Environmental and auxin regulation of wood formation involves members of the Aux/IAA gene family in hybrid aspen

Authors

  • Richard Moyle,

    1. Umeå Plant Science Centre, Swedish University of Agricultural Sciences, Department of Forest Genetics and Plant Physiology, 90183 Umeå, Sweden
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Jarmo Schrader,

    1. Umeå Plant Science Centre, Swedish University of Agricultural Sciences, Department of Forest Genetics and Plant Physiology, 90183 Umeå, Sweden
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Anneli Stenberg,

    1. Umeå Plant Science Centre, Swedish University of Agricultural Sciences, Department of Forest Genetics and Plant Physiology, 90183 Umeå, Sweden
    Search for more papers by this author
  • Olof Olsson,

    1. Department of Molecular Biology, Lundberg Laboratory, Göteborg University, 405 30 Göteborg, Sweden
    Search for more papers by this author
  • Sangeeta Saxena,

    1. Centre for Plant Molecular Biology, National Botanical Research Institute, 226 001 Lucknow, India
    Search for more papers by this author
  • Göran Sandberg,

    1. Umeå Plant Science Centre, Swedish University of Agricultural Sciences, Department of Forest Genetics and Plant Physiology, 90183 Umeå, Sweden
    Search for more papers by this author
  • Rishikesh P Bhalerao

    1. Umeå Plant Science Centre, Swedish University of Agricultural Sciences, Department of Forest Genetics and Plant Physiology, 90183 Umeå, Sweden
    Search for more papers by this author

*For correspondence (fax +46 90 7865901; e-mail Rishi.Bhalerao@genfys.slu.se).

Summary

Indole acetic acid (IAA/auxin) profoundly affects wood formation but the molecular mechanism of auxin action in this process remains poorly understood. We have cloned cDNAs for eight members of the Aux/IAA gene family from hybrid aspen (Populus tremula L. × Populus tremuloides Michx.) that encode potential mediators of the auxin signal transduction pathway. These genes designated as PttIAA1-PttIAA8 are auxin inducible but differ in their requirement of de novo protein synthesis for auxin induction. The auxin induction of the PttIAA genes is also developmentally controlled as evidenced by the loss of their auxin inducibility during leaf maturation. The PttIAA genes are differentially expressed in the cell types of a developmental gradient comprising the wood-forming tissues. Interestingly, the expression of the PttIAA genes is downregulated during transition of the active cambium into dormancy, a process in which meristematic cells of the cambium lose their sensitivity to auxin. Auxin-regulated developmental reprogramming of wood formation during the induction of tension wood is accompanied by changes in the expression of PttIAA genes. The distinct tissue-specific expression patterns of the auxin inducible PttIAA genes in the cambial region together with the change in expression during dormancy transition and tension wood formation suggest a role for these genes in mediating cambial responses to auxin and xylem development.

Introduction

The role of indole acetic acid (IAA/auxin) as an important regulator of growth-related processes such as cell division, elongation and differentiation has been well established over the years (Davies, 1995). Auxin regulation of diverse developmental processes is thought to involve changes in both cellular auxin content and the sensitivity of cells towards the hormone as well as changes in polar auxin transport (Davies, 1995; Little and Bonga, 1974; Muday, 2001). In recent years genetic approaches in the model plant Arabidopsis have considerably increased our knowledge of auxin action (Leyser, 2001), nevertheless our picture of auxin regulation is far from complete. One reason has been the difficulty to accurately measure alterations of endogenous hormone levels, especially with cellular resolution. This technical limitation has so far made it difficult to distinguish whether a cellular response stems from changes in cellular hormone concentration or alterations in auxin sensitivity.

Auxin plays a key role in regulating wood formation through its effects on cambial activity and xylem development (Sundberg et al., 2000). Auxin is required for maintaining the cambium in a meristematic state as depleting the cambium of auxin leads to differentiation of cambial cells to axial parenchyma (Savidge, 1983). The wood-forming tissue also provides an experimental system in which cellular sensitivity to auxin is strongly modulated by environmental signals. It has been shown that the sensitivity of the cambium to auxin is lost when trees enter dormancy at the end of the growing season (Little and Bonga, 1974). This phase of dormancy, known as rest, can be overcome by giving a chilling treatment to the tree upon which the cambium makes the transition from rest to quiescence and regains responsiveness to IAA. Auxin has further been proposed to be an important regulator of xylem development (Uggla et al., 1996). Therefore, analysing the role of auxin in wood formation will allow a better understanding of how auxin distribution regulates the sequence of developmental events and how environmental signals can modulate cellular sensitivity to auxin.

