SEARCH

SEARCH BY CITATION

Keywords:

  • wood development;
  • plant hormones;
  • cell growth;
  • lignin;
  • secondary xylem

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Both indole acetic acid (IAA) and gibberellins (GAs) stimulate cell and organ growth. We have examined GA/IAA cross-talk in cambial growth of hybrid aspen (Populus tremula×tremuloides). Decapitated trees were fed with IAA and GA, alone and in combination. Endogenous hormone levels after feeding were measured, by mass spectrometry, in the stem tissues below the point of application. These stem tissues with defined hormone balances were also used for global transcriptome analysis, and the abundance of selected transcripts was measured by real-time reverse-transcription polymerase chain reaction. By feeding isotope-labeled IAA, we demonstrated that GA increases auxin levels in the stem by stimulating polar auxin transport. This finding adds a new dimension to the concept that the endogenous GA/IAA balance in plants is determined by cross-talk between the two hormones. We also show that GA has a common transcriptome with auxin, including many transcripts related to cell growth. This finding provides molecular support to physiological experiments demonstrating that either hormone can induce growth if the other hormone is absent/deficient because of mutations or experimental treatments. It also highlights the potential for extensive cross-talk between GA- and auxin-induced responses in vegetative growth of the intact plant. The role of endogenous IAA and GA in wood development is discussed.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Auxin and gibberellin (GA) are key signals in plant growth, and are often observed to act synergistically. Such synergism may be due to additive effects on independent or common response pathways, and/or to one hormone affecting the concentrations of the other in the target tissue. Interactions between auxin and GA have been intensively studied in the elongation growth of primary stems, a process that is stimulated by both substances (Ross et al., 2003; Cleland, 2004; Reid et al., 2004). An early hypothesis suggested that auxin was the active hormone in the elongation response, and that the observed GA effect was mediated by increases in auxin biosynthesis or auxin transport. Indeed, several studies later found that increases in GA levels following its application to intact plants, or over-expression in transgenic trees, led to significant increases in indole acetic acid (IAA) levels (Law and Hamilton, 1984; Israelsson et al., 2003; Ross et al., 2003). However, more recent studies have demonstrated that GA biosynthesis genes (GA20ox and/or GA3ox) are induced when auxin is supplied to decapitated plants of various species (Ross et al., 2000; Wolbang and Ross, 2001; Wolbang et al., 2004) or Arabidopsis seedlings (Frigerio et al., 2006). These findings imply that biosynthesis of the endogenous GA required for elongation growth is dependent on the supply of polarly transported auxin. In addition to this hypothesis, it has also been suggested that polarly transported auxin facilitates GA-induced degradation of DELLA proteins, and is therefore required for GA stimulation of root elongation (Fu and Harberd, 2003). In contrast, research on mutants deficient in GA biosynthesis and/or auxin responses has demonstrated that both hormones can stimulate elongation growth of pea internodes and Arabidopsis hypocotyls (Yang et al., 1996; Barratt and Davies, 1997; Collett et al., 2000).

Similar to elongation growth in primary stems, auxin and GA have profound effects on cambial growth in secondary stems. Early studies on auxin demonstrated that an exogenous source of IAA could replace the apical shoot in inducing cambial cell division and secondary xylem differentiation in sunflower (Snow, 1935; Gouwentak, 1941). These experiments have since been repeated and extended in numerous gymnosperm and angiosperm species, including model species such as Arabidopsis and Populus (Sundberg et al., 2000; Little et al., 2002). From such studies, it was concluded that the apical shoot is a major source of polarly transported IAA in secondary stems (Sundberg and Uggla, 1998), and that auxin supplied to the cambial tissues by polar transport maintains the tissue integrity and cell division activity of the cambial meristem and stimulates xylogenesis of cambial derivatives, and (hence) promotes cambial growth (Sundberg et al., 2000). Furthermore, a concentration gradient of auxin is maintained across developing wood tissues and has been postulated to provide positional information for meristem development (Uggla et al., 1996; Tuominen et al., 1997), a concept later developed and confirmed for Arabidopsis root and shoot meristems (for review, see Bhalerao and Bennett, 2003; Swarup and Bennett, 2003). The auxin concentration peaks in the vascular cambium, which is the major compartment for polarly transported auxin (Sundberg et al., 2000). Schrader et al. (2003) cloned two PIN genes, PttPIN1 and PttPIN2, and localized their expression to the cambial meristem and expanding xylem derivatives. These PIN proteins are likely to play key roles in the polar transport of auxin in the Populus stem (Schrader et al., 2003), while other putative auxin efflux and influx proteins that are expressed in later stages of vascular development have been suggested to be important in maintaining the auxin gradient and/or cell-type-specific auxin concentrations, as demonstrated for PIN proteins in Arabidopsis root and shoot meristems (Benkova et al., 2003; Leyser, 2005).

When GA was experimentally applied to decapitated stems in a similar manner to auxin, division and expansion of cambial zone cells, but not xylem differentiation, were found to be stimulated in various angiosperm trees, herbaceous species and (commonly but not always) conifers (Wareing et al., 1964; Little and Savidge, 1987). When both auxin and GA were applied together, however, a pronounced synergistic effect on cambial growth was observed, similar to that observed in primary stems (Wareing, 1958; Wareing et al., 1964; Little and Savidge, 1987). Digby and Wareing (1966) described this synergism in more detail in Populus.

