Transitions in the functioning of the shoot apical meristem in birch (Betula pendula) involve ethylene


  • Raili Ruonala,

    1. Plant Biology, Department of Biological and Environmental Sciences, University of Helsinki, FI-00014 Helsinki, Finland,
    Search for more papers by this author
  • Päivi L.H. Rinne,

    1. Department of Plant and Environmental Sciences, Norwegian University of Life Sciences, N-1432 Ås, Norway, and
    Search for more papers by this author
  • Mourad Baghour,

    1. Umeå Plant Science Centre, Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, S-90187 Umeå, Sweden
    Search for more papers by this author
  • Thomas Moritz,

    1. Umeå Plant Science Centre, Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, S-90187 Umeå, Sweden
    Search for more papers by this author
  • Hannele Tuominen,

    1. Plant Biology, Department of Biological and Environmental Sciences, University of Helsinki, FI-00014 Helsinki, Finland,
    Search for more papers by this author
    • Current address: Umeå Plant Science Centre, Department of Plant Physiology, Umeå University, S-90187 Umeå, Sweden.

  • Jaakko Kangasjärvi

    Corresponding author
    1. Plant Biology, Department of Biological and Environmental Sciences, University of Helsinki, FI-00014 Helsinki, Finland,
    Search for more papers by this author

*(fax +358 9 19159552; e-mail


In many trees, a short photoperiod (SD) triggers substantial physiological adjustments necessary for over-wintering. We have used transgenic ethylene-insensitive birches (Betula pendula), which express the Arabidopsis ethylene receptor gene ETR1 carrying the dominant mutation etr1-1, to investigate the role of ethylene in SD-induced responses in the shoot apical meristem (SAM). Under SD, the ethylene-insensitive trees ceased elongation growth comparably to the wild-type. In contrast, the formation of terminal buds, which in trees is typically induced by SD, was abolished. However, although delayed, endo-dormancy did eventually develop in the ethylene-insensitive trees. This, together with the rapid resumption of growth in the ethylene-insensitive trees after transfer from non-permissive to permissive conditions suggests that ethylene facilitates the SD-induced terminal bud formation, as well as growth arrest. In addition, apical buds of the ethylene-insensitive birch did not accumulate abscisic acid (ABA) under SD, suggesting interaction between ethylene and ABA signalling in the bud. Alterations in SAM functioning were further exemplified by reduced apical dominance and early flowering in ethylene-insensitive birches. Gene expression analysis of shoot apices revealed that the ethylene-insensitive birch lacked the marked increase in expression of a beta-xylosidase gene typical to the SD-exposed wild-type. The ethylene-dependent beta-xylosidase gene expression is hypothesized to relate to modification of cell walls in terminal buds during SD-induced growth cessation. Our results suggest that ethylene is involved in terminal bud formation and in the timely suppression of SAM activity, not only in the shoot apex, but also in axillary and reproductive meristems.


Plants monitor critical environmental changes to time important developmental events. For example, in many species, flowering and dormancy are triggered by perception of a critical day length (Espinosa-Ruiz et al., 2004; Mouradov et al., 2002; Rohde et al., 2000; Vince-Prue, 1994; Wareing, 1956). In photoperiodically induced dormancy, the perception of short days (SD) results in cessation of growth, formation of a terminal bud, and the development of a dormant and freezing-tolerant state. Perception of SD takes place in the leaves, from where a signal is believed to be transported to the shoot apex (Vince-Prue, 1994) where it triggers developmental processes, such as drastic changes in the organization of the shoot apical meristem (SAM) (Rinne et al., 2001) that ultimately lead to bud dormancy.

Changes in plant hormone levels are often associated with alterations in meristem activity and bud dormancy. Abscisic acid (ABA) has traditionally been viewed as a central player in the regulation of dormancy (Eagles and Wareing, 1963; Horvath et al., 2003). Recent studies, however, suggest additional roles for ABA in buds, for example as a regulator of acclimatization to low temperature and dehydration in the buds that are entering the dormant state (Arora et al., 2003; Rinne et al., 1998; Rohde et al., 2000; Welling et al., 1997).

The gaseous plant hormone ethylene also has a central role in various developmental and stress responses in plants (Bleecker and Kende, 2000), and it is known to interact with ABA in some of these processes (Stepanova and Alonso, 2005). For example, mature seeds of ethylene-insensitive mutants in Arabidopsis (e.g. etr1) have an elevated sensitivity to exogenous ABA, which more effectively inhibits germination of etr1 seeds (Ghassemian et al., 2000). In potato, application of ethylene has been shown to induce endo-dormancy in microtubers grown in tissue culture (Suttle, 1998), although it is not known whether endogenous ethylene signalling is involved in SD-induced tuber formation. However, there is no information available as to whether a similar interaction between ethylene and ABA is involved in bud dormancy of perennial plants.

As suggested previously (Rinne and van der Schoot, 2003), dormancy can also be viewed as a defence or stress response as it involves mechanisms that primarily function in defence against various biotic and abiotic stresses. For example, deposition of 1,3-β-d-glucan (callose), which characteristically blocks the plasmodesmata channel during endo-dormancy (Rinne and van der Schoot, 1998; Rinne et al., 2001), is also induced during the hypersensitive response triggered by pathogen invasion (Blumwald et al., 1998) and during prolonged exposure to abiotic factors such as aluminium (Sivaguru et al., 2000) or ozone (Schraudner et al., 1992). As ABA and ethylene are implicated in a variety of defence and stress responses, they might also function in concert during dormancy development.