The wood forming tissue in trees as an experimental system to investigate auxin action, has several advantages in addition to the unique role of auxin in wood formation. The large size of the cambial meristem with its spatially separate zones of dividing, elongating and secondary wall forming cells allows the isolation of nearly pure cell layers in distinct stages of development (Uggla et al., 1996). The establishment of micro techniques for measuring both auxin content and gene expression make it feasible to investigate alterations in the endogenous auxin level, gene expression and the attendant developmental changes in an auxin-regulated phenomenon at almost cellular resolution (Hertzberg et al., 2001; Tuominen et al., 1997). These advantages of wood-forming tissues in trees have been successfully used earlier to demonstrate an overlap of an auxin gradient and a developmental gradient in the wood forming tissues, thereby suggesting a role for auxin as a positional signal in wood formation (Uggla et al., 1996).

Despite a critical role for auxin in regulating diverse aspects of wood formation as well as the obvious advantages in studying auxin action in wood formation, little is known about the molecular basis of this process. One of the best-characterized auxin responses is the induction of gene expression by auxins. Among these auxin-induced genes, the Aux/IAA family encoding 19–36 kDa short-lived nuclear-localized proteins is probably the most well studied, and several independent lines of evidence have suggested their role in mediating auxin responses (For a review see Reed, 2001). Given the role of Aux/IAA genes in mediating auxin responses we have focused our efforts to understand auxin action in wood formation by cloning and characterizing the cDNAs for Aux/IAA genes in the woody plant hybrid aspen (Populus tremula L.×Populus tremuloides Michx.). The results presented here indicate that Aux/IAA genes in hybrid aspen, referred to as PttIAA genes, form a multigene family comprising at least 8 members that are expressed differentially in the tissues of the cambial region. Furthermore, PttIAA genes are auxin inducible and their induction by IAA is developmentally regulated. The PttIAA genes are downregulated during cambial transition into dormancy and their expression is altered by treatments leading to tension wood formation. Taken together these data suggest a role for PttIAA genes in mediating auxin responses during various aspects of wood formation.

Results

Cloning of PttIAA genes and their relationship to the Arabidopsis sequences

We have identified eight different Aux/IAA gene homologues in the Populus Database (populusDB) (http://www.biochem.kth.se/PopulusDB/) constructed from a collection of cDNA sequences generated during the course of an ongoing EST (expressed sequence tag) sequencing project (Sterky et al., 1998). Sequence comparisons of the hybrid aspen ESTs with the Aux/IAA homologues from other plants indicated that half of these ESTs were truncated at the 5′ end. Full-length cDNA clones corresponding to these genes were therefore obtained using a 5′ RACE strategy as described in Experimental procedures. The eight full-length cDNA clones presented in this paper were named PttIAA1–PttIAA8 for Populus tremula × tremuloides Aux/IAA gene. A more detailed comparison of the derived amino acid sequences of PttIAA genes with those of their closest Arabidopsis homologues is shown in Figure 1. All eight PttIAA proteins contain the conserved domains I–IV, a characteristic of the Aux/IAA gene family. But even outside the domains I–IV there are considerable stretches of homology between the hybrid aspen and the most closely related Arabidopsis Aux/IAA genes.

Figure 1.

Sequence alignment of Arabidopsis and poplar Aux/IAA genes.

A multiple sequence alignment obtained by the PILEUP function in the Wisconsin Package (Genetics Computer Group, Madison, WI, USA) was manually improved. Sequence conservation in the entire family is indicated by shading in orange and yellow as follows: Dark orange absolutely conserved, light orange similar in at least 70% (12 of 16), and yellow similar in more than 50% (8 of 16) of the sequences. The consensus line shows the residues for identical and 70% conserved regions in upper and lower case letters, respectively. Blue shading indicates additional similarity in different subgroups, which were inferred from a phylogenetic tree. The positions of the conserved domains I to IV are marked by roman numerals and lines below the sequence. Purple bars indicate the position of conserved basic residues functioning as a NLS. Green shading marks a duplication of domain II in PttIAA6.

While the PttIAA proteins have all the conserved features of the Aux/IAA proteins there are certain novel properties that distinguish some of them from either their putative Arabidopsis homologues or other members of the hybrid aspen PttIAA gene family. Firstly, PttIAA3 and PttIAA4 are most similar to IAA20 and two predicted genes At3g62100 and At3g17600 from Arabidopsis, both of which appear to lack the bipartite nuclear localisation signal (NLS) that is formed by a part of domain II and the conserved KR residues between domains I and II in nearly all of the other Aux/IAA genes (Abel and Theologis, 1995). Secondly, the glycine of the qvvgWPPirs motif in domain II known to be important for regulating the stability of Aux/IAA proteins (Ramos et al., 2001), is replaced with an aspartic acid residue in PttIAA3 and PttIAA4. Another interesting feature unique to PttIAA proteins is the duplication of domain II in PttIAA6. This kind of tandem repeat of domain II has not been reported in any other member of the Aux/IAA or PttIAA gene family to date. In PttIAA6 the consensus domain II is followed by a nearly identical 26 amino acid long duplication, albeit with two differences, valine99 and alanine113 in the first copy are replaced with glutamic acid and valine in the duplication, respectively.