In this study, we have revisited the experiment by Digby and Wareing (1966) to further explore the cross-talk between auxin and GAs in cambial growth using modern methods. Gas chromatography/mass spectrometric (GC/MS) measurements of hormone concentrations in the stem tissues resulting from apical feeding, combined with gene expression analysis, have allowed us to demonstrate that GA stimulates polar auxin transport. This novel finding adds another dimension to our knowledge of GA/auxin interactions in plants. We also analyzed the transcriptomal changes (using a poplar microarray) induced by various defined GA/IAA ratios in the experimental stem tissues. The results revealed that most of the transcriptomal changes induced by increases in GA levels (in an auxin-depleted background) were also induced by increases in auxin levels (with a constant GA level). This observation provides molecular data to complement earlier experiments showing that both hormones can induce growth if the other hormone is absent or deficient because of mutations or experimental treatments.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Auxin supplied to stem tissues via the polar transport pathway stimulates cambial growth in a dose-dependent manner

Numerous experiments in herbaceous and woody species have demonstrated that labeled auxin apically applied to decapitated stems, or stem segments, is basipetally transported in a manner characteristic of polar auxin transport (Goldsmith, 1977; Little and Savidge, 1987; Sundberg et al., 2000). In woody species, microautoradiography has shown that the label is localized to the cambial tissues (Lachaud and Bonnemain,1984), and, by using tangential cryo-dissectioning in combination with GC/MS analysis, we similarly found, in preliminary experiments, that the applied auxin is distributed across the cambial tissues in a manner that mimics that of endogenous IAA in intact trees (data not shown). Apical feeding to decapitated stems has been extensively exploited to evaluate the role of auxin in cambial growth and xylogenesis. However, the resulting internal hormone level in the underlying tissues has rarely been measured, although such characterization is essential for correct interpretation of molecular and physiological responses. In an initial experiment, we investigated internal IAA levels resulting from apical feeding in relation to the resulting cambial growth (Figure 1). The cambial growth response was examined 3 weeks later, to allow time for secondary xylem development, and the auxin concentrations in the stem segment below the auxin source were measured after 1 and 3 weeks. In lanolin-treated trees (lanolin control), IAA was depleted in the stem tissues and no xylem production was observed (Figure 2b). Apical auxin treatment resulted in cambial growth and xylem development (Figure 2c), and increases in IAA concentrations in the stem tissues. However, it should be noted that only the highest auxin concentration (10 mg g−1) resulted in similar concentrations to those in intact trees after a week, and even with this high concentration of applied auxin, the internal IAA level decreased to sub-optimal levels during the 3-week experimental period. The experiment demonstrated dose-dependent, correlated responses in internal IAA levels and cambial growth (Figure 1), although their exact relationship cannot be determined as the IAA concentrations (and probably growth rates) changed during the experiment. Nevertheless, the results provide quantitative data to support the concept that the amount of IAA supplied to the cambial tissues by polar transport is a key regulator of secondary xylem (wood) production.

image

Figure 1.  Internal indole acetic acid (IAA) concentrations and xylem growth resulting from application of IAA to decapitated Populus trees. Trees were decapitated at a position just below the point of initiation of secondary xylem growth. IAA (0.3, 1 or 10 mg) in lanolin was applied to the cut end. (a) Width of the radial xylem after 3 weeks of treatment, measured 1.5 cm below the decapitation position. Intact control trees (IC) were measured at a corresponding position. The width of the xylem in the lanolin-treated control trees was not significantly different from its width at the start of the experiment (data not shown). (b) Internal IAA levels measured after 1 and 3 weeks in the extra-xylary tissues 5–25 mm below decapitation, and in the intact trees (IC) at a corresponding position. LC, lanolin control; IC, intact control, 0.3, 1, 10 = 0.3, 1 or 10 mg IAA per g lanolin. Values are means ± SD of five independent biological replicates (each replicate sampled from one individual tree).

Download figure to PowerPoint

image

Figure 2.  Anatomy of stem tissues after application of indole acetic acid (IAA) and gibberellins (GA) to decapitated Populus trees. Trees were decapitated at a position just below the point of initiation of secondary xylem growth, and 1-mg IAA or 0.5 mg GA4 per g lanolin were applied to the cut end alone and in combination. Transverse sections were obtained 10 mm below the decapitation position after 3 weeks of treatment. (a) Intact control trees (the developing expanding xylem is crushed). (b) Lanolin-treated control trees. Note the absence of cambial growth indicated by the lack of developing xylem cells. (c) IAA-treated trees. Note the small vessels (arrow) formed in response to sub-optimal IAA supply. (d) GA-treated trees. Note the absence of induced xylogenesis. (e) IAA- and GA-treated trees. Note the enhanced cambial growth compared with the IAA-treated trees. (f) IAA-treated trees; paraffin-embedded section stained with phoroglucinol to visualize lignin, showing the delayed lignification of fibers. X, secondary xylem; P, phloem; VC, vascular cambium; V, vessel; F, fiber; IG, induced growth in response to hormone treatment; IXG, induced secondary xylem growth in response to hormone treatment; LF, lignified fibers; NLF, non-lignified secondary wall fibers; LF, lignified vessel. Scale bar = 100 μm.

Download figure to PowerPoint

Anatomically and histochemically, the xylem that formed under sub-optimal auxin concentrations resulting from auxin feeding differed in several respects from that in the intact plants (Figure 2c). The diameter of the vessel elements was clearly much smaller following auxin feeding. Lignification of the fibers was delayed, resulting in islands of lignified vessels in the early stages of xylem development (Figure 2f) (Digby and Wareing, 1966). The xylem that formed was also less organized, with more variable cell sizes, than in control trees.

GA stimulates auxin polar transport

The synergistic effect of GA on auxin-induced cambial growth in Populus was investigated in an experiment in which 1 mg g−1 of IAA and/or 0.5 mg ml−1 GA4 were applied in lanolin. GA4 was used because, together with GA1, it is the major bioactive GA in Populus (Israelsson et al., 2004). GA alone stimulates cell division in the cambial zone but not xylogenesis (Figure 2d), and resulted in the apparent loss of an easily distinguished cambial meristem. Possible the auxin deficiency in these stems results in de-differentiation and loss of meristem identity of the newly formed cells. Application of GA together with auxin had the expected synergistic effect on cambial growth (Figure 2e) (Digby and Wareing, 1966). The xylem that formed was similar in structure to that formed when IAA was applied alone.

