•The formation of secondary cell walls in cell types such as tracheary elements and fibers is a defining characteristic of vascular plants. The Arabidopsis transcription factor KNAT7 is a component of a transcription network that regulates secondary cell wall biosynthesis, but its function has remained unclear.
•We conducted anatomical, biochemical and molecular phenotypic analyses of Arabidopsis knat7 loss-of-function alleles, KNAT7 over-expression lines and knat7 lines expressing poplar KNAT7.
•KNAT7 was strongly expressed in concert with secondary wall formation in Arabidopsis and poplar. Arabidopsis knat7 loss-of-function alleles exhibited irregular xylem phenotypes, but also showed increased secondary cell wall thickness in fibers. Increased commitment to secondary cell wall biosynthesis was accompanied by increased lignin content and elevated expression of secondary cell wall biosynthetic genes. KNAT7 over-expression resulted in thinner interfascicular fiber cell walls.
•Taken together with data demonstrating that KNAT7 is a transcriptional repressor, we hypothesize that KNAT7 is a negative regulator of secondary wall biosynthesis, and functions in a negative feedback loop that represses metabolically inappropriate commitment to secondary wall formation, thereby maintaining metabolic homeostasis. The conservation of the KNAT7 regulatory module in poplar suggests new ways to manipulate secondary cell wall deposition for improvement of bioenergy traits in this tree.
Plant secondary cell walls are deposited in specialized cell types, such as vessels, tracheids and fiber cells, during cell maturation after cell expansion has ceased. Secondary cell walls are complex matrices of cellulose, hemicelluloses and lignin, and are a considerable sink for fixed carbon derived from photosynthesis (Pauly & Keegstra, 2008). In Arabidopsis thaliana (Arabidopsis), significant metabolic commitment to secondary wall deposition occurs during the maturation of xylem vessels, xylary fiber cells and interfascicular fiber cells in the developing inflorescence stem. Wood formation in trees such as Populus sp. (poplar) involves massive deposition of secondary walls in secondary xylem as vessel and fiber cells derived from the cambium cease to expand and mature into functional cells (Mellerowicz et al., 2001; Plomion et al., 2001; Savidge, 2001; Jansson & Douglas, 2007).
The Arabidopsis inflorescence stem, root and in vitro secondary cell wall induction systems have all been used as models to explore the regulatory networks underlying the coordinated control of genes encoding enzymes required for secondary cell wall polymer biosynthesis (Kubo et al., 2005; Zhong & Ye, 2007; Zhong et al., 2008; Yamaguchi et al., 2011). A set of NAC domain transcription factors has been shown to act as a positive master regulator of secondary cell wall biosynthesis in Arabidopsis, and homologs of these are found in other plant species, for example Populus (Demura & Ye, 2010; Zhong et al., 2010a). These positive regulators include the functionally redundant VASCULAR-RELATED NAC-DOMAIN7 (VND7) and VND6 transcription factors associated with secondary cell wall formation in protoxylem and metaxylem vessels (Kubo et al., 2005; Yamaguchi et al., 2008, 2011) and SECONDARY WALL-ASSOCIATED NAC DOMAIN1 (SND1) and NAC SECONDARY CELL WALL THICKENING PROMOTING FACTOR1 (NST1) transcription factors that activate secondary cell wall formation in interfascicular fiber and silique valve cells (Zhong et al., 2006, 2008; Mitsuda et al., 2007; Ohashi-Ito et al., 2010).
The Class II KNAT7 gene has gained attention for its potential role in the transcriptional network regulating secondary cell wall biosynthesis. KNAT7 expression is strongly up-regulated in concert with secondary cell wall development in the Arabidopsis inflorescence stem (Ehlting et al., 2005; Zhong et al., 2008), is highly co-expressed with secondary cell wall biosynthetic enzymes (Brown et al., 2005; Ehlting et al., 2005; Persson et al., 2005) and is one of the direct targets of both SND1/VND6 (Zhong et al., 2008) and MYB46 (Ko et al., 2009). Further evidence for the role of KNAT7 as a regulator of secondary wall biosynthesis comes from the irregular xylem (irx) phenotype of a knat7 loss-of-function mutant (Brown et al., 2005), which led to its earlier designation as IRX11 (Brown et al., 2005). In addition, expression of an engineered dominant transcriptional repression variant of KNAT7 in transgenic Arabidopsis resulted in thinner interfascicular cell walls, but no irx phenotype (Zhong et al., 2008).
Other than these data, the role played by the KNAT7 transcription factor in the regulatory network governing secondary cell wall biosynthesis has remained unclear, and detailed analyses of knat7/irx11 loss-of-function mutants, beyond the preliminary results of Brown et al. (2005), have not been reported. Recent clues with regard to KNAT7 function come from the demonstration that it acts as a potent transcriptional repressor in Arabidopsis leaf protoplasts, and that repression is enhanced by physical interaction with OVATE FAMILY PROTEIN4 (OFP4) (Li et al., 2011). ofp4 loss-of-function alleles have an irx phenotype and thinner interfascicular fiber cell walls (Li et al., 2011). KNAT7 also physically interacts with MYB75, which regulates plant secondary cell wall biosynthesis (Bhargava et al., 2010). myb75 loss-of-function mutants have a higher lignin content and increased expression of a suite of lignin, cellulose and hemicellulose biosynthetic genes, suggesting that it may play a role in limiting carbon flux into these pathways (Bhargava et al., 2010).
