing the FxVTxK motif, missing from the PttCESA8 structure, was predicted to extend across the entrance to the active site. With the absence of a PilZ domain in plant sequences, it is not clear which other plant sequences might interact with the FxVTxK sequence and substitute for the PilZ regulatory action. However, flexible sequences downstream of the coiled-coil of P-CR that also align in the entrance might be a good candidate for interaction with the FxVTxK motif in regulation of substrate entry through interaction with the plant version of the gating loop (Figure 6B). Consistent with a role in catalytic activity, mutations to a conserved PWP motif in this downstream P-CR region of the P. alba secondary wall cellulose synthase PalCesA4 resulted in severe decreases in cellulose content (Nayeri et al., 2022).

Remaining questions

Although the trimerization model is attractive based on expression of three isoforms of primary and secondary wall CESAs, the ability of the AtCesA1 transgene with a mutated R₄₅₇ residue to fully complement the WT phenotype in rsw1 indicates this residue is not critical in vivo. Other amino acids might participate to stabilize the trimer structure in the absence of the Arg, but an alternative hypothesis is that the trimer is a property only of a detergent-solubilized and reconstituted protein and other combinations of CESA isoforms might be possible.

The experiments described here consider only the potential roles of the P-CR and CSR in coupling of individual CESAs, and do not address how these assemble further into a rosette complex. Early genetic experiments showed by domain-swap experiments that specific isoforms interact through their C-termini (Wang et al. 2006). However, a comprehensive analysis employing domain-swap constructs of three secondary cell wall AtCESAs showed that class specificity extended throughout the protein and in not only localized to the Zn-finger, P-CR, and CSR domains (Kumar et al. 2017). Different CESA isoforms varied in the degree to which they exhibit class specificity. For example, AtCESA7 was found to be more constrained and least able to complement in domain-swaps, whereas AtCESA8 was least constrained. A role for N-terminal Zn-finger domains in dimerization of CESAs was also proposed (Kurek et al., 2002). The PttCESA8 trimer structures inferred by cryo-electron microscopy placed the Zn-finger trimers in an extended structure into the cytoplasm (Purushotham et al., 2020). The Zn-finger domains of CESAs are known to form dimers specifically, and not trimers (Kurek et al., 2002; Olek et al., 2014), how these domains might function in coupling CESAs into the full complex awaits further experiments. Advances in cryo-electron microscopy and cryo-electron tomography promise to resolve these issues in determining the structure in situ of native CesA complexes.

Materials and methods

Cloning and expression of OsCesA8-CatD

A pBluescript SKII vector (Stratagene, La Jolla, CA) containing the rice (O. sativa) CELLULOSE SYNTHASE8 cDNA, OsCesA8 (LOC_Os07g10770, formerly called OsCesA13) was used as a template to clone the catalytic site of OsCesA8 (CatD). The 1525-bp CatD fragment (encoding amino acid residues 341–846) was ligated into a modified pETM11 (EMBL) vector containing an N-terminal 6× His-tag and a Tobacco Etch Virus (TEV) protease cleavage site (ENLYF(Q/G)A; Supplemental Table S3; Olek et al., 2014). The native pETM11 plasmid was obtained from EMBL (http://www.embl-heidelberg.de). Plasmid constructs were transformed by heat shock into competent Escherichia coli (E. coli) (Rosetta2 [DE3], Novagen). The 500-mL cultures in LB media were initiated with the 5-mL overnight culture and grown to OD₆₀₀ of 0.6–0.8 (about 4 h), cooled to 16°C, and recombinant protein production was induced with 0.2 mM isopropyl β-D-1-thiogalactopyranoside (IPTG; IB02100; IBI Scientific) for 13 h before harvesting.

