Cellulose synthesis in Physcomitrella patens : gene expression and mutational analysis

A detailed analysis of cellulose synthesis in nonvascular plants can contribute to a better understanding of the evolution of this important process. In this study, the nonvascular plant Physcomitrella patens was used as a model system to investigate the roles of the different isoforms of cellulose synthase (CESA). PpCESA gene expression was quantified through Reverse Transcription quantitative (RT-q) PCR and localized through construction and analysis of promoter::reporter lines to determine the roles of the PpCESAs throughout development. Physcomitrella patens CESA genes are ubiquitously expressed in the filamentous protonema stage. All of the PpCESAs are expressed in the gametophore as well, with PpCESA4 and PpCESA10 mainly expressed in the axillary hairs. This broad expression is unique to non-vascular plants, in contrast to vascular plants in which CESA expression is restricted to cells depositing either primary cell walls or secondary cell walls during development. Upregulation under osmotic stress induced by mannitol may indicate a role for cellulose under high osmotic stress. PpCESA6, PpCESA7, and PpCESA8 were hypothesized to be responsible for osmotic stress-induced cellulose synthesis based on mannitol-induced upregulation of expression as indicated by analysis of microarray data. The roles of CESAs in development and stress tolerance were assessed by producing knockout mutants of PpCESA6, PpCESA7, and PpCESA8. Ppcesa8 knockout (KO) and ppcesa6/7 KO mutants do not have dramatic developmental phenotypes. However, ppcesa6/7KO mutants show sensitivity towards high salinity, indicating that cellulose is important under abiotic stress. Currently, only ppcesa5 KO mutants show a phenotype in the gametophore and no single KO mutants have phenotypes in the protonema. Cellulose synthesis inhibitors were used to examine the role of cellulose in the protonema. Results show that protonemal tissue is relatively insensitive to cellulose inhibitors, since only high concentration of the cellulose synthesis inhibitor DCB had any effect. DCB caused rupturing of tips, indicating that cellulose is necessary in tip growth. Results also indicate that cellulose synthase-like D (CSLD) proteins may contribute to the synthesis of cellulose in moss protonema. Since single and double KO mutants of PpCESA6, PpCESA7, and PpCESA8 do not produce a phenotype and PpCESA expression is ubiquitous. PpCESAs maybe be redundant in function such that another PpCESA may compensate loss of a single PpCESA. PpCESAs are highly similar in sequence and may have not fully subfunctionalized [1] and therefore, the PpCESAs isoforms may be more interchangeable than those of seed plants. Other cell wall components, such as hemicelluloses, pectins, and arabinogalactan proteins, may also compensate for lack of cellulose. These cell wall components were also examined through immunolabeling of regenerating protoplasts. The results showed the highest abundance of crystalline cellulose and moderate levels of callose, mannan, 1,5-α-L-arabinan and arabinogalactan proteins. Very low levels of 1,4-β-D-galactan and no homogalacturonans were detected.

x

INTRODUCTION
Cellulose is an abundant biopolymer that has many economic uses, such as biofuels, lumber, and textiles. It is essential in plant development. Cellulose consists of approximately 18-36 β1, 4-glucan chain, bundled together forming a single microfibril unit. Microfibrils can further associate to form macrofibrils that make up the backbone of the plant's cell wall [2].
Cellulose synthase A protein family (CESAs) have been identified as key proteins in the synthesis of cellulose [3]. There are multiple CESA isoforms found in both vascular and nonvascular plants [4,5]. These form rosette structures known as cellulose synthase complexes (CSC) [6]. The CSCs in vascular plants are heterooligomeric, consisting of three different CESAs, which are specific for either primary or secondary cell wall formation. Primary cell walls are flexible and found in growing tissues, while secondary cell walls are rigid due to the aromatic polymer lignin and deposited in maturing cells in vascular and support tissue [7]. The CSC composition of the vascular plant Arabidopsis was discovered through CESA expression analysis, including promoter-reporter constructs and RNA in situ hybridization, and confirmed through co-immunoprecipitation [8,9]. AtCESA null mutants are characterized by either defects in vascular development or embryo lethality [8,9]. Phylogenetic analysis and functional analysis in other seed plants indicated that hetero-oligomeric CSCs evolved early in seed plant evolution [1].
Physcomitrella patens, a nonvascular land plant, also has multiple CESA isoforms but does not have primary and secondary cell wall formation like vascular plants. The roles of the multiple CESA isoforms in P. patens are still unknown [4].
Understanding how the roles of the PpCESAs differ from those of the vascular plant CESAs and whether PpCESAs form hetero-oligomeric or homo-oligomeric rosettes CSCs will provide insight into the roles of the distinct CESA isoforms in the assembly and function of seed plant CSCs. An advantage of using P. patens is its ability to be genetically manipulated due to its high rate of homologous recombination. With this unique property, genes of interest can be investigated through knockout mutations and gene expression analysis through a gene reporter system [10].

