Asymmetric expression of a poplar ACC oxidase controls ethylene production during gravitational induction of tension wood


For correspondence (fax +46 90 786 59 01; e-mail


Ethylene is produced in wood-forming tissues, and when applied exogenously, it has been shown to cause profound effects on the pattern and rate of wood development. However, the molecular regulation of ethylene biosynthesis during wood formation is poorly understood. We have characterised an abundant 1-aminocyclopropane-1-carboxylic acid (ACC) oxidase gene (PttACO1) in the wood-forming tissues of Populus tremula (L.) × P. tremuloides (Michx). PttACO1 is primarily expressed in developing secondary xylem, and is specifically upregulated during secondary wall formation. Nevertheless, according to GC–MS analysis combined with tangential cryosectioning, the distribution of ACC was found to be fairly uniform across the cambial-region tissues. Gravitational stimulation, which causes tension wood to form on the upper side of the stem, resulted in a strong induction of PttACO1 expression and ACC oxidase activity in the tension wood-forming tissues. The ACC levels increased in parallel to the PttACO1 expression. However, the increase on the upper (tension wood) side was only minor, whereas large amounts of both ACC and its hydrolysable conjugates accumulated on the lower (opposite) side of the stem. This suggests that the relatively low level of ACC on the tension wood side is a result of its conversion to ethylene by the highly upregulated PttACO1, and the concurrent accumulation of ACC on the opposite side of the wood is because of the low PttACO1 levels. We conclude that PttACO1 and ACC oxidase activity, but not ACC availability, are important in the control of the asymmetric ethylene production within the poplar stem when tension wood is induced by gravitational stimulation.


Ethylene mediates developmental control and modifies growth patterns in response to a range of stresses and environmental cues during the life cycle of plants (Abeles et al., 1992; Hyodo, 1991). It is synthesised from S-adenosyl methionine via 1-aminocyclopropane-1-carboxylic acid (ACC) by the action of ACC synthase and ACC oxidase, which belong to the multigene families (Yang and Hoffman, 1984). Developmental and spatial induction patterns of ethylene production are established by both transcriptional and post-transcriptional control of the ethylene biosynthetic enzymes (Fluhr and Mattoo, 1996; Johnson and Ecker, 1998; Wang et al., 2002). Levels of ACC are normally low in vegetative tissues and limiting for high rates of ethylene production. An increase in ethylene biosynthesis therefore requires the initial induction of ACC synthase. This commonly results in a burst of ethylene because of a constitutive expression of ACC oxidase and positive feedback effects of ethylene on ACC oxidase (Peck and Kende, 1995; Petruzzelli et al., 2000). A well-documented type of localised ethylene induction that seems to proceed as outlined above occurs in the lower portions of gravistimulated above-ground organs, such as flower stalks, hypocotyls, coleoptyles and apical shoots (Clifford et al., 1983; Kaufman et al., 1995; Philosoph-Hadas et al., 1996, 2001; Prasad et al., 1989; Wheeler et al., 1986; Woltering, 1991). The initial induction of ACC synthase in the gravitational response is commonly attributed to an asymmetric redistribution of auxin towards the lower side of the affected organ. However, there are many examples of more complex networks regulating the expression of ACC synthase and ACC oxidase. Furthermore, following the cloning of ACC oxidase, its differential expression and role in controlling ethylene production is becoming increasingly recognised (Barry et al., 1996; Blume and Grierson, 1997; English et al., 1995; Gong and McManus, 2000; Kende and Zeewart, 1997; Kim et al., 1998; Lasserre et al., 1996). However, there are a few cases where ACC oxidase activity, rather than ACC availability, has been implicated as the controlling step in time- and space-specific ethylene induction (Dunlap and Robacker, 1994; Vriezen et al., 1999), and evidence to demonstrate that differential ACC oxidase expression controls spatial induction of ethylene production has previously not been presented.

The formation of reaction wood in stems of woody species is another example of a gravitational response that involves ethylene induction (Scurfield, 1973; Timell, 1986; Wilson and Archer, 1977). Reaction wood is induced when a stem or branch is displaced from its normal position. This process occurs in both gymnosperm and angiosperm trees, where the reaction wood is denoted as compression and tension wood, respectively. The response of the reaction wood is unilateral, and involves accelerated growth of the vascular cambium and differentiation of the modified wood. The differential growth results in physical strains that force back the stem or branch towards its original position. Like the gravitational response in primary tissues, the asymmetric growth pattern in reaction wood is associated with a differential localisation of ethylene production, in which more intense production occurs on the side of the reaction-wood formation than on the opposite side (Little and Eklund, 1999; Nelson and Hillis, 1978). In angiosperm trees, such as poplar, tension wood develops on the upper side of the gravistimulated stem. The formation of wood on the opposite side is either inhibited or strongly retarded. Tension wood is characterised by fibres with cell walls that have an inner gelatinous cell wall layer (G-fibres), rich in cellulose and devoid of lignin. It also has a few vessels with small radial diameters (Haygreen and Bowyer, 1996; Jourez et al., 2001). Tension wood is easily induced by leaning or bending the stem, and therefore offers a convenient experimental system.