Auxin responsiveness of the PttIAA genes

We evaluated auxin responsiveness of the PttIAA gene family and examined the dependence of their auxin induction on protein synthesis to investigate the potential differences between individual members. Young leaves and mature leaves were treated with IAA, cycloheximide, and IAA plus cycloheximide and the expression of PttIAA genes was determined (Figure 2). The auxin induction experiments were performed in leaves as auxin applications can be performed with greater ease in leaves than in stems. Treatment with IAA for 30 min resulted in the up-regulation of all but PttIAA3, thus confirming these genes to be auxin inducible (Figure 2). To test whether the genes could be classified as early response genes, their response to treatment with cycloheximide was tested. In the case of PttIAA1, and PttIAA5–PttIAA8 pre-treatment with cycloheximide diminished the auxin response, indicating that protein synthesis is required for the auxin induction of these genes, thus making them part of the late response group. There is however, an inductive effect of cycloheximide alone on the expression of this group of PttIAA genes which could be due to mRNA stabilisation in the presence of cycloheximide (Gil and Green, 1996; Koshiba et al., 1995). PttIAA4 was the only gene where auxin induction was unaffected by cycloheximide pre-treatment, indicating that it is an early response gene. A completely different pattern was observed for PttIAA3. The induction of this gene is only detectable by the combination of cycloheximide and IAA treatments. Since the observed signal is weak, it is possible that this gene is induced in the presence of IAA alone, and a cumulative effect of cycloheximide and IAA raises the signal over the detection limit. Based on this observation we propose to also classify PttIAA3 as a primary response gene.

Figure 2.

Auxin induction of PttIAA genes.

Young (< 2 cm) and mature leaves were subject for 1 h to the following treatments: MS-medium (MS), 10 µm IAA (IAA), 50 µm cycloheximide (CHX) or pre-treatment with CHX for 30 min followed by IAA treatment (CHX + IAA). Untreated leaves served as controls (CNT). About 0.6 µg of poly(A)+ mRNA per lane were separated on an Agarose gel, blotted onto a nylon membrane and hybridized with 32P labelled probes.

Developmental regulation of the auxin induction of the PttIAA genes

While Aux/IAA genes are differentially expressed, as has been shown in this and previous studies (Wong et al., 1996; Wyatt et al., 1993), the role of developmental control in regulating auxin inducibility of the Aux/IAA genes is less well studied. For example it is important to known whether all Aux/IAA genes retain their auxin inducibility independently of tissue and developmental background as has been suggested for AtIAA4 (Wyatt et al., 1993). In order to address the question of developmental control of auxin inducibility we compared the auxin response of PttIAA genes in young and mature leaves. The results shown in Figure 2 indicate a nearly complete absence of induction of any PttIAA gene in mature leaves. There are several possible explanations for this observation. It could be that mature leaves do not take up as much IAA as young leaves or that they could metabolise it quickly. Measurements of hormone levels before and after auxin treatment, however, rule out this explanation showing that mature leaf tissue accumulates about twice as much IAA as the young leaves (Table 1). Another explanation is that since these genes are only expressed at very low levels in leaves, they might play no role in general leaf responses to IAA. However, this does not appear to be the case as expression of the majority of these genes is undetectable in both young and mature leaves but the auxin induction is affected only in mature leaves. It seems therefore more likely that loss of inducibility of PttIAA genes in mature leaves of hybrid aspen could be part of a wider mechanism that causes a change in auxin response of leaves during the course of leaf development.

Table 1.  Auxin content of young and mature leaves before and after auxin treatment
TreatmentUntreatedMSMS + 10 mm IAA
  1. Young and mature leaves were treated for 1 h with either MS or MS + 10 mm IAA. The tissues used are the same as for Northern analysis in Figure 2. The data represents the average of two measurements.

 IAA (pg mg−1 FW)
Young leaves91821047
Mature leaves32242005

Tissue specific expression of PttIAA genes in the wood-forming zone

We investigated PttIAA gene expression in cells comprising the developmental gradient in the wood-forming tissues. As auxin distribution in these cells is known (Tuominen et al., 1997) it should be possible to examine how auxin distribution can regulate the expression of PttIAA genes in the context of its proposed role in xylem development. The results shown in Figure 3 indicate that PttIAA genes display a large variety in their expression patterns, which can be grouped into five different classes. PttIAA1 and PttIAA2 are expressed in most of the cambial region tissues from functional phloem to the expanding xylem cells. PttIAA3 and PttIAA4 display a more specific expression in the cambium and the adjacent zones of dividing and expanding xylem cells, with a peak in the division zone. A good correlation between expression and auxin maxima can be observed for these two genes. PttIAA6 and PttIAA7 resemble PttIAA3 and PttIAA4 in their expression pattern in the vicinity of the cambium, but in addition they show a strong signal in the outer cortex layers of the stem. A common denominator for the expression of these genes is that they are expressed in tissues that contain dividing cells. This is obvious for the cambium, which produces large numbers of new phloem, and above all, xylem cells, but also the cells in the cortex have to undergo divisions to account for the increase in diameter in an actively growing stem. A similar correlation between PSIAA4/5 expression and dividing cells has been noted before (Wong et al., 1996). It is therefore tempting to speculate about a role for PttIAA6 and PttIAA7 in dividing cells, in the case of PttIAA3 and PttIAA4 this role would be restricted to the cells in the cambium.