PpCESAs maybe important developmentally and also for response to osmotic stress
Prior experiments have indicated that PpCESAs are involved in certain stages of development and may play a role in stress response. PpCESA5 has a developmental role in gametophore formation based on mutation analysis [11]. PpCESA6 and PpCESA7 knockout mutants do not show obvious developmental phenotype impairment [12], but the encoded proteins may play a role when under osmotic stress.
When P. patens is subjected to osmotic stress via addition of mannitol to the culture medium, cellulose deposition is upregulated [13]. Expression of PpCESA6, PpCESA7, and PpCESA8 is increased under these conditions based on analysis of microarray data [14]. PpCESA6, PpCESA7, and PpCESA8 mutants have not been tested for upregulation of cellulose under osmotic stress and the effects of mannitol on expression of these genes have not been confirmed by RT-qPCR.
Physcomitrella patens upregulates cellulose under osmotic stress [13], while vascular plants downregulate cellulose [15]. These responses might be due to differences in water uptake and dehydration tolerance mechanisms. Poikilohydric mosses depend on external surface water for hydration and homeohydric vascular plants maintain a constant hydrated state through mechanisms that prevent dehydration. When P. patens is dehydrated, losing 95% of its water weight, it is still able to survive [16]. In Arabidopsis, atcesa8 mutants shows drought tolerance, which is presumed to be due to the collapsed xylem that prevents water loss [17].
Physcomitrella patens uses a different mechanism to combat drought because it does not contain a vascular system and does not need to maintain a constant amount of water [18].
Physcomitrella patens is also tolerant to salt stress, surviving under 350 mM NaCl [16,19]. Analysis of microarray data suggests that PpCESAs are upregulated in response to salt [14]. In Arabidopsis, cellulose deficient mutants have shown impairment in growth under high salinity conditions [20]. Cellulose synthase like D5, atcsld5, mutant also has decreased osmotic stress tolerance under drought, high salinity, and mannitol [21]. On this basis, PpCESAs are also predicted to be involved in salinity stress response and impaired survival in response to salinity is expected if these genes are deleted.

Cell wall composition in P. patens
With the exception of ppcesa5KO [11], single PpCESA KO mutations have not produced phenotypes [12,22]. This indicates that cell wall components other than cellulose may play leading roles in structural development in P. patens. In vascular plants, some cells are still able to expand and divide when cellulose is absent or reduced due to upregulation of other cell wall polysaccharides, such as pectin and callose [23]. Non-cellulosic cell wall components, such as pectin, hemicellulose, and arabinogalactan proteins (AGPs), have been generally analyzed in P. patens using carbohydrate microarrays [24] and some immunofluorescent staining [25][26][27][28]. Pectin adds flexibility to the plant cell wall, hemi-cellulose cross-links cellulose, and AGPs interacts with pectin, which is found to be important in cell extension [29].
Currently, it is known that hemicelluloses xylan and xyloglucan are found in gametophores [26], while mannan was found in protonema with deposition concentrated in cell junctions [27]. Callose was found in early developing spores [28] and cellulose was found in developing gametophore buds [11]. AGPs, which crosslink pectins, were found to be essential in tip growth of protonemal cells [25]. No comprehensive study of cell wall composition has been done in all tissues and few immuno and affinity histochemical studies have been done in protoplasts.

Affect of cellulose synthesis inhibitors isoxaben and DCB
Cellulose synthesis was previously shown to be important in stabilizing tip growth. Both pollen tubes and root hairs extend through tip growth [30], similarly to P. patens protonemal tissue [31]. Inhibition of cellulose synthesis in both cell types that extend by tip growth is highly disruptive. Petunia and lily pollen tubes were treated with cellulose synthesis inhibitor 2,6-dichlorobenzonitrile (DCB), which caused irregular cell wall deposition and rupturing of tips, indicating that the presence of cellulose is essential in pollen tube tip growth [32]. Similarly, Lilium and Solanum pollen tubes were grown in the presence of cellulase and cellulose crystallation inhibitor CGA (1-cyclohexyl-5-(2,3,4,5,6-pentafluorophenoxy)-1,4,2,4,6-thiatriazin-3amine), where low concentration caused irregular pollen tube size and direction and rupturing of tips at high concentrations of either the cellulase or CGA [33].
Root hair tip growth has been studied most extensively in Arabidopsis, including effects of DCB and isoxaben. Isoxaben treatment resulted in clearance of YFP::AtCESA6 from the membrane seen in the YFP::AtCESA6 rescued atcesa6 mutant. In contrast, DCB was seen to cause hyperaccumulation of YFP::CESA6 in the cell cortex and inhibited motility [34]. Interestingly, in the moss Funaria hygrometrica, rosette structures decreased under DCB treatments as visualized with freeze fracture electron microscopy. At high concentrations of DCB, Funaria protonemal tips ruptured [35].