Although it is established that ethylene is produced during wood formation and that it influences wood development when applied exogenously (Little and Pharis, 1995; Little and Savidge, 1987), genes encoding ACC synthase and ACC oxidase in cambial tissues have not yet been characterised. However, there is one previous report of an ACC oxidase protein that is upregulated during compression wood formation in Maritime pine (Plomion et al., 2000). Within the Swedish poplar expressed sequence tag (EST) sequencing project (, several homologues to ACC synthase and ACC oxidase genes were found in the cDNA libraries from the wood-forming tissues. Here, we report on the cloning and characterisation of PttACO1, which is an abundant ACC oxidase in these tissues. We present results of a high-resolution analysis of PttACO1 expression and ACC distribution across cambial-region tissues using tangential cryosectioning and GC–MS, and we provide evidence showing the induction of PttACO1 expression, but not ACC availability, controls the asymmetric ethylene induction during tension wood formation.


Cloning of an ACC oxidase from wood-forming tissues

Three identical ACC oxidase homologues were present in the EST library from the cambial-region tissues of hybrid aspen described by Sterky et al. (1998). Seven clones homologous to these ESTs were also found in a cDNA library from developing tension wood sequenced within the poplar EST initiative. Using one of these ESTs as a probe, a full-length ACC oxidase clone (designated as PttACO1, GenBank accession no. AY167040) was isolated from the developing tension wood cDNA library. The full-length clone contained a predicted coding region of 933 nucleotides (310 amino acids), as well as a 93-nucleotide 5′-untranslated region and a 285-nucleotide 3′-untranslated region.

The predicted amino acid sequence of PttACO1 showed highest identity (72%) to a putative ACC oxidase from Arabidopsis thaliana identified within the Arabidopsis Genome Initiative, protein ID number At1g77330. A multiple alignment of the deduced amino acid sequence of PttACO1 with ACC oxidases from Arabidopsis and the well-characterised families from muskmelon (Lasserre et al., 1996) and tomato (Barry et al., 1996; Nakatsuka et al., 1998) was carried out. This revealed that PttACO1 contained amino acid motifs that were conserved among all members of the the Fe(II)-dependent oxygenase/oxidase family (Figure 1a), including the HXD…H motif shown to be involved in metal ligation during catalysis (Zhang et al., 1997). Furthermore, phylogenetic analysis of amino acid sequences from various Fe(II)-dependent oxygenases/oxidases showed that PttACO1 clustered with ACC oxidases, supporting the function of PttACO1 as an ACC oxidase (Figure 1b).

Figure 1.

Amino acid sequence alignment and phylogenetic analysis of PttACO1 with related sequences.

(a) Comparison of the deduced amino acid sequence of PttACO1 with ACC oxidases from Arabidopsis, muskmelon and tomato. Sequences used in the alignment corresponded to Arabidopsis thaliana, At1g77330 (NP_565154); Cucumis melo, CM-ACO1 (S66175), CM-ACO2 (S66175), CM-ACO3 (S66176); Lycopersicon esculentum, LE-ACO1 (CAA41212), LE-ACO2 (CAA68538), LE-ACO3 (CAA90904), LE-ACO4 (BAA34924). Identical and similar amino acids are shown in black and grey backgrounds, respectively. Asterisks indicate amino acids conserved across all members of the Fe(II)-dependent oxygenase/oxidase family. The plus sign indicate likely Fe(II) binding sites. The GenBank accession number for each sequence is shown in parenthesis.