Ethylene is also involved in other developmental phenomena, such as flowering in Arabidopsis. For example, the ethylene-insensitive mutants etr1 and ein2 have a late-flowering phenotype (Guzmán and Ecker, 1990; Kieber et al., 1993; Smalle and Van Der Straeten, 1997), suggesting that ethylene may facilitate the transition to flowering, a process that requires reprogramming of the vegetative SAM. Meristem reprogramming also occurs during growth cessation and the development of bud dormancy. However, the role of ethylene in the development of bud dormancy is poorly understood. This might be partly due to the fact that annuals such as Arabidopsis do not assume dormancy, and thus cannot be used as a model to study the dormancy regulation typical of perennial plants.

We have investigated the role of ethylene signalling in the growth and development of birch (Betula pendula Roth), as well as in the responses to various stresses, by using transgenic ethylene-insensitive trees (Vahala et al., 2003). Ethylene insensitivity in these trees is due to expression of the Arabidopsis ethylene receptor gene carrying the dominant etr1-1 mutation (Chang et al., 1993a). Previously we have addressed ozone responses (Vahala et al., 2003) and herbivore resistance (Alatalo, I., Haviola, S., Ruonala, R., Kangasjärvi J. and Saloniemi, I., University of Turku and University of Helsinki, Finland, unpublished data) of these trees, and found that ethylene insensitivity reduced ozone-induced cell death and improved the leaf quality as food for an insect herbivore.

Here we address the question of whether ethylene signalling influences the functioning of the SAM, and, in particular, whether ethylene is involved in transitions in SAM responses and function triggered by photoperiodic changes. We show that ethylene-insensitive trees grew normally under long-day conditions (LD) but differed from the wild-type in their response to SD: although elongation growth ceased in the ethylene-insensitive trees, the formation of terminal buds, as well as the development of bud dormancy, were different than in the wild-type. In addition, buds of the ethylene-insensitive birches were impaired in SD-induced ABA accumulation, suggesting that ethylene promotes ABA accumulation at an early stage during the photoperiodic induction of bud dormancy. Alterations in SAM functioning were further evident from reduced apical dominance and the early flowering in the ethylene-insensitive birch. Taken together, our findings suggest novel roles for ethylene in SAM-specific phenomena that are associated with bud dormancy and flowering in perennial plants.


Ethylene insensitivity leads to impaired terminal bud formation under natural inductive conditions

Growth and development of birch shoot tips was monitored under LD in five ‘wild-type’ birch clones and in up to four transgenic ethylene-insensitive lines in each clonal background. No differences were found in stem elongation between the transgenic lines and the corresponding wild-types (Figure 1a), but radial growth of the stem was reduced in some of the ethylene-insensitive lines (Figure 1b).

Figure 1.

 Growth of wild-type birch and ethylene-insensitive lines under long-day conditions.
(a) Cumulative length of 10 internodes and (b) stem diameter 2 cm above the base were measured for each individual after three months of growth in the greenhouse. Transgenic, ethylene-insensitive lines in five clonal backgrounds are shown. First on the left in each background is the wild-type (WT, black bars), which is followed by the transgenic lines. The mean value of 7–9 measurements is shown, and vertical bars represent standard deviation. Asterisks indicate significant differences between a transgenic line and the corresponding background clone (anova: *0.01 ≤ P ≤ 0.05; **0.001 ≤ P < 0.01; ***P < 0.001).

To test the responses to a combination of SD and low temperature (LT), wild-type and transgenic trees were exposed to natural outdoors SD conditions (60°N; 24°E, Helsinki, Finland), beginning in late August. During the 14-week experiment, the clones were exposed to their critical inductive photoperiod and low temperature; day length shortened from 15.5 to 7.0 h, and the average daily temperature decreased from 18°C to −1°C, varying between 21°C and −2°C. During the experiment, cessation of elongation growth took place in all wild-type trees and transgenic lines, and all wild-type trees developed normal terminal apical buds. In contrast, apical bud development was impaired in the ethylene-insensitive transgenic trees. Under SD− conditions, they did not form terminal buds, i.e. SAMs covered by bud scales (Figure 2), instead they formed leaves, which, however, did not fully expand when growth ceased. This was evident in the shoot apex as well as in the apices of branches of all transgenic ethylene-insensitive lines tested in all clonal backgrounds. In addition, naturally induced leaf senescence was significantly delayed in the strongly ethylene-insensitive lines, as judged by chlorophyll breakdown (Figure 3a,b). This was most pronounced in the JR1/4 background. Leaf abscission was also affected by ethylene insensitivity; the first-season leaves remained attached to the transgenic trees after over-wintering (Figure 3c). In summary, this indicates that ethylene is involved in terminal bud formation and in the successive events of autumn leaf senescence and leaf abscission.

Figure 2.

 Phenotype of wild-type and ethylene-insensitive birch after bud set.
The plants were exposed to natural outdoor conditions for 6 weeks before photographing. (a) Ethylene-insensitive line (BPetr1-1-35) on the left and the corresponding background clone (V5834) on the right, with arrows pointing to the apices. Close-up of the apex of (b) BPetr1-1-35 and (c) V5834. WT, wild-type; etr1-1, ethylene-insensitive.

Figure 3.

 Natural senescence and abscission of leaves in birch.
Wild-type and transgenic ethylene-insensitive lines were exposed to ambient day length and temperature for (a) 5 weeks or (b) 7 weeks, and young, fully expanded leaves from three individuals of each line were collected for determination of chlorophyll content. The average values of wild-type (JR1/4, black bars) and the transgenic lines in this background are shown. Vertical bars represent standard deviation. Asterisks indicate significant differences between transgenic line and wild-type (anova: *0.01 ≤ P ≤ 0.05; **0.001 ≤ P < 0.01). Similar results were recorded for leaves representing older developmental stages of the same trees (data not shown). (c) JR1/4 and two ethylene-insensitive lines in this background photographed at the beginning of the second growth season. WT, wild-type.