Thesis Outline:
PpCESA expression was examined through construction and analysis of promoter-reporter constructs and RT-qPCR. PpCESA promoters were fused to βglucuronidase (GUS) for localization of expression. Since no ppcesa KO mutants produced a drastic protonemal phenotype, the importance of cellulose synthesis in tip growth of protonemal filaments was tested through cellulose synthesis inhibitors. Protonemal filaments were treated with cellulose synthesis inhibitors isoxaben and DCB and assayed for tip growth rate and morphology. Both isoxaben and DCB had no effect on tip growth rate. However, at a high concentration of DCB, rupturing of tips was observed. Rupturing of tips caused by DCB indicated the importance of cellulose synthesis in tip growth, but it also indicated that PpCESAs may not be the only contributors to cellulose synthesis.
Cellulose synthase like D proteins are also affected by DCB treatments, indicating that CSLDs may potentially produce cellulose [36].
My work demonstrates that cellulose synthesis is important in tip growth in the

Background:
Cellulose is an abundant biopolymer with many commercial applications, yet the mechanism of its biosynthesis is still being understood. Cellulose synthases (CESAs) have been identified as key proteins in the synthesis of cellulose in plants [3]. There are multiple CESA isoforms found in both vascular and nonvascular plants [2,4] and in both cases the CESAs form rosette structures known as cellulose synthase complexes (CSC) [6,37].
In vascular plants, CSCs are specific for either primary or secondary cell wall formation and are hetero-oligomeric, consisting of three different CESAs [8,9,38].
Along with phenotype analysis of CESA mutants and protein-protein interaction studies, analysis of CESA coexpression patterns provided critical information for understanding the composition of the hetero-oligomeric Arabidopsis CSCs. Initially AtCESA4, ACESA7, and AtCESA8 mutants in Arabidopsis were discovered to have the same phenotype with reduction of cellulose and irregular xylem. Coexpression of these genes was demonstrated by northern blot [39] and microarray analysis [40].
Later, coimmunopreciptation experiments confirmed that the encoded proteins interact, forming the CSCs that synthesize cellulose in the secondary cell wall [38].
Similarly, AtCESA1 and AtCESA3 are essential in primary cell wall formation, along with AtCESA6-like proteins [9]. These AtCESAs are coexpressed based on promoterreporter analysis. Co-immunoprecipitation and bimolecular complementation experiments showed that AtCESA1, AtCESA3, and AtCESA6 interact to form the CSC in the primary cell wall [9,40,41]. No studies on CESA composition of the CSCs of nonvascular plants like P. patens have been done.
Physcomitrella patens, a nonvascular land plant, also has multiple CESA isoforms and rosette CSCs, but it does not have lignified secondary cell walls like vascular plants so the divergence of primary and secondary type CESA may not be expected [1]. Seven CESA genes have been identified in P. patens and phylogenetic analysis indicates that they diversified independently from the seed plant CESAs [4]. However, their individual functions are still unknown [1]. Understanding how the functions of the PpCESAs differ from those of vascular plant CESAs will provide insight into the roles of the distinct CESA isoforms in CSC assembly and function in moss.
Since the P. patens genome has been fully sequenced and can be easily genetically manipulated, it is possible to investigate gene function through knockout mutations and gene expression through promoter-reporter analysis [42]. Prior experiments in P.
patens have indicated that PpCESAs might be involved in different stages of development. PpCESA5 has a developmental role in gametophore formation [11].
Although PpCESA6 and PpCESA7 single knockout mutants do not show obvious developmental phenotypes, double knockouts of PpCESA6 and PpCESA7 have shorter gametophores [12]. With the exception of an analysis of EST abundances in various P.
In this study, we examined PpCESA expression through relative quantitative RT-PCR, analysis of lines transformed with promoter::β-glucuronidase (GUS) reporters, and hierarchical cluster analysis of public microarray data to test for tissue and developmental stage specific CESA expression. We also aimed to determine whether specific P. patens CESAs are coexpressed as in vascular plants such as Arabidopsis and determine whether there are any unique patterns suggesting potential PpCESA functions and interactions.

Materials and Methods:
Vector construction Genomic DNA was extracted from P. patens protonemal tissue grown on solid BCDAT medium as described previously [44]. PpCESA genomic sequences were downloaded from http://www.cosmoss.org/ (

Results:
Promoter::GUS localization young gametophores with 6-10 leaves, and mature gametophores that had stopped producing new leaves. In the young gametophore buds, all CESA promoters were active except for proCESA4 and proCESA10 ( Figure 2). As the buds matured and produced leaves, all CESA promoters except proCESA4 and proCESA10 were active in the apical meristem as shown for proCESA3 ( Figure 2I). In contrast, proCESA4 and proCESA10 showed activity mainly in the axillary hairs of gametophores with 2 to 3 leaves as shown for proCESA4 ( Figure 2J). In two-week-old gametophores with 6 to 10 leaves, all promoters were active in the axillary hairs (