(b) Phylogenetic analysis of amino acid sequences of some members of the Fe(II)-dependent oxygenase/oxidase family. clustalw was used to align the protein sequences. The tree was obtained by the neighbour-joining method using 1000-bootstrap replicates and were visualised using tree viewppc (version 1.6.6). The scale represents 0.1 substitutions per site. The GenBank accession number of each sequence is shown in parenthesis. A. thaliana, At1g77330 (NP_565154); C. melo, CM-ACO1 (S66175), CM-ACO2 (S66175), CM-ACO3 (S66176); L. esculentum, LE-ACO1 (CAA41212), LE-ACO2 (CAA68538), LE-ACO3 (CAA90904), LE-ACO4 (BAA34924), Le3OH3 (BAA34125), LeE8 (T06406); Lactuca satiga, Ls3OH3 (BAB12439); Malus domestica, ApGA20OX (BAB20975); Nicotiana tabacum, NtGA20OX (T01749); Pisum stavitum, Ps3OH3 (T06243); Populus tremula × P. tremuloides, PttACO1 (AY167040), PttGA20OX (CAC00709).

PttACO1 is the major ACC oxidase in wood-forming tissues and is induced during tension wood formation

The expression pattern of PttACO1 was examined by Northern blot analysis in greenhouse-grown trees (Figure 2). The PttACO1 transcript was preferentially found in the developing xylem fraction from mature stems, and to a lesser extent in the phloem/cambium fraction and root-tip samples. To further visualise the PttACO1 expression across the cambial-region tissues, a high-resolution dot blot analysis was performed by combining tangential cryosectioning with quantitative PCR amplification of mRNA. This revealed a strong tissue-specific upregulation of the transcript in developing xylem cells at the early stage of secondary wall formation, confirming results obtained by microarray analysis (Hertzberg et al., 2001). When greenhouse-grown trees were gravistimulated by leaning, the PttACO1 transcript was specifically induced in developing xylem on the tension wood side (Figure 3b). The upregulation of PttACO1 was associated with a large increase in ACC oxidase activity in the same tissue (Figure 3a). In phloem/cambium tissues, the activity of ACC oxidase was lower than in xylem tissues, and no increase was observed after tension wood induction (data not shown).

Figure 2.

Expression analysis of PttACO1 in vegetative tissues of greenhouse-grown poplars.

(a) Northern blot analysis of PttACO1 transcript in 10 µg of total RNA extracted from different organs and tissues. Numbering of leaves and internodes starts from the first leaf longer than 1 cm.

(b) Dot blot analysis of the PttACO1 transcript in different stem tissues. Messenger RNA extracted from tangential sections was amplified by PCR and dotted onto a nylon membrane. The stage of secondary wall formation, where PttACO1 was upregulated, is indicated in the anatomical picture. The bar corresponds to 210 µm. CO, cortex; PF, phloem fibres; Ph, phloem; CZ, cambial zone; Ex1, fibres in the stage of early expansion; Ex2, fibres in the stage of late expansion; SW, early stage of secondary wall formation; MX, maturing xylem.

Figure 3.

The 1-aminocyclopropane-1-carboxylic acid (ACC) oxidase activity assay and Northern blot analysis of PttACO1 in poplars induced to form tension wood by gravistimulation.

The assay was done in greenhouse-grown poplars that were either upright or gravistimulated for 21 days by leaning and actively producing tension wood.

(a) Activity of ACC oxidase in developing xylem. Mean ± SD, n = 4. Asterisks indicate statistically significant differences (t-test) compared to control levels in upright trees. **P < 0.01; ***P < 0.001.

(b) Detection of PttACO1 transcript in 20 µg of total RNA extracted from three independent trees. C = upright control trees, TW = tension wood side, OW = opposite wood side, P = phloem/cambium, X = developing xylem.

Two other putative ACC oxidases were identified from cDNA libraries obtained from wood-forming tissues. One EST present in the cambial-region library was highly abundant in the EST database from floral tissues, and another gene represented by two EST copies in the tension wood library was highly abundant in the EST database from senescing leaves. However, the expression of these ACC oxidases was not increased by the induction of tension wood formation (data not shown). Therefore, we attributed the majority of the ACC oxidase activity induced by gravistimulation to the expression of PttACO1.