The internal concentration of IAA resulting from each of the feeding treatments was measured 2 and 6 days later (shorter times than in the experiment described above because preliminary experiments showed that the internal IAA concentration was highest during the first days of feeding). The results showed that, after 2 days of IAA treatment, the internal IAA levels were similar to endogenous levels in intact trees, and then decreased by day 6 to similar levels to those observed earlier (Figure 3a). However, a much less expected finding was that the combined feeding of GA and IAA resulted in internal IAA levels after 2 days that were about twice as high as when IAA was applied alone, and they were still slightly higher than when IAA was applied alone after 6 days (Figure 3a). The GA-induced increase in IAA levels could have been due either to increased IAA biosynthesis in the stem or increased transport of IAA from the apical source. To distinguish between these possibilities, we performed an experiment in which the stable isotope [13C6]IAA was used as the apical auxin source, allowing the IAA originating from the apical source ([13C6]IAA) and the non-labeled IAA originating from endogenous biosynthesis ([12C6]IAA) to be separately quantified. IAA was measured in the stem, this time 24 h after decapitation and feeding, and again the level was found to be about twice as high when auxin was applied in combination with GA compared with application of auxin alone (Figure 3b). As almost all of the IAA detected was in the form of the heavy [13C6] isotope, it can be concluded that GA stimulated the transport of IAA from the applied source.

image

Figure 3.  Internal indole acetic acid (IAA) concentration resulting from application of IAA and gibberellins (GA) to decapitated Populus trees. Trees were decapitated at a position just below the point of initiation of secondary xylem growth. Internal hormone levels were measured in the extra-xylary tissues 5–25 mm below the decapitation position (and at the corresponding position of intact control plants). IAA (1 mg) or GA4 (0.5 mg) per g lanolin were applied to the cut end alone and in combination. (a) Concentrations of IAA after 2 and 6 days of apical feeding. The IAA levels significantly differed between IAA- and IAA/GA-treated trees at both day 2 and day 6 (t-test, ≤ 0.001). (b) Concentrations of [12C6]IAA (of endogenous origin) and [13C6]IAA (originating from apical feeding) after 1 day of treatment. The sum of [12C6]IAA and [13C6]IAA makes up the total free IAA in the measured tissues. The [13C6]IAA levels significantly differed between IAA- and IAA/GA-treated trees (t-test, P ≤ 0.001). Values are means ± SD of five independent biological replicates (each replicate sampled from one individual tree).

Download figure to PowerPoint

PttPIN1 is synergistically induced by GA and IAA

The finding that GA enhanced the amount of polarly transported auxin from the apical source led us to investigate the expression of PIN genes under the various experimental conditions. We focused on PttPIN1 and 2 (putative auxin efflux transporters), which have been identified as the most prominent PIN genes expressed in the cambial tissues of the Populus stem (Schrader et al., 2003), the major compartment for polar auxin transport (Sundberg et al., 2000). Real-time reverse transcription polymerase chain reaction (RT-PCR) analysis revealed that levels of PttPIN1 transcripts were reduced in lanolin-treated and auxin-depleted tissues, but induced by IAA application (Figure 4). We also found that PttPIN1 was 50% induced by the GA treatment, but observed a prominent synergistic effect on its expression when GA was applied together with IAA. Using Northern dot blot analyses, Schrader et al. (2003) found that PttPIN2 was co-expressed in the cambium with PttPIN1. Real-time RT-PCR analysis with gene-specific primers revealed, however, that it was much more weakly expressed than PttPIN1, and its expression did not vary much following the various treatments (data not shown). Therefore, we conclude that PttPIN1 is the major PIN gene expressed in the cambial meristem. Our data show that GA stimulates PttPIN1 expression, but to a lesser extent than IAA. It can be hypothesized that this small induction by GA of the PttPIN1 transcript results in a small increase in IAA transport and IAA level, which in turn further stimulates PttPIN1 expression, finally resulting in the large increase in PttPIN1 observed after combined GA/IAA feeding. However, the possibility cannot be excluded that GA stimulation of auxin transport acts through some other mechanism, and the increase in PttPIN1 observed after combined GA/IAA feeding is solely due to the increase in IAA.

image

Figure 4.  Abundance of transcripts involved in auxin transport and gibberellin (GA) biosynthesis/signaling in Populus stem tissues with different GA/indole acetic acid (IAA) balances. Experimental conditions were as described in the legend to Figure 3. Real-time reverse transcription polymerase chain reaction was used to estimate the transcript abundance of genes after 1 day of treatment with external hormone sources. 18S rRNA was used as an internal control. For each treatment, two independent biological replicates (consisting of pooled tissues from two and three trees, respectively) were analyzed. Values are means ± SD of three technical replicates. IC, intact control; LC, lanolin control; IAA, 1-mg IAA g−1 lanolin; GA, 0.5 mg GA g−1 lanolin; IAA/GA, combined treatment with 1-mg IAA and 0.5 mg GA g−1 lanolin.

Download figure to PowerPoint

Resulting GA levels after various feeding treatments

We also wished to document the internal stem levels of GAs under our experimental conditions to confirm that the applied GA4 was transported down the stem and affected the GA concentrations in the analyzed tissues. For this purpose, we measured the two major bioactive GAs, GA4 and GA1, together with GA34 (the de-activation product of GA4) 1 and 2 days after decapitation and apical feeding with IAA and GA alone and in combination (Figure 5). The results showed that decapitation did not significantly affect the endogenous levels of bioactive GAs, and that the application of GA4 increased the internal levels of both GA4 and GA34 about fivefold. In addition, the level of GA1 was increased after 2 days, indicating that some of the applied GA4 was metabolized to GA1. The application of auxin also tended (albeit not statistically significantly) to increase the GA4 concentration to higher levels than those in lanolin-treated trees. Taken together, the results show that apically applied GA4 is transported basipetally and induces increases in biologically active GAs to levels well above those in control trees.

image

Figure 5.  Internal gibberellins (GA) concentration resulting from application of indole acetic acid (IAA) and GA to decapitated Populus trees. Experimental conditions were as described in the legend to Figure 3. Concentrations of GA4, GA1 and GA34 after 1 and 2 days of apical feeding are shown. The values are means ± SD of five biological replicates (each replicate sampled from one individual tree).