CESA expression levels measured by RT-qPCR
Because some of the PpCESA sequences are very similar, all primers were tested for specificity by PCR using plasmids containing full-length PpCESA cDNA clones as templates and analysis by gel electrophoresis. All primer pairs amplified fragments of the expected size when paired with their corresponding cDNA template and no amplification was seen when primers were paired with other CESA cDNA templates or in no template control reactions (Additional File 1). All primer pairs had efficiencies of 90% to 110% (Table 1). Despite repeated attempts, we were unable to design efficient primers that specifically amplified PpCESA6, which is nearly identical to PpCESA7 throughout the CDS and UTR sequences [12].
RT-qPCR was performed on cultured protonemal tissue and on leafy gametophores isolated by dissection to measure the expression levels in these tissues. The results show that PpCESA10 (P<0.0001) is more highly expressed in the protonemal tissue and PpCESA3, PpCESA5, and PpCESA7 are more highly expressed in the gametophores (P<0.005) ( Figure 6).
To test whether differences in PpCESA expression extrapolated from analysis of EST abundances [4,43,51] are valid, CESA expression was measured by RT-qPCR in tissues that had been induced to differentiate on media containing different nitrogen sources and hormone supplements ( Figure 7). Homogenized protonema was grown for 7 d on medium containing ammonium and nitrate as nitrogen sources (BCDAT), which stimulates protonemal growth, and medium containing only nitrate as a nitrogen source (BCD), which promotes gametophore development. Physcomitrella patens was also grown for 7 d on BCD with added cytokinin, which promotes over-production gametophores, and auxin, which promotes over-production of rhizoids [43,52].
RT-qPCR revealed that PpCESA8 is upregulated in tissue cultured on BCD vs.
BCDAT medium, whereas all other PpCESAs are expressed at equal levels on both media ( Figure 7). This extends the analysis of EST abundance in which only PpCESA3 and PpCESA8 were represented in cDNA libraries from tissue grown on BCDAT medium [4]. The auxin treatment resulted in significant upregulation of PpCESA8 were downregulated in the dark.

Discussion:
Rosette CSC structures are visible with freeze fracture electron microscopy in protonemal tips of P. patens and the related species Funaria hygrometrica [1,37], indicating that cellulose is synthesized in protonemal filaments. High expression of PpCESA10 in the protonema detected by RT-qPCR indicates that PpCESA10 is important for protonemal cellulose synthesis. This is consistent with microarray data [49] analyzed through Genevestigator (Nebion AG), which show high PpCESA10 expression in the protonema (Additional File 2). However, promoter-GUS analysis indicates that all PpCESAs participate in deposition of the protonemal cell wall.
Most PpCESAs appear to participate in cell wall synthesis in gametophores, with expression varying in select locations. All PpCESAs, except for PpCESA10 were moderately to highly expressed in gametophore development through microarray data [49]. Measured by RT-qPCR, PpCESA3, PpCESA5, and PpCESA7 had significantly higher expression in gametophores compared to protonema. Previously, analysis of ESTs suggested that PpCESA4, PpCESA5, PpCESA6 and PpCESA7 are overrepresented in libraries when treated with cytokinin, which promotes gametophore development [4,43]. RT-qPCR confirmed upregulation of these genes, and also PpCESA3 and PpCESA8, under treatment with cytokinin. Mutational analysis has confirmed the role of PpCESA5 in gametophore development [11]. High expression of PpCESA4 under cytokinin treatment is also seen in our RT-qPCR and an overrepresentation of PpCESA4 and PpCESA10 ESTs under cytokinin treatment [4,43] suggests that PpCESA4 and PpCESA10 are also expressed in the gametophore, but very little expression is seen in our promoter::GUS analysis.
The cellulose of the bud and stem of gametophores appears to be predominately PpCESA6, PpCESA7 and PpCESA8 appear to be the main contributors to cellulose synthesis in the rhizoids. This is supported by upregulation of PpCESA7 and PpCESA8 by auxin compared to BCD based on RT-qPCR, and over representation of PpCESA6 and PpCESA8 ESTs in libraries from auxin-treated tissue [4]. Strong histochemical staining of proCESA6::GUS lines in rhizoid tissue and PpCESA6-GFP localization in rhizoids [12], also suggest that PpCESA6 is involved in cellulose synthesis in rhizoids. However, ppcesa6KO, ppcesa7KO, [12] and our analysis of ppcesa8KO and have shown no defect in rhizoid development indicating that PpCESA6, PpCESA7, and PpCESA8 may function redundantly in cellulose synthesis in the rhizoids [12].
Beyond localizing and quantifying gene expression throughout P. patens development, expression analysis may help elucidate interactions between the PpCESAs within hetero-oligomeric CSCs, as in Arapidopsis [9,40]. than ppcesa5KO, indicate that the PpCESAs are interchangeable [12]. Phylogenetic tree of CESAs indicate that PpCESAs are not orthologous in specialized functionality to seed plant CESAs, where they form hetero-oligomeric complexes. Since CESAs are originally homo-oligomeric, it is possible that PpCESAs still form homo-oligomeric, making PpCESAs more functional redundant [1].

Conclusion:
Complex expression of PpCESAs at different times and in different tissues within the gametophore is consistent with functional development of multiple cell types and the need for structural support [1]. Overlapping expression and lack of phenotypes in single PpCESA knockout lines, except for ppcesa5 KO [11], indicates that some of the           Expression levels were measured using qPCR and normalized to PpVhpp and PpACT. Three independent samples were assayed in duplicate for the gametophore and protonema qPCR. Stars indicate significant difference, where * P<0.005 and ** is P<0.0001. CESA10 is most highly expressed in protonemal tissue (blue bars). CESA3, CESA5, and CESA7 are more highly expressed in gametophores (black bars). CESA8 has similar expression in both the gametophores and protonema.

