, Volume 224, Issue 4, pp 828–837

Xylem-specific and tension stress-responsive coexpression of KORRIGAN endoglucanase and three secondary wall-associated cellulose synthase genes in aspen trees


  • Suchita Bhandari
    • Biotechnology Research Center, School of Forest Resources and Environmental SciencesMichigan Technological University
    • BILT
  • Takeshi Fujino
    • Biotechnology Research Center, School of Forest Resources and Environmental SciencesMichigan Technological University
  • Shiv Thammanagowda
    • Biotechnology Research Center, School of Forest Resources and Environmental SciencesMichigan Technological University
  • Dongyan Zhang
    • Biotechnology Research Center, School of Forest Resources and Environmental SciencesMichigan Technological University
  • Fuyu Xu
    • Biotechnology Research Center, School of Forest Resources and Environmental SciencesMichigan Technological University
    • Biotechnology Research Center, School of Forest Resources and Environmental SciencesMichigan Technological University
Original Article

DOI: 10.1007/s00425-006-0269-1

Cite this article as:
Bhandari, S., Fujino, T., Thammanagowda, S. et al. Planta (2006) 224: 828. doi:10.1007/s00425-006-0269-1


In nature, angiosperm trees develop tension wood on the upper side of their leaning trunks and drooping branches. Development of tension wood is one of the straightening mechanisms by which trees counteract leaning or bending of stem and resume upward growth. Tension wood is characterized by the development of a highly crystalline cellulose-enriched gelatinous layer next to the lumen of the tension wood fibers. Thus experimental induction of tension wood provides a system to understand the process of cellulose biosynthesis in trees. Since KORRIGAN endoglucanases (KOR) appear to play an important role in cellulose biosynthesis in Arabidopsis, we cloned PtrKOR, a full-length KOR cDNA from aspen xylem. Using RT-PCR, in situ hybridization, and tissue-print assays, we show that PtrKOR gene expression is significantly elevated on the upper side of the bent aspen stem in response to tension stress while KOR expression is significantly suppressed on the opposite side experiencing compression stress. Moreover, three previously reported aspen cellulose synthase genes, namely, PtrCesA1, PtrCesA2, and PtrCesA3 that are closely associated with secondary cell wall development in the xylem cells exhibited similar tension stress-responsive behavior. Our results suggest that coexpression of these four proteins is important for the biosynthesis of highly crystalline cellulose typically present in tension wood fibers. Their simultaneous genetic manipulation may lead to industrially relevant improvement of cellulose in transgenic crops and trees.


AspenCelluloseCellulose synthaseKORRIGAN endoglucanaseTension woodTrees



Cellulose synthase


Korrigan endoglucanase


Degree of polymerization




Since cellulose is an important component of plant cell walls, genetic modulation of cellulose synthesis can have a direct impact on several aspects of plant growth and development (Somerville et al. 2004) including, but not limited to, cell division and expansion, plant morphogenesis, and plant responses to the environment. Therefore, a thorough understanding of the intricacies of the cellulose biosynthetic processes is pivotal for future experiments targeting genetic improvement of cellulose production in economically important plants (Delmer 1999; Doblin et al. 2002; Williamson et al. 2002; Joshi et al. 2004). However, this field has recently witnessed many interesting yet contrasting situations. Although cellulose is the most abundant biopolymer on earth, the first plant cellulose synthase (CesA) gene was reported only recently (Pear et al. 1996). Chemically, cellulose is a polymer of β-1,4–linked glucose and this simple structure is conserved among all plants; yet at least two different groups of CesA proteins are involved in cellulose deposition during primary and secondary wall formation (Doblin et al. 2002). These two types of cell walls also greatly differ in the quantity and quality of cellulose they possess in terms of degree of polymerization (DP), crystallinity, and microfibril orientation (Joshi et al. 2004). Therefore, it is possible that CesA proteins play a major role in determining cellulose quantity as well as quality in plant cell walls, and that other coexpressed proteins might also be important in cellulose biosynthetic processes. Furthermore, multiple and distinct CesAs (most probably three) are necessary for cellulose biosynthesis in both the primary and secondary cell walls of plants, and in some cases, a mutation in any one of these CesA isoforms is deleterious to the entire cellulose biosynthetic process active in that particular type of cell wall (e.g. Schieble et al. 2001; Taylor et al. 2003). Finally, whereas endoglucanases (EGases; EC are generally associated with the degradation of cellulose, KORRIGAN (KOR)(Nicol et al. 1998) a membrane-anchored EGase, appears to be involved in the cellulose biosynthesis of both primary and secondary cell walls in Arabidopsis (Lane et al. 2001; Szyjanowicz et al. 2004). However, the precise role of KOR proteins in the cellulose biosynthetic process remains largely unclear.