Cloning and heterologous protein expression

A pBluescript SKII vector (Stratagene) containing the rice CELLULOSE SYNTHASE8 cDNA (OsCesA8) was used as a template to clone the P-CR, encoding A₃₉₉ to V₅₂₃. The 375 bp P-CR DNA insert was amplified by PCR using a forward primer (5′-GCGCCATGACTAGTGACTCGGCAATGCTGACTT-3′) and reverse primer (5′- GTGGTGCTCGAGTTACTAGACCAGGCGGGGAA-3′) obtained from Integrated DNA Technologies, Inc. The P-CR was ligated into a modified pETM-11 (EMBL) vector, containing an N-terminal 6× His-tag, using SpeI and XhoI restriction sites. The empty pETM-11 plasmid was obtained from EMBL (http://www.embl-heidelberg.de). The OsCesA8 P-CR pETM-11 plasmid construct was transformed into competent E. coli (Rosetta2 [DE3], Novagen) cells using heat shock. Positive transformants were selected with 50 µg mL⁻¹ kanamycin and 25 µg mL⁻¹ chloramphenicol and then used to inoculate 5 mL Luria–Bertani media (LB) for overnight growth. The 5-mL overnight culture was used to inoculate a 500-mL LB culture that was grown to 0.6–0.8 OD₆₀₀ then cooled to 16°C before recombinant protein induction with 0.2 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) for 13 h.

Isolation and purification of recombinant proteins

Recombinant rice OsCESA8-CatD proteins were isolated as described previously (Olek et al., 2014). Briefly, frozen cells suspended in homogenization buffer were burst by sonication. Cell debris was pelleted, and a suspension of Ni⁺ 6-Fast Flow resin microbeads (GE Healthcare, 17-5318-01) was added to supernatant liquids. After a 1-h incubation, the suspension was then microfuged to collect resins with bound protein. After washes in low-imidazole buffer, the proteins were eluted in 0.5-mL batches a buffer containing 300 mM imidazole. In some experiments, the His-tag was excised from CatD by TEV protease. Proteins in 10 mM DTT eluted from the Ni⁺-column (∼0.6 mg mL⁻¹) were digested with 30 units of AcTEV protease itself tagged with Hisx6 (Invitrogen, 12575-015) at ambient temperature for 6 h with gentle rocking. The Hisx6 tag had no influence on dimerization or UDP-Glc binding.

Proteins were separated on a column of Superdex 200, 10/300GL (GE Healthcare, 17-5175-01). Protein elution was monitored by absorbance at 280 nm, and 0.5-mL fractions were collected. In experiments to reduce DTT, fractions of protein from column in 10 mM DTT were reduced to near zero by solvent exchange during concentration using Amicon Ultra Centrifugal Filters with 30K membranes (Millipore, UFC503024). Protein content after concentration was approximately 0.5 mg mL⁻¹.

Site-directed mutagenesis and Arabidopsis transformation

For Arabidopsis transgenes, an Agrobacterium tumefaciens pFGC5941 binary vector (TAIR: CD3-447) containing the Arabidopsis AtCesA1 coding sequence driven by its native promoter, created in previous work (Slabaugh et al., 2014), was used for these experiments. Primer design and site-directed mutagenesis were done using the Q5 Site-Directed Mutagenesis Kit (New England Biolabs, Ipswich, Massachusetts) and NEBaseChanger algorithm (http://nebasechanger.neb.com/) for primer design (Supplemental Table S2). Vectors were transformed into E. coli DH5α cells using heat shock, and positive bacterial transformants were selected with 50 μg mL⁻¹ kanamycin. Following sequence confirmation by sequencing and analytical restriction enzyme digest, vectors were transformed into Agrobacterium strain GV3101 by electroporation. Positive transformants were selected on plates containing 50 µg mL⁻¹ kanamycin, 35 µg mL⁻¹ gentamycin, and 10 µg mL⁻¹ rifampicin. Seeds of the rsw1-1 mutant were obtained from TAIR (https://www.arabidopsis.org/; TAIR: CS6554) and transformed using the floral dip method (Clough and Bent, 1998).