Introduction:
The plant cell wall is very dynamic and has many components, such as pectin, hemicelluloses, arabinogalactan proteins (AGPs), and cellulose. In previous studies, examining the different cell wall polysaccharides has led to the understanding how plant cells grow, divide, and interact with neighboring cells [54]. Cellulose acts as a structural scaffold to the cell wall, while hemicelluloses associate with cellulose to help cell plants growth [29]. Cellulose has been shown to be essential in plant growth from mutational analysis [2]. Hemicellulose links cellulose microfibrils and matrix together through hydrogen bonding and has also shown a role in storage of carbohydrates as mannan [27,55]. Pectin synthesis and modification is highly regulated and been associated with cell wall flexibility [29]. AGPs have been shown to covalently bond to pectin and are necessary for apical cell extension in P. patens [25,54].
Immuno and affinity histochemical techniques using carbohydrate-binding modules (CBMs) and monoclonal antibodies are beneficial in studying these cell wall components in different plant organs and developmental stages [54]. Limited characterization of P. patens cell wall components through immuno-histochemical staining was done in caulonema, chloronema, rhizoids, and gametophores. AGPs and 1, 5-α arabinin are found throughout filamentous tissue and slightly concentrated at tips [25]. Protoplasts were also seen to have 1,5-α arabinan. Low amounts of xylan were seen in throughout gametophores, but more concentrated in axillary hairs.
Xyloglucan was also abundant throughout gametophores [26]. Mannan was found in protonema with deposition concentrated in cell junctions [27]. Callose was found in early developing spores [28] and cellulose was found in developing gametophore buds [11]. No comprehensive study of cell wall composition has been done in all the tissues and few immuno and affinity histochemical studies have been done in protoplasts.
In addition to histochemical staining, high throughput microarrays have been used by others to profile the abundance of cell wall components in different tissues of P.
patens [24]. In protonema high levels of mannan, arabinan, and crystalline cellulose and moderate to low amounts of galactan, nonfucosylated xyloglucan, xylan, and AGPs were detected. The highest amounts of cellulose, arabinan, and AGPs were seen in gametophores and sporophytes [24].
Protoplasts of P. patens represent the simplest form of a plant with only one cell type, consisting of only a single cell with no cell wall. Using this uniform cell type allowed us to quantify labeling intensity efficiently through flow cytometry without needing to sort cell types for data analysis. We allowed the protoplasts to regenerate its cell wall for 24 h. At this point, the protoplasts had deposited their cell walls and some had begun to divide and form a filament. This allowed us to examine the first cell wall components that were deposited during regeneration and the initiation of cell division.
With flow cytometry, we also determined relative abundance of each cell wall components by capturing the mean fluorescent intensity per cell in high volumes.
These results were compared with carbohydrate microarrays [24] Materials and Methods: Moss protonemal tissue was digested into protoplasts and protoplasts were grown on PRMB media for 24 h, according to [44].

Results:
Flow cytometry data.
Populations of round protoplasts were selected based on FSC and SSC and cellular debris was omitted which has very low FSC at 10 2 and SSC at 10 1  Microscopy.
Immuno-labelled protoplasts were examined with a compound microscope.
Protoplasts were sorted as immature, mature round, and dividing protoplasts as well as developing filament.
Crystalline cellulose (CBM3a) is strongly labelled throughout immature, mature round, and dividing protoplasts as well as developing filament ( Figure 3E-H

Discussion:
The most abundant polysaccharide in regenerated protoplasts is crystalline cellulose, which is deposited as the cell matures. Crystalline cellulose content appeared greater in protoplasts than protonema and gametophore tissue from microarray analysis [24].
One of the reasons why crystalline cellulose may appear to be more abundant is that protoplasts, devoid of a cell wall, are grown under high osmotic conditions to prevent rupturing of the cell. The high osmotic media may cause the upregulation of cellulose [13]. Another polysaccharide that is abundant in protoplasts is callose. Callose has been shown to be important in abiotic stress and plant development [57] and was found to be the second most abundant polysaccharide in protoplast in our study.
Protoplasts undergo high levels of abiotic stress from the removal of the cell wall using Drislease enzymes and washing in osmotic media; therefore, high levels of callose are expected.
Among hemicelluloses, mannan is the most abundant in protoplasts. Moderate levels of mannan have been previously seen in both chloronema and caulonema tissue with a stronger deposition between cell junctions [24,27]. Currently, the exact function of mannan is unknown, but there are indications of its role in cell differentiation [54].
Low levels of nonfucosylated xyloglucan were detected in protoplasts with LM15.
Low levels were also seen in protonemal tissue when microarrays were probed with LM15 [24]. However, the shoot axis of the gametophore labelled strongly using CCRC-M88 [26]. CCRC-M88 (National Center for Biomedical Glycomics, Athens, GA, USA) has a higher cross reactivity with XXGG xyloglucan [58] compared to LM15, which has a higher cross reactivity with XXXG xyloglucan [59].
Physcomitrella patens xyloglucan was found to be the XXGG type [60]. Both branched and unbranched xylan was found at low levels in chloronemal tissue [24] but no xylan was found in protoplasts. Only axillary hairs show strong xylan labeling [24,26]. moderate levels of nonesterified homogalacturonan (mAb2F4) in protonemal tissue [24]. Based on our results and microarray data, early stages of P. patens do not contain high levels of homogalacturonan. High levels of α 1, 5-arabinan (LM6) are deposited in mature and dividing protoplasts, which can occur as side chains of either pectin or AGP. There are only low amounts of galactan in young protoplasts. Less galactan is seen in older protoplasts. Because anti-arabinan (LM6) is also believed to associate with AGPs, our data matches very well with high levels of AGPs detected by JIM13 and moderate levels of AGPs through LM2 [25]. Based on microscopy, LM2 labels more strongly than JIM13. Since LM2 and JIM13 recognize different epitopes, different types of AGPs are recognized by the different antibodies and are expected to reveal slightly different profiles of AGPs [61,62].