Our knowledge regarding the Arabidopsis KOR has rapidly progressed since its first report in 1998 (review: Robert et al. 2004) but little is known about the KOR proteins from other plants. A number of point and null mutations in ArabidopsisKOR genes that affect the primary cell wall development are reported (e.g. kor1-1, kor1-2, acw1, rsw2). Collectively, these mutants exhibited abnormal plant morphology, defects in cell wall formation, reduced cellulose content, increased pectin synthesis, and aberrant cell divisions (Nicol et al. 1998; Zuo et al. 2000; Sato et al. 2001; Lane et al. 2001). Recently, Szyjanowicz et al. (2004) reported two additional irregular xylem mutations (irx2-1 and irx2-2) that were defective in secondary cell wall deposition similar to other irx mutants (Turner and Somerville 1997). These irx2 mutants have collapsed xylem cell walls due to reduced cellulose synthesis caused by two independent mutations in the KOR gene. Thus, mutated KOR proteins lead to defects in cellulose production in Arabidopsis. The KOR orthologs have also been reported from a few other plants (e.g. tomato: Brummell et al. 1997; rape: Molhoj et al. 2001; hybrid poplar: Master et al. 2004) and based on the EST analyses, they appear to be expressed in many higher plants (Joshi et al. 2004; Geiser-Lee et al. 2006). However, it is not clear whether the existing knowledge about Arabidopsis KOR gene expression and its putative functions can directly be applied to economically important trees.

Tree biology differs from that of Arabidopsis and it is currently unknown whether KORs from different plant species are functionally equivalent. Therefore, cloning and detailed characterization of KOR genes and their expression patterns from trees is a prerequisite for pursuing such experiments. Trees are slow growing and long-living plants that increase stem diameter every year through secondary growth via reactivation of cambium activity. These processes add woody materials to tree stems that not only confer mechanical strength and water transporting capacities to trees but also provide raw materials for forest product industries. Moreover, angiosperm trees develop tension wood on the upper side of their leaning trunks and drooping branches. Formation of tension wood is a mechanism by which trees counteract tension stress due to leaning or bending and correct their growth direction by pulling the spatially displaced stems/branches away from the ground (Timell 1969; Wu et al. 2000). Tension wood is also characterized by the deposition of a highly crystalline cellulose-enriched gelatinous (G) layer deposited in the interior cell wall of tension wood fibers. This process can be experimentally induced and provides a model system to understand the process of cellulose biosynthesis in trees.

To investigate the molecular basis of cellulose biosynthesis in trees, we first cloned a number of CesA genes from aspen (Populus tremuloides) that are involved in the wood development (review: Joshi et al. 2004). To-date, we have characterized seven distinct types of full-length CesA cDNAs, designated PtrCesA1 to PtrCesA7, from aspen xylem (Wu et al. 2000; Samuga and Joshi 2002, 2004; Kalluri and Joshi 2003, 2004). Of these, PtrCesA1, PtrCesA2 and PtrCesA3 share a high degree of similarity with the three secondary wall associated CesAs (secondary CesAs) from Arabidopsis which are suggested to form a functional complex during cellulose biosynthesis (Gardiner et al. 2003; Taylor et al. 2003). We have also reported molecular evidence at both the transcriptional and translational levels that these three secondary CesAs are coordinately expressed during xylem/phloem fiber development in aspen trees (Joshi et al. 2004). The same three genes might also be regulating the elevated levels of crystallinity of G-layer cellulose found in tension wood. Therefore, we are also investigating whether these three secondary CesAs are coordinately upregulated in response to tension stress.

Since KOR appears to be an important gene for cellulose biosynthesis in Arabidopsis, we have cloned a KOR cDNA, PtrKOR from aspen xylem. Here, we characterize PtrKOR expression in a variety of aspen tissues. Using RT-PCR, in situ hybridization and immunolocalization experiments, we examined whether PtrKOR gene is specifically upregulated in xylem tissues and whether it shows typical tension stress-responsive behavior along with three secondary wall associated CesA genes from aspen.