Seeds of Arabidopsis wild type (Col-0), rsw1-1, and transgenic mutated AtCesA1 lines were surface-sterilized by mixing in 10% (v/v) Clorox bleach (6.25% [w/v] NaClO) and 0.05% (w/v) Tween 20 solution with seeds for 10 min followed by five washes with a > 10:1 volume of sterile water. The final wash was decanted and seeds cold-treated in the dark at 4°C for at least 2 d before planting or plating. Seeds were sown on soil and grown under standard Arabidopsis germination and growth conditions in growth rooms and greenhouses maintained near 22°C. Healthy transgenic seedlings (∼2 weeks post-sowing) were selected for the bar gene on the pFGC5941 vector by spraying with 20 mg L⁻¹ glufosinate-ammonium. For initial screening of recombinant gene expression and complementation experiments, individual seeds were sown in a grid pattern on agar plates containing autoclaved one-half-strength Murashige and Skoog (MS) media (Sigma-Aldrich M5519) or 2.15 mg L⁻¹, 2% (w/v) sucrose, 0.5 g L⁻¹ MES, 8 g L⁻¹ agar at pH 5.7. For gene expression assays, plates were placed vertically in a growth room for 7 d at 22°C under a 16/8 h light/dark photoperiod. For complementation experiments, plants were grown vertically for 4 d at 22°C and 3 d at 31°C with ∼60% humidity, under a 16/8 h light/dark photoperiod at 150 μmol m⁻² s⁻¹ light intensity in a growth chamber. Independent lines carrying a transgene were self-crossed to homozygosity by the T3 generation. For additional experiments, seeds were surface sterilized with 70% (v/v) ethanol in water, then 10% (v/v) Clorox bleach (6.25% [w/v] NaClO), and rinsed with several changes of sterile water. They were chilled at 4°C for 3 d. Seeds were grown on plates with one half-strength MS media or 2.15 mg L⁻¹, 1% (w/v) sucrose, and 1% (w/v) agar.

Transgene expression

Seeds of Arabidopsis wild type (Col-0), rsw1-1, and the transgenic lines of mutated AtCesA1, were sown on MS plates to create a single biological replicate. This was iterated to create a total of three biological replicates. Seven-day-old seedlings of each plate were harvested a few hours into the light phase of the light/dark cycle, flash-frozen using liquid nitrogen, and homogenized with steel beads. RNA was extracted using the TRIzol™ Reagent (Invitrogen) protocol and DNase-treated with the TURBO DNA-free™ kit (Invitrogen). cDNA was synthesized from RNA templates using the iScript cDNA synthesis kit (Bio-Rad). RT-qPCR was done using custom primers and SYBR™ Green PCR Master Mix (Applied Biosystems) with a StepOnePlus™ Real-Time PCR system (Applied Biosystems). For each of the three biological replicates, CT values were measured and averaged for three technical replicates. Expression was averaged across the biological replicates as a fold change ratio compared to the nuclear cap-binding protein (CBP20; At5g44200) reference control gene (Czechowski et al., 2005; Kim et al., 2008). No substantial amplification occurred using the RT-qPCR primers in the wild type or rsw1-1 backgrounds.

Plant materials

For root growth and morphology assays, seeds of Arabidopsis (A. thaliana) wild type, rsw1, and transgenic rsw1 transformed with AtCesA1 constructs that encode mutated amino acids were soaked in deionized water for 3 d at 4°C to prime germination and sown in rows on plates of 1% (w/v) agar supplemented with one-half strength MS salts and grown upright at 22°C for 4 d under continuous light at 150 µmole m⁻² s⁻¹ Photosynthetic Active Radiation (PAR). Plates were photographed and returned to 22°C (permissive temperature) or transferred to a chamber held at 31°C (restrictive temperature) for up to 24 h. Plates were re-photographed at 8 h and at 24 h for determination of root lengths using ImageJ software (https://imagej.nih.gov/ij/; Schneider et al., 2012). Micrographs of root-tip morphology were also taken using a Nikon SMZ1500 stereo microscope, with a SPOT4 digital camera.