Here, we describe in detail the phenotypes of knat7 loss-of-function mutants. In addition to irx phenotypes, these mutants have significantly thicker interfascicular fiber cell walls, increased lignin content and enhanced expression of representative secondary cell wall cellulose, lignin and xylan biosynthetic genes. Furthermore, a poplar KNAT7 ortholog with the same function was identified. These results are consistent with previous observations showing that KNAT7 is a transcriptional repressor. We hypothesize that KNAT7, in concert with interacting proteins, forms a negative feedback loop that maintains cellular homeostasis during developmental commitment to secondary cell wall deposition. By balancing the powerful feed-forward transcriptional activation activity of other transcription factors in the regulatory network, KNAT7 and interacting proteins may provide a mechanism that helps match metabolic demands with resource availability.
Materials and Methods
Arabidopsis thaliana (L.) Heynh. (Col-0) plants were grown and transgenic plants were generated as described previously (Li et al., 2011). Inflorescence stems at different ages of development were sampled at 0–3, 3–5, 5–7 and 7–9 cm from the inflorescence stem apex, as described previously (Ehlting et al., 2005), and for some phenotypic analyses were sampled 2 cm from the bottom of 6–8-wk-old stems.
Cuttings of Populus trichocarpa Torr. & Gray × Populus deltoides Bartr genotype H11-11 were propagated in the glasshouse as described previously (Ralph et al., 2006). Trees, 60–70 cm in height, were used after 1 month of outdoor growth. Field-grown Populus balsamifera trees were sampled in the spring of 2003 in the University of New Brunswick Forest, Fredericton, NB, Canada during the months of May and June.
T-DNA insertion alleles for selected Arabidopsis genes were identified using the SIGnal database (http://signal.salk.edu/), obtained from the Arabidopsis Resource Center (ABRC, Columbia, OH, USA). T-DNA insertions were confirmed using flanking gene-specific primers (Supporting Information Table S1) and SALK line T-DNA left border LBb1 (5′TCAAACAGGATTTTCGCCTGCT-3′) primer. To verify the T-DNA insertion locations and effects of gene expression, PCR products were sequenced and reverse transcription-polymerase chain reaction (RT-PCR) analysis of KNAT7 expression was performed using the primers 62990F (5′-TGCAATCCTGTCTCCTCCACCAAT-3′) and 62990R (5′-GCAAATGGCCTTACCCTACGGAA-3′). Homozygous lines were identified for phenotype characterization.
RNA isolation and RT-PCR gene expression analysis
Poplar RNA for quantitative RT-PCR was isolated from the same sets of pooled tissues harvested from poplar sapling internodes or developing secondary xylem following the protocol of Kolosova et al. (2004). For quantitative RT-PCR of poplar tissues, PCR amplification was performed with gene-specific primers (Table S2), threshold detection cycles (CT) were normalized using the reference gene C672 (Ralph et al., 2006) and CT values were used to generate ΔCT values. ΔCT values for each gene were compared with the CT value obtained for xylem sample ‘G’ or internode sample ‘T5’ to generate ΔΔC(T). Each quantitative RT-PCR assay was reproduced at least three times using independently generated cDNA templates.
Arabidopsis total RNA for quantitative RT-PCR was isolated from the lower half of inflorescence stems with three biological replicates (each consisting of pooled stems from eight to ten plants) and reverse transcribed as described by Li et al. (2011). PCR amplification was performed with gene-specific primers (Table S3) and ACTIN1 (At2g37620) was used as the reference gene. Differences in gene expression, expressed as the fold change relative to the control, were calculated as described previously (Bhargava et al., 2010). Each measurement was carried out in triplicate.
β-Glucuronidase (GUS) expression assays
A 2-kb Arabidopsis genomic DNA fragment, 5′ to the KNAT7 translation start site, was amplified by PCR using gene-specific primers, and was cloned into the binary vector pCambia 1305.1 (Genbank Accession AF354045, Cambia, Brisbane, Australia) to create promKNAT7::GUS. Following Agrobacterium-mediated transformation, six independent prom-KNAT-GUS transgenic lines were analyzed. GUS activity was assayed histochemically as described by Li et al. (2011). For root expression analysis of promKNAT7::GUS, the root from a stained 4-d-old seedling was embedded in Spurr’s resin and sectioned with a microtome to obtain 10-μm-thick sections.
Hand sections from inflorescence stems of 6–8-wk-old Arabidopsis plants were stained with 1% toluidine blue for OLYMPUS AX70 light microscopy (Olympus Canada, Richmond Hill, ON, Canada). Poplar internodes were sectioned using a vibratome, and UV autofluorescence was examined using a Zeiss Axioplan microscope.
Tissue embedding for light and transmission electron microscopy was carried out as described by Li et al. (2011). Samples were taken 2 cm from the bases of inflorescence stems from 8-wk-old plants. For secondary cell wall thickness determination, measurements were taken from 100 separate cells in light micrograph images from 1-μm inflorescence stem cross-sections of wild-type and knat7-1 inflorescence stem bases. The stages from springtime resumption of cell division activity to maturation of earlywood elements in poplar secondary xylem were determined by bright-field microscopy examination of hand sections prepared from both ends of stem segments a few minutes before bark peeling of the segments to expose cambium (surface of inner bark face) and developing xylem (surface of wood), which were then scraped from the surfaces and frozen in liquid nitrogen.
