Phloem transdifferentiation from immature xylem cells during bark regeneration after girdling in Eucommia ulmoides Oliv

Abstract

Eucommia ulmoides Oliv. (Eucommiaceae), a traditional Chinese medicinal plant, was used to study phloem cell differentiation during bark regeneration after girdling on a large scale. Here it is shown that new sieve elements (SEs) appeared in the regenerated tissues before the formation of wound cambium during bark regeneration after girdling, and they could originate from the transdifferentiation of immature/differentiating axial xylem cells left on the trunk. Assays of water-cultured twigs revealed that girdling blocked sucrose transport until the formation of new SEs, and the regeneration of the functional SEs was not dependent on the substance provided by the axis system outside the girdled areas, while exogenous indole acetic acid (IAA) applied on the wound surface accelerated SE differentiation. The experiments suggest that the immature xylem cells can transdifferentiate into phloem cells under certain conditions, which means xylem and phloem cells might share some identical features at the beginning of their differentiation pathway. This study also showed that the bark regeneration system could provide a novel method for studying xylem and phloem cell differentiation.

Bark regeneration, Eucommia ulmoides Oliv., immature xylem cells, sieve elements, transdifferentiation

Introduction

Tissue regeneration after wounding has been studied not only in mice hair follicles (Ito et al., 2007) and human blood vessels (Ludatscher, 1981), but also in plant bark (Brown and Sax, 1962; Grunwald et al., 2002; Noel, 1970; Stobbe et al., 2002). However, girdling on a large area was thought too destructive to woody plants to pursue; only a strip or a piece of bark was removed from trees in previous reports. Subsequently, newly formed periderm and the cambium were developed from the callus on the inner side of the bark or the surface of the xylem, and it was generally considered that the new phloem was derived from the wound cambium in this case (Grunwald et al., 2002; Stobbe et al., 2002). However, it would be a different case for bark girdling on a large scale (as much as 1–2 m) in trees. Previous girdling experiments in Eucommia ulmoides Oliv. and many other species of plants showed that, when girdled at a suitable stage, trees will grow new bark within 1–2 months and continue to grow normally when the girdled part is wrapped in plastic sheeting (Li et al., 1982; Lu et al., 1987; Li and Cui, 1988). Could trees really suffer from a month of starvation without transportation of nutriments by the phloem? How and where the nutriments are transported during the process of bark regeneration are still open questions.

Significant progress has been made in our understanding of vascular differentiation and pattern formation based upon the studies of the Arabidopsis and Zinnia cell culture system (Ye, 2002; Fukuda, 2004; Carlsbecker and Helariutta, 2005). It was suggested that the control of differentiation of primary phloem and xylem is highly integrated and the developmental framework that regulates the decision of the fate of primary xylem or phloem has been established (Carlsbecker and Helariutta, 2005). However, little is known about the decision of the fate of secondary phloem and secondary xylem cells. Both secondary xylem and secondary phloem are products of the vascular cambium. Conceptually, the cambial initial cells remain meristematic, and produce phloem mother cells destined to become phloem and xylem mother cells destined to become xylem (Larson, 1994). However, there is no evidence for their determination among initials, phloem mother cells, and xylem mother cells; therefore, the term cambial cells is used to denote both the initials and the mother cells (Mellerowicz, 2001; Savidge, 2001). Recently, transcript profile research identified several marker genes for early stages of cambial cell differentiation in poplar (Schrader et al., 2004), providing the basic molecular information on secondary xylem and phloem differentiation.

Eucommia ulmoides is a multipurpose tree indigenous to China, especially widely used as Chinese Medicine. Thus, it is frequently subjected to radical peeling of its bark, which contains the bioactive substances used for treating some human diseases. The optimal conditions for recovery of bark after girdling in E. ulmoides had been established by Li et al. (1981). The anatomical changes that occurred during bark reconstitution have already been described (Li et al., 1981; Li and Cui, 1988) and the change in endogenous phytohormones during the process was also studied recently (Mwange et al., 2003). However, no study has paid attention to the phloem regeneration on the wounded area, although the phloem is responsible for the distribution of photoassimilates and nutrients among various organs of higher plants (van Bel, 1993; van Bel et al., 2002). Here it is shown that newly formed phloem cells (including sieve tube members and companion cells) could result from the transdifferentiation of immature/differentiating axial xylem cells left on the trunk during bark regeneration in E. ulmoides after girdling. Furthermore, using assays of water-cultured twigs, it was found that the regeneration of wound phloem was not dependent on the substance provided by the axial system outside the girdled areas. The experiments suggest that immature xylem cells can transdifferentiate into phloem cells under certain conditions. The results suggest that xylem and phloem cells may share some identical features during their differentiation pathway at the beginning, and this bark regeneration system could provide a novel method for studying xylem and phloem cell differentiation.