For cellulose synthesis determinations, seedlings were grown for 5 d at 22°C before transfer of one-half of the plates to 31°C for 6 h before addition of 8 µCi (270 M Becquerel) of ¹⁴C-D-glucose (290 mCi mmole⁻¹ [10.7 GBq mmol⁻¹]; Amersham Biosciences) in 150 µL of water was added to the base of the seedlings. For labeling, plates containing 20–25 seedlings each were transferred to sealed plastic chambers containing KOH bicarbonate traps for evolved ¹⁴CO₂. After the 18-h labeling, the chambers were out-gassed in a fume hood, and seedlings were harvested into liquid nitrogen and were homogenized in a sintered glass-glass grinder in ∼5 mL of 50 mM Tris[HCl], pH 7.2, supplemented with 1% (w/v) sodium dodecyl sulfate (SDS), transferred to 15-mL Corning tubes, volumes of SDS-buffer adjusted to 12 mL, and the homogenates incubated at 60°C for 20 min before centrifugation at 1,200 × g at r_max for 5 min at ambient temperature. The liquid supernatants were aspirated from the pellets, and 12 mL of SDS-buffer added for a second incubation at 60°C for 20 min. After centrifugation, the pellets were washed twice with 12 mL of water each, and 3 times with 12 mL of 50% (v/v) ethanol in water. The pellets were returned to water washes until radiolabel in the supernatants neared background. The pellets suspended in 1 mL of water were transferred to 5-mL Pierce Reacti-vials (Thermo-Aldrich), adjusted to 5-mL, and centrifuged at 4,000 × g at r_max for 5 min. The pellets were suspended in 2 mL of acetic–nitric reagent (acetic acid:nitric acid:water, 8:1:2 [v/v/v]) and heated to 100°C for 1 h (Updegraff, 1969). After cooling to ambient temperature, the samples were centrifuged at 4,000 × g at r_max for 5 min. One-milliliter of the supernatant was taken for liquid scintillation counting; the pellet was washed twice with 5 mL of water, and suspended in 1 mL of water for counting. Both preparations were brought to 20 mL in Eco-Lite counting solution, and radioactivity was determined using a Beckman 9600 liquid scintillation counter with quench correction for efficiency losses in the acid solutions. The proportion of radiolabel in the cellulose pellet was then determined compared to the acetic–nitric soluble noncellulosic materials.

For proteomics experiments, seeds of Arabidopsis wild type (Col-0), rsw1, and transgenic rsw1 complemented with AtCesA1 with mutated amino acids were soaked in deionized water for 3 d at 4°C to prime germination and sown in #499 pots containing Metromix360 (Sungro). Plants were then grown in a growth room with 16-h–light/8-h–dark photoperiod at 22°C. Light at 150 µmole m⁻² s⁻¹ PAR was delivered by VYPR 23 lamps (Fluence Bioengineering; spectral distribution available at https://fluence.science/store/vypr-series/vypr-23/). The seedlings were grown for 21–22 d before harvest.

Statistical analysis

Analysis of variance (ANOVA) for the root length and cellulose synthesis data was performed using the “Anova Single Factor” function in Microsoft Excel using an alpha value of 0.05. A t test was used classify each mutated AtCesA1 line vs. wild type (Col-0) and rsw1 at both permissive and restrictive temperatures for similarity at P ≤ 0.05 (Matlab version 2021; Mathworks; https://www.mathworks.com/). Those that were significantly different from both Col-0 and rsw1 were then tested against additional lines to determine groups exhibiting similar phenotypes.