Download figure to PowerPoint

Auxin induces GA biosynthesis genes and inhibits GA degradation genes

We have previously identified the major GA metabolism genes in the Populus stem (Israelsson et al., 2005), and here we performed real-time RT-PCR measurements of these transcripts 1 day after decapitation and hormone feeding (Figure 4). Expression of PttCPS1, which encodes ent-copalyl diphosphate synthase (CPS), the first enzyme committed to GA biosynthesis, decreased after decapitation and was induced by IAA but not by GA. Expression levels of both the PttGA20ox1 and PttGA20ox4 genes, which encode the multi-functional enzyme GA20-oxidase, decreased strongly in the decapitated and auxin-depleted tissues. PttGA20ox1 expression was induced by IAA and decreased in tissues with increased GA levels when GA was added in combination with auxin, supporting earlier results suggesting that PttGA20ox1 is under feedback control (Eriksson and Moritz, 2002). PttGA20ox4 was similarly induced by IAA, but was not affected by the increased GA levels resulting from the combined IAA/GA feeding. PttGA3ox encodes an enzyme that catalyzes the last step in the biosynthesis of bioactive GAs. This gene is expressed at very low levels in the Populus stem (Israelsson et al., 2005), and no reproducible expression results were obtained. Finally, the PttGA2ox1 and PttGA2ox2 genes encoding GA deactivation enzymes showed different, complementary patterns. Expression of PttGA2ox1 was weaker in the decapitated stem than in control trees, and was only restored to wild-type levels by combined IAA/GA feeding, whereas PttGA2ox2 was not expressed in the intact trees but strongly induced by decapitation. GA feeding reduced its expression, but levels in GA-supplied trees were still higher than those in control trees. IAA application reduced the expression level to that observed in control trees. In conclusion, all of these results are consistent with the idea that auxin supplied by polar transport stimulates the expression of GA biosynthesis genes in stem tissues (Ross et al., 2000; Frigerio et al., 2006).

PttDELLA-like1 has been identified as a putative component of GA response pathways in Populus, and its expression has been found to be co-localized with endogenous GA1 and GA4 in the cambial region tissues (Israelsson et al., 2005). The expression of PttDELLA-like1 was little affected by decapitation and slightly induced by applications of either GA or IAA. However, its expression was markedly increased by combined IAA/GA feeding.

Global gene expression analysis reveals a common transcriptome for GA and auxin

The Populus microarray (Andersson et al., 2004) was used to analyze the global gene expression in stem tissues sampled 24 h after treatment with various defined hormone balances (Figures 3 and 5). In the lanolin controls, stem tissues were depleted in IAA and GA levels were low. IAA treatment increased auxin contents to levels similar to those of intact trees, but, despite the observed auxin induction of GA20ox transcripts, the GA level was not increased in the IAA-treated stems after 24 h. GA treatment increased GA levels about fivefold in an auxin-depleted background. Initially, the supervised multivariate projection method [partial least-square discriminant analysis (PLS-DA); Trygg and Wold, 2002] using normalized signal values as the x-matrix was used to obtain an overview of the dataset. PLS-DA was used to maximize the information relating to the differences between the classes of samples (lanolin control, intact control, GA, IAA and GA/IAA). The results show that decapitation has a major impact on gene expression, and all treated trees were separated from the intact controls in the first component (Figure 6a). The second component separates the hormone-treated trees from the lanolin-treated trees (the IAA-treated trees most strongly). The IAA- and IAA/GA-treated trees do not separate until the fourth component (Figure 6b). The similarity between the IAA- and IAA/GA-induced transcriptomes suggests that GA induces few transcripts that are not also induced by IAA.

image

Figure 6.  Partial least-squares discriminant analysis of the global transcriptome resulting from indole acetic acid and gibberellins feeding to Populus stems. Experimental conditions were as described in the legend to Figure 3. mRNA was extracted from the extra-xylary tissues 5–25 mm below the decapitation position after 1 day of treatment and used for microarray analysis. The score plot shows the distribution of the five sample classes for (a) the first two components [t2] versus [t1], and (b) components 1 and 4, t[4] versus t[1]. The three included components describe 11.5% (component 1), 20.0% (component 2) and 11.5 (component 4) of the variation in the microarray data. Each point represents data from one hybridization.

Download figure to PowerPoint

To estimate the number of transcripts whose abundance changed following the IAA and GA treatments, the microarray datasets were normalized against the lanolin control, and significant differences were calculated using B-statistics, applying false-discovery rate corrections for P values. The threshold P level was set to 0.001. The complete dataset is available in the form of Excel sheets in Supplementary Table S1. In agreement with the PLS-DA analyses, about three times as many gene models were found to be affected by IAA treatment (1299) than by GA treatment (388). A striking observation from these two sets of differentially abundant transcripts is that 83% of those genes affected by GA treatment were also affected by IAA treatment (Figure 7, Supplementary Table S2). Furthermore, with a threshold of P < 0.05, 95% of the transcriptome induced by GA treatment was significantly affected by IAA treatment, and 98 of the 100 transcripts that were most strongly increased or decreased by the GA treatment were increased and decreased, respectively, by the IAA treatment. Thus we conclude that the set of transcripts modified by increased GA levels in an auxin-depleted background is, with very few exceptions (see Supplementary Table S3), similarly modified by increasing the IAA level in a constant GA background. We acknowledge that the detected transcriptomal changes resulting from the hormone feeding reflect the abundance of the measured transcripts at a specific hormone balance, rather than information on their dose–response behavior. However, the strengths of the analysis were that the endogenous hormone levels in the experimental tissues were determined, and the hormone balances had physiological relevance.

image

Figure 7.  Venn diagram of gibberellins (GA)- and indole acetic acid (IAA)-induced transcripts, showing the numbers of transcripts induced by increased GA and IAA levels, and the number that are significantly (P ≤ 0.001) affected by both hormones.