Poplar KNAT7::GFP construction, complementation and green fluorescence protein (GFP) visualization
To generate a poplar KNAT7 fusion under the control of the 35S promoter (35S::PoptrKNAT::GFP), the full-length open-reading frame of poplar KNAT7 (POPTR_0001s08550; http://www.phytozome.net) was amplified from poplar (P. trichocarpa) cDNA using the primers CDSPtKNAT7F (5′-ATGCAAGAACCAAACTTGGGCA-3′) and CDSPtKNAT7R (5′-CTACCTTTTGCGCTTGGACTTC-3′), the amplicon was cloned into pCR8/GW/TOPO (Invitrogen, http://www.invitrogen.com) and subcloned into Gateway plant transformation destination binary vector pGWB6 by LR recombination reactions (Invitrogen). The construct was transformed into knat7-1 by Agrobacterium-mediated transformation. Four representative transgenic lines were used for further studies. Phenotypes were examined in hand cross-sections prepared from inflorescence stems. To visualize GFP localization, roots from transgenic Arabidopsis 7-d-old seedlings were stained for 1–2 min in an aqueous solution containing 2 mg ml−1 propidium iodide. The roots were rinsed and mounted in distilled water and viewed using a Zeiss LSM5 PASCAL confocal laser scanning microscope equipped with 488-nm argon and 543-nm helium–neon lasers. Images were managed with ImageJ software (http://rsb.info.nih.gov/ij/index.html).
Expression analysis of KNAT7
Data from microarray experiments show that KNAT7 is expressed in a stele-specific manner in roots, and up-regulated during the course of vascular and fiber differentiation in the inflorescence stem (Birnbaum et al., 2003; Ehlting et al., 2005). Other transcriptional profiling experiments have shown that KNAT7 is co-expressed with CELLULOSE SYNTHASE A (CESA) genes involved in secondary cell wall biosynthesis (Brown et al., 2005; Persson et al., 2005), and is preferentially expressed in fiber and xylem cells in the Arabidopsis inflorescence stem (Zhong et al., 2008) in coordination with transcription factors of the NAC and MYB classes that regulate secondary wall biosynthesis (Zhong et al., 2008). Expression of a promKNAT7::GUS transgene in Arabidopsis showed that the KNAT7 promoter directs GUS expression in vascular tissue in leaves, in cortex cells adjacent to interfascicular fibers and in developing vascular bundles in inflorescence stems (Li et al., 2011). We extended these results in multiple lines expressing promKNAT7::GUS. In seedlings, strong GUS activity was observed in the vascular systems of roots, hypocotyl and cotyledons, and in the stele of young roots (Fig. 1a–c). In inflorescence stems, GUS activity was associated with developing metaxylem in vascular bundles near the apex of the stems, but was also consistently observed in the phloem and cambial regions and in cortical cells (Fig. 1d,e). This expression pattern was retained in more mature regions of inflorescence stems in which interfascicular fibers were differentiating (Fig. 1f,h), as well as in mature stems with fully differentiated interfascicular fibers (Fig. 1g,i). These results confirm a correlation between secondary wall formation and KNAT7 expression, but also indicate that the KNAT7 promoter is active in cells adjacent to fiber cells as they develop thick secondary cell walls.
KNAT7 loss-of-function mutants
The irx phenotype of a KNAT7 T-DNA insertion mutant SALK_002098 (knat7-1) in Arabidopsis inflorescence stems has been described previously (Brown et al., 2005). In addition, ectopic expression of a dominant repression KNAT7 derivative, generated by fusion of a KNAT7 cDNA to the strong EAR transcription repression domain, resulted in plants with thinner interfascicular and xylary fiber secondary cell walls (Zhong et al., 2008). However, interpretation of the latter results may be confounded by the recent observation that KNAT7 is a strong transcriptional repressor (Li et al., 2011). In order to analyze the loss-of-function phenotype of KNAT7 in more detail, we identified five independent T-DNA insertion mutant alleles of KNAT7 (At1g62990). These included SALK_002098 (knat7-1), WiscDsLox367A5 (knat7-2), SALK_110899 (knat7-3), SAIL_757 (knat7-4) and WiscDsLox461 (knat7-5). Insertions in knat7-2 and knat7-3 were located within the second intron of KNAT7, whereas the T-DNA insertions in other alleles, including knat7-1, were located in the fourth intron (Fig. 2). The results from RT-PCR analysis indicated that the full transcript of KNAT7 was undetectable in plants homozygous for the knat7-1 allele (Fig. 2a), and similar results were obtained for other alleles (Fig. 2b), with very low to undetectable levels of KNAT7 mRNA of the appropriate length seen. This is consistent with the analysis of Ülker et al. (2008), who found that T-DNA insertions in introns can lead to moderate to severe decreases in wild-type mRNA abundance. For knat7-1, we also failed to detect any truncated KNAT7 transcripts (data not shown). These data suggest that these five alleles are strong loss-of-function alleles of KNAT7.
knat7 mutants exhibit both irx and enhanced fiber cell wall thickness phenotypes
As reported previously (Brown et al., 2005), all plants homozygous for knat7-1 showed a strong irx phenotype in hand sections taken from the base of inflorescence stems of 6-wk-old plants (Fig. 3b). As plants homozygous for all additional knat7 alleles showed a similar irx phenotype (Fig. 3c–f), knat7-1 was used for all further phenotypic analyses.