Isolation of Golgi membrane fractions for proteomics analyses

Golgi membranes were obtained by flotation centrifugation essentially as described by Okekeogbu et al. (2019). About 10–14 g of fresh Arabidopsis leaves at were placed in a chilled mortar and overlaid with an equal volume of ice-cold 100 mM HEPES (2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid; Sigma-Aldrich), 20 mM KCl, and 20 mM ascorbic acid, adjusted to pH 7.6 with BTP (1,3-bis[tris(hydroxymethyl)-methylamino]propane; Sigma-Aldrich), and brought to 84% (w/v) Suc. The leaves were stirred gently in homogenization buffer for 5 min before mashing in a chilled mortar and pestle. The homogenates were squeezed through a nylon mesh (45-µm² pores) and adjusted to ∼42% (w/v) Suc if necessary. Twenty milliliters of the homogenate were pipetted into each 38.5-mL centrifuge tube (Ultraclear; Beckman) and overlaid with 8 mL each of 35%, and 29% (w/v) Suc, and tubes were topped off with 18% (w/v) Suc, in a gradient buffer containing 20 mM HEPES[BTP], pH7.2, supplemented with 40 mM KCl. After flotation centrifugation at 27,000 rpm (131, 5,003 × g at r_max) in an SW28 rotor (Beckman) for 90 min at 4°C, the 29%/35% (w/v) Suc interfaces containing Golgi membranes were precipitated using five volumes (v/v) of cold (–20°C) 100% acetone in 50-mL Falcon tubes (source), and held at –20°C for 18 h. After three washes with 50-fold excess chilled 80% (v/v) acetone (–20°C), the pellets were transferred to 2-mL Eppendorf tubes, and washed with additional 80% (v/v) acetone (–20°C), and the Suc-free pellets were air-dried for proteome analysis.

The protein samples were subjected to proteomics analysis as described previously (Mohallem and Aryal, 2020; Zembroski et al., 2021). Proteins in the dried pellets were re-suspended in 100 mM ammonium bicarbonate buffer to determine protein concentration using Bicinchoninic (BCA) assay with BSA as a standard. Then, 50 μg (equivalent volume) was precipitated overnight (−20°C) with 4 volumes of cold acetone, and dried pellet was dissolved in 8 M urea containing 10 mM DTT and incubated at 37°C for 1 h for reduction followed by alkylation using alkylating reagent (195 uL acetonitrile, 1 µL triethylphosphine, and 4 µL of iodoethanol) and incubated for 1 h at 37°C. After reduction and alkylation, samples were dried in a vacuum centrifuge and digested using trypsin/LysC mix at 1:25 enzyme-to-protein ratio. Digestion was performed at high pressure using a Barocycler (50°C; 60 cycles:50 s at 20 kPSI and 10 s at 1 ATM) as described previously (Connelly et al., 2018). Digested peptides were desalted using MicroSpin columns (C18 silica; The Nest Group), dried purified peptides in a vacuum centrifuge, and stored samples at 80°C until LC-MS/MS. Peptides were then re-suspended in 3% (v/v) acetonitrile/0.1% (v/v) formic acid to a final concentration of 0.2 µg/µL and 5 µL (1 µg) was loaded to HPLC system

LC-MS/MS analysis for peptide sequencing

Peptides were analyzed in a Dionex UltiMate 3000 RSLC nano System (Thermo Fisher Scientific, Odense, Denmark) coupled on-line to an ETD-enabled Orbitrap Fusion Lumos Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) as described previously (Barabas et al., 2019). Briefly, reverse phase peptide separation was accomplished using a trap column (300 µm ID × 5 mm) packed with 5 µm 100 Å PepMap C18 medium coupled to a 50-cm long × 75 µm inner diameter analytical column packed with 2 µm 100 Å PepMap C18 silica (Thermo Fisher Scientific). The column temperature was maintained at 50°C. Sample was loaded to the trap column in a loading buffer (3% (v/v) acetonitrile, 0.1% (v/v) formic acid) at a flow rate of 5 uL min⁻¹ for 5 min and eluted from the analytical column at a flow rate of 200 nL min⁻¹ using a 160-min LC gradient. Column was washed and equilibrated by using three 30-min LC gradient before injecting next sample. All data were acquired in the Orbitrap mass analyzer and data were collected using an HCD fragmentation scheme. For MS scans, the scan range was from 350 to 1,600 m/z at a resolution of 120,000, the automatic gain control target was set at 4 × 10⁵, maximum injection time 50 ms, dynamic exclusion 30s, and intensity threshold 5.0 × 10⁴. MS data were acquired in Data Dependent mode with cycle time of 5 s per scan. MS/MS data were collected at a resolution of 15,000.