Download figure to PowerPoint

A cluster analysis of the genes affected by both the GA and IAA treatments showed that many of the transcripts that were increased or decreased by GA treatment were either similarly or further affected in the same direction by the auxin treatment, whereas few transcripts were less strongly increased or decreased by the auxin treatment than by the GA treatment (Supplementary Figure S1). Despite the many transcriptomal changes induced by both the GA and IAA treatments (hereafter denoted ‘common transcriptome’ for convenience) few of the affected transcripts showed synergistic changes in the combined GA/IAA treatment (Supplementary Figure S1). In order to identify the transcripts with the most obviously synergistic responses, we filtered out transcripts that were significantly more strongly increased or decreased by the GA/IAA treatment compared to the GA and IAA treatments. This left 13 transcripts that were further increased or decreased by the combined hormone feeding (Supplementary Table S4). Interestingly, 10 of these transcripts were decreased by the combined GA/IAA feeding. It should be noted that the changes in transcript levels of genes identified as having synergistic responses cannot necessarily be attributed to effects of GA, because the auxin levels were higher in GA/IAA-treated tissues than in tissues treated with IAA alone, as a result of GA-induced auxin transport.

Both GA and IAA treatment induced division and expansion of cambial zone cells. Indeed, the common GA- and IAA-induced transcriptome includes many Populus carbohydrate-active enzymes putatively involved in both loosening of the cell wall and supplying carbohydrate building blocks (Supplementary Table S5), several of which are abundant in the poplar stem tissues (Geisler-Lee et al., 2006). Transcripts of some hormone signaling genes were also found in the common GA and IAA-induced transcriptome, including transcripts encoding an IAA-Aux (PttIAA8) previously demonstrated to be auxin-induced and specifically expressed in expanding cambial derivatives (Moyle et al., 2002), an ARF similar to MONOPTEROS, which is putatively involved in vascular development (Hardtke and Berleth, 1998), and both DELLA-like and EIN3-like proteins. Although the GA treatment did not induce any xylem development (or lignification), there was a significant induction of several lignification genes by GA (P < 0.05, Supplementary Table S6). On the other hand, all known lignin biosynthesis genes except PttC4H2 and PttHCT2 were IAA-induced, and those that were induced by GA were induced more strongly by IAA.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Auxin supplied by polar transport is essential for maintaining and controlling cambial growth (Sundberg et al., 2000; Little et al., 2002). Varying amounts of auxin have been experimentally supplied to the vascular cambium of plants of diverse species through the polar transport system by apical feeding. However, resulting internal levels of auxin in an angiosperm tree following apical feeding have not been monitored previously. Here we demonstrated a quantitative relationship between auxin supply and cambial growth in Populus (Figure 1). More unexpectedly, we found that the IAA concentration in the stem tissues was higher when IAA was applied in combination with GA than when IAA was applied alone. The results of feeding experiments with stable isotope-labeled IAA conclusively demonstrate that this increase in IAA can be attributed to GA stimulation of transported IAA (Figure 3). In Arabidopsis, PIN proteins encoded by AttPin genes are involved in cellular auxin efflux, and hence in auxin polar transport and distribution (Friml and Palme, 2002; Leyser, 2005; Petrasek et al., 2006). In Populus, PttPIN1 is the most important PIN gene expressed in the cambial meristem and expanding xylem derivatives (Schrader et al., 2003), the major pathway for polar transport in the stem (Uggla et al., 1996). It is well established that auxin stimulates its own transport by feed-forward effects on PIN genes (Schrader et al., 2003; Vieten et al., 2005). This was also observed in our experiments, but in addition we demonstrate that the combined GA/IAA treatment increased PttPIN1 expression levels well above those induced by IAA alone (Figure 4). PttPIN1 was also induced by approximately 50% by GA alone, as compared to lanolin control. GA-induced expression of auxin transporters has also been demonstrated for PsPIN1 in pea (Chawla and DeMason, 2004) and for AtPin2 and AtPIN7 in germinating seeds of Arabidopsis (Ogawa et al., 2003). We hypothesize that the GA-induced increase in auxin supply in poplar stems is due to the increased expression of PttPIN1. However, the possibility cannot be excluded that GA stimulates auxin transport by some other mechanism, and the increase in PttPIN1 transcripts observed after the combined feeding is solely an effect of increased IAA. In addition to the novel finding of GA involvement in the regulation of auxin transport, we also found that auxin stimulates GA biosynthesis genes, as previously demonstrated in several species (Figure 4) (Ross et al., 2000; Wolbang and Ross, 2001; Frigerio et al., 2006). Frigerio et al. (2006) demonstrated that auxin stimulation of GA biosynthesis genes involves Aux/IAA-ARF signaling elements and is independent of the GA-mediated regulation of its biosynthesis through its effects on the DELLA proteins GAI and RGA. Whether GA- and auxin-stimulated PIN gene expression is similarly independently regulated remains to be demonstrated. However, the data presented here, together with current knowledge, imply that the GA/auxin balance is established by intricate cross-talk and self-control mechanisms involving the expression of auxin transport and GA metabolism genes.

Access to stem tissues with various defined hormonal balances provided opportunities to investigate and compare changes in global gene expression patterns resulting from increases in GA and IAA both independently and in combination (Supplementary Table S1). Conceivably, a majority of the identified transcripts can be attributed to late-response genes maintained by the increase in respective hormone, and reflect biological processes that are affected by them. The results indicate that auxin supplied through the polar transport pathway affects not only the expression of GA biosynthesis genes, but also more than a thousand other genes of the 9700 unique gene models represented on the microarray. This is not surprising given that auxin stimulates both cambial cell division and xylogenesis. GA treatment, which stimulated division and expansion of cambial zone cells but not xylogenesis, also affected the expression of a large number of genes in the stem tissues under auxin-depleted conditions, although only about a third of the number affected by IAA treatment. Comparison of the lists of GA-affected and IAA-affected genes led to the striking observation that almost all of the genes that were significantly affected (in terms of transcript levels) after GA treatment were similarly affected by auxin (Figure 7 and Supplementary Table S2). Our results therefore demonstrate that GA has a common transcriptome with auxin. This is perhaps not surprising because both hormones stimulate cell division and expansion; however, the almost complete lack of transcripts specifically induced by an increase in GA is noteworthy. The finding of a common GA- and IAA-induced transcriptome agrees well with the finding that GA stimulates cell division and expansion in the cambial zone under auxin depletion (Figure 2d), and findings in primary stems that auxin stimulates elongation growth in mutants deficient in GA biosynthesis and that GA stimulates elongation growth in auxin-depleted internodes and auxin signaling mutants (Yang et al., 1996; Barratt and Davies, 1997; Collett et al., 2000; Ross et al., 2003). Furthermore, different time-course patterns have been observed in growth responses to GA and IAA in pea internodes, suggesting that there are independent early-response mechanisms to the hormones (Yang et al., 1996). Taken together, our transcriptome data provide molecular support to physiological experiments demonstrating that either hormone can induce growth if the other hormone is absent/deficient because of mutations or experimental treatments. They also show the potential for extensive cross-talk between GA and auxin in vegetative growth of the intact plant.