Figure 3.

Expression of PttIAA genes in stem tissues.

(a) Cross section through a hybrid aspen stem. The curve indicates the IAA concentration gradient as described by Tuominen et al. (1997). Horizontal bars indicate the location of tissue samples from different developmental zones. CO cortex, PF phloem fibres, PH phloem, CZ cambium, DX dividing xylem mother cells, EX expanding xylem, SW secondary wall formation. Scale bar in upper right hand corner 200 µm.

(b) Expression of PttIAA genes in the different developmental zones of the cambial region. Messenger RNA extracted from tangential sections through the stem was PCR amplified, spotted onto a nylon membrane and hybridized with 32P-labelled probes from PttIAA or Ubiquitin ESTs.

Of the PttIAA genes described here, PttIAA5 is unique in terms of its strong expression in the secondary wall-forming xylem cells. In addition it is also expressed in phloem fibres, which likewise build secondary walls, suggesting a role for this PttIAA gene in secondary wall formation. The expression pattern of PttIAA5 is intriguing in terms of the relatively low auxin content of the cell types in which it is expressed. It is interesting to note that a hybrid aspen homologue of the auxin influx carrier AUX1 is also localised to the very same cell layers as PttIAA5 (J. Schrader, unpublished results). Whether this indicates that at some stage the secondary wall forming cells experience an influx of IAA and that PttIAA5 has a role in mediating auxin responses in these cells during such an influx, remains to be investigated. Of all the genes characterized here, PttIAA8 displays the highest specificity in expression. PttIAA8 was expressed almost exclusively in the expansion zone with some residual signal also in the adjacent zone of dividing xylem mother cells. This exclusive expression of PttIAA8 suggests a function in processes such as elongation and/or transition of xylem mother cells from division to elongation. Alternatively, strict expression of genes such as PttIAA8 could be indicative of their role in setting up the boundaries between different developmental zones.

Effect of tension wood formation on PttIAA gene expression

Bending trees leads to the formation of tension wood on the upper side of the stem in angiosperms, and auxin is one of the key regulators of this process (Morey and Cronshaw, 1968; Timell, 1969), but its mode of action is not known. Therefore we investigated the expression of PttIAA genes and measured the auxin levels in bent hybrid aspen trees to obtain a better understanding of the role of PttIAA genes and auxin in this process (Figure 4). The first detectable alteration in gene expression could be observed after 6 h in PttIAA1 and PttIAA2 with the expression of these genes being reduced to about 50% of control levels. For both genes the signal reappears after 24 h but remains low after 5 and 11 days. PttIAA3 and PttIAA4 are unaffected by the bending treatment. A different pattern of expression is observed in the case of PttIAA7 with the expression increasing after 24 h, reaching its maximum at the end of the experimental period. Given the overlap of PttIAA1, PttIAA2 and PttIAA7 genes in the wood-forming zone, the induction of PttIAA7 after 5 days of bending must be specific for this gene and cannot be attributed to an increase in auxin levels in a specific tissue because this should lead to an induction of all three genes. This data is therefore suggestive of a role for PttIAA7 in regulating specific aspects of tension wood formation.

Figure 4.

Response to bending treatment.

Tension wood formation was induced in 9- to 11-year-old wild grown trees by bending to an angle of 45 degrees for the times indicated. Unbent trees served as a control. Total RNA was extracted from cambial scrapings of the upper, tension wood forming side; 25 µg RNA per lane were separated on an Agarose gel, blotted onto a nylon membrane and hybridized with 32P-labelled probes. The bottom lane shows an ethidium bromide stained gel of RNA aliquots prior to loading.

PttIAA gene expression during cambial dormancy

In order to understand the molecular basis of the loss of cambial sensitivity to auxin during dormancy in trees we analysed the regulation of the PttIAA genes during this process. The results in Figure 5 show that the expression of all the PttIAA genes is downregulated in the stems of dormant compared to actively growing plants. However, the individual PttIAA genes respond differentially to the dormancy-inducing signal. While most PttIAA genes are undetectable in dormant tissue (like PttIAA3 and PttIAA6–PttIAA8), others still show significant expression at this stage, markedly PttIAA2, which retains about 15% of its signal. To date most of the work on dormancy and auxin responses has focused on buds and not much is known about the applicability of these results to other tissues. Our results presented here showing differences in the extent of downregulation of PttIAA genes in dormancy despite overlapping expression patterns is indicative of a differential regulation by dormancy inducing conditions of genes with similar tissue specific expression patterns.

Figure 5.

Response to dormancy inducing conditions.

Trees were treated with short days (8 h light) for 11 weeks, by which time growth had ceased completely and leaves were shed. Expression of PttIAA genes was measured by relative RT-PCR in active (A) and dormant (D) stem tissue. For each gene the upper band is the gene specific signal while the lower one comes from the 18S ribosomal RNA internal standard.