ACC increases on the opposite wood side after bending but shows no tissue-specific accumulation across the cambial region

We wanted to investigate the occurrence and distribution pattern of ACC, the substrate for ACC oxidase, in cambial-region tissues with the same resolution as used for the PttACO1 expression. Commonly, ACC in plant tissues is estimated after conversion to ethylene. However, this method is not sensitive enough to detect tissue-specific levels of ACC, which occurs in low amounts in non-stressed vegetative tissues. Therefore, we used mass spectrometry based on selected reaction monitoring (Edlund et al., 1995), which can measure picogram amounts of ACC in plant extracts with good accuracy and precision (Hellgren, Sundberg and Moritz, unpublished). This method was used to study the concentration and distribution pattern of ACC in cambial-region tissues. In initial experiments using greenhouse-grown trees, samples covering all cambial-region tissues were analysed for ACC both in upright trees and in trees that were induced to form tension wood by leaning. In the upright trees, the ACC concentration was about 20 ng g−1 fresh weight (FW) (Figure 4a), which is comparable to the estimates in leaves and internodes of annual plants (Chauvaux et al., 1997; Hall et al., 1993). A 15-fold increase in ACC concentration was seen in trees forming tension wood. Interestingly, this huge increase was only observed on the opposite wood side, whereas the concentration on the tension wood side was only slightly increased. We attributed the lower concentration of ACC on the tension wood side to the strong induction of PttACO1 expression and ACC oxidase activity in these tissues. Measurements of IAA showed that the auxin concentration was not increased in trees induced to form tension wood (Figure 4b), demonstrating that the ACC induction was not mediated through an auxin-induced ACC synthase.

Figure 4.

Levels of 1-aminocyclopropane-1-carboxylic acid (ACC) and IAA in poplar stems induced to form tension wood by gravistimulation.

(a) ACC was measured in the cambial-region xylem of the greenhouse-grown poplars that were either upright or gravistimulated for 21 days by leaning and actively producing tension wood. Mean ± SD, n = 4.

(b) IAA was measured in the cambial-region of field-grown poplars that were either upright or gravistimulated for 11 days by bending and actively producing tension wood. Mean ± SD, n = 8.

Asterisks indicate statistically significant differences (t-test) compared to control levels in upright trees. **P < 0.01; ***P < 0.001. C = upright control trees; TW = tension wood side; OW = opposite wood side.

We also measured the concentration of ACC in the xylem sap of field-grown trees that were upright and trees that were forming tension wood after 11 days of bending. In upright trees, the concentration was 0.9 ng ml−1, and in accordance with the previous results from the cambial-region tissues, bending caused a sixfold increase in ACC concentration (5.2 ng ml−1) in the xylem sap.

The sensitive GC–MS technique was used in combination with tangential cryosectioning to obtain a high-resolution map of ACC distribution across the cambial-region tissues. Surprisingly, the tissue-specific expression of PttACO1 transcripts in maturing xylem cells was not reflected in the distribution of ACC. ACC levels were found to be similar across all cambial-region tissues, from the functional phloem, across the cambial meristem and throughout the xylem, to the annual ring (Figure 5a). The ACC distribution was also analysed in stems forming tension wood after 11 days of bending (Figure 5b,c). On the tension wood side, the ACC concentration was low in all tissues in spite of the strong tissue-specific upregulation of the PttACO1 transcript in the developing xylem. In accordance with the earlier measurements of cambial-region tissues, the levels of ACC were also much higher on the opposite wood side, and in this case, the levels were also similar throughout all the cambial-region tissues.

Figure 5.

Levels and distribution patterns of 1-aminocyclopropane-1-carboxylic acid (ACC) across cambial-region tissues in upright and gravistimulated poplars.

The radial distribution pattern of ACC was measured across the cambial-region tissues in upright control trees and at the tension wood side and opposite wood side of field-grown trees actively forming tension wood after 11 days of bending. ACC was measured in 30 µm tangential sections obtained by a cryomicrotome. Each column represents one sample. The amount of ACC cm−2-section is indicated with a black dot, and the shading of the column reflects the composition of tissues in the sample, as explained in the figure.

(a) Upright control trees.

(b) Tension wood side.

(c) Opposite wood side.

In (a) and (b), each sample consists of three pooled 30 µm sections, and in (c), each sample consists of one section. Note the different scale in (c). For each treatment, three independent trees were measured. Ph = phloem, CZ = cambial zone, Ex = expanding fibres, MX = maturing and mature xylem.

Time-course of ACC and ACC-conjugate concentrations and of PttACO1 expression during tension wood formation

To investigate the timing of the increase in ACC and the induction of PttACO1 during tension wood induction, a time-course experiment was performed with field-grown trees. Stimulation of cambial growth on the tension wood side was established after 11 days of bending (Figure 6a), whereas the characteristic gelatinous walls within the fibres could be observed after 5 days. There were parallel increases in the concentration of ACC on the opposite wood side and the expression of PttACO1 on the tension wood side, with the increases being detectable after 1 day and clearly established after 3 days (Figure 6b,c). The expression of both PttACO1 and ACC concentrations continued to increase until day 11. Again, the PttACO1 was mainly expressed in the developing xylem. Its induction was also observed in phloem/cambium fractions, but its expression was much weaker in these tissues, although it followed a similar time-course (data not shown). On the tension wood side, a small increase in ACC was observed, but with only three replicates and with a rather high variation for each time-point, the exact timing for this increase was not established. We also observed a slight transient increase in the expression of PttACO1 after 1 day on the opposite wood side, but thereafter, the expression declined. The significance of this is difficult to evaluate without more detailed information.