To view the anatomy and morphology of developing knat7-1 inflorescence stems relative to the wild-type, thin sections from the bases of inflorescence stems were examined. As expected, knat7-1 xylem vessels at the base of inflorescence stems exhibited a collapsed perimeter morphology (Fig. 4c,d), relative to the normally round shapes of such vessels in wild-type plants (Fig. 4a,b). In addition, knat7-1 mutant xylary fibers appeared to have a collapsed morphology (Fig. 4d), a phenotype not described previously. Furthermore, comparison of cross-sections taken from knat7-1 and wild-type plants revealed a noticeable increase in the thickness of interfascicular fiber secondary cell walls in the mutant relative to those in sections of wild-type stems at a similar developmental stage (Fig. 4d).
To obtain a more complete picture of the cell morphology, we prepared longitudinal sections from the base of wild-type and knat7-1 inflorescence stems, focusing on the interfascicular fiber (Fig. 4e,g) and vascular bundle (Fig. 4f,h) regions of the stem. In these sections, knat7-1 interfascicular fiber cell walls also appeared to be thicker, and the fiber cells appeared to be less regular in organization. However, spiral secondary cell wall thickenings could be observed in knat7-1 xylem vessels, despite the predominant irx phenotype of the mutant, suggesting that secondary wall formation is qualitatively similar to that of the wild-type.
To assess the potential changes in cell wall thickness and morphology at higher resolution, we used transmission electron microscopy. These analyses, shown in Fig. 5, confirmed the knat7 phenotypes observed using light microscopy (Fig. 4a–d). Both vessel elements and xylary fibers showed obvious collapsed cell morphologies in the knat7-1 mutant (Fig. 5g,h). By contrast, interfascicular fibers showed no evidence of changes in morphology, but revealed significantly enhanced secondary cell wall thickness (Fig. 5e,f).
We performed measurements on multiple cells in high-magnification light microscopy images obtained from thin sections of wild-type and knat7 mutant plants to quantify the interfascicular fiber, xylary fiber and vessel cell wall thicknesses. This analysis (Table 1) demonstrated a c. 60% increase in cell wall thickness in the knat7-1 interfascicular fibers relative to the wild-type, and xylary fiber walls were c. 30% thicker than in the wild-type. By contrast, the cell walls of collapsed vessel elements were significantly thinner in the mutant relative to the wild-type. Examination of stem cross-sections of plants homozygous for additional knat7 alleles (Fig. 3) revealed that they also exhibited thicker interfascicular fiber cell walls compared with wild-type plants.
Table 1. Secondary cell wall thickness in wild-type and knat7-1 Arabidopsis thaliana stems
Interfascicular fiber wall thickness (μm)1
Vessel wall thickness (μm)1
Xylary fiber wall thickness (μm)1
1Data are means ± SD from at least 100 cells measured from light micrographs of the bases of primary inflorescence stems.
2Cell wall thickness differences in knat7-1 were significant (P <0.001; Student’s t-test).
0.95 ± 0.17
0.97 ± 0.21
0.61 ± 0.15
1.50 ± 0.172
0.63 ± 0.152
0.80 ± 0.172
KNAT7 over-expression phenotype
We next studied the effects of the over-expression of KNAT7 under the control of the parsley 4CL1 promoter (prom4CL1:KNAT7), which directs expression to cells with secondary cell wall thickening (Hauffe et al., 1991). Plant growth and development inflorescence stem anatomy phenotypes relative to wild-type plants were assayed in T2 lines obtained following Agrobacterium-mediated transformation. Although no obvious developmental phenotypes were observed in prom4CL1:KNAT7 lines with enhanced KNAT7 expression, a pronounced anatomical phenotype was observed in inflorescence stems relative to the wild-type in five individuals from each of two such lines (Fig. 6). Vessels in plants of prom4CL1:KNAT7 lines appeared normal (Fig. 6a,c), but there was a striking decrease in the secondary wall thickness of the interfascicular fibers (Fig. 6d) compared with the wild-type (Fig. 6b). Measurements of cell wall thickness in these images supported this observation (wild-type interfascicular fiber walls, 1.84 ± 0.29 μm vs prom4CL1:KNAT7 interfascicular fiber walls, 0.80 ± 0.21 μm; n ≥ 50 cells from two individual transgenic lines). No changes in thickness were observed in the secondary walls of xylary fibers and vessels (Fig. 6a,c). This interfascicular fiber cell wall phenotype is opposite to that of KNAT7 loss-of-function mutants, and is similar to the phenotype reported for plants expressing a KNAT7-EAR domain fusion with enhanced transcriptional repression (Zhong et al., 2008).