Discussion

Auxin has been shown to influence diverse aspects of plant growth and development but its mode of action is still not completely understood. We have cloned cDNAs for 8 members of the Aux/IAA gene family from hybrid aspen, a woody angiosperm, and characterized their regulation in order to understand the role of this class of genes and auxin signalling in regulating wood formation. Our results point towards a complex modulation of PttIAA gene expression depending on the tissue, developmental and environmental context.

The PttIAA gene family in hybrid aspen

Sequence analysis of the PttIAA genes confirms the overall conservation of motifs in the Aux/IAA gene family, but at the same time identifies some novel features and supports the high degree of diversity and the existence of distinct subgroups within the Aux/IAA proteins. One such subgroup is formed by PttIAA3, PttIAA4, AtIAA20, At3g62100 and At3g17600, all of which are lacking the bipartite NLS formed by a conserved KR and a part of domain II. This bipartite NLS has been identified as the essential nuclear localisation signal in PS-IAA4 in a whole protein context (Abel and Theologis, 1995). Questions therefore arise on how its absence in PttIAA3, PttIAA4 and the other Aux/IAA proteins of this subgroup will affect their nuclear localisation. The lack of this bipartite NLS might allow either conditional nuclear localisation or affect the kinetics of this process.

An unusual feature found in PttIAA proteins is the alteration of conserved residues in domain II (PttIAA3 and PttIAA4) and in one case (PttIAA6) the duplication of this motif. Whether the replacement of glycine in the qvvgWPPirs motif of domain II with aspartic acid in PttIAA3 and PttIAA4 leads to higher stability of these proteins and if this has a functional relevance needs to be established. Since domain II has been implicated in the degradation of Aux/IAA proteins (Ramos et al., 2001; Worley et al., 2000), the duplication of this motif in PttIAA6 could potentially provide a second target for the degradation machinery. It is important to note however, that the second valine in the qvvgWPPi motif, which is conserved in 19 of the Arabidopsis Aux/IAA proteins and therefore is probably functionally important, is replaced with glutamic acid in the duplication in PttIAA6 raising questions regarding the functionality of this second copy of domain II in PttIAA6. The unusual features in some of the PttIAA proteins described above may be indicative of novel modes of regulation of these proteins adding to the overall complexity of auxin signalling.

PttIAA genes are auxin inducible

Like the majority of the Arabidopsis genes, all of the PttIAA genes investigated here are auxin inducible, but differences were observed with respect to the requirement of protein synthesis for auxin induction. As can be seen in Figure 2, with the exception of PttIAA3 and PttIAA4, most of the PttIAA genes appear to be secondary response genes requiring protein synthesis for their auxin induction. Interestingly, cycloheximide has a stimulatory effect on the expression of several PttIAA genes and a possible explanation for this is that cycloheximide acts through stabilising the mRNA of these genes. Such an mRNA stabilising effect of cycloheximide has also been noted for other Aux/IAA genes (Koshiba et al., 1995) and in case of SAUR-AC1, an auxin-regulated gene from Arabidopsis it has been shown that an mRNA destabilising element plays an important role in its regulation (Gil and Green, 1996).

Auxin induction of PttIAA genes is developmentally regulated

The finding that auxin inducibility of PttIAA1–PttIAA8 is abolished in mature leaves indicates a developmental regulation of their auxin induction. This loss of induction might be related to a possible detrimental effect of inappropriate expression of these genes on plant development. For example, inappropriate expression of Aux/IAA genes as shown to occur in the axr2 and axr3 mutants (Abel and Theologis, 1995; Ouellet et al., 2001) results in altered responses to auxin which in turn lead to aberrant development (Leyser et al., 1996; Lincoln et al., 1993; Sabatini et al., 1999; Tian and Reed, 1999). In the case of mature leaves it has been shown that they have a low turnover of auxin (Ljung et al., 2001; G. Sandberg, unpublished results). Mature leaves are unable to counteract rapid increases in auxin levels by up-regulation of conjugation or catabolism. Preventing auxin induction of genes such as those of the Aux/IAA gene family would serve to block at least a part of the auxin signalling pathway in mature leaves and this could serve to protect them from any harmful side-effects resulting from sudden increases in auxin levels.

The control of auxin induction in a developmental stage specific manner may provide the switch regulating the competence of a cell or tissue to respond to a developmental signal, thereby creating a sensing mechanism regulating developmental transitions. In this respect, it is interesting to consider the mechanism of vascular differentiation in leaves, where the pattern of vascular strands is fixed early in development. In Arabidopsis, application of auxin transport inhibitors and the subsequent auxin accumulation can alter vascular patterns only during the very young stages of leaf development, while mature leaves become progressively insensitive to the changes in auxin status (Mattsson et al., 1999). One way to attain this tight control of vascular differentiation is to terminate the expression or inducibility of the mediators of auxin signalling at a later stage in development. Whether the lack of auxin response in mature leaves is a phenomenon common to all Aux/IAA genes remains to be seen but, our recent finding of a similar loss of auxin induction in mature leaves of Arabidopsis indicates that at least a subgroup of the gene family shows the same behaviour (J. Schrader and R.P. Bhalerao, unpublished results).