The synergistic responses to GA and IAA clearly reflect mutual effects at the respective hormone level. However, from a simplistic perspective, it could also be predicted that combined treatment would amplify the transcription of genes induced by both hormones independently. Indeed, such expression patterns were observed for a number of transcripts (Supplementary Figure S1). However, in most cases the amplification was weak, and few genes passed the statistical filter used for detecting synergistic effects on gene expression (Supplementary Table S4). However, the real-time RT-PCR data for the PttPin1 and PttDELLA-like1 transcripts demonstrate that the combined GA/IAA treatment had a more than additive effect on the expression of these genes (Figure 4). While PttPIN1 expression may be part of the synergistic response, DELLA proteins act as inhibitors of GA responses (Alvey and Harberd, 2005), so the increased transcription of PttDELLA-like1 may reflect a feedback response to increased GA and IAA levels. In support of this hypothesis, the expression of PttDELLA-like1 has been demonstrated to be co-localized with endogenous biologically active GAs in expanding xylem cells in poplar (Israelsson et al., 2005).

The results of the application experiments demonstrate that both GA and IAA stimulate division and expansion of cambial zone cells and induce many genes that play important roles in cell growth, suggesting they are endogenous regulators of these processes. However, when inferring the in planta functions of GA and IAA from feeding experiments, it is important to consider their tissue localization under experimental and natural conditions. In intact poplar trees, there are high auxin concentrations in the cambial meristem and its expanding daughter cells, whereas bioactive GAs show peak levels in the expanding cambial derivatives, and very low levels in the cambial meristem (Figure 8) (Tuominen et al., 1997; Israelsson et al., 2005). Under the experimental conditions, the applied IAA enters the cambial tissues through its natural transport pathways by polar transport, while the transport pathway of the GA4 applied in this experiment is not known, but it probably enters the adjacent stem tissues through the phloem stream. Thus, active GA was supplied to the cambium through an unnatural compartment, conceivably resulting in unnaturally high levels in the cambial meristem. The stimulation of cambial cell division resulting from the GA feeding therefore demonstrates a potential function of the hormone, but may not reflect its role in normal growth conditions. The synergistic effects on cambial growth observed when GA was applied with IAA under our experimental conditions may therefore be less important in the intact tree. Auxin is more likely to be the primary regulator of cambial growth, and correlations between cambial growth and endogenous auxin have indeed been demonstrated in Scots pine (Sundberg and Little, 1990; Uggla et al., 1998). However, GA may indirectly stimulate cambial growth in the intact tree through enhancement of polar auxin transport. PttPIN1 expression, for example, is not strictly localized to the cambial meristem, but also occurs in expanding cambial derivatives where GA levels are high (Figure 8; Schrader et al., 2003). In support of this view, increased cambial growth and increased auxin levels have been found in transgenic poplar trees with increased levels of bioactive GAs (Israelsson et al., 2003).

image

Figure 8.  Schematic description of the distribution of endogenous auxin and bioactive gibberellins across the cambial region of Populus. Data from Tuominen et al. (1997) and Israelsson et al. (2005). P, phloem; VC, vascular cambium; X, xylem.

Download figure to PowerPoint

The growth of cambial xylem derivatives in the post-meristematic phase involves both radial expansion of fibers and vessel elements by symplastic growth and elongation growth of fibers by intrusive growth (Mellerowicz et al., 2001). At this developmental stage, both IAA and GA are present at high concentrations (Figure 8). The stimulation by IAA of GA biosynthesis suggests that auxin is required for GA biosynthesis in these tissues. However, the non-overlapping distributions between IAA and GA also imply that auxin is not the only substance that determines the site of GA biosynthesis. GA is generally accepted to play a role in fiber elongation (Digby and Wareing, 1966; Ridoutt et al., 1996; Eriksson et al., 2000). Studies on radial expansion of cambial derivatives have mainly focused on vessel elements, and a common conclusion from dose–response feeding experiments is that auxin is a key regulator of the expansion of vessel elements (Digby and Wareing, 1966; reviewed in Sundberg et al., 2000) and cell expansion in general (Cleland, 2004). The reduced vessel diameters that developed in the xylem following IAA feeding can therefore be attributed to the lower than normal IAA concentration during the experimental period (Figures 1 and 3a). However, information on the role of auxin in intrusive fiber growth, and the role of GA in symplastic radial cell growth, is limited or non-existent. The finding that GA and IAA induce similar changes in the expression of many genes that play important roles in cell growth suggests that both GA and IAA should be considered important signals for the expansion of cambial derivatives. The identification of hormone-specific signaling pathways and appropriate mutant analysis would shed more light on whether or not this is true (and, if so, the mechanisms involved).

Lignification of fibers takes place late in xylem development when both IAA and GA concentrations are low (Figure 8). Interestingly, all strongly expressed lignin biosynthesis genes, except PttC4H2 and PttHCT2, were induced by IAA (Supplementary Table S6). Therefore, we attribute the delayed lignification of xylem fibers in the IAA-treated trees (Figure 2f) to the low auxin supply during the course of the feeding experiment. This notion is also supported by the reported absence of lignification in inter-fascicular fiber secondary walls in the revoluta/ifl1 mutant, in which polar auxin transport is deficient (Zhong and Ye, 2001; Lev-Yadun et al., 2004). Although GA does not induce lignification in the absence of auxin, it has been suggested that it may modify the lignin syringyl/guaiacyl (S/G) ratio, as deduced from feeding experiments in Coleus (Aloni et al., 1990) and GA over-producing trees (Israelsson et al., 2003). However, all GA-induced lignin synthesis genes were even more strongly induced by IAA (Supplementary Table S6). It is therefore possible that the lignin modifications observed by Israelsson et al. (2003) in GA-over-producing trees were indirectly caused by the increased auxin levels in those trees.