Materials and methods

Plant materials

Shoot apices, young leaves, and developing xylem scrapings from vertical stems were collected for total RNA and protein isolation from field grown aspen (Populus tremuloides) plants (∼5 ft tall) near the School of Forest Resources and Environmental Sciences, Michigan Technological University. Xylem from tension stressed and compression stressed saplings were collected from individuals of the same group of plants that were bent for 3 days. Developing xylem scrapings from the upper and lower side of the bent aspen stem were designated as tension wood and opposite wood samples. Generally, aspen trees of 1–3 in. stem diameter were cut and debarked, and exposed developing xylem tissue was immediately scrapped into liquid nitrogen. This tissue was used as developing xylem tissue enriched in secondary cell wall forming cells. All other plant materials were similarly frozen and stored in liquid nitrogen until use.

For tension stressed samples that were used for in situ mRNA hybridization, greenhouse grown aspen plants (∼2 ft tall) were bent for 7 days by first attaching one end of a string at the 33rd internode counted from the bottom, pulling the plant down without damaging and tying the other end to a firm support near the base of the potted plant. The maximum tension stress (indicted by arched stem) was developed at the 25th internode (counted up from the bottom of the plant). The unbent, vertical aspen plant of similar height was used as a control for such experiments.

cDNA library screening

A 700 bp long aspen EST (GenBank accession # CA926262) sharing a high percent identity with ArabidopsisKOR cDNA was available in our laboratory as a part of aspen EST project (Ranjan et al. 2004). This aspen KOR EST was used to screen ∼100,000 clones of aspen xylem cDNA library following the methods described earlier (Samuga and Joshi 2002). Six positive clones were identified from the primary screening and were further used for secondary screening. Ten single-clones were selected from the secondary screen and eighteen positive clones were identified during the tertiary screen. Sequencing of many rescued clones resulted in a full-length cDNA clone (PtrKOR) that was completely sequenced from both directions and analyzed with the GCG sequence analysis software package (Genetics Computer Group, Madison, WI).

RT-PCR analysis

Total RNA samples were isolated from five types of aspen tissues (shoot apices, leaves, developing xylem, tension wood and opposite wood) using the methods described earlier (Samuga and Joshi 2002). The following 5′UTR (untranslated region)-specific primers from PtrKOR and 5.8S rRNA-specific primers as internal control (Samuga and Joshi 2002) were used for RT-PCR:
  • PtrKOR 5′UTR-specific primers



  • 5.8S rRNA-specific primers

  • Forward Primer: 5′ CTAAACGACTCTCGGCAAC 3′

  • Reverse Primer: 5′ TCAAAGACTCGATGGTTCAC ′

We performed a two-step RT-PCR with GeneAmp Gold RNA PCR Kit (PE Applied Biosystems, Foster City, CA) as described earlier (Samuga and Joshi 2002; Kalluri and Joshi 2004). Briefly, for RT-PCR, the mRNA was first reverse transcribed to cDNA followed by PCR using the gene-specific reverse and forward primers. The two 5.8S rRNA-specific primers were used as an internal control to confirm that equal quantities of total RNA were used as template for RT-PCR experiments. All these experiments were repeated at least twice with the same outcome.

In situ mRNA hybridization

For each of the in situ mRNA hybridization experiments, two similar sized aspen plants (∼2 ft tall) were selected in the greenhouse. One plant was bent for 7 days and the other one was kept vertical and used as a control. The detailed in situ hybridization procedure was described earlier (Kalluri and Joshi 2004). Briefly, aspen stem tissues were fixed with 4% paraformaldehyde in 10 mM phosphate buffer, pH 7.2. These were dehydrated with water/ethanol/t-butanol series, and then embedded with paraplast (Sigma). The KOR transcripts were detected by in situ hybridization with antisense riboprobes from the 3′UTR region of the PtrKOR cDNA as described earlier (Kalluri and Joshi 2004). Briefly, the 3′UTR region from aspen KOR cDNA was amplified using gene-specific primers and subcloned into the pGEM T-easy vector system II (Promega). This construct was used to produce antisense and sense digoxygenin (DIG) labeled transcripts with T7 and SP6 RNA polymerases (DIG RNA Labeling Kit, Roche). Hybridization was done with DIG-labeled and denatured (70°C for 5 min in formamide) T7 and SP6 probes, respectively. Slides were viewed with a Nikon Eclipse 400 microscope and images taken using a Sony DKC-5000 digital camera. The details of PtrCesA1, PtrCesA2, and PtrCesA3 antisense and sense probe preparations have been described elsewhere (Kalluri and Joshi 2003; Samuga 2003; Kalluri 2003). All these experiments were repeated at least twice with the same outcome.