Growth resumption of ethylene-insensitive birch is rapid after a transient exposure to short day length

The impaired terminal bud formation in the ethylene-insensitive trees prompted us to study the responses to SD independently of the effect of LT. Actively growing wild-type (V5834) and transgenic ethylene-insensitive lines in this background (BPetr1-1-35 and BPetr1-1-86) were placed under SD conditions at a constant temperature (18°C), and transferred back to LD at regular intervals. Under constant LD, the rate of stem elongation and leaf formation did not differ between the wild-type and the ethylene-insensitive trees (data not shown). Four days of SD was enough to reduce stem elongation and leaf formation in all lines, but, after return to LD, growth resumed rapidly with no differences between the wild-type and transgenic lines (data not shown). In response to 12 days of SD, growth cessation was slightly faster in the wild-type when compared to the ethylene-insensitive lines (Figure 4a). After transfer to LD, the small unexpanded leaves of the transgenic lines started to grow, and the formation of new leaves resumed immediately, while in the wild-type there was a lag of more than 4 days before formation of new leaves continued (Figure 4b). Resumption of leaf formation after 16 and 20 days of SD showed a similar pattern (data not shown). After 24 days under SD, the wild-type trees resumed growth only 2 weeks after the transfer to LD (Figure 4c,d), whereas the transgenic trees formed new primordia immediately (Figure 4d).

Figure 4.

 Growth rate of birch in response to transient short days.
Wild-type (V5834) and ethylene-insensitive lines (BPetr1-1-35 and BPetr1-1-86) were grown in the greenhouse before transferring to growth cabinets. For the SD exposure (filled bar), day length was shortened to 12 h for 12 days (a, b) or 24 days (c, d). After SD, the plants were transferred back to LD (open bar) to follow growth resumption. Throughout the experiment, stem elongation (a, c) as well as formation of leaves (b, d) were followed. The growth rate of the SD-exposed plants is presented as percentage of the growth of plants under constant LD. Two to four plants in each line were followed during each exposure and the mean values were used for the final calculations. The experiment was repeated with independent material, and similar results were obtained. WT, wild-type.

A distinct difference was also observed in the appearance of the apical bud under SD, whereby wild-type trees formed a terminal bud after 12 days of SD, while the transgenic trees did not form a terminal bud and the youngest leaves continued growing, although they did not expand fully (not shown). This is identical to the response to a combined SD and LT exposure under natural conditions (Figure 2b). In birch, the small leaf below the terminal bud is normally committed to die, while the stipules eventually expand to form scales covering the bud containing approximately nine embryonic leaves with stipules and two primordia. In contrast, leaves in the apical region of the ethylene-insensitive trees were not aborted; instead they were somewhat more unfolded. Nonetheless, the number of organs in the apex did not differ from that of the wild-type (data not shown). Taken together, exposure of the transgenic trees to SD without LT induced growth cessation that was slightly slower, and recovery after return to LD was faster than in the wild-type.

Endo-dormancy develops late in ethylene-insensitive birch

The more rapid resumption of growth after a short-term exposure to SD suggests that the entry to dormancy is compromised in the ethylene-insensitive trees. To investigate whether ethylene signalling affects endo-dormancy, we transferred trees to SD for 2-7 weeks and tested the bud burst capacity of single node cuttings under growth-promoting conditions. At the onset of SD exposure, bud burst capacity did not differ between the wild-type and the ethylene-insensitive trees (Figure 5). In the wild-type trees, bud burst was inhibited by increasing duration of SD: only 20% of the buds burst after 5 weeks under SD and 0% burst after 6 weeks under SD (Figure 5). In contrast, an exposure to SD of up to 5 weeks had no effect on bud burst capacity in the strongly ethylene-insensitive line BPetr1-1-35, in which the bud burst percentage did not decrease during this time (Figure 5). However, after 7 weeks under SD, even the line BPetr1-1-35 had entered endo-dormancy and bud burst capacity was completely abolished (Figure 5). Similar results were obtained with intact plants after a 10-week SD treatment: neither the ethylene-insensitive trees nor the wild-type could initiate growth without a dormancy-removing cold exposure (data not shown). In conclusion, in birch buds, ethylene appears to be involved in the entry to dormancy rather than the establishment of dormancy as such.

Figure 5.

 Endo-dormancy development in birch.
Wild-type (V5834) and ethylene-insensitive lines (BPetr1-1-35 and BPetr1-1-86) were exposed to SD for dormancy development. After 2–7 weeks of SD, single node cuttings were collected from three to five trees, and incubated in water at 18°C under constant light. The number of burst buds was recorded after 3 weeks of incubation, and expressed as a percentage of the total number of buds (n = 18–30). Mean values are shown, and vertical bars represent standard deviation. Asterisks indicate significant differences between transgenic line and the corresponding background clone at the given time point (anova: ***P < 0.001). The experiment was repeated with independent material, and similar results were obtained. WT, wild-type.

Endogenous abscisic acid levels do not increase in the ethylene-insensitive birch during SD-induced growth cessation

Elevated levels of endogenous ABA have been associated with growth cessation and bud formation in birch (Rinne et al., 1998; Welling et al., 1997) and Populustremula x P. alba (Rohde et al., 2002). We measured ABA concentrations in the apices of the wild-type and ethylene-insensitive birch trees at 3- to 4-day intervals from the start of an SD treatment. As earlier reports have shown that ABA concentration increases transiently in birch before growth cessation takes place (Rinne et al., 1998; Welling et al., 1997), the experimental period was limited to 2 weeks, which is enough for growth cessation to begin under the conditions used (see Figure 4). A transient, threefold increase in ABA content was found in the wild-type V5834 after 10 days under SD, whereas ABA levels remained unchanged in the transgenic lines BPetr1-1-35 and BPetr1-1-86 during the 2 weeks under SD (Figure 6a). The peak in the endogenous ABA level in the clone V5834 coincided with the beginning of growth cessation, which was initiated one week after the beginning of SD treatment (Figure 6b–d).