Gene expression profiling of knat7 plants
We targeted two sets of genes for expression profiling in the knat7-1 background relative to the wild-type. The first set (Fig. 7a) included a suite of structural genes involved in cellulose and hemicellulose biosynthesis (CESA, IRX8, IRX9, IRX10 and FRA8). The expression of CESA genes involved in primary cell wall cellulose biosynthesis (CESA1, CESA3, CESA6) was hardly affected by a loss-of-KNAT7 function. However, significant up-regulation of at least two CESA genes involved in secondary cell wall biosynthesis (CESA7 and CESA8) was observed, with more modest up-regulation of CESA4. Three of four genes involved in hemicellulose biosynthesis (IRX8, IRX9 and FRA8) also showed increased transcript abundance in knat7 plants. However, carbohydrate analysis of knat7-1 inflorescence stems failed to reveal major differences in hemicellulose sugar levels (Supplementary Information Fig. S1).
A second set included representative genes encoding enzymes involved in monolignol biosynthesis (Fig. 7b). Most of these genes were up-regulated in the knat7 mutant, suggesting that a primary effect of KNAT7 may be to directly or indirectly regulate genes in this pathway. To further test this hypothesis, we analyzed the lignin content in knat7-1 and wild-type inflorescence stems (Fig. 8). This analysis showed that knat7 mutant stems had higher lignin content than wild-type stems, and further suggests that one function of KNAT7 is to repress lignin biosynthesis.
Poplar ortholog of Arabidopsis KNAT7
The results presented above, and those of Li et al. (2011), suggest that KNAT7 is a transcriptional repressor that functions to down-regulate secondary cell wall biosynthesis and, more specifically, lignin biosynthesis. This finding is of potential significance with respect to the modification of secondary cell walls in biomass crops, including Populus spp. (poplar, cottonwood, aspen). The KNOX gene family in the fully sequenced genome of P. trichocarpa has been shown previously (Groover et al., 2006; Du et al., 2009) to contain a single potential KNAT7 ortholog within the Class II clade of KNOX genes. We independently retrieved KNAT7 homologs from version 1.1 of the P. trichocarpa genome sequence using AtKNAT7 as a query in a BLASTP search. After alignment of the retrieved and annotated poplar and Arabidopsis KNOX protein amino acid sequences, we generated a rooted maximum likelihood tree that, in agreement with Groover et al. (2006), shows a single Class II poplar KNOX gene that is most closely related to Arabidopsis KNAT7 (Fig. 9). The systematic P. trichocarpa genome version 2.2 (http://www.phytozome.net/) name for this gene is POPTR_0001s08550, which we designate as poplar KNAT7.
To determine whether poplar KNAT7 expression is up-regulated over the course of secondary cell wall synthesis during secondary xylem and wood formation, in concert with other regulatory genes, we used RNA isolated from stem sections taken from the apices of poplar shoots undergoing the transition from primary to secondary growth, as well as RNA isolated from secondary xylem tissue isolated at different times following the post-dormancy reactivation of the vascular cambium.
In the first experiment, we sampled internodes from the apices of 4-month-old poplar (P. trichocarpa × P. deltoides, clone H11) as shown in Fig. 10(a). The internode subtended by the first fully expanded leaf was designated internode ‘T1’, and internodes above were numbered sequentially from T2 to T5; T5 was reproducibly the youngest macroscopic internode that could be isolated. Hand sections from multiple ramets were taken from the centers of nodes T2–T5 and examined with UV excitation to detect autofluorescence associated with lignification. Rings of extensively lignified secondary xylem were apparent in internode T1 and older nodes (not shown), and the second internode above the fully expanded leaf (T2, Fig. 10a) was the first internode in which secondary xylem was clearly obvious. Sections from internodes T4 and T5, nearest the shoot apex, showed only individual vascular bundles without a vascular cambium or secondary xylem (Fig. 10a), whereas, in section T3, an incipient vascular cambium was observed, but little if any secondary xylem was obvious. Thus, the onset of the developmental change from primary to secondary growth occurred mainly between internodes T3 and T4.
In the second experiment, we sampled developing secondary xylem during and after spring re-activation of cambial cell division activity in wild naturally grown P. balsamifera trees. In these trees, progression of secondary growth from dividing fusiform cambial cells to radially expanding primary-walled cambial derivatives, to cells actively producing secondary walls and lignifying as they differentiated into fibers and vessel elements, was observed (Fig. 10b). By performing microscopy of live cambium and sampling cambial derivatives on the wood surface following bark peeling at different times and positions within trees, we were able to isolate cambial derivatives at different stages of xylem cell development. Samples enriched for cells with different cell wall and cell expansion characteristics were isolated from separate trees as shown in Table 2. Samples G, O, M and CC were enriched for cambial derivatives along a developmental gradient from those with primary walls only (sample G) to those with lignifying secondary cell walls in both developing fibers and developing vessel elements (sample CC).