Protein extraction, western blotting and tissue printing

The same five aspen tissue samples namely, leaf, apex, developing xylem, xylem tension stressed for 7 days (tension wood), and xylem compression stressed for 7 days (opposite wood) that were used for RT-PCR experiments were also used for western blotting. These tissues (100–200 mg) were ground in liquid nitrogen to fine powder and an equal volume of protein extraction buffer [100 mM Tris–HCl, pH 7.5, 25 mM Ascorbic acid, 5% Polyvinylpolypyrrolidone, 15% Ethylene Glycol, 1.5% TritonX-100, 1 mM DTT, and 1% protease inhibitor cocktail (Sigma)] was added and mixed vigorously. After centrifugation at 15,000g, the clarified supernatants were used as the protein extracts. Protein concentration for each sample was determined using Protein Assay Reagent (Bio-Rad). Ten microgram of protein from each tissue sample was separated on 10% SDS/PAGE, viewed by duplicate commassie blue staining, and transferred onto a nitrocellulose membrane (Schleicher & Schuell). The protein immunoblotting was performed according to standard protocols (Sambrook and Russell 2001). For detection of KOR proteins, a KOR rabbit antiserum (kindly provided by Professor Hermann Hofte and Dr. Samantha Vernettes, INRA) (Nicol et al. 1998) was used at 1:1000 dilution, and a secondary Alkaline phosphatase conjugate anti-rabbit IgG was used at 1:5000 dilution. Unrelated preimmune rabbit serum diluted 1:1000 was used as a control that did not show any specific interaction with KOR proteins. All these experiments were repeated at least twice with the same outcome.

The detailed procedure of tissue printing that we used was described previously (Pérez-Garcia et al. 1998). Briefly, a series of stem cross-sections of 1–2 mm thickness were taken using a sharp double-edged razor blade. These sections were pressed onto a pre-wetted nitrocellulose membrane for 5 s. Tissue prints were transferred to TBST buffer (10 mM Tris–HCl, pH 8.0 and 150 mM NaCl with 0.05% Tween 20) for 10 min at room temperature. The blocking of blotted membrane was carried out by incubation in TBST containing 10% non-fat dry milk for 2 h at room temperature or overnight at 4°C. These membranes were incubated at 4°C overnight with antibodies raised against Arabidopsis KOR as described previously (Nicol et al. 1998). These antibodies were diluted at 1:1000 in TBST containing 1% BSA. Nitrocellulose membranes were subsequently washed three times with TBST for 15 min at room temperature and incubated with anti-rabbit IgG serum conjugated to Alkaline phosphatase (Sigma) at a 1:1000 dilution in TBST containing 1% BSA for 1 h at room temperature. Unrelated preimmune rabbit serum diluted 1:1000 was used as a control that did not show any specific interaction with KOR proteins.


Cloning and characterization of KORRIGAN endoglucanase cDNA from aspen

In order to obtain a homologous KOR probe for screening the aspen xylem cDNA library, we first performed a computer search of our ASPENDB database covering over 12,000 aspen ESTs (Ranjan et al. 2004) with the coding region of the ArabidopsisKOR gene (GenBank accession # AF073875). Of the five significant hits, only one (MTU6CR.P11.H12) contained the cDNA sequence corresponding to the putative N-terminal region with a single transmembrane domain that is characteristic of KORRIGAN-like membrane-associated EGases in other plants (Molhoj et al. 2002). Therefore, we screened ∼100,000 clones from an aspen xylem cDNA library with MTU6CR.P11.H12 EST and obtained a full-length clone, PtrKOR (GenBank accession # AY535003). This 2,562 bp long cDNA contained 262 bp of 5′ untranslated region (UTR) and 426 bp long 3′UTR followed by 17 bp of poly-A tail. Both UTRs are long for plant genes (Joshi 1987a, b) although the importance of unusually long UTRs is still unclear.