These results lead to the question of a possible mechanism that may account for the lack of auxin inducibility of the Aux/IAA genes in mature hybrid aspen leaves. Some clues regarding this come from comparing the lack of inducibility in mature leaves with altered expression of the Aux/IAA genes in mutants of Arabidopsis. It has been shown that various auxin response mutants affect the expression of Aux/IAA genes differentially (Abel et al., 1995). In the axr1 mutant for instance, the steady state levels of most of the Aux/IAA genes are reduced, but the induction of the genes is still functional. The axr2 mutant in contrast not only shows altered steady state levels, it is also impaired in the induction of the Aux/IAA genes (Abel et al., 1995). Thus, the loss of auxin inducibility of PttIAA genes resembles the situation in the axr2 mutant. It has been shown that AXR2 functions as a transcriptional repressor and dominant mutations in AXR2 which stabilise the mutated protein further enhance its ability to serve as a repressor (Gray and Estelle, 2000; Gray et al., 2001; Ruegger et al., 1998; Tiwari et al., 2001). Whether the repression of PttIAA genes described here results from stabilisation of a candidate PttIAA repressor in mature leaves needs to be investigated. However, if this is the case, then this particular PttIAA repressor must be expressed only in mature leaves and should be resistant to AXR1-TIR1 mediated turnover even after the addition of auxin.

The presence of a conserved auxin response element in the promoters of many Aux/IAA genes suggests an alternative mechanism for the lack of auxin inducibility of PttIAA1–PttIAA8 in mature leaves. This auxin response element is required for imparting auxin inducibility on auxin responsive genes like PS-IAA4/5 and GH3 (Abel and Theologis, 1996; Guilfoyle et al., 1998). It can be bound by members of the ARF family of transcription factors leading to either activation or repression of transcription (Ulmasov et al., 1999). Thus, if ARF transcription factors regulate the activity of PttIAA genes by binding to their upstream elements, then either the lack of a relevant ARF activator or the presence of an ARF repressor in mature leaves, compared to young leaves, could explain the loss of induction of the PttIAA genes in mature leaves. We have cloned three ARFs from hybrid aspen and are currently analysing their expression in young and mature leaves to test this hypothesis.

PttIAA genes are differentially expressed in the wood-forming zone

We investigated the tissue- and cell-type specific expression patterns of the PttIAA genes in the wood-forming zone to understand their roles during developmental transitions. The distinct spatial expression patterns of the PttIAA genes in the wood-forming zone (Figure 3) is interesting in view of the proposed role of auxin as a positional signal in xylem development (Uggla et al., 1996). The overlap between a developmental gradient of dividing, elongating and secondary wall-forming cells in xylem and the concentration gradient of auxin has led to the suggestion for auxin as a positional signal regulating sequential aspects of xylem development.

The hypothesis of auxin as a positional signal requires the existence of a mechanism consisting of regulatory genes that utilise information from the auxin concentration gradient to orchestrate the sequential execution of specific gene expression programmes leading to the formation of the developmental gradient in xylem. Several lines of evidence argue strongly in favour of PttIAA genes as potential regulators acting downstream of the auxin concentration gradient to control sequential aspects of gene expression in developing xylem. Firstly, Aux/IAA proteins form homo- and heterodimers, interact with ARFs and are able to influence transcription (Kim et al., 1997; Ouellet et al., 2001; Ulmasov et al., 1997). Secondly, several of the Aux/IAA genes have been shown to repress transcription and their stability and therefore their ability to repress transcription is regulated by auxin in a dose-dependent manner (Gray et al., 2001; Tiwari et al., 2001). Thirdly, PttIAA genes display distinct expression patterns in the tissues that comprise the developmental gradient in xylem as shown here. The existence of a large number of Aux/IAA and ARF proteins in Arabidopsis and hybrid aspen suggests the possibility of an equally large variety of transcription factor complexes that can form between these two classes of proteins. All these features of Aux/IAA and ARFs suggest the possibility of generating a large number of tissue specific Aux/IAA-ARF transcriptional regulators that could function in an auxin dose dependent manner, thereby linking the auxin concentration gradient to sequential and tissue specific gene expression and leading to the formation of the developmental gradient in the xylem. In view of the proposed role of PttIAA genes in xylem development one would expect that expression of PttIAA genes would be affected when the normal course of xylem development is altered. In this respect it is worth mentioning that alteration in the expression of the PttIAA genes occurs during tension wood formation which has also been shown to cause altered xylem development (see below).