Data analysis

LC-MS/MS data were analyzed using MaxQuant software (version 1.6.3.4) against the A. thaliana protein database downloaded from Uniprot ((https://www.uniprot.org/) using the following parameters: precursor mass tolerance of 10 ppm; fragment mass tolerance of 20 ppm; enzyme specificity of Trypsin/Lys-C allowing up to two missed cleavages; oxidation of methionine (M) as a variable modification and iodoethanol (C) as a fixed modification. False discovery rate of peptide spectral match and protein identification was set to 0.01. Proteins with LFQ (label-free quantitation) # 0 and MS/MS (spectral counts) ≥2 were only considered as true identification and used for further analysis and comparison across samples.

Protein modeling

Four classes of protein structural comparison algorithms were employed to match P-CR structural features against PDB: carbon-α (Cα) comparisons in Dali (Holm and Laakso, 2016) and FATCAT (Ye and Godzik, 2004), secondary structural element (SSE) in deconSTRUCT (Zhang et al., 2010) and PDBeFold (Krissinel and Henrick, 2004), 3D surface comparisons (surface) in 3D-SURFER 2.0 (La et al., 2009) and PDBePISA (Krissinel, 2007), and protein fragment sequence and geometry (Seq/Geo) comparisons in BioXGEM 3D-BLAST (Tung et al., 2007) and iSARST (Lo et al., 2009). The P-CR surface structure was also characterized using the VisGrid (Li et al., 2008) and Fpocket (Schmidtke et al., 2010) algorithms. P-CR modeling was carried out using the threading-based modeling algorithms of RaptorX (http://raptorx.uchicago.edu/StructurePrediction/predict/), Phyre2 (http://www.sbg.bio.ic.ac.uk/~phyre2/), HHpred (https://toolkit.tuebingen.mpg.de/), and MUSTER (https://zhanglab.ccmb.med.umich.edu/MUSTER/).

The full-length PttCESA8 sequence (PDB: 6WLB; Purushotham et al., 2020) was modeled using AlphaFold (Jumper et al., 2021) to show the relative position of a 21-amino acid sequence containing the FxVTxK motif in the PttCESA8 monomer. Alphafold was obtained from their Github: https://github.com/deepmind/alphafold and run locally. Alphafold versions earlier than v2.0.1 originally run did not output the confidence score (pLDDT score), but version 2.2.2 was run to obtain confidence scores. Scores for the F×VT×K motif was 46.70 on average for the six residues, with individual scores of F (46.19) – T (45.54) – V (45.10) – T (47.08) – × (46.96) – K (49.35). These scores were deemed reasonable for a loop region without secondary structure. The sequence database used was a combination of the following databases (https://pubmed.ncbi.nlm.nih.gov/34293799/); UniRef90 v.2020_01 (https://ftp.ebi.ac.uk/pub/databases/uniprot/previous_releases/release-2020_01/uniref/); BFD (https://bfd.mmseqs.com); Uniclust30 v.2018_08 (https://wwwuser.gwdg.de/~compbiol/uniclust/2018_08/); and MGnify clusters v.2018_12 (https://ftp.ebi.ac.uk/pub/databases/metagenomics/peptide_database/2018_12/).

Bioinformatics

CesA gene sequences from 33 species representing a diverse collection of angiosperms were obtained from Phytozome plant genomics database (https://phytozome-next.jgi.doe.gov/), and the CesA genes closest in sequence to the six Arabidopsis AtCesA genes associated with primary and secondary wall formation were determined by protein–protein BLAST (Altschul et al., 1990). These genes were truncated to P-CR and CSR sequences for MultAlin analysis according to algorithms using online software at (http://multalin.toulouse.inra.fr/multalin; Corpet et al., 1988). Sequence logos (WebLogos) that graphically represent conservation of amino acids in multiple sequence alignments were obtained according to Crooks et al. (2004) at a public online software (https://weblogo.berkeley.edu/).

Predictor of Natural Disordered Regions (PONDR) analysis was performed using on-line software (http://www.pondr.com/) using the VLXT predictor that integrates the three neural networks, the VL1 predictor (Romero et al., 1997), the N-terminus predictor, and the C-terminus predictor, together called XT (Li et al., 1999).