A single open reading frame of 1,857 bp encodes an aspen KOR protein of 619 amino acids with the predicted molecular weight of 68,421 D and a pI value of 8.76. A single transmembrane domain was detected between amino acids 72–94, a feature that is similar to other known KOR proteins from plants (Molhoj et al. 2002). Aspen KOR also contains many amino acid motifs conserved in other KORs such as two putative polarized targeting sequences, LL and YXXΦ (where Φ is any hydrophobic amino acid and X is any amino acid) in the N-terminal cytoplasmic tail and two glycosyl hydrolase family 9 signature motifs near the C-terminus, all similar to Arabidopsis KOR (del Campillo 1999; Zuo et al.2000). A computer search of the PROSITE database with aspen KOR also predicted the presence of 8 N-glycosylation sites and 15 phosphorylation sites, suggesting that this protein might be highly modified after translation. In addition, a proline-rich region near the C-terminus of aspen KOR that is conserved in other KOR proteins but absent in non-KOR EGases (Nicol et al. 1998). It has been recently suggested that based on the comparison of amino acid sequences among a variety of plant EGases, KOR proteins form a separate group of EGases designated as γ-EGases (Libertini et al. 2004).

Gene expression pattern of PtrKOR gene

The expression pattern of the PtrKOR gene was examined using total RNA from five types of plant tissues namely growing shoot apices and young leaves enriched in expanding primary wall containing cells and developing xylem tissues from vertical aspen stems, tension-stressed tension wood and compression-stressed opposite wood enriched in thickening secondary wall containing tissues. RT-PCR experiments were performed using a gene-specific primer pair flanking a portion of the 5′UTR from PtrKOR cDNA (Fig. 1 lanes 1, 3, 5, 7 and 9). Internal control reactions (Fig. 1 lanes, 2, 4, 6, 8 and 10) showed similar level of amplification products suggestive of equal quantities of starting mRNA for all these reactions.
Fig. 1

RT-PCR analysis of PtrKOR expression using total RNAs from shoot apices (Lanes 1 and 2), young leaves (lanes 3 and 4), developing xylem (lanes 5 and 6), opposite wood (lanes 7 and 8), and tension wood (lanes 9 and 10). The PCR amplification products using the 5′UTR-specific primers from PtrKOR (lanes 1, 3, 5, 7 and 9) and rRNA-specific primers (lanes 2, 4, 6, 8, 10) as internal control are shown. The internal control reactions show similar levels of amplifications in all five samples while PtrKOR-specific reactions show higher expression in the xylem samples as compared to apex and leaves and lower expression in the opposite wood as compared to tension wood samples. The molecular weight marker indicated by M shows 500 bp DNA fragment as indicated by an arrow

The aspen KOR gene is expressed in all five tissues examined (Fig. 1). The amounts of RT-PCR products appeared to be lower in the shoot apices and leaves than in developing xylem samples. Tension side xylem showed higher expression of KOR gene than opposite side. Similar to ArabidopsisKOR expression reported by Nicol et al. (1998), the aspen KOR gene is expressed in all tissues examined but secondary wall forming tissues have a higher level of expression than primary wall forming tissues. Moreover, tension wood fibers are known to have higher amounts of almost pure cellulose in their G-layers (Timell 1969) and similar to our previous report of PtrCesA1 upregulation in response to tension stress (Wu et al. 2000) the KOR gene is also upregulated in tension wood and downregulated in opposite wood (Fig. 1). This is the first indication that KOR and CesA genes might be coexpressed in response to tension stress in hardwoods.

In situ hybridization confirms coexpression of PtrKOR and PtrCesA1 in xylem fibers

RT-PCR experiments although useful in providing the overall picture of gene expression patterns do not provide an accurate picture of cell-specific expression of genes due to tissue homogenization for the isolation of RNA. We have described the use of in situ mRNA hybridization to clarify cell-specific expression of cellulose biosynthesis related genes in aspen (Wu et al. 2000; Samuga and Joshi 2004; Kalluri and Joshi 2004). Use of the 3′UTR from aspen PtrKOR for preparation of antisense riboprobes clearly delineated the upregulation of KOR gene in the first 1–2 developing xylem cell layers compared to growing phloem and cortex cells that showed uniformly low expression of KOR (Fig. 2a). Similar results were obtained by using the 5′UTR region of PtrKOR (data not shown). Within developing xylem tissues elongating fibers and expanding ray cells express the KOR gene and weaker expression was observed in developing phloem fiber cells (Fig. 2c). Hybridization of similar sections with sense KOR riboprobes did not show any hybridization as expected (Fig. 2e). Stem sections were also hybridized with antisense PtrCesA1-specific riboprobes. Strong hybridization signals were detected almost exclusively in the first 3–4 developing xylem layers, some primary xylem tissues towards the pith, and developing phloem fiber cells (Fig. 2b). All these cells are characterized by thickening secondary cell walls (Fig. 2d). Again as expected, the control hybridization with sense PtrCesA1 HVRI-specific probes did not show any signal in xylem or phloem fiber cells (Fig. 2f). In conclusion, we have observed a coordinate expression of PtrKOR and PtrCesA1 genes in developing xylem fibers.
Fig. 2