Table 2. Populus balsamifera tissues used for RNA isolation and expression analysis
Earliest stage of primary wall expansion of cambial derivatives that will become the first vessel elements
Cambial derivatives adjoining mature secondary xylem latewood in the process of primary wall radial expansion
Secondary wall polysaccharide deposition underway in enlarged vessel elements and younger cambial derivatives are undergoing primary wall expansion. Lignification not started and fiber differentiation yet to commence in adjoining cambial derivatives
A mixture of primary-walled radially expanding cambial derivatives, developing fibers and vessel elements actively producing lignifying secondary walls, and protoplasmically autolyzing elements having fully lignified secondary walls. Lignification at this stage is most obvious in vessel elements
We assayed the expression of KNAT7 and selected MYB transcription factor genes in these two developmental series. The two poplar MYB genes chosen, MYB018 (POPTR_0004s08480) and MYB028 (POPTR_0005s09930), are homologs of Arabidopsis MYB20/MYB43 and MYB58/MYB63, respectively (Wilkins et al., 2009), both of which have been implicated as positive regulators of secondary wall formation and targets of the master regulator SND1 (Zhong et al., 2008; Zhou et al., 2009). Increases in expression of all three genes were observed in coordination with the onset of secondary wall formation in both experiments (Fig. 10c). Very little expression was observed in samples with minimal commitment to secondary wall formation (samples G and T5), and increased to a maximum in those with greatest commitment to secondary wall formation (samples M, CC, T3, T2).
Poplar KNAT7::GFP expression rescues the knat7 phenotype and is localized to the nucleus in Arabidopsis
To establish the subcellular localization of poplar KNAT7, and to determine whether poplar KNAT7 can replace the Arabidopsis KNAT7 function, we prepared a C-terminal fusion of GFP to poplar KNAT7 and placed it under the control of the 35S promoter to generate 35S::PoptrKNAT7::GFP. This transgene was introduced into the knat7-1 mutant background. Fig. 11(a) shows the expression levels in four representative lines, where line 3 shows the highest expression level, and transgene expression is barely detectable in line 1. F2 plants from the four lines were identified that were homozygous for the knat7-1 mutation and contained 35S::PoptrKNAT7::GFP. Phenotypic analysis of mature parts of inflorescence stems taken from such individuals from lines 1 and 3 are shown in Fig. 11(b). Although no rescue of the knat7 irx phenotype was observed in line 1, the irx phenotype was abolished in line 3, which expresses the transgene at a higher level. Thus, poplar KNAT7 appears to be capable of replacing the Arabidopsis KNAT7 function, suggesting that the genes are true orthologs. Furthermore, this result demonstrates that the poplar KNAT7-GFP fusion protein retains KNAT7 activity. We monitored GFP fluorescence by confocal microscopy in the roots of Arabidopsis plants transformed with PoptrKNAT7::GFP. Fig. 11(c) shows that GFP fluorescence was restricted to the nucleus of root cells, indicating that the poplar transcription factor is targeted to the nucleus and presumably functions in that subcellular compartment.
A role for KNAT7 in the regulation of secondary wall formation was originally proposed by Brown et al. (2005), based on the irx phenotype of a knat7/irx11 mutant, and the observation that KNAT7 is co-expressed with CESA genes implicated it in secondary cell wall biosynthesis. Parallel and subsequent studies (Ehlting et al., 2005; Persson et al., 2005; Zhong et al., 2008) also implicated KNAT7 as a member of a transcription network regulating secondary wall formation in Arabidopsis. However, the specific function of KNAT7 in this process has remained unclear. The knat7 loss-of-function irx phenotype suggests a defect in secondary cell wall biosynthesis. Although knat7 plants exhibit a slight decrease in cellulose content when measured at the base of inflorescence stems, this decrease is not as striking as in other irx mutants (irx6, irx7, irx8, irx9; Brown et al., 2005), and the noncellulosic carbohydrate composition (e.g. xylose and glucose) is similar in knat7 and wild-type stems (Brown et al., 2005).
Our data, combined with those of Li et al. (2011), suggest that KNAT7 plays a unique role in the secondary cell wall regulatory network, functioning in a regulatory module to repress secondary wall biosynthesis, rather than to activate it, as has been assumed previously (Zhong et al., 2008). Thus, the KNAT7 transcription factor may contribute to a negative feedback loop that functions to fine tune metabolic commitment to secondary cell wall biosynthesis during development. Key data in support of this hypothesis are as follows. First, that KNAT7 functions as a transcriptional repressor rather than as an activator, and functionally interacts with potent OFP transcriptional repressors (Li et al., 2011). Second, interfascicular fiber secondary cell walls in knat7 mutants are significantly thicker than in the wild-type (Figs 3, 4), whereas KNAT7 over-expression lines exhibit thinner interfascicular fiber secondary cell walls (Fig. 6). Third, the expression of a suite of lignin biosynthetic genes, as well as certain CesA and hemicellulose biosynthetic genes involved in secondary wall deposition, is up-regulated in knat7 (Fig. 7). Fourth, the lignin content in knat7 inflorescence stems is significantly higher than in a knat7 mutant (Fig. 8). Finally, the identification of a poplar KNAT7 ortholog (Figs 9, 11) that is strongly up-regulated in concert with secondary wall formation during secondary growth in poplar (Fig. 10), suggests that KNAT7-mediated fine tuning of secondary cell wall deposition is part of an evolutionarily conserved mechanism governing secondary wall deposition.