Figure 6.

 ABA content of apical buds and growth responses of birch under SD.
For the SD-exposure, wild-type (V5834) and transgenic ethylene-insensitive lines (BPetr1-1-35 and BPetr1-1-86) were transferred to a greenhouse compartment with a photoperiod of 12 h. At given time points, apices were collected to analyse the ABA content (a). Three apices were pooled for a single analysis, and each data point shows the mean value of three such analyses. During the same experiment, stem elongation (b), number of new leaves formed (c), and bursting of axillary buds (d) were recorded. The cumulative mean values of seven individuals are shown. Vertical bars represent standard deviation. WT, wild-type.

The ethylene-insensitive lines display impaired ABA responsiveness

As exogenous ABA inhibits the release of dormancy in axillary buds of willow (Salix viminalis; Barros and Neill, 1986) and birch (Rinne et al., 1998), we tested the effect of exogenous ABA on the wild-type and the transgenic lines with single node cuttings. Cuttings were collected from dormant trees at the beginning of December close to the shortest annual photoperiod when birch buds are responsive to ABA (Rinne et al., 1998). As anticipated, ABA delayed bud burst in the wild-type V5834 (Figure 7a) when compared to the water-treated controls (Figure 7b). In contrast, the ethylene-insensitive trees showed a reduced ABA inhibition of bud burst. In the experiment shown in Figure 7, 80% of ABA-treated buds in the transgenic line BPetr1-1-35 had burst after 7 days of incubation under growth-promoting conditions, while bud burst in the wild-type was prevented under identical conditions (Figure 7a). This indicates that, in comparison to the wild-type, the buds of the ethylene-insensitive trees were less sensitive to exogenously applied ABA.

Figure 7.

 ABA responsiveness of dormant birch buds.
The effect of ABA on bud burst was tested on wild-type (V5834) and ethylene-insensitive lines (BPetr1-1-35 and BPetr1-1-86) after 3 months of exposure to outdoor conditions. Single node cuttings were collected from five individuals at the beginning of December, and incubated in 50 μm ABA (a) or water (b), in growth-promoting conditions. Bud bursting was recorded over a period of five 5–12 days. The values represent the number of burst buds as a percentage of the total number of buds (n = 10). Similar results, albeit with somewhat different timing of bud burst, were reproduced three times with independent material. WT, wild-type.

Reduced apical dominance and early flowering in the ethylene-insensitive trees

Altered SAM functioning in the ethylene-insensitive trees raised the question of whether ethylene also controls other meristems than SAM. Apical dominance, which reflects the activity of axillary meristems (Chatfield et al., 2000; McSteen and Leyser, 2005), was reduced in the ethylene-insensitive birch, which produced approximately three times more branches than the wild-type (Figure 8a). A similar trait was evident in all transgenic lines observed in all five backgrounds. Even the axillary buds of branches (second-order buds) were released acropetally in the ethylene-insensitive trees during the first growing season. This was never observed in the wild-type. The size and form of the axillary buds did not differ from the wild-type. Furthermore, all the apices of the ethylene-insensitive trees remained viable (Figure 8b). In contrast, in the wild-type,, a phenomenon called self-pruning led to abortion of apical buds of the branches (Figure 8c) and the death of the oldest branches at the end of the growing season. This, together with reduced apical dominance, led to a bush-like appearance in the ethylene-insensitive trees (Figure 8d).

Figure 8.

 Apical dominance and fate of apices and axillary buds of branches in wild-type and ethylene-insensitive birch.
(a) The degree of apical dominance in wild-type (V5834, black bar) and ethylene-insensitive lines (BPetr1-1-35 and BPetr1-1-86) was quantified as the number of branches in individuals grown for 4 months under LD. Mean values of eight to ten individuals of each line are shown, and vertical bars represent standard deviation. A branch of an ethylene-insensitive line (BPetr1-1-86) (b) and a wild-type (V5834) tree (c) with an arrow pointing at the aborted apex. The branches were of the same developmental age. (d) Wild-type (JR1/4) and an ethylene-insensitive line (BPetr1-1-44) at the age of two years. Close-up of a male (e) and a female (f) inflorescence of ethylene-insensitive lines (BPetr1-1-55 and BPetr1-1-201) at the beginning of the second growth season. WT, wild-type; etr1-1, ethylene-insensitive.

The fact that the axillary buds had already developed branches during the first growing season, and that the apices of the branches of the transgenic trees did not abscise as in the wild-type, suggests that the functioning of both SAM and axillary meristems is drastically altered in the ethylene-insensitive trees. This was further evident from the formation of male inflorescences during the second growth season (Figure 8e), followed by the emergence of female inflorescences from axillary buds during the subsequent season (Figure 8f). We have grown the wild-type V5834 under greenhouse conditions now for five years, and during this time it has never produced inflorescences. Under outdoor conditions, birch trees usually flower at the age of 10–15 years (Perala and Alm, 1990). Early flowering of the ethylene-insensitive trees has been observed in all transgenic lines in all five backgrounds. In the most pronounced case, line BPetr1-1-55 in the JR1/4 background, the transition from vegetative to reproductive development occasionally occurred already during the first growing season. Crosses between the transgenic lines produced viable seeds. The resulting progenies retained ethylene insensitivity as judged by a visual screen that scored (i) the capacity to form an apical bud under inductive conditions, (ii) the degree of apical dominance, and (iii) the degree to which lateral shoot tips abscised. The early-flowering phenotype of the ethylene-insensitive birch suggests that ethylene is involved in facilitating the transition to flowering in birch.