In situ mRNA hybridization of aspen stem cross-sections (internode #25th from the bottom) with PtrKOR-specific (a, c,e) and PtrCesA1-specific (b, d, f) antisense riboprobes. Stem sections in e and f were hybridized with corresponding sense riboprobes and do not show any cell-specific hybridization as expected and serve as the negative controls. Note the developing xylem-specific (marked by letter X) and phloem fiber-specific (marked by letter P) accumulation of PtrKOR (a) and PtrCesA1 (b) transcripts. A weak expression of PtrKOR is also observed in cortex cells (marked by C) in a. In contrast, PtrCesA1 transcripts are not visible in the cortex of b. PtrCesA1 transcripts are also visible in primary xylem (px) tissues. Same sections are enlarged in c and d. The scale bar 100 μm

The RT-PCR experiments also suggested that both PtrKOR and PtrCesA1 genes are coordinately yet differentially expressed on the upper and lower side of the bent aspen stem. In order to identify cells involved in this response, we performed in situ hybridization using tension stressed aspen stem. For the in situ hybridization, we included two additional secondary wall associated aspen CesA genes namely PtrCesA2 and PtrCesA3 that have been shown to be upregulated in response to tension stress (Samuga 2003; Kalluri 2003). As shown in Fig. 3, PtrKOR, PtrCesA1, PtrCesA2, and PtrCesA3 are coordinately upregulated on the upper side of the bent aspen stem under tension stress (tension wood) while being coordinately downregulated on the lower side of bent stem experiencing compression stress (opposite wood). While the PtrKOR gene was strongly expressed in developing xylem and phloem fiber cells on the upper side of the stem with only weak hybridization signal across the entire stem section, the three PtrCesA genes were highly expressed only in developing xylem and phloem fiber cells. PtrKOR is also highly expressed in ray cells on both the upper and lower sides of the bent stem. Thus while both KOR and CesA genes are coordinately upregulated in secondary wall forming fiber cells; they maintain their normal expression patterns elsewhere. PtrKOR was weakly expressed in primary cell wall forming expanding cells while three PtrCesAs were highly expressed mostly in the secondary wall forming cells. Entire aspen stem sections hybridized with either DIG-labeled antisense (Fig. 4a) or sense (Fig. 4b) PtrCesA1 probes exhibits a progression of PtrCesA1 expression across the aspen stem in response to tension stress. Similar hybridization patterns were observed with PtrCesA2, PtrCesA3, and PtrKOR probes. It must be noted that some inner xylem cells towards the pith show some non-specific binding of the DIG labeled probes on tension wood side as indicated in Fig. 4b. It is possible that thickened G-layers in tension wood show some non-specific binding with the DIG label but the actual hybridization signals are far stronger than the background noise.
Fig. 3

In situ mRNA hybridization of aspen stem cross-sections (internode 25th from the bottom) bent for 7 days with gene-specific antisense riboprobes from PtrCesA1 (a, e, i, m), PtrCesA2 (b, f, j, n), PtrCesA3 (c, g, k, o) and PtrKOR (d, h, l, p) genes. Tension side of the bent aspen stem is shown in ad while compression side is shown in eh. Note coordinate up-regulation of three CesA and KOR genes from aspen on the tension side and down-regulation of same genes on the lower side of the aspen stem. Same eight sections (ah) are enlarged in ip for clarity. The scale bar 100 μm
Fig. 4

Whole-mount of the transverse section of aspen stem bent for 7 days and hybridized with antisense PtrCesA1 RNA probe (a) and sense PtrCesA1 RNA probe (b). Red arrow indicates the side that experienced maximum tension stress. Note the differential expression of PtrCesA1 transcripts on tension stress side and suppression of PtrCesA1 expression on opposite side. The black arrow indicates some weak non-specific binding of the DIG-labeled probe in the tension wood region of the control section (b). Scale bar 1 mm