Figure 6.

Time-course of growth response, 1-aminocyclopropane-1-carboxylic acid (ACC) levels and PttACO1 expression in poplar stems during gravistimulation.

Field-grown trees were induced to form tension wood by bending, and sampled during a 26-day period. (a) Growth response and (b) ACC levels in cambial-region tissues were measured on the tension wood (white bars) and opposite wood (black bars) sides. Mean ± SD; n = 3. Expression of the PttACO1 transcript in developing xylem on the (c) tension wood side (TW) and (d) opposite wood side (OW). Each lane was loaded with 20 µl of total RNA extracted and pooled from three independent trees.

Bending of field-grown poplars was repeated in the following growing season to perform ACC-conjugate analysis during tension wood formation in order to increase our understanding of the ACC dynamics. The results confirmed the stimulation of growth on the tension wood side, and provided evidence that the growth on the opposite side was inhibited in the bent trees in comparison to the upright ones (Figure 7a). Moreover, a statistically significant increase in ACC on both sides could be established after 1 day (Figure 7b). The increase on the tension wood side was transient, but with support provided by the data from the greenhouse experiment (Figure 4), we believe that the small increase in ACC levels observed on the tension wood side was consistent and real. ACC-conjugate concentrations were about 2 µg g−1 FW in upright trees (Figure 7c). This is in agreement with the much higher concentration of conjugates compared to ACC normally observed in vegetative tissues (see, for instance, Chauvaux et al., 1997). We also observed that the increase in the concentration of the conjugates paralleled that in ACC. This agrees with the broadly accepted idea that excess ACC is irreversibly conjugated to N-malonyl-ACC (Yang and Hoffman, 1984).

Figure 7.

1-aminocyclopropane-1-carboxylic acid (ACC) and ACC-conjugate levels, and growth response in gravistimulated poplar stems.

Field-grown trees were induced to form tension wood by bending.

(a) Growth (b) ACC and (c) ACC conjugates in the cambial-region tissues were measured after 1, 5 and 11 days in the tension wood side (white bars) and opposite wood side (black bars) and in upright control trees (grey bars). Mean ± SE, n = 6. Significance of treatments was tested for each time-point by analysis of variance followed by the Duncan's test at P-value less than or equal to 0.05. Different letters indicate significant differences.


To further our understanding of the biosynthesis and function of ethylene in the secondary xylem development, we identified a major ACC oxidase (PttACO1) expressed in the wood-forming tissues of poplar. PttACO1 was identified from the cambial-region and tension wood cDNA libraries created for the Swedish Poplar EST program ( After sequencing 100 000 poplar ESTs from 20 cDNA libraries representing different tissues, organs and environmental conditions, the only other library in which an EST for this gene had been found was a root library. The full-length PttACO1 gene was cloned and sequenced, and expression analysis confirmed its specific expression in developing secondary xylem, its strong induction during tension wood formation, and its presence in root tips (Figures 2, 3 and 6).

The unilateral induction of PttACO1 on the tension wood side added to a long list of cases, including many species and conditions, in which differential expression of ACC oxidases (and their essential involvement in increased ethylene production) had been demonstrated. It is generally accepted, however, that because of the low abundance of ACC during vegetative growth and the short lifetime of the ACC synthase transcripts and enzymes, the conversion of S-adenosyl methionine to ACC is the controlling step in ethylene production (Wang et al., 2002). Indeed, induced ethylene production has been repeatedly observed in parallel with an increase in ACC levels (Hyodo, 1991; Yang and Hoffman, 1984). Our gravistimulation of poplar stems did not provide an exception to these observations, but unexpectedly, there was only a very slight increase in ACC on the tension wood side of the stem, where PttACO1 expression and ACC oxidase activity were induced, while the opposite wood side accumulated high levels of ACC (Figures 4, 6 and 7). The opposite side also exhibited an increase in conjugated ACC, which is interesting, as conjugation serves to inactivate surplus ACC (Yang and Hoffman, 1984). Despite the increase in ACC on the opposite side, the expression of PttACO1 was maintained at low levels (Figure 6b). Therefore, we conclude that differential expression of PttACO1 plays an essential role in controlling the asymmetric production of ethylene induced during tension wood formation, and the results also strongly suggest that ACC availability is not an important element of this control.