Conclusion:
P. patens is becoming an ideal model plant for studying the cell wall due to its fully sequenced genome, quick regeneration time, and ability to be genetically manipulated [42]. However, currently there has not been a comprehensive study in the cell wall    Physcomitrella patens is a nonvascular plant whose genome has been fully sequenced [3,4]. It is considered to be a good model organism because of its ability to be genetically manipulated due to its unusually high rate of homologous recombination [5,6]. Physcomitrella patens is a simple moss plant with two haploid stages, a filamentous protonemal stage and gametophore stage, where it produces small leafy stalks [7].
Seven CESAs isoforms have been identified in P. patens [8]. As of now, PpCESA5 [9], PpCESA6, and PpCESA7 [10]  with shorter gametophores seen in the double knockout [10]. The roles of the other PpCESA isoforms are currently unknown; however, none of the single PpCESA knockouts appears to impair the development of the protonemal [11].
The protonemal tissue is a filamentous stage of the moss. Previously, carbohydrate microarrays showed moderate amounts of cellulose in the P. patens cell wall [12]. Cellulose is concentrated at the protonemal filament tips observed through microscopy using cellulose binding module 3 (CBM3a) affinity cytochemistry [13].
Cellulose is also implicated to be important in protonemal growth based on the abundance of rosette cellulose synthase complexes (CSCs) in protonemal tips of Funaria hygrometrica [14] and P. patens [15].
The protonemal filament extends by tip growth similarly to pollen tubes and root hairs of many other species [16,17]. The effect of the cellulose synthesis inhibitor, 2, 6-dichlorobenzonitrile (DCB) on various pollen tubes, such as lily, petunia [18], and Pinus bungeana [19], includes distortion of cell walls and changes in cell wall components, such as an increase in pectin [18,19]. It also causes rupturing of the tips at very high concentrations [18]. Treatment with the cellulose synthesis inhibitor isoxaben caused shorter tips, as well as tip swelling in conifer pollen tubes [20]. These results indicate that although pollen tubes have very little cellulose content, cellulose is necessary in tip growth and development. Arabidopsis root hairs treated with DCB have also caused rupturing of tips and with isoxaben treatments, it caused retarded growth .
Both isoxaben and DCB have been well characterized to inhibit cellulose synthesis, but the mechanism of inhibition is very different. DCB treatment immobilizes AtCESA6-YFP in the plasma membrane, while isoxaben causes accumulation of AtCESA6-YFP in the Golgi vesicles below the membrane [21]. DCB treatment of the protonemal filaments of the moss Funaria hygrometrica caused no changes in tip growth rate and rupturing of tips at high concentrations. Rosette CSCs visualized by freeze fracture electron microscopy also showed irregular distribution after 10 min of treatment in F. hygrometrica [22]. In contrast to results from live cell imaging in Arabidopsis, freeze fracture electron microscopy in F. hygrometrica indicated that rosette CSCs decrease in the plasma membrane with DCB treatment.
Prior experiments have indicated that PpCESAs may be involved in stress responses. Physcomitrella patens upregulates cellulose deposition when subjected to osmotic stress through the addition of mannitol to the culture medium [13]. The thickening of the cell wall has also been observed under drought conditions [23].
Microarray data showed PpCESA6, PpCESA7, and PpCESA8 were expressed at higher levels under mannitol stress [1]. Interestingly, vascular plants downregulate cellulose under osmotic stress [24], and Arabidopsis cesa8 mutants are drought tolerant [25]. This downregulation of cellulose may be beneficial to vascular plants under drought stress.
These opposite responses of osmotic stress-induced cellulose upregulation in P. patens and downregulation in Arabidopsis might be due to differences in water uptake and dehydration tolerance mechanisms. Mosses are poikilohydric, while vascular plants are homeohydric [26]. Poikilohydric mosses, including P. patens, depend on external water to provide hydration. Physcomitrella patens is able to survive becoming dehydrated under water stress [27]. Homeohydric vascular plants, on the other hand, are adapted to maintain a constantly hydrated state through mechanisms that prevent water loss. Physcomitrella patens does not have a vascular system and does not need to maintain a constant water level, so it uses a different mechanism to combat drought [15].
Ppcesa6/7KO and ppcesa8KO mutants were produced and used to investigate the roles of the mutated genes in protonemal development and stress response. Since cellulose upregulation is seen under osmotic stress, the influence of mannitol induced osmotic stress and sensitivity to high salinity treatments in the mutants was analyzed.
Mutant lines were also assayed for developmental phenotypes. None of the mutants had a dramatic phenotype in the protonema. Wildtype protonemal tissues were treated with cellulose synthesis inhibitors, isoxaben and DCB, to examine the role of cellulose in tip growth, since no dramatic phenotype was observed in the protonema of any of the ppcesaKO single knockouts. Protonemal tissue was assessed for tip growth rate, swelling, and rupturing of tips as seen previously in pollen tubes.