Ethylene-dependent gene expression in birch apices under SD

To identify ethylene-dependent genes that show differential expression in birch buds under SD, we analysed the expression of approximately 6300 genes with a microarray containing Populus euphratica ESTs (Broschéet al., 2005) of branches. We selected for further studies a putative beta-xylosidase gene (PeBXL1, AJ776591), which had a clearly altered expression pattern in the ethylene-insensitive trees in two independent experiments (data not shown). Xylosidases are also interesting as they are implicated in processes such as modification of cell walls, and therefore could play a role in growth cessation and dormancy (Collins et al., 2005; Goujon et al., 2003; Martínez et al., 2004; Rinne and van der Schoot, 1998, 2003; Rinne et al., 2001; Ruperti et al., 2002; Van Zhong and Burns, 2003; de Vries and Visser, 2001). The deduced amino acid sequence of PeBXL1 is 95% identical to the predicted amino acid sequence of a Populus trichocarpa BXL (estExt_fgenesh4_pm.C_LG_VIII0450), which belongs to a gene family of 11 members ( The amino acid sequences of P. trichocarpa and P. euphratica BXLs are 83 and 80% identical, respectively, to an Arabidopsis beta-xylosidase with demonstrated enzyme activity (Goujon et al., 2003). The identities were determined over a 121-amino acid sequence that encloses the two glycosyl-hydrolase domains. Moreover, the P. trichocarpaBXL genes do not have similarity to any other gene families and therefore can all be regarded as beta-xylosidase genes.

We identified in birch a BXL EST cluster, consisting of two EST clones (BpBXL1, Genbank numbers DW986525, DW986526), that was most similar to PeBXL1. The deduced amino acid sequences of BpBXL1 and PeBXL1 were 90% identical. Quantitative RT-PCR showed that the expression of BpBXL1 was significantly increased after 12 SD in the apices of wild-type V5834 (Figure 9), whereas it remained low in the transgenic trees (Figure 9). The difference was reduced by 16 SD and further by 20 SD (Figure 9). The fact that the SD-induced activation of BpBXL1 expression coincides with the most intense growth retardation in the wild-type (see e.g. Figure 4a), and that both BpBXL1 activation and growth retardation are altered in the ethylene-insensitive birch, suggests that BpBXL1 could be involved in regulation of growth cessation under SD in an ethylene-dependent manner.

Figure 9.

 Expression of beta-xylosidase under SD and LD in apices of birch branches.
Wild-type (V5834) and ethylene-insensitive lines (BPetr1-1-35 and BPetr1-1-86) were exposed to SD, and apices of four individuals were pooled as one sample for RNA isolation at each time point. Material for the LD data point was collected at the first two time points from actively growing trees that were of the same age as the corresponding SD-exposed trees. The relative abundance of BpBXL1 mRNA was determined by quantitative RT-PCR as described in Experimental procedures. Vertical bars represent standard deviation among three to four samples. WT, wild-type.


Ethylene is involved in arresting the activity of the shoot apical meristem in birch

In photoperiodically induced dormancy, perception of a critical day length leads to growth cessation, terminal bud formation and dormancy (Goffinet and Larson, 1981; Rohde et al., 2000; Welling et al., 1997). The entry to dormancy in the SAM coincides with the formation of a terminal bud, the scales of which act as protective structures for the SAM. Our results show that SD-induced growth cessation was similar in the ethylene-insensitive and wild-type birch (Figure 4). However, ethylene appears to control terminal bud formation and is involved in the entry of the buds into dormancy, as these processes were either missing or delayed in the ethylene-insensitive trees (Figures 2 and 5). Similar to our results, over-expression of ABSCISIC ACID-INSENSITIVE3 (ABI3) in Populustremula x alba resulted in apical buds with large embryonic leaves and small bud scales under SD (Rohde et al., 2002). Furthermore, deficiency in ABA biosynthesis delayed or reduced bud maturation in birch (A. Welling, University of Helsinki, Finland, personal communication). Therefore, it seems that both ethylene and ABA are involved in the photoperiodic control of terminal bud formation in trees.

Our results confirm previous observations that growth cessation and bud set are distinct developmental phenomena (Rinne et al., 2001; Rohde et al., 2002). Moreover, we show here that the formation of a terminal bud and endo-dormancy are also distinct developmental events, as the ethylene-insensitive birch eventually developed endo-dormancy but did not form terminal buds. The capacity of the ethylene-insensitive trees to enter endo-dormancy, although delayed, suggests that ethylene modifies but is not required for endo-dormancy. This resembles ethylene's role in the regulation of senescence, which is delayed but not prevented in ethylene-insensitive Arabidopsis mutants (Grbic and Bleecker, 1995). However, it remains unclear whether ethylene acts directly on the shoot apex itself or whether ethylene effects are mediated by the leaves, for example. Nevertheless, our results suggest that ethylene influences several different aspects of SAM activity in birch, and that ethylene might therefore also have an important role for survival in natural conditions.