Few other studies have given indications that ethylene production may be controlled by ACC oxidase activity rather than ACC availability. In pre-climacteric muskmelon fruit, the conversion of ACC to ethylene was concluded to determine the spatial and temporal distribution of wound-induced ethylene (Dunlap and Robacker, 1994). However, the cited study was based on measurements of ACC and ethylene, and did not provide any gene expression data. In another study on submerged Rumex palustris, ethylene biosynthesis was shown to be limited at the level of ACC oxidase activity (Vriezen et al., 1999). Investigations of differential growth during the formation and opening of the apical hook in germinating seedlings of pea and Arabidopsis have provided further examples of correlations between asymmetric induction of ACC oxidase and differential growth patterns (Peck et al., 1998; Raz and Ecker, 1999). However, in both of these studies, ACC oxidase was used as a marker for ethylene sensitivity, and its role in controlling ethylene production was not investigated.

When primary above-ground tissues of snapdragon were gravistimulated, observed increases in ACC were attributed to an auxin-induced ACC synthase, resulting from an auxin redistribution and increased levels on the lower side of the organ (Philosoph-Hadas et al., 2001). During our gravistimulation of poplar stems, auxin concentrations decreased both at the tension wood and at the opposite wood side compared to the levels in upright trees. This indicated that an auxin-inducible ACC synthase was not responsible for the overall increase in ACC in gravistimulated poplar, and that the response mechanism in gravistimulated poplar seemed to differ from that in primary above-ground tissues. It is also noted that while gravitational ethylene induction in primary stems was observed within hours, no significant induction of PttACO1 expression was detected until after 3 days (Figure 6). As the level of ACC in the xylem sap was found to increase in trees forming tension wood, the possibility could not be excluded that the ACC in the cambial region originated from transport in the xylem, but it was most likely that the synthesis took place within the poplar stem. Identification and characterisation of ACC synthases active in the poplar stem will be required to resolve this issue.

The expression of the PttACO1 transcript was specifically upregulated in xylem cells forming secondary walls, and the induction of ACC oxidase activity during gravistimulation was specific to the developing xylem tissues (Figures 2 and 3). These were intriguing observations suggesting that there were large differences in ethylene concentration across cambial-region tissues, which could be maintained because diffusion of ethylene in the aqueous phase was slow, and there was a lack of air spaces between cambial cells (Ingemarsson et al., 1991). In contrast to the expression of PttACO1, ACC levels were fairly similar across the cambial-region tissues (Figure 5) and its distribution patterns did not indicate that its conversion to ethylene was specifically localised in the developing xylem. A clearer understanding of the localised ethylene production within the cambial-region tissues will require immunolocalisation of the PttACO1 combined with high-resolution assays of ACC oxidase activity.

In addition to gravitational stimuli, mechanical strain imposed by rubbing, touching, flexing or bending of internodes/stems can induce ethylene biosynthesis in annual herbs as well as in forest trees (Brown and Leopold, 1973; Jaffe, 1980; Morgan and Drew, 1997; Telewski and Jaffe, 1986). Furthermore, bending of both Vigna and Arabidospsis leaves has been demonstrated to induce ACC synthase (Arteca and Arteca, 1999; Botella et al., 1995). In our study, mechanical strain was imposed on both the upper and lower sides of the poplar stems by bending. Despite this, PttACO1 expression was only induced on the upper side (Figure 6). This suggested that the induction of ethylene by bending is a gravitational response rather than one induced by mechanical strains. The gravitational nature of the unilateral induction of PttACO1 was also indicated in greenhouse-grown trees that were leaned but well supported to avoid mechanical strain (Figure 3). Similarly, Robitaille and Leopold (1974) concluded that the induction of ethylene by bending is a response to gravitational stimulation rather than to mechanical strain in experiments with primary apple shoots. These observations raise questions about whether the experimental induction of ethylene by, for example, touch, flexing and bending, and its environmental induction by wind, normally attributed to mechanical strain, are, in fact, gravitational responses resulting from the displacement of internodes from their equilibrium position. This issue is not easily solved, and will require very careful experimentation.