KNAT7 loss-of-function and gain-of-function phenotypes
Our results confirm the knat7 irx loss-of-function phenotype in multiple independent T-DNA insertion alleles (Fig. 3), as originally described by Brown et al. (2005). However, we also observed a consistent and contrasting phenotype of thicker secondary cell walls in interfascicular fiber cells in these mutants, as exemplified by knat7-1 (Figs 3, 4; Table 1), and thinner interfascicular fiber cell walls in transgenic lines strongly expressing a prom4CL1::KNAT7 construct (Fig. 6). Given that KNAT7 functions as a transcriptional repressor (Li et al., 2011), one interpretation of these results is that KNAT7 represses, rather than activates, secondary cell wall biosynthesis in interfascicular fiber secondary walls, such that KNAT7 loss of function results in excess wall deposition, whereas gain of function (KNAT7 over-expression) leads to reduced wall deposition.
Our results are in contrast with those of Zhong et al. (2008), who suggested that KNAT7 is an activator of secondary cell wall formation, based on their observation that Arabidopsis lines expressing a dominant repression version of KNAT7, generated by fusion of the protein to an EAR transcriptional repressor domain, exhibited thinner interfascicular fiber cell walls. However, no irx phenotype, a hallmark of knat7 loss of function, was evident in these lines (Zhong et al., 2008). The most probable explanation for this discrepancy is that, as a transcriptional repressor (Li et al., 2011), the EAR domain enhances the KNAT7 repression function, generating a gain-of-function phenotype similar to that observed in prom4CL1::KNAT7-expressing lines when expressed in transgenic Arabidopsis, rather than a loss-of-function phenotype.
Considering that the loss of KNAT7 function appears to affect vessel secondary cell wall integrity (irx phenotype), and causes a dramatic increase in the thickness of interfascicular fiber secondary walls, it appears that secondary wall synthesis in both fibers and vessels is regulated by KNAT7. This is consistent with data showing that KNAT7 is the direct target of NAC domain transcription factors, including SND1, NST1, VND6 and VND7, that collectively activate secondary cell wall deposition in both vessel and fiber cells (Zhong et al., 2008). However, the contrasting knat7 loss-of-function phenotypes observed in vessels (irx; decreased cell wall thickness) and interfascicular fibers (increased cell wall thickness) are difficult to explain on the basis of current data. However, our data suggest that KNAT7 may regulate different aspects of secondary cell wall deposition in different cell types, perhaps by targeting different genes for regulation. KNAT7 is known to possess a number of interaction partners, including OFP4, MYB75, BELL-LIKE HOMEODOMAIN (BLH) proteins and other KNAT proteins (Hackbusch et al., 2005; Bhargava et al., 2010; Li et al., 2011). One intriguing possibility is that KNAT7 interacts with different partner proteins in different cell types to form functionally distinct complexes. If this were the case in vessel and fiber cells, loss of KNAT7 function in different cells could lead to distinct regulatory functions in specific cell types.
Conserved regulatory pathways are thought to be involved in xylem and secondary wall formation in herbaceous plants, such as Arabidopsis and trees (Groover, 2005). We confirmed that poplar KNAT7 is a true ortholog of the Arabidopsis KNAT7 gene, based on the ability of an ectopically expressed poplar KNAT7::GFP fusion protein to complement the Arabidopsis knat7 loss-of-function mutant phenotype, and that it is expressed during the onset of wood formation in coordination with poplar homologs of Arabidopsis MYB transcription factors that positively regulate secondary wall biosynthesis. Thus, both KNAT7 and MYB transcription factors could be used as tools to modulate secondary cell wall properties to improve bioenergy phenotypes of this species.
KNAT7 expression pattern
In accordance with the role proposed for KNAT7 in the regulation of secondary cell wall biosynthesis (Brown et al., 2005; Persson et al., 2005), GUS expression directed by promKNAT7::GUS fusions in transgenic Arabidopsis suggest that KNAT7 is specifically expressed in the vascular system in seedling roots and shoots and in the stele of roots (Fig. 1). In inflorescence stems, expression was mainly evident in cortex cells adjacent to interfascicular fibers, in cambial regions near xylem in stems and in the vicinity of developing vascular bundles (Fig. 1; Li et al., 2011). The pattern of promKNAT7-directed GUS expression in inflorescence stems is very similar to patterns directed by the Arabidopsis MYB75 and OFP4 promoters (Bhargava et al., 2010; Li et al., 2011), genes that encode transcriptional factors known to physically interact with KNAT7 and to regulate secondary wall formation (Bhargava et al., 2010; Li et al., 2011). Thus, it appears that, in addition to possible expression in interfascicular fibers themselves (Zhong et al., 2008), KNAT7 and the genes encoding these interaction partners are all expressed in cortex cells adjacent to developing interfascicular fibers, and that KNAT7 interaction with MYB75 and OFP4 in these cells may negatively regulate secondary cell wall deposition in the adjacent interfascicular fiber cells.