Interaction between ethylene and ABA signalling pathways in birch

Interactions between ethylene and ABA signalling pathways are well-known (Gazzarrini and McCourt, 2003; Stepanova and Alonso, 2005). For example, ethylene-insensitive Arabidopsis mutants are also affected in ABA responses and metabolism (Beaudoin et al., 2000; Chiwocha et al., 2005; Ghassemian et al., 2000). The ein2 mutant overproduces ABA (Ghassemian et al., 2000), and in etr1-2, the high levels of ABA may contribute to a deeper seed dormancy compared to the wild-type (Chiwocha et al., 2005). The interplay between ethylene signalling and ABA in trees, however, has received little experimental attention due to the lack of proper tools, such as mutants and transgenic plants deficient in these processes. Although the role of ABA in bud dormancy remains to be established (see e.g. Arora et al., 2003), a transient increase in ABA content accompanying growth arrest in trees has been documented in birch (Rinne et al., 1998; Welling et al., 1997) as well as in poplar (Rohde et al., 2002). Our analyses on the wild-type birch (clone V5834) were in agreement with these earlier findings in showing that ABA content is transiently elevated during the initiation of growth arrest. In contrast, we did not detect a similar ABA increase in apices of the ethylene-insensitive birch (Figure 6a), suggesting that ethylene is required for ABA accumulation in birch apices during SD-induced growth cessation. However, it is also possible that the transient increase in ABA content was delayed beyond the time frame of our analysis. Altogether, our results suggest that, in birch, ethylene and ABA signalling overlap or interact in several biological processes, such as the timely and proper formation of a bud, the release from bud dormancy, and the accumulation of ABA in apices in response to SD.

Ethylene plays a role in apical dominance and flowering in birch

Ethylene not only affected the behaviour of terminal buds in birch, but also that of the axillary buds, which consistently developed into branches in the ethylene-insensitive lines. Such changes in the degree of apical dominance have not been reported in ethylene-insensitive tobacco (Knoester et al., 1998) or Arabidopsis (Cline, 1991). Axillary bud outgrowth is believed to be inhibited by apically derived auxin, although auxin itself does not enter the bud (Hall and Hillman, 1975; Morris, 1977). The indirect effect of auxin appears not to be relayed by ethylene in Arabidopsis (Chatfield et al., 2000). In contrast, in birch, ethylene is clearly required for the suppression of the axillary meristem activity (see e.g. Figure 8d). This difference between Arabidopsis and birch might reflect the differential growth habits of annuals and perennials or differences in axillary bud ontogeny between Arabidopsis and birch. The axillary buds of Arabidopsis develop unrelated to the age of the subtending leaf (Schnittger et al., 1996; Stirnberg et al., 1999), while in birch the axillary meristems arise at a short distance from the SAM.

In birch, apical buds of branches normally abort at the end of the growing season (Maillette, 1982; see also Figure 8c). In contrast to this, the shoot tips in the branches of the ethylene-insensitive trees did not abscise but remained vegetative or formed male inflorescences. This finding is indicative of the involvement of endogenous ethylene in the abscission of shoot tips in birch, and raises the question as to whether the apical buds of branches are programmed to die in the wild-type via an ethylene-dependent process in order to prevent the formation of inflorescences until the proper developmental age of the plant is reached. Alternatively, a constant flowering signal might be produced in the ethylene-insensitive birch, which not only leads to early flowering but might also inhibit SD-induced dormancy in vegetative meristems.

Considering that ethylene-insensitive Arabidopsis flowers late, the opposite phenotype of ethylene-insensitive birch may seem surprising. However, as evident from other studies, plant hormones can have different roles in perennials and annuals. For example, gibberellins promote flowering in Arabidopsis but inhibit flowering in deciduous trees (see Brunner and Nilsson, 2004). Furthermore, the roots of ethylene-insensitive Arabidopsis are insensitive to ABA whereas the seeds are hypersensitive (reviewed in Gazzarrini and McCourt, 2003). Our results also indicated that the buds of the ethylene-insensitive birch are insensitive to ABA (Figure 7), whereas leaves appear to be hypersensitive (unpublished data). It could also be possible that the SAMs of a perennial tree and a short-cycle annual rosette plant have different responses to ethylene reflecting the different life strategies of the species. However, the basic mechanisms and relevance of these findings remain to be elucidated.

Ethylene-dependent expression of a xylosidase increases during growth cessation

A transient increase in the expression level of BpBXL1 was detected in apices of the wild-type under SD but not in the apices of the ethylene-insensitive trees. Beta-xylosidases are enzymes that are commonly secreted by cell wall-degrading pathogens, such as bacteria and fungi, and act together with endoxylanases to hydrolyse xylan, a hemicellulosic component that influences cell wall elasticity (Chanliaud et al., 2004; Collins et al., 2005; Fry, 2004; Somerville et al., 2004; de Vries and Visser, 2001). Endoxylanases cleave the xylan backbone into xylose oligomers, which are subsequently degraded by beta-xylosidases (Collins et al., 2005). In Arabidopsis, beta-xylosidase activity has been localized to the secondary cell wall, primarily in stems, flowers and siliques, and down-regulation of a beta-xylosidase by an antisense strategy suggested that it is involved in the normal development of leaves and siliques (Goujon et al., 2003).

Two possible explanations for the differential expression of BpBXL1 shown here can be envisaged. First, it is possible that the increase in expression of BpBXL1 under SD serves to modify the secondary cell walls in order to slow down growth in the wild-type. The expression of BpBXL1 differed most at 12 days of SD (Figure 9), the time at which a slight difference in the rate of stem elongation was first detected between the transgenic lines and the wild-type (Figure 4a,c).

An alternative explanation is that the increase in the mRNA level of BpBXL1 in the wild-type under SD might be related to the sealing of the cell walls and plasmodesmata that occurs during entry to dormancy in birch (Rinne and van der Schoot, 1998; Rinne et al., 2001). Similarly, a modulation of cell wall properties as well as induction of a callose synthase gene has been documented for poplar during cambial dormancy (Schrader et al., 2004). Accumulation of callose at plasmodesmata can be induced by xylanases (Iglesias and Meins, 2000), which, as described above, act in concert with xylosidases to degrade xylan. It is conceivable that the activities of xylanases and xylosidases are positively interconnected, in which case the low mRNA level of BpBXL1 suggests reduced callose accumulation in the ethylene-insensitive trees. In either case, assessment of the expression pattern of BpBXL1 over a more extended period, as well as cellular localization of the protein in birch apices, could reveal the possible function of xylosidases in SD-induced cessation of growth and bud dormancy.