Materials and Methods:
Vector construction Double KO mutations of PpCESA6 and PpCESA7 were made instead of single KO mutations because these genes only differ by only 2 amino acids [8,10]. To construct a vector to knockout both PpCESA6 and PpCESA7, sequences upstream of PpCESA6 and downstream of PpCESA7 were amplified from P. patens genomic DNA extracted from wildtype protonemal tissue as previously described [28]. The 5' UTR region of CESA6 was amplified with 0.5 μM primers flanked with attB1 and attB4 sites ( Bioscience) and prepared for transformation into P. patens as described previously [28].
For the cesa8KO vector construction, a hygromycin selection cassette was inserted into a PpCESA8 cDNA clone (Goss & Roberts, 2009). Again, final vector was digested with EcoRI and NsiI (New England Bioscience) and precipitated for transformation [29].
Transformation and genotyping of ppcesa8KO and ppcesa6/7KO. ppcesa8KO and ppcesa6/7KO vectors were transformed into wild-type Grandsen 2011 moss as previously described [28]. Genomic DNA was extracted from stably transformed colonies that survived two rounds of hygromycin selection as described previously [28]. Image stacks were assembled into kymographs using Image J (National Institutes of Health, USA) and distance of tip growth was measured at the base of the slope. The distance of tip growth was divided by time for tip growth rate [33]. All experiments were performed in triplicate with three different biological replicates.

Statistical analysis
ANOVA was used for overall comparison between the lines, and Tukey Kramer's unpaired t-test was used for pairwise comparison if the ANOVA p-value was significant.

RT-qPCR of PpCSLDs and PpCESAs
Four-d-old protonemal tissue plates were split into 3 parts, where 1) has no treatment control, 2) is transferred to BCDAT, and 3) is transferred to PRMB (BCDAT+mannitol). RNA was extracted from all treatments, converted to cDNA, and analyzed using RT-qPCR, as described in Chapter 2.
To test CSLD expression in the protonema, 100 mg of 4-d-old protonema tissue was collected for RNA extraction and converted to cDNA according to Chapter 2.
ppcesa8KO protonema have high solidity Three lines of ppcesa6/7KO mutants, 3 lines of ppcesa8KO mutants, and 3 biological replicates of wildtype tissue were assayed for defects in caulonema, gametophore, and rhizoid development. Both ppcesa6/7KO and ppcesa8KO mutants produced straight caulonema filaments with no difference in length compared to wildtype (P>0.05, Figure 3, Table 2, and Table 3). Gametophores were observed to grow similarly to wildtype in all 3 ppcesa8KO lines ( Figure 5). ppcesa6/7KO mutants all develop gametophores. However, ppcesa6/7KO 7B appears to have shorter and fewer gametophores than wildtype. Since only 1 out of 3 lines has this phenotype, the dwarf gametophore phenotype is not likely due to the KO of PpCESA6 and PpCESA7 ( Figure 6). Rhizoids produced by ppcesa6/7KO and ppcesa8KO mutants were similar to wildtype (Figure 7 and Figure 8).
Three lines of ppcesa6/7KO, 3 lines of cesa8KO, and 3 biological replicates of wildtype tissue were grown from protoplasts to 6-d-old filaments and measured for area, solidity, and perimeter as previously described [32]. Solidity was scored from 0 to 1. The lowest solidity with the highest branching of the filaments was scored 0 and the highest solidity possible with less branching of filaments was scored 1 [32].
ppcesa6/7KO mutants and wildtype lines had no difference in area, solidity, or perimeter growth (Figure 9). However, ppcesa8KO (4C, 7C, and 10C) protonema lines have significantly higher solidity than those of wildtype (P<0.05, Anova and P<0.05, t test for individual lines compared to wildtype), with no difference in area growth and perimeter ( Figure 10).
ppcesa8KO and ppcesa6/7KO mutants show no defect in cellulose upregulation under osmotic stress Three replicates of ppcesa8KO, ppcesa6/7KO, and wildtype lines were grown on BCDAT medium and PRMB, which is BCDAT medium supplemented with mannitol.
Fluorescence intensity was measured for cells grown on mannitol to cells grown on no mannitol, which were distinguished based on cell diameter where mannitol treated cells were much shorter, and ratios were calculated. Fluorescent intensity ratios showed no differences between ppcesa8KO or ppcesa6/7KO mutants to wildtype moss using ANOVA (Figure 11).

PpCESA expression does not change under mannitol treatment
Three replicates of protonema tissue treated with mannitol showed no difference in PpCESA expression compared to no transfer control and no mannitol control ( Figure   12.  Figure 13). However, after 7 d of treatment of NaCl, ppcesa6/7KO showed greater sensitivity to salt compared to wildtype tissue, where chlorophyll content in ppcesa6/7KO decreased dramatically compared to wildtype moss ( Figure 14).