Not much has been concluded about the exact role of ethylene in cambial growth and wood development (Little and Pharis, 1995; Little and Savidge, 1987), and its requirement in any of the processes of division, expansion and maturation of cambial derivatives has yet to be conclusively established. This is in marked contrast to auxin, which has been demonstrated to be required for cambial cell division, cell identity of cambial initials and tracheary differentiation (Sundberg et al., 2001). It is possible that ethylene is produced in the wood-forming tissues only in response to various stresses such as strong wind, extreme temperature and water deficiency, and its function is to modulate cambial growth patterns according to these stresses. Numerous experiments in which ethylene or ethylene precursors have been applied to cambial tissues of conifer and angiosperm trees have demonstrated its potential to modify many aspects of cambial growth patterns and differentiation of xylem cells (Eklund and Little, 1996; Kalev and Aloni, 1999; Little and Pharis, 1995). However, application of ethylene has never been observed to induce typical tension or compression wood. The most consistent response to exogenous ethylene is the stimulation of cambial cell division and radial growth. Therefore, an important function of the ethylene induced during the wood-response reaction may be to stimulate cambial growth, and at least in poplar, this stimulation appears to be independent of an increase in auxin.

Experimental procedures

Plant material

Potted clonal plants of hybrid aspen (Populus tremula L. × P. tremuloides Michx.) were raised in a greenhouse to a height of 1–2 m, under a 22/17°C (day/night) temperature regime and a photoperiod of 18 h. Natural daylight was supplemented with HQI-TS 400 W TH−1 metal halogen lamps (Osram, Munich, Germany). The plants were watered daily and fertilised once a week with a 1 : 100 dilution of SUPERBA S (Hydro Superba AB, Landskrona, Sweden). Tension wood was induced by leaning and supporting the trees at an angle of about 30°.

Field-grown aspen trees (P. tremula L.) were selected from a natural stand near Umeå, Sweden (63°50′N, 20°20′E). Tension wood was induced by bending and fixing the trees with strings, so that the midpoint of the stem was at an angle of about 45°. Bending was induced during the most active period of cambial growth. Stem pieces were collected at the midpoint of the stem. For time-course experiments, the trees used were about 5 m high and 3 cm in diameter (breast height). Stem pieces for ACC analysis across cambial-region tissues were obtained from trees about 7 m high and 7 cm in diameter (breast height) after 11 days of bending.

Plant materials intended for molecular and chemical analysis were immediately frozen in liquid nitrogen and stored at −70°C. Materials collected in the field were transported to the laboratory on dry ice. Plant materials intended for anatomical analysis were frozen on dry ice and stored at −20°C. Samples for analysis were obtained from the upper (tension wood side) and lower (opposite wood side) quarters of the circumference.

Cambial-region tissues were collected by peeling the bark and scraping the exposed surfaces with a scalpel. The fraction on the xylem side (denoted by developing xylem) consisted of radially expanding, primary-walled xylem elements in the later stage of expansion and xylem elements in the stage of secondary wall formation. The fraction on the bark side (denoted phloem/cambium) consisted of xylem elements in the early stage of expansion, cambial zone cells and differentiating and mature phloem elements. For the analysis of ACC and ACC conjugates, the fractions were combined and denoted as cambial-region tissues. For Northern blot analysis, each fraction was used separately. Anatomical investigations were performed on transverse sections obtained from fresh material by cryosectioning, and was stained with toluidine blue O. Annual ring width was measured under a microscope at five positions.

For analysis of ACC across the cambial-region tissues, frozen samples were trimmed into 2 mm (tangential) × 10 mm (radial) × 15 mm (vertical) blocks, consisting of phloem, cambium and xylem. Samples for analysis consisted of 30 µm consecutive tangential sections obtained from the blocks using an HM 505E cryomicrotome (Microm Laborgeräte, Walldorf, Germany) at −20°C, according to Uggla and Sundberg (2001). The radial position of the tangential sections was determined in cross-sections sampled after every third tangential section.

Xylem sap was collected from 12 cm long stem pieces. Each piece was placed vertically and connected to a rubber tube. The tube was then filled with coloured water, creating pressure that forced the water through the xylem, thereby displacing the xylem sap and allowing it to be collected. Sampling was terminated when the coloured water had passed through the stem sample.

cDNA library and screening

Developing xylem tissues collected from greenhouse-grown trees were induced to form tension wood by leaning, and were used to construct a cDNA library using the Uni-ZAP XR vector system (Stratagene, La Jolla, CA, USA). Approximately 2.4 × 105 plaques were transferred onto Hybond–N filters (Amersham, Little Chalfont, UK) and screened with a [α-32P]-dCTP-labelled random-primed PCR fragment (0.5 kb) amplified from the ACC oxidase-like EST A018P61U (Sterky et al., 1998). After two rounds of plaque purification and hybridisation, eight positive clones were excised in vivo with ExAssist helper phage (Stratagene, La Jolla, CA, USA), according to the manufacturer's recommendations. Three of the cDNAs appeared to be of full length and were fully sequenced in both directions. Based on the nucleotide sequence, we concluded that all these cDNAs corresponded to the same gene, designated as PttACO1.