Accession numbers

Sequence data from this article can be found in the GenBank/EMBL libraries, and the Phytozome plant genomics database (https://phytozome-next.jgi.doe.gov/), under the following accession numbers: AtCESA1 (AT4G32410); AtCESA3 (AT5G05170) ; AtCESA6 (AT5G64740); AtCESA4 (AT5G44030); AtCESA7 (AT5G17420); AtCESA8 (AT4G18780); OsCeESA8 (OS07G10770); PttCESA8 (AAT09896.2).

Data availability

All proteomics raw LC-MS data are deposited and available in massIVE data repository site (massive.ucsd.edu) under ID MSV000089874.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Multiple alignment of P-CR and CSR sequences of AtCESA1 and OsCESA8.

Supplemental Figure S2. Substitution of C₆₆₂ with Ala results in loss of the ability of recombinant rice OsCESA8-CatD to dimerize in vitro.

Supplemental Figure S3. Time-course of floral stem initiation in plants expressing mutated AtCesA1trans-genes in the rsw1 background.

Supplemental Figure S4. Relative expression of the mutated AtCesA1 trans-genes in rsw1.

Supplemental Figure S5. Root tip phenotypes and root elongation of representative lines of rsw1 expressing AtCesA1trans-genes modified to substitute Ala or Ser residues for targeted amino acids in the CSR.

Supplemental Figure S6. Multiple alignment of the CSR domains of homologs closest in sequence to primary wall AtCESA1, AtCESA3, and AtCESA6 for 33 species of diverse angiosperms.

Supplemental Figure S7. Conservation of sequence in the CSR among homologs closest in sequence to primary wall AtCESA1, AtCESA3, and AtCESA6.

Supplemental Figure S8. Multiple alignment of the CSR domains of homologs closest in sequence to secondary wall AtCESA4, AtCESA7, and AtCESA8 for 33 species of diverse angiosperms.

Supplemental Figure S9. Conservation of sequence in the CSR among homologs closest in sequence to secondary wall AtCESA4, AtCESA7, and AtCESA8.

Supplemental Figure S10. Patterns of predicted intrinsic disorder vary within CSRs of secondary cell wall AtCESAs in a class-specific manner.

Supplemental Figure S11. Protein structural comparisons of the coiled-coil domains of the P-CRs of AtCESA1 and OsCESA8 compared to an iron transport 2 domain of the voltage gated sodium channel protein (VGIC).

Supplemental Figure S12. Relative expression in the rsw1 background of the mutated AtCesA1 trans-genes.

Supplemental Figure S13. Root tip phenotypes and root elongation of representative lines expressing in rsw1 background AtCesA1 transgenes modified to substitute Ala, Ser, or Lys residues for targeted amino acids in the P-CR.

Supplemental Figure S14. Time-course of floral stem initiation in plants expressing mutated AtCesA1trans-genes in the rsw1 background.

Supplemental Table S1. Predicted functions of amino acids of the coiled-coil domain of the Arabidopsis AtCESA1 P-CR.

Supplemental Table S2. Top five Dali matches to the AtCESA1 P-CR coiled-coil domain ranked by Z-score.

Supplemental Table S3. Primers used to generate mutations in OsCesA8-CatD constructs and AtCesA1trans-genes, and to measure AtCesA1 transgene expression by RT-qPCR.

 

N.C.C., A.T.O., and M.C.M designed the experiments. A.T.O. generated the recombinant OsCesA8-CatD lines, and most of the AtCesA1 mutated CSR and P-CR transgenic lines. P.S.R generated some of the AtCesA1 mutated P-CR lines. A.T.O characterized growth and expression of all transgenic lines, and, with N.C.C., performed growth rate and cellulose determinations on high-expression lines. P.S.R. and D.K. performed modeling studies of the P-CR, and P.S.R. and C.V.S. designed mutations in the P-CR domain. U.K.A. performed proteomic analyses. N.C.C. performed MultAlin and PONDR analyses. Z.Z. and D.K. performed AlphaFold modeling of the PttCESA8, and P.C. reconstructed models of the AtCESA3-CatD and the PttCESA8 monomer and trimer. N.C.C. wrote the manuscript with contributions from all coauthors.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is: Nicholas C. Carpita (nick.carpita@nrel.gov).