Protonemal growth rate is not inhibited by DCB or isoxaben
Protonemal tissues were treated with isoxaben and DCB for 5 d. No differences were seen in overall colony growth, as seen in Arabidopsis ( Figure 15). Tip growth rate was measured before rupturing of tips through kymographs from 20 μM isoxaben and 20 μM DCB treatments. There is no effect on tip growth rate in treated protonemal tissue (P>0.05) when compared to negative controls of PNO3 with and without diluents ( Figure 16 and Figure 15). However, treatment with 20 μM DCB resulted in rupturing of the protonema tips after less than 20 min of treatment ( Figure 17). No rupturing of tips was seen with treatments of isoxaben ( Figure 17).  [34,35]. Isoxaben did not cause rupturing of tips even at very high concentrations. These differences maybe be due to the different mechanism of inhibiting cellulose where DCB inhibits motility in the plasma membrane causing an accumulation of AtCESA6-YFP particles [21]. Interestingly, DCB has similar affect on protonemal filaments in Funaria in causing rupturing of tips. However, rosette structures that are believed to be the CSCs were decreased in the membrane based on freeze fractured experiments after 10 min of treatment [22]. Although the effects of DCB are the same in Arabidopsis and the moss Funaria, the mechanism in its affect on CSCs may be different. We are currently planning future experiments using ppcesa8KO rescued with GFP-PpCESA8 lines to monitor PpCESA trafficking under treatments of DCB and isoxaben. These experiments will inform us if the mechanism of cellulose inhibition of DCB and isoxaben is the same in Arabidopsis.

Discussion
Another explanation for the rupturing of tips with DCB, but not isoxaben, may be that DCB affects cellulose synthase like D proteins (CSLDs) whereas isoxaben has no affect on CSLDs. Recent studies have shown CSLD's potential in producing cellulose [36]. Atcsld3KO mutants have shown rupturing of tips in root hairs [36], while atcsld1 and atcsld4 mutants shown rupturing in pollen tubes [37], similar to the DCB treatments. Some atcsld mutations have also resulted in decreased cellulose content based on S4B staining [36,37]. Furthermore, the catalytic subunit of CSLDs have been shown to be interchangable with AtCESA6, suggesting that CSLDs are able to synthesize cellulose [36].
PpCSLDs are highly expressed in the protonema as seen by RT-qPCR, particularly PpCSLD1 ( Figure 18). CSLDs might be responsible for cellulose synthesis in tip growth based on high expression of CSLDs and rupturing of tips after treatments with DCB, which inhibits CSLDs as well as CESAs. These results indicate the importance of cellulose in protonemal tip growth and that CSLDs may possibly have a role in cellulose synthesis.

Conclusions:
These data suggest that P. patens is very resilient to the absence of cellulose in protonemal development, since PpCESA knockout mutants have yet to cause lethality in protonema tissue. In contrast, Arabidopsis single knockout mutants atcesa1 and atcesa3 have caused gametophytic lethals [38] and atcesa4, atcesa7, and atcesa8 cause collasped xylem [39,40]. The cellulose synthesis inhibitor isoxaben also causes no defects to protonemal tip growth. Only treatment with a high concentration of DCB had an effect, where the protonemal tips ruptured. This suggests the PpCESAs and/or PpCSLDs can compensate for each other. PpCSLDs needs to be further investigated for its role in synthesizing cellulose in the protonema. PpCESA single KO mutants should also be investigated for upregulation of other PpCESA isoforms and multiple PpCESA KO mutations may be necessary to produce a phenotype in the protonemal tissue.            . No morphology differences seen in ppcesa6/7KO compared to wildtype moss ppcesa6/7KO mutant lines and wildtype moss were analyzed for (A) area, (B) perimeter growth, and (C) solidity, according to [31]. Solidity scale of 1 represents the highest solidity with 0 as the lowest solidity with the highest branching. Error bars display standard error of the mean between each data set. No significant differences were found (P>0.05). P<0.05 Figure 10. ppcesa8KO mutants have higher solidity than wildtype moss ppcesa8KO mutant lines and wildtype were analyzed for (A) area, (B) perimeter growth, and (C) solidity, according to [31]. Solidity scale of 1 represents the highest solidity with 0 as the lowest solidity with the highest branching. Error bars display standard error of the mean between each data set. No significant differences were found in area and perimeter growth (P>0.05). ppcesa8KO lines: 4C, 7C, and 10C have significantly higher solidity compared to wildtype (Anova P value < 0.05). Three lines of (A) ppcesa8KO mutants, (B) ppcesa6/7KO mutants, and (C) wildtype protoplasts were grown on PRMB for 2 days and transferred to BCDAT for another 2 days. Regenerated filaments were collected and fix according toAW Roberts, CS Dimos, MJ Budziszek, Jr., CA Goss and V Lai [28]. Filaments were then stained with cellulose binding affinity antibody, CBM3A, and with AlexaFluor488 secondary. Images of filaments were captured and (D) ratios of fluorescences were measured where they show no significant differences between mutants and wildtype (Anova P value >0.05).

Figure 12. PpCESAs expression do not upregulation in mannitol
Expression levels were measured using qPCR and were normalized to PpVhpp and PpACT. Three independent samples were assayed in duplicate. No significant differences were seen in mannitol treated tissue compared BCDAT and no transfer control (Anova P<0.05).

Figure 18. Expression of CSLDs in protonemal tissue
Expression levels of CSLDs in protonemal tissue were measured using qPCR and were normalized to PpC45 and PpUB1. Three independent samples were assayed in duplicate. Results show very high expression of CSLD1 in the protonemal tissue and low to moderate expression of the other CSLDs.