The differential expression of BpBXL1 between apices of wild-type and ethylene-insensitive birch under SD gains further significance in the light of a potential overlap in bud dormancy and plant defence responses, both of which may involve ethylene and cell wall modifications. Whether the other ethylene-dependent processes in perennial trees revealed in this study, such as inhibition of axillary bud outgrowth and stimulation of apical bud abortion of the branches, relate to modification of cell walls remains to be established.

Experimental procedures

Plant material and growth conditions

The generation of the transgenic ethylene-insensitive birch (Betula pendula Roth) lines has been described previously (Vahala et al., 2003). All lines presented in Figure 1 display impaired leaf abscission and impaired induction of gene expression in response to exogenously applied ethylene (Vahala et al., 2003; data not shown). Of the ethylene-insensitive lines for which additional data is shown, the lines BPetr1-1-35, BPetr1-1-44 and BPetr1-1-55 are considered strongly transgenic whereas the lines BPetr1-1-24, BPetr1-1-54 and BPetr1-1-86 are viewed as weakly transgenic, which is in agreement with the copy number and expression of the transgene etr1-1 in these lines (Vahala et al., 2003; data not shown). Two birch clones, V5834 and JR1/4, as well as two transgenic lines in both backgrounds were used for all experiments. For clarity, only the data for V5834 are shown, unless background-related variation was observed.

The in vitro-micropropagated shoots were planted in a mixture of peat, sand and vermiculite (6:2:1, v/v/v), fertilized with 2.5 g l−1 (w/v) of Osmocote Exact Hi-Start (Scotts, Heerlen, The Netherlands; N:P:K 15:4:8) and grown for 7–10 weeks in the greenhouse prior to the experiments, unless otherwise indicated. Only well-rooted and vigorously growing individuals without any visible symptoms of infection were selected for the experiments. Growing conditions were identical for all individuals. Plant material was sampled during the middle of the photoperiod.

Growing conditions in the greenhouse were set as follows: natural light was supplemented with high-pressure sodium lamps (Lucalox LU400/HO/T/40, GE Lighting, Budapest, Hungary) to give a minimum of 200 μmol m−2 sec−1 photosynthetic photon flux (PPF) and a day length of 18 h, with a temperature of 18°C and a relative humidity of 70%. During SD experiments, the photoperiod was shortened to 12 h using curtains. In parallel to each SD experiment, plants of the same age were grown under LD as controls and analysed in the same way as the SD-exposed plants. All conditions in the greenhouse compartments were regulated and monitored by a computer-controlled system (Priva Intégro, Canadian Climatrol Systems Ltd, Surrey, BC, Canada). The experiments that were performed in growth cabinets (Weiss Bio1300; Weiss Umweltstechnik, Reiskirchen-Lindenstruth, Germany) were supported by light from fluorescent tubes (Osram 36W/31-830 and 36W/77; Osram, Munich, Germany) at 200–250 μmol m−2 sec−1 PPF, with other conditions being equal to those in the greenhouse. The plants were allowed to acclimatize to the conditions in the cabinets for 1 week before the beginning of the experiments. The potential effects of microclimate inside greenhouse compartments and differences among separate growth cabinets were avoided by regularly varying the position of the plants. For some experiments, the plants were exposed to natural outdoor light and temperature conditions, where the temperature was recorded with a computer-controlled system (Priva Intégro).

Chlorophyll content of the leaves

Greenhouse-grown, 12-week-old plants were exposed to natural outdoor inductive conditions (60°N, 24°E, Helsinki, Finland) to follow autumn leaf senescence. At given time points, the two youngest, fully expanded leaves were collected, frozen in liquid nitrogen, and stored at −80°C until analysis. In addition, two leaves were taken from the lower part of the shoot at about one-third of the height of the plant and processed similarly.

Total chlorophyll content of the senescing leaves was determined according to the method described by Richardson et al. (2002), with minor modifications. Briefly, 1 ml of DMSO (Fluka, Buchs, Switzerland) was added to approximately 60 mg of crushed leaf material, and chlorophyll was extracted at 65°C for 1 h. Absorbance of the extract was measured at 645 and 663 nm using a spectrophotometer (Agilent 8453; Agilent Technologies, Waldbronn, Germany), and the total chlorophyll concentration was calculated using Arnon's (1949) equation: total chlorophyll (g l−1) = 0.0202 A645 + 0.00802 A663. Leaf chlorophyll content was calculated (μg mg−1 FW) from the chlorophyll concentration of the extract.

Abscisic acid content of buds

All visible leaves surrounding the apex were dissected with a scalpel and the stem was cut approximately 3 mm below the shoot apex. The samples were frozen in liquid nitrogen, weighed, and stored at −80°C until analysis. Frozen samples were extracted in 500 μl 50 mm sodium phosphate buffer, pH 7.0, with 0.02% w/v Na-diethyldithiocarbamate as an antioxidant and [2H7]ABA as an internal standard in 1.5 ml Eppendorf tubes. Extraction was performed using an MM 301 Vibration Mill (Retsch GmbH & Co. KG, Haan, Germany) at a frequency of 30 Hz for 3 min after addition of 3 mm tungsten carbide beads (Retsch GmbH & Co. KG) to each tube to increase the extraction efficiency. After centrifugation in an Eppendorf centrifuge for 10 min at 16 000 g, the supernatant was adjusted to pH 3.0 with 1 m HCl and applied to a pre-equilibrated 100 mg C8-EC ISOLUTE cartridge (Sorbent AB, V. Frölunda, Sweden). The column was washed with 2 ml of 1% aqueous acetic acid, and then ABA was eluted with 2 ml 80% MeOH.