RNA extraction and expression analysis

Samples were homogenised in liquid nitrogen using a mortar and pestle. Total RNA was prepared using an RNeasy Plant Mini Kit (Qiagene, Hilden, Germany). For cortex, fully expanded leaves and roots, the method described by Chang et al. (1993) was used. After precipitation with LiCl overnight, the RNeasy kit was used for the washing and elution steps.

Total RNA was separated on a formaldehyde agarose gel according to Sambrook et al. (1989), and blotted onto a Hybond-N nylon filter (Amersham, Little Chalfont, UK). A PttACO1-specific probe was created from a PCR-amplified fragment (0.6 kb) including the 5′-untranslated region. Radiolabelling was performed with [α-32P] dATP using a Strip-EZ™ DNA (Ambion, Austin, TX, USA) probe synthesis kit, following the supplier's instructions. Separation of the labelled probe from unincorporated nucleotides was performed by Nick columns (Pharmacia Biotech, Uppsala, Sweden) using standard techniques. Hybridisation was performed overnight at 42°C in Ultrahyb™ solution (Ambion, Austin, TX, USA). Final washing was carried out in 0.1× SSC, 0.1% SDS at 42°C. The radioactivity on the membrane was detected using a GS-525 Molecular Imager (Bio-Rad, Solna, Sweden).

Expression analysis across the cambial region was carried out using dot blots as previously described by Moyle et al. (2002). Poly(A)+ mRNA was extracted from specific cambial-region tissues obtained by tangential cryosectioning, reverse-transcribed and PCR-amplified. For each dot, about 500 ng of PCR product was added and the nylon membrane was hybridised with the PttACO1-specific probe described above.

Measurement of ACC oxidase activity

ACC oxidase activity was measured according to Veveridis and John (1991) with some modifications. Samples were homogenised in liquid nitrogen with a mortar and pestle. One gram of homogenised material was extracted in 2 ml of buffer (100 mm Tris–HCl, pH 7.2, 30 mm Na-ascorbate and 10% glycerol). After centrifugation at 15 000 g for 10 min at 4°C, the supernatant was removed and used in the assay. The reaction mixture, containing 1.1 ml of the extraction buffer, 50 µl of 40 mm ACC, 50 µl of 2 mm FeSO4 and 800 µl of the extract, was incubated at 30°C for 2 h with gentle agitation. One millilitre of the head-space was then removed and analysed with a Shimadzu GC-8 A gas chromatograph.

Quantification of ACC

One to fifteen milligrams of each sample was extracted in 80% methanol, with [2H4]ACC (1-amino-[2,2,3,3–2H4]cyclopropane-1-carboxylic acid (Sigma, St. Louis, MO, USA) as an internal standard. The extract was then purified by SCX–cation exchange chromatography. Samples for ACC-conjugate analysis were hydrolysed in 4 m HCl for 1 h at 100°C, prior to purification. The ACC was converted to an n-benzoyl-n-propyl derivative using a protocol modified from the one used in the study by Hall et al. (1989). After derivatisation, the sample was further purified on a C18 solid-phase cartridge. The samples were analysed by GC–MS-selected reaction monitoring (SRM; Edlund et al., 1995) using a JEOL MStation mass spectrometer (JEOL, Tokyo, Japan). A more detailed description of the method will be published elsewhere. The accuracy of the analysis of both ACC and conjugated ACC was established by successive approximation (Reeve and Crozier, 1980), and the identity of the quantified peaks was established by obtaining full-scan mass spectra, including the characteristic masses for the n-benzoyl-n-propyl derivatives of ions m/z 247, m/z 205, m/z 187, m/z 159 and m/z 105, in the same ratios as in spectra from standard samples (data not shown).


We would like to thank Drs Ewa Mellerowicz and Jaakko Kangasjärvi for valuable comments on the manuscript, Jarmo Schrader for providing the filter for dot-blot analysis and Kjell Olofsson for technical assistance. This research was supported by Formas, the Swedish Research Council and the Foundation for Strategic Research.

Accession number for EMBL sequence database: AY167040.