The data presented in Figure 3 are to our knowledge the first investigation of the expression of Aux/IAA genes in cell types in which the endogenous auxin levels are known. The results indicate that the expression of the PttIAA genes is not always restricted to cells with high auxin content as exemplified by the expression of PttIAA5 exclusively in secondary wall-forming tissues that have low auxin levels. This observation corroborates similar findings in Arabidopsis (Wyatt et al., 1993) indicating that the upstream regulatory elements of PttIAA genes must combine elements conferring tissue specificity with elements responsible for auxin inducibility. Such a composite promoter has been described for the GH3 gene from soybean (Ulmasov et al., 1995) suggesting that a similar mechanism could be responsible for the control of PttIAA expression. The potential interactions between the cis-regulatory elements directing tissue specificity and auxin response deserves further investigation in order to better understand how plants integrate information relating to tissue context and auxin status.

Tension wood induction is accompanied by alterations in PttIAA gene expression

In tension wood formation, cambial cell division and xylem development are altered, resulting in the production of wood with modified properties. Auxin is known to be involved in this process (Timell, 1969), however, it is not known whether stimuli that induce tension wood formation affect auxin distribution or if they modulate the sensitivity of the xylem cells to auxin. Furthermore, it is not known how changes in auxin distribution or sensitivity could bring about the observed changes in the xylem that accompany the formation of tension wood. The regulation of PttIAA genes by auxin and their distinctive expression patterns in the wood-forming zone prompted us to investigate whether this class of genes might be involved in tension wood formation.

The rapid downregulation of PttIAA genes reported here must be one of the earliest known responses during tension wood formation as to date most of the studies on tension wood have investigated this process after several days when the effects of bending are already apparent. Further, most of these studies have focused on changes in anatomical features rather than the underlying molecular control. Our data suggests that even after 24 h of bending there is no measurable change in auxin levels of xylem or phloem (Table 2) and yet the expression of specific PttIAA genes is differentially altered. This data would mean that it is not a redistribution of auxin upon bending that causes the initiation of diverse aspects of tension wood formation. However it is possible that following bending the auxin sensitivity of specific cells could be altered leading to the initiation of tension wood formation and this change in auxin sensitivity could be reflected in the changes taking place in the expression of the PttIAA genes. While the data reported here is far from explaining all the features of tension wood formation, it implicates at least a subset of PttIAA genes and therefore auxin signalling intermediates in influencing wood formation.

Table 2.  Auxin content of stem tissues after induction of tension wood by bending
Time after bending030 min1 h24 h11 d
  1. Trees were bent to 45 degrees and stem sections were sampled at the times indicated. After peeling off the bark the inside of the bark and the outside of the stem were scraped to provide phloem and xylem samples, respectively. Tissue from three trees per time point was pooled and IAA content is reported as average and sd of five measurements per time point.

 IAA (pg mg−1 FW) 
Phloem683 ± 42616 ± 23570 ± 19521 ± 29535 ± 57
Xylem299 ± 7272 ± 16257 ± 13342 ± 44281 ± 20

PttIAA gene expression is downregulated during cambial dormancy

Growth cessation during cambial transition into dormancy is accompanied by the loss of cambial sensitivity to auxin but the molecular mechanism of this phenomenon is not well understood (Little and Bonga, 1974). Auxin promotes and is required for continued cell division in the cambium (Little and Bonga, 1974; Savidge, 1983). However, since auxin levels in the cambium proper do not undergo a reduction upon transition into dormancy (Uggla et al., 1996), it cannot be a lack of auxin that causes the cell division to cease. Hence, it has been proposed that the entry into dormancy involves a loss of sensitivity of the cambium to auxin (Little and Bonga, 1974). There exists, however, no data on how this change in sensitivity is achieved by the plant. One possible explanation comes from the temporal correlation between the downregulation of the PttIAA genes, which are key components of auxin signalling, and the transient insensitivity of the cambium to applied IAA. Given that auxin levels are not reduced in the cambium during dormancy, it appears unlikely that the reduction in the expression of PttIAA genes, such as that observed for PttIAA3, which is expressed only in the cambium, results from reduced auxin levels. It can be rather suggested that the downregulation of PttIAA3 must be brought about by some other mechanism that is specific for the transition of the cambium into dormancy and maintaining it in the dormant state. In this context it is important to note that the reduction in the PttIAA transcript levels is not simply due to a general reduction in transcription upon the transition of the cambium into dormancy. Control experiments using several other genes, e.g. CDC2A, Retinoblastoma, and even a hybrid aspen TIR1 homologue showed virtually no change in the transcript levels between active and dormant plants (Saxena et al., unpublished results). Hence, it is possible that the downregulation of PttIAA genes is part of a specific mechanism that renders the cambium insensitive to applied IAA during dormancy.

While PttIAA genes are downregulated in dormant plants, how this downregulation is connected to cessation of cell division is still unclear. There is no data as yet on the regulation of cell cycle genes in auxin resistant mutant such axr2, axr3 and shy2 that are mutated in Aux/IAA genes. However our data indicates that several cell cycle genes are regulated by auxin in hybrid aspen (Schmidt et al., unpublished results) and similar results have been obtained in Arabidopsis (Doerner et al., 1996). Additionally, several of the PttIAA genes are expressed in the dividing cells and may be involved in linking the regulation of cell division to auxin signalling during dormancy.