Immunological studies bolster results from RT-PCR and in situ hybridization

The transcriptional modulation studies are generally sufficient to gain a picture of gene regulation in plants. Since this study is mainly about aspen KOR proteins, we wanted to see whether RNA expression correlated with protein accumulation or distribution. We used polyclonal antibodies against Arabidopsis KOR (kindly provided by Drs. Hermann Hofte and Samantha Vernhettes, INRA, France) for western blot analysis and tissue print assays. Since Arabidopsis KOR proteins share 83% identity and 87% similarity with aspen KOR proteins, we expected that Arabidopsis KOR antibodies would cross react with aspen KOR proteins. As shown in Fig. 5 and similar to our RT-PCR results, aspen KOR proteins are much more abundant in xylem tissues than in leaf or apex tissues. Moreover, KOR proteins are more abundant in tension wood tissues than in opposite wood samples. It is worth noting that the apparent molecular mass of the reacting protein band is between 80 and 100 kD and the predicted molecular weight of aspen KOR based on the amino acid sequence is only about 68 kD. This observation suggests that aspen KOR proteins that are similar to Arabidopsis KOR proteins can be detected using Arabidopsis KOR antibodies, and they might be undergoing post-translational modifications similar to rape KOR (Molhoj et al. 2001). As stated before, aspen KOR protein has eight such predicted N-glycosylation sites. However, the use of these antibodies for immunolocalization of KOR proteins on thin stem sections did not show strong hybridization. This may be due to the unstable nature of this protein or due to the low concentrations of KOR protein. Similar observations were also noted by Zuo et al. (2000) who could perform western blots with their Arabidopsis KOR antibodies but could not use them for immunolocalization. Therefore, we used Arabidopsis KOR antibodies for tissue print experiments.
Fig. 5

Western blot analysis of proteins from young leaf, shoot apex, developing xylem, tension wood and opposite wood (lanes 15, respectively) samples hybridized with Arabidopsis KOR antibodies

The use of Arabidopsis KOR antibodies for tissue prints also showed that in the younger stem internodes (31–32 from the bottom) KOR proteins accumulate in both phloem and xylem fibers (Fig. 6a). While in maturing internodes (24–25 from the bottom) KOR continues to express in xylem and phloem fibers (Fig. 6b), KOR proteins in the mature internodes (<20 from the bottom) are associated only with developing xylem tissue (Fig. 6c). When aspen stems were bent for ∼7 days, KOR expression was predominant on the upper or tension side of the bent stem (Fig. 6d). In short, examination of KOR protein expression in aspen stem supports our transcriptional studies suggesting that the KOR gene is highly expressed during secondary wall development as well as in response to tension stress where upregulation of cellulose synthesis is known to occur.
Fig. 6

Tissue print analysis of aspen stems probed with Arabidopsis KOR antibodies. ac show aspen stem sections from 32nd, 24th, and 20th internodes (counted from the bottom) respectively. Aspen stem in d was bent for 7 days. The tension side is indicated by an arrow. Note that the xylem and phloem fiber-specific KOR expression (indicated by X and P in ad) in the internode 24 from the bottom, loss of phloem fiber-specific KOR expression in the older internode 19, and the higher KOR expression on the tension side than the compression side in response to 7 days tension (bending) stress


Earlier investigations with Arabidopsis KOR indicated that having an intact and unmodified KOR protein is important for primary cell wall development (Robert et al. 2004). Several mutations in the ArabidopsisKOR gene affect primary cell wall development (kor1-1: Nicol et al. 1998; kor1-2: Zuo et al. 2000; acw1: Sato et al. 2001; rsw2: Lane et al. 2001). Complementation by wild type ArabidopsisKOR gene eliminated mutant phenotypes, confirming the importance of KOR for cell wall development in all these mutants. Recently, Szyjanowicz et al. (2004) reported two additional irx2 mutants that had defective secondary cell walls. Previously, three other cellulose-deficient irx mutations, irx1, irx3, and irx5 were reported that were caused by defects in three distinct CesA proteins, namely, AtCesA8, AtCesA7, and AtCesA4, respectively (Taylor et al. 1999, 2000, 2003). Thus KOR and three CesA proteins appear to be important for cellulose biosynthesis in Arabidopsis and our present results lend support to the importance of the same 4 genes for cellulose biosynthesis in wood forming tissues of aspen trees.