In summary, we have demonstrated that GA stimulates polar transport of IAA in poplar stems. We have also shown that GA has a common transcriptome with auxin, including many transcripts related to cell growth. Interpretation of available information from feeding experiments and endogenous hormone distributions in poplar suggests that, whereas auxin is a key regulator of cambial cell division, both hormones play important roles in the post-meristematic expansion of cambial derivatives. Our data also suggest that auxin depletion results in delayed fiber lignification.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant material and growth conditions

Hybrid aspen (Populus tremula L. ×tremuloides Michx.; clone T89) trees were propagated in vitro, potted in fertilized peat and grown in a greenhouse under a 18 h light/6 h dark photoperiod with natural light supplemented with metal halogen lamps. The temperature was approximately 22°C/15°C (day/night), and the trees were watered daily and fertilized once a week with a nutrient solution (Superba, Yara).

Hormone treatments

Lanolin paste was used as a carrier for the exogenous hormones, which were IAA (Sigma-Aldrich, http://www.sigmaaldrich.com/), [13C6]IAA (Cambridge Isotope Laboratories) and GA4 (Sigma-Aldrich). The hormones were dissolved in small amounts of 99% ethanol, and thoroughly stirred into melted lanolin. About 1 g of the lanolin paste was poured into small plastic molds that were stored at −70°C. The hormone sources were applied to approximately 1 m tall trees, decapitated just above the youngest fully expanded leaf where a vascular cambium had been initiated. In long-term experiments, the hormone source was changed three times a week, and a fresh cut surface was prepared. Axillary buds were removed to avoid additional hormone sources complicating interpretation of the results.

Sampling and quantification of IAA and GAs

Sampled stem pieces were directly frozen in liquid nitrogen and stored at −70°C. To avoid contamination from the external hormone sources, the top 5 mm of the stem was carefully removed and discarded. Samples for analysis consisted of the whole bark peel and developing xylem scraped from the exposed wood surface collected from the segment 0.5–2.5 cm below the point of application of the hormones. The bark peel and the scrapings were immediately frozen in liquid nitrogen.

Tissues were homogenized in liquid nitrogen using a mortal and pestle. Samples (about 20 mg FW) were extracted with 500 μl of MeOH:H2O:CH3COOH (800:190:10) after adding the following internal standards: 1 ng 13C6-IAA for 12C6-IAA analyses, [D5]IAA (C/D/N Isotopes; http://www.cdnisotopes.com) for 13C6-IAA analyses, and 50 pg 2H2-GAs (purchased from Prof. L. Mander, Australia University, Canberra) for GA analyses. IAA samples were processed and quantified using an isotope dilution mass spectrometry technique as described by Edlund et al. (1995), and GAs were analyzed as described by Eriksson et al. (2000). Quantification was performed using a GC/MS-SRM instrument (JMS-MStation 700, JEOL; http://www.jeol.com).

Anatomy

Samples intended for growth analysis were fixed in FAA, dehydrated in an ethanol series and embedded in paraffin. Transverse sections (10 μm) of the whole-stem segments were cut on a microtome, and stained with phloroglucinol (Sigma-Aldrich; http://www.sigmaaldrich.com). The width of the xylem was measured at ten positions around the circumference of the stem.

Samples intended for anatomical investigations were fixed in 4% glutaraldehyde in 25 mm phosphate buffer, pH 7.2, for at least 1 day. The fixed samples were then dehydrated in an ascending series of acetone concentrations and embedded in a methacrylate resin (purchased from Professor S. Fink, Albert-Ludwigs-Universita˝t Freiburg, Germany). Transverse sections of embedded samples were cut with a Leica-2065-microtome (Leica Instruments GmbH) using a diamond knife. The sections were stained, mounted in Entellan (Merck; http://www.merck.com), and then examined under a Zeiss-Axiophot light microscope (Carl Zeiss, http://www.zeiss.com/).

RNA extraction

Total RNA was extracted from portions of the homogenized samples used for hormone analyses, purified using an RNeasy Plant Mini Kit (Qiagen, http://www.qiagen.com/) according to the manufacturer’s instructions, and DNA was removed using DNAse (Ambion; http://www.ambion.com) according to the manufacturer’s recommended protocols. RNA was quantified using a nanodrop machine (ND-1000, NanoDrop; http://www.nanodrop.com).

Microarray analysis

mRNA was extracted from total RNA using Dynabeads (Dynal AS; http://www.invitrogen.com) according to the manufacturer’s instructions. The mRNA was fragmented using a Hamilton syringe, and 150 ng of the fragmented mRNA were amplified by PCR as described by Hertzberg et al. (2001). Labeling with fluorescent dyes (Cy3 and Cy5) and hybridization to the POP1 arrays (described in Andersson et al., 2004) was performed according to the method described by Hertzberg et al. (2001). Four replicate hybridizations for each hormone treatment and the lanolin control, two of which were dye swapped, were performed using the RNA from the intact control trees as a reference. Slides were scanned using a ScanArray 4000 scanner (Perkin Elmer Life Sciences; http://www.perkinelmer.com) at a resolution of 5 μm. Laser settings and the photo-multiplier tube were adjusted to obtain similar overall signal strengths in the two channels and 1–2% of saturated spots. Spot intensities were quantified using Genepix Pro 4.1 software (Axon Instruments; http://www.axon.com). Spots with high backgrounds or dust specks were manually flagged as bad. The median background pixel intensity was subtracted from the mean spot pixel intensity. To calculate the background intensity, the local method was chosen. The data were normalized and analyzed using base software (http://base.thep.lu.se/). The mean signal intensity was estimated after median background corrections by the R plug-in. Print tip lowess using the Limma package was used for within-array normalization, and the Aquantile package (R plug-in) for between-array normalization. Data were deposited in the UPSC-BASE database as experiment number 29 (Sjodin et al., 2006).