In the future, it will be important to investigate the kinetics of change in PttIAA gene expression upon transfer to dormancy-inducing conditions, as it is possible that not all genes respond uniformly. It will also be necessary to analyse whether it is possible to induce the PttIAA genes by IAA following their downregulation in dormant tissues to get a better understanding of how auxin responsiveness is regulated by environmental clues.

Experimental procedures

Expression in the cambial region

A 17.0 × 2.5 mm block from the cambial region of a hybrid aspen stem was cut into 30 µm thick tangential sections. The sample sections were pooled into the following groups representing different developmental zones: CO, cortex, between cork and the first fibre bundles; PF, phloem fibres; PH, functional phloem; CZ, cambial zone with differentiating phloem; DX, dividing xylem mother cells; EX, expanding xylem; SW, zone of visible secondary wall formation. From each pool of sections poly(A)+ mRNA was extracted, Reverse-transcribed and PCR amplified essentially as described by Hertzberg and Olsson (1998). Approximately 500 ng of PCR product were dotted onto a HybondN nylon membrane (Amersham Pharmacia Biotech, Uppsala, Sweden) and hybridized with 32P probes derived from the PttIAA EST clones.

Induction by IAA and cycloheximide

Young leaves (first 2–3 leaves with a size of < 15 mm) and mature leaves (below the 25th leaf) were cut into small pieces of < 1 cm2. The leaf samples were subject to the following treatments: (i) Untreated (ii) Control treated 1 h in ½ strength Murashige and Skoog medium (½ MS) (iii) 1 h in ½ MS with 10 µm IAA (iv) 1 h in ½ MS with 50 µm cycloheximide (v) 30 min in ½ MS with 50 µm cycloheximide after which IAA was added to 10 µm and incubation continued for 1 h. The samples were frozen in liquid nitrogen and mRNA was extracted from crude tissue extracts using paramagnetic oligo (dT)25 Dynabeads (Dynal AS, Norway) according to the manufacturer's instructions

Approximately 600 ng of poly(A)+ mRNA was loaded on a 1% formaldehyde/formamide agarose gel, separated by electrophoresis and blotted onto a Hybond N nylon membrane and probed as described above.

Auxin content of control, MS and IAA treated leaves was measured in two 10 mg samples for each treatment using a isotope dilution mass spectrometry method described by Edlund et al. (1995).

Analysis of PttIAA expression in active and dormant stems

Plants from Populus tremula×tremuloides clone T89 were transferred from the greenhouse where they were growing under 18 h light at 25°C and 6 h darkness at 15° with at least 300 µmol s−1 m−2 of light to a growth chamber with 8 h light of approximately 500 µmol s−1 m−2 and temperatures of 20°C daytime and 15°C at night. At the time of transfer the plants were approximately 60 cm tall. After 6 weeks in the growth chamber the temperature was dropped to 15°C day and 10°C night. Samples were taken on the day of transfer and after a total of 11 weeks in short day conditions.

For expression analysis total RNA was isolated from internodes 8–16 with a method described by Chang et al. (1993) followed by an additional purification step using RNEASY columns (Qiagen, Germany). 2 µg of total RNA was transcribed into cDNA with a first strand cDNA synthesis kit (Amersham Pharmacia Biotech, Uppsala, Sweden). PttIAA gene expression was determined by relative PCR with help of the QUANTUM 18S RNA internal standard kit (Ambion, Austin, TX, USA) following the manufacturer's recommendations. The PCR bands were separated on 1.5% agarose gels and visualized by ethidium bromide staining.

Tension wood analysis

Wild grown, 9- to 11-year-old, Populus tremula trees were bent at an angle of approximate 45 degrees, and kept in this position in order to induce tension wood. Stem pieces from the bent region were sampled after 0.5 h, 1 h, 3 h, 6 h, 12 h, 24 h, 5 days and 11 days. The cambial region tissue was collected from the upper, tension wood forming side of each of the bent, as well as from non-bent (control), stem pieces. The collection was done by peeling off the bark and scraping the inside of the bark (phloem) and the outside of the stem (xylem) separately. Tissue from three trees per time point was pooled and total RNA was then extracted from the scrapings according to Chang et al. (1993). About 12,5 µg of xylem and phloem RNA were pooled, denatured and separated on a 1% glyoxal/agarose gel. Separated RNA was transferred to a nylon filter (Hybond-N, Amersham, UK) and hybridized with radiolabelled probes as described above under Expression in the cambial region. An ethidium bromide-stained agarose-gel was used as a loading control.

IAA content was measured for control 0.5 h, 1 h, 24 h and 11 day samples. Five 10 mg samples of xylem and phloem scraping for each time point were analysed for their IAA content using an isotope dilution mass spectrometry method described by Edlund et al. (1995).

Acknowledgements

This work was supported by grants from NFR and EU project PopWood to R.P.B. We thank R. Granbom for auxin measurements and I. Sandström for technical assistance.

Ancillary