The irx2 mutants also had severe crystalline cellulose reduction (down to 30%) that was associated with defect in secondary cell wall development as compared to wild-type plants. Interestingly, unlike kor1-1, irx2 showed neither reduced hypocotyl growth in the dark nor the radial swelling typically observed in acw1 and rsw2 mutants. Other cellular anomalies like abnormal cell shape or incomplete cell plates that were commonly reported in kor mutants also were absent in irx2 mutant. On the other hand, stems of kor1-1 that were reported to be primary cell wall mutants also showed severely collapsed xylem cells just like irx2 (Nicol et al. 1998). Thus, the kor1-1 mutation also affects secondary wall formation. The situation regarding secondary wall development in other kor mutants was not discussed in the respective publications, however, close inspection of published stem section images of kor1-2 (Zuo et al. 2000) and rsw2 (Lane et al. 2001) appears to reveal deformed xylem cells. Thus, the KOR gene appears to play a major role in cellulose biosynthesis in both primary as well as secondary walls, and it appears to be highly expressed during secondary than primary cell wall formation to cope with the increased rates of cellulose synthesis. Here, we have observed similar expression patterns of aspen KOR gene in both primary as well as secondary cell wall forming cells based on RT-PCR and in situ hybridization results. Therefore, KOR indeed appears to be an important protein that is essential for cell wall formation in growing plant cells and it is upregulated during xylem development. Genetic engineering of KOR expression levels might significantly affect wood and fiber products produced in transgenic plants. This inference can further be bolstered by observed coordinate upregulation of CesA and KOR proteins during secondary wall-rich xylem development.

Trees also offer an interesting system for understanding cellulose synthesis via tension wood formation. In response to tension stress, relatively pure cellulose is produced in G layers of tension wood fibers and simple bending or tilting of angiosperm trees can mimic the tension stress naturally experienced by hardwood trees. At least four lines of evidence presented here suggest that the aspen KOR gene is upregulated in response to tension stress similar to three aspen secondary CesA genes. First, RT-PCR results show increased accumulation of KOR transcripts in response to tension stress and significant downregulation on the opposite side in response to perceived compression stress. Second, in situ mRNA hybridization experiments confirmed such upregulation of KOR genes on the upper side of bent aspen stem concomitant with upregulation of three secondary CesA genes. Third, KOR protein accumulation patterns observed through western blot analysis confirmed the authenticity of transcript accumulation patterns. Finally, tissue prints showed increased accumulation of KOR proteins in response to tension stress.

In plants, multienzyme complexes consisting of CesA and possibly other interacting proteins organized in the form of hexameric rosettes are suggested to orchestrate cellulose biosynthesis (Delmer 1999). Read and Bacic (2002) have even suggested that KOR might be an integral part of the elementary particles in such rosettes. In addition, Molhoj et al. (2002) have suggested that in all KOR substitution mutants in Arabidopsis, the substituted amino acid are always located on the surface of the KOR protein, and it is therefore possible that these mutations affect the interactions of KOR with other proteins rather than affecting catalytic activity of KOR itself. Therefore, it is likely that CesA and KOR proteins physically interact with each other. Szyjanowicz et al. (2004) have recently addressed this possibility by using pull-down assays similar to the interactions among detergent-solubilized secondary CesA proteins as they have demonstrated previously (Taylor et al. 2003). However, under similar conditions, KOR did not co-purify with the secondary CesA proteins, suggesting that KOR may not be a part of the same complex. One may explain this observation in two different ways. First, it is likely that CesA-KOR interaction is weak or transient and there was not enough KOR protein that could be detected under their reaction conditions. Alternatively, KOR might be indirectly interacting with CesAs via some other proteins that were not stable in their reaction conditions. Identification of KOR interacting proteins, therefore, needs to be pursued to fully understand how KOR participates in cellulose biosynthetic processes.


We wish to thank Drs. Herman Hofte and Samantha Vernhettes, INRA, France for generously providing Arabidopsis KOR antibodies, Dr. Scott Harding for critical reading of the manuscript, and Dr. Chung-Jui Tsai for providing aspen KOR EST. BILT’s financial support for SB’s visit is also gratefully acknowledged. The research in Joshi lab is supported by NSF-CAREER award (IBN-0236492), USDA-NRI grants (1999-35103-7986) (2005-35103-15256), and USDA-McIntire Stennis Forestry Research Program.

Copyright information

© Springer-Verlag 2006