Data analysis

Partial least-squares with a discriminant function was used for multivariate sample classification, as suggested by Sjostrom et al. (1986), based on all the recorded spot intensities simultaneously. Each sample was assigned to an individual pre-defined class (lanolin control, intact control, GA, IAA, IAA/GA), and PLS regression was used to correlate the spot intensities to a matrix containing the class identity. Cross-validation according to the method described by Wold (1978) was used to calculate the model complexity (the number of significant PLS components) and to decide the model’s predictive ability. PLS-DA models were calculated for each pairwise comparison between sample classes. All multivariate analyses were carried out using simca-p+11 software (Umetrics AB; http://www.umetrics.com). Prior to any multivariate modeling, the data were mean-centered and scaled to unit variance.

The microarray data were analyzed using B-statistics (R package provided by base software; http://base.thep.lu.se/), applying a false-discovery rate correction (1%) for P values. Clustering was performed using K-means with tigr mev software (Saeed et al., 2003).

Gene models and annotations for each PU number are according to PopulusDB (http://www.populus.db.umu.se/; Sjodin et al., 2006). Genes relevant to the discussion have been checked and annotated manually.

Real-time RT-PCR

cDNA was synthesized according to the Iscript cDNa Synthesis protocol (Bio-Rad, http://www.bio-rad.com/) starting with up to 1 μg total RNA. PCR reactions were performed in a 96-well plate using a Bio-Rad iCycler iQ™ real-time PCR detection system. Products were detected by binding the fluorescent DNA dye SYBR Green (SYBR® Green PCR kit, Bio-Rad) to them, and measuring the resulting fluorescence. All assays were carried out in triplicate. Each PCR reaction mixture contained 5 μl of 20-fold diluted cDNA sample, 12.5 μl SYBR Green Master Mix, 0.2 μm of forward and reverse primers (Invitrogen, http://www.invitrogen.com/) and 0.25 pmol fluorescein calibration dye (Bio-Rad) in a total volume of 25 μl. Thermal cycling conditions were as follows: initial activation step, 5 min at 95°C, followed by 45 cycles of 15 sec at 95°C, 30 sec at 57°C, 30 sec at 72°C, a melting curve program (80 cycles, 10 sec each, with 0.5°C elevations starting at 54°C), and a cooling step to 4°C. The presence of one product per assay was confirmed by analysis of the dissociation curves. The icycler iq™ software 3.0 (Bio-Rad) was used to calculate the significant fluorescence signal in the middle of the exponential phase (CT). The PCR efficiency for each amplicon was determined using pooled cDNA samples from all samples in a fivefold dilution series and icycler iq™ software 3.0. The relative transcript levels were calculated as follows: gene-specific factor ×2-(CT Target - CT Control), thus normalizing target gene expression to that of the control gene (ribosomal 18S). Gene-specific factors were: PttGA20ox1, 109; PtGA20ox4, 1010; PttGA2ox1, 1014; PttGA2ox2, 109; PttDELLA-like1, 1011; PttPIN1, 108; PttCPS1, 1011.

The primers used were: PttGA20ox1 forward, 5′-GCACAAGTTCTTCGACACCAG-3′ and reverse, 5′-GCAGCAACAGGGTTACCAGAG-3′; PtGA20ox4 forward, 5′-TGGCACTCCGTTACTCCTG-3′ and reverse, 5′-AAGCGAGGGATTTCCTTACT-3′; PttGA2ox2 forward, 5′-TTCTTCTCATTACCGCTCTCTG-3′ and reverse, 5′-TCTACCCAGCCCACATCAC-3′; PttDELLA-like1 forward, 5′-GCAAGTCGAGTCCACGTTATC-3′ and reverse, 5′-AATTCCCGTCAGCCGAAATG-3′; PttPIN1 forward, 5′- GCAATGTTCAGTCTTGGTCT-3′ and reverse, GGAATCTCACAGCCATAGAA; PttPIN2 forward, 5′-TAGCTCCAGAGGCTCCCTTGA-3′ and reverse, 5′-C CAAATACCCCCAAGTCCTCCAC-3′; 18S forward, 5′-TCAACTTTCGATGGTAGGATAGTG-3′ and reverse, 5′-CCGTGTCAGGATTGGGTA ATTT-3′.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

This work was supported by grants from the Swedish Research Council Formas and the Swedish Research Council. We thank Dr Rishikesh Bahlerao for valuable comments on the manuscript, and Mr Kjell Olofsson for technical assistance.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supplementary material

Table S1. The complete dataset for all gene models represented on the microarray. For gene models represented by several spots (PU numbers) the most significant is shown.

Table S2. Transcripts significantly affected by GA identified using a threshold P-value≤0.001. Transcripts that were also significantly (P≤0.001) affected by the IAA treatment are marked in red.

Table S3. Transcripts specifically affected by GA. These transcripts fulfill the following criteria: (i) significantly affected by GA (P≤0.001), (ii) not significantly affected by IAA (P≤0.001), and (iii) significantly differing in abundance when comparing the GA with the IAA treatments (P≤0.001) and (P≤0.01).

Table S4. Transcripts synergistically affected by the GA/IAA treatment. Both GA and IAA significantly affected these transcripts (P≤0.001), and their abundance significantly differed between the GA/IAA treatments and both the GA (P≤0.001) and IAA treatments (P≤0.001) and (≤0.01).

Table S5. GA- and IAA-induced transcripts with putative importance in cell growth and hormone signaling.

Table S6. GA- and IAA-induced transcripts with putative importance in lignification.

Fig. 1S. Cluster analysis of the common GA- and IAA-induced transcriptome.

FilenameFormatSizeDescription
TPJ_3250_sm_FigS1.tif66KSupporting info item
TPJ_3250_sm_TabS1_4.xls5464KSupporting info item
TPJ_3250_sm_TabS5.doc72KSupporting info item
TPJ_3250_sm_TabS6.doc63KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.