Cell wall properties regulated by KNAT7
The irx and increased interfascicular fiber cell wall thickness phenotypes of knat7 T-DNA insertion mutants are indicative of possible changes in the chemistry of the corresponding secondary cell walls. However, the impact of KNAT7 loss of function on secondary cell wall chemistry has remained unclear. Consistent with Brown et al. (2005), our analysis (Fig. S1) showed no significant changes in cell wall sugar levels in knat7, with the exception of glucose, which was marginally higher. Although Zhong et al. (2008) reported modest decreases in several cell wall sugars in Arabidopsis lines strongly expressing a KNAT7-EAR domain repression construct, these lines seem more likely to be gain-of-function derivatives, as discussed above, and these results could thus suggest modest decreases in hemicellulose sugars in KNAT7 gain-of-function plants. Although three genes involved in xyloglucan biosynthesis were up-regulated in knat7-1, we failed to observe significant changes in xylan or other hemicellulose sugar content in these plants. The reason for this discrepancy between transcript levels and product levels is unclear, but it is possible that the entire suite of genes required for xyloglucan biosynthesis is not up-regulated, or that there are differential effects on hemicellulose content in different cell types or developmental stages that are obscured when using whole stems for chemical analysis. These questions could be addressed in the future by genome-wide expression profiling and by employing antibodies specific to cell wall constituents.
We observed a significant increase in the lignin content of inflorescence stems from knat7 plants (Fig. 8). Thus, KNAT7 may play a key role in the negative regulation of lignin biosynthesis and deposition. Consistent with this interpretation, the expression of a suite of lignin biosynthetic genes was up-regulated in the knat7 background (Fig. 7b), suggesting partial deregulation of the lignin biosynthesis pathway. In addition, although cellulose and hemicellulose levels are apparently little changed in the knat7 background (Fig. S1; Brown et al., 2005), increases in the expression of secondary cell wall CESA genes (Fig. 7a) and three representative genes involved in hemicellulose biosynthesis in secondary walls (Fig. 7a), in knat7, suggest deregulation of secondary cell wall biosynthesis.
Interestingly, these results are similar to those reported recently for a myb75 loss-of-function mutant (Bhargava et al., 2010). Like knat7, myb75 plants exhibit thicker interfascicular fiber cell walls, increased lignin deposition and increases in expression of lignin biosynthetic genes, secondary cell wall CESA genes and genes implicated in secondary cell wall hemicellulose biosynthesis. Coupled with the finding that MYB75 and KNAT7 interact via protein–protein interactions and share a common expression pattern (Bhargava et al., 2010), these data indicate that MYB75 and KNAT7 may cooperate in repressing secondary wall biosynthesis. However, although the sets of secondary cell wall biosynthetic genes up-regulated in knat7 and myb75 mutant plants are overlapping, they are also distinct, suggesting that they may cooperate in a complementary manner to negatively regulate secondary cell wall biosynthesis.
Model for KNAT7 function in the regulation of secondary cell wall biosynthesis
Recent studies on the network of NAC and MYB domain transcription factors that regulate secondary wall biosynthesis have focused on the role of these proteins as positive regulators of the transcription of secondary cell wall biosynthetic genes, such as CESA and lignin biosynthetic genes (Zhong et al., 2007; Ko et al., 2009; McCarthy et al., 2009; Zhou et al., 2009). Recent models have suggested that this network is responsible for the positive regulation of carbon flux into secondary cell wall polymer biosynthesis (Demura & Ye, 2010; Zhong et al., 2010a), with feed-forward loops serving to rapidly activate secondary cell wall biosynthesis in appropriate cell types (McCarthy et al., 2009).
In contrast with this pattern, the data presented here and elsewhere (Bhargava et al., 2010; Li et al., 2011) suggest that KNAT7 acts as a strong repressor of secondary cell wall biosynthesis, apparently in concert with OFP4, MYB75 and, potentially, other transcription factors (Hackbusch et al., 2005) with which KNAT7 forms one or more transcription repression complexes. However, KNAT7 expression is strongly up-regulated in concert with secondary wall formation in Arabidopsis and poplar, and Arabidopsis KNAT7 is the direct target of NAC domain and MYB domain proteins that are strong positive regulators of secondary wall biosynthesis (Zhong et al., 2008; Ko et al., 2009; McCarthy et al., 2009). On the basis of these data, we propose that KNAT7 acts in a negative feedback loop within these regulatory networks governing secondary cell wall biosynthesis, working antagonistically to NAC domain and MYB domain positive regulators (Fig. 12). This negative feedback loop would allow the fine tuning of secondary wall biosynthesis, providing a mechanism for balancing metabolic commitment to secondary wall biosynthesis with available carbon and other physiological and/or environmental variables, and thereby maintaining cellular homeostasis with respect to carbon acquisition and allocation. In support of this model, recent data have shown that KNAT7 is also a direct target of MYB61, a transcriptional regulator that links whole-plant resource acquisition with resource allocation in sink tissues, such as xylem and seeds (J. M. Romano et al., unpublished). Thus, the KNAT7 negative regulatory node may receive multiple inputs that allow it to fine tune metabolic commitment to biosynthetic processes. The apparent conservation of this negative cell wall regulatory loop in poplar suggests potential new ways for manipulation of secondary cell wall deposition for improvement of bioenergy traits in this tree.
We thank Margaret Ellis (University of British Columbia (UBC) Michael Smith Laboratories) for phylogenetic analyses, Alex Skyba (UBC Department of Wood Science) for input on chemical analyses, Etienne Grienenberger (UBC Department of Botany) for helpful discussions, and the UBC Bioimaging Facility for transmission electron microscopy support. This work was supported by Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grants to C.J.D., B.E.E., S.D.M. and R.A.S., and by the NSERC Green Crops Network (funds to C.J.D. and B.E.E.).