After evaporation to dryness, the samples were methylated with ethereal diazomethane, dried and dissolved in heptane. Samples were injected in the splitless mode into an HP 6810 gas chromatograph (Agilent, Palo Alto, CA, USA) fitted with a fused silica glass capillary column (15 m long, 0.25 mm internal diameter) with a chemically bonded 0.25 μm DB-5MS stationary phase (J&W Scientific, Folsom, CA, USA). The column effluent was introduced into the ion source of a JMS-MStation mass spectrometer (JEOL, Tokyo, Japan). The acceleration voltage was 10 kV, and ions were generated with 70 eV at an emission current of 300 μA. For quantification, samples were analysed in selected ion-monitoring mode (SIM) (Moritz and Olsen, 1995) at a resolution of 10 000. For the ABA analyses, calibration curves were recorded from 10 to 200 pg ABA with 50 pg 2H7-ABA as an internal standard.

Dormancy assessment and abscisic acid responsiveness of buds

Dormancy was estimated in bud burst experiments with single node cuttings in water, and expressed as a percentage of burst buds (see Rinne et al., 2001). To test whether ABA could suppress the capacity for bud burst in dormant birch, plants were exposed for 3 months to ambient outdoors conditions during autumn (60°N, 24°E, Helsinki, Finland). During the exposure, the day length shortened from 14.5 to 6.5 h, while the average ambient temperature decreased from 15 to −3°C and varied between 20 and −7°C. At the end of the exposure, a single twig from the middle part of each plant was selected and cut into single node cuttings. Cuttings were placed in 50 μmol of ABA or water. The hydroponic cultures were incubated under growth-promoting conditions of a 16 h photoperiod at 23°C, and the solutions were changed twice a week. Buds were recorded as ‘bursting’ when the first true leaf was clearly visible.

RNA preparation, microarray and quantitative RT-PCR analysis

To study the gene expression patterns under SD and LD, plants were grown in controlled conditions as described above. Apices of branches of three or four plants were dissected approximately 3 mm below the shoot apex, frozen, and pooled in liquid nitrogen for each replicate sample. Total RNA was isolated from the apices as described previously (Chang et al., 1993b). For the microarray analysis, 15 μg of the RNA was reverse-transcribed into aminoallyl-labelled cDNA using SuperScript III (Invitrogen, Carlsbad, CA, USA) and MMLV-RT H(-) Point Mutant (Promega, Madison, WI, USA). Subsequent labelling of the cDNA, array hybridization and image analyses were performed as described by Broschéet al. (2005). To prepare the templates for quantitative RT-PCR, the RNA samples were first treated with DNase I (Amersham, Little Chalfont, UK), extracted with phenol:chloroform (1:1, v/v) and precipitated. A 3 μg aliquot of the DNase I-treated total RNA was reverse transcribed into cDNA using Superscript III (Invitrogen), and random decamers (RETROscript, Ambion, Cambridgeshire, UK) as primers.

Primers for quantitative RT-PCR were designed using Primer Express 2.0 software (Applied Biosystems, Foster City, CA, USA) to amplify products of 100 bp, and were as follows: for BpBXL1 (Genbank number DW986525) forward 5′-GGGCTGAGCAACGTGTTAGG and reverse 5′-AAGGTTATGTTCCCCCATTGGA, and for birch actin (Genbank number DW986527) forward 5′-TGGTCAAGGCTGGGTTTGC and reverse 5′-CTGACCCATCCCAACCATGA. PCR was performed using qPCR Mastermix Plus for SYBR green I (Eurogentec, Seraing, Belgium) with 5 pmol of each primer and 1/50 of the cDNA as template. The default program of ABI Prism 7000 (Applied Biosystems) was used for amplification and dissociation of the gene products. For each sample, three technical repeats were performed to gain an average value for the threshold cycle (Ct) of BpBXL1 and actin. To obtain the normalized ΔCt value, the average Ct value of actin was subtracted from the corresponding Ct value of BpBXL1. For the final calculations, the relative expression levels for each data point were calculated using the formula 2−ΔCt (User Bulletin #2, Applied Biosystems) and the means of three to four replicate samples are shown.

Statistical analysis

Student's t-test and one-way analysis of variance (anova) were used to detect significant differences between wild-type and the corresponding transgenic lines as indicated, and Dunnett's two-sided test was used for post-hoc comparisons. The data were checked for normality and heterogeneity of variances. Chlorophyll contents were log10(X + 1)-transformed to meet the assumptions of anova. Analyses were performed with SPSS version 12.0.1 or Sigma Stat for Windows version 3.0.1 (SPSS Inc., Chicago, IL, USA) software packages.


We sincerely thank Dr Mikael Brosché (University of Helsinki, Finland) for guidance in microarray techniques and Dr Jorma Vahala (University of Helsinki, Finland) for help in sequence data mining. Mirva Tirkkonen, Airi Lamminmäki, Pinja Jaspers and Marja Tomell (University of Helsinki, Finland) are acknowledged for technical assistance, and IngaBritt Carlsson (Umeå Plant Science Centre, Sweden) for assistance in ABA measurements. Mika Korva and Leena Laakso (University of Helsinki, Finland) are acknowledged for plant care, and Sakari Silvennoinen and Dr Matti Rousi (Finnish Forest Research Institute, Punkaharju Research Station, Finland) for advice on birch crossing techniques. We also thank Drs Annikki Welling (University of Helsinki, Finland) and Rishikesh Bhalerao (Umeå Plant Science Centre, Sweden) for helpful remarks on the manuscript. This work was financially supported by the Academy of Finland (Finnish Centre of Excellence programme 2000–2005), the Finnish Graduate School in Forest Sciences, the Norwegian Research Council (P.R.; no. 155041/140), the Swedish Research Council and EU-RTN (T.M.; project HPRN-CT-2000-00090).