Materials and methods

Girdling and sampling

Eight- to 10-year old E. ulmoides trees growing at Tongxian Forestry Station in northern China (39°39′ N, 116°46′ E) were employed in the study. Fifty trees were girdled in July 2004–2006 as described by Li et al. (1981). Samples were collected at 0, 2, 4, 6, 9, 10, 12, 15, 18, 21, and 28 days after girdling (DAG). Blocks (2×4 cm²) were taken following the method of Uggla and Sundberg (2002) and were delimited by the exposed surface of progressively regenerated bark and mature xylem. Small blocks (1×2 mm²) were taken from the larger blocks, and fixed in 3% glutaraldehyde and 4% paraformaldehyde in 0.01 M phosphate-buffered saline (PBS; pH 7.2) for microscopic observation. Small strips (3×20 mm²) were fixed in 70% ethanol for hand cutting. Tissues scraped from the surfaces of the trunks were immediately frozen in liquid N₂ and stored at –70 °C.

Histological sections and immunochemistry

To confirm sieve elements (SEs), hand-cut sections were stained with 0.005% aniline blue in 0.15 M K₃PO₄ (pH 8.2), then observed under UV light by fluorescence light microscopy (Zeiss Axioskop2 plus, Germany). Histological sections of 4 μm thickness were cut from small blocks embedded in Spurr's resin (SPI, USA) on a microtome (Leitz 1512, Germany), and stained with toluidine blue O (TBO).

To observe proliferation, girdled areas were treated with 2 mg ml⁻¹ IdU (Sigma) in 0.01 M PBST (PBS containing 0.05% Tween-20, pH 7.2) for 2 d. Sampling tissues were fixed in 4% paraformaldehyde in 0.01 M PBS for 16 h, dehydrated in an ethanol series, infiltrated with a ethanol/L.R.White mixture, and embedded in L.R.White (Sigma). Longitudinal 5 μm thick sections were cut on a microtome (Leitz 1512, Germany), and immersed in 0.01 M PBST for 10 min. The immunochemistry procedure was carried out according to Yu et al. (1999) with some modifications. Sections were blocked with 5% normal horse serum in PBST for 1 h, then incubated for 2 h floating in monoclonal antibody IU-4 (Caltag, CA, USA) diluted 1:50 in PBST containing 1.5% normal horse serum at 37 °C. The sections were rinsed five times with PBS, and treated for 1 h with alkaline phosphatase (AP)-conjugated horse anti-mouse IgG (Vector, USA) diluted 1:200 in PBS. After three rinses, the AP signal was detected using a BCIP/NBT detection kit (Roche).

Scanning electron microscopy

Samples were fixed in 3% glutaraldehyde and 4% paraformaldehyde in 0.01 M PBS overnight and treated with 10% NaClO for 10 min before dehydration in a graded ethanol series. Then, the samples were transferred to isoamyl acetate and dried in a critical-point drier with liquid CO₂. The dried specimens were coated with gold in an ion sputter coater and observed with a scanning electron microscope (1910 FEM, Amray, USA).

Twig girdling assays

Eighty twigs (2 cm in diameter, 50 cm in length) with 35 cm girdling in length at the mid-portion were divided into four groups, with application of lanolin alone, lanolin with 1 mg g⁻¹ indole acetic acid (IAA), lanolin with 1 mg g⁻¹ IAA and 25 mM sucrose, or lanolin with 25 mM sucrose, respectively, on the top of the branches, then covered with transparent plastic sheets and cultured in water. Samples were collected from the girdled areas at 6, 8, 10, and 12 DAG. Hand-cut sections were used for SE detection with aniline blue.

Thirty twigs were entirely girdled and separated into two groups. One group (d-IAA) were dipped in 80 μM IAA solution in 20 mM MES buffer (pH 5.0) for 15 min. The other group was dipped in MES buffer as a control. All the twigs were covered with transparent plastic sheets and cultured in water. Samples were collected from the girdled areas at 6, 9, and 12 DAG. SEs were detected with aniline blue as above.

[¹⁴C]Sucrose transport of E. ulmoides twigs

Twenty-four mid-portion girdled twigs were cultured in water. Droplets of 37 kBq [¹⁴C]sucrose in 10 mM MES (containing 10 mM sucrose, pH 5.5) were introduced on the inner bark. After 2 d, segments (2 cm in length) were sampled at the positions Bb (bark), Bx (xylem), Cx (girdled xylem), and Dx (girdled xylem) from each twig (Fig. 4a). The central parts were used for autoradiography and the others for counting the levels of radioactivity. After drying, 50 mg of sample powder was detected by CPM (counts per minute) using a low background α and β measuring instrument (BH1216, China).

Gene expression analysis

Expression of the genes EuABP1, EuIAA2, EuIAR3, and EuSUT2 was determined by reverse transcription-PCR (RT-PCR) using EuUbiquitin E2 as the internal standard gene. Scraped samples at each harvesting time were pulverized in liquid nitrogen using a mortar and pestle. Total RNA was extracted with TRIzol Reagent (Invitrogen) according to the manufacturer's protocol. A 5 μg aliquot of total RNA was reverse-transcribed into cDNA with a RevertAid H Minus First Strand cDNA Synthesis Kit (Fermentas), and subsequently diluted for RT-PCR. The specific sense primer 5′-GTGGTACTAAAGGGCAGTG-3′ and antisense primer 5′-CTCGTATATAAACACTTTGAC-3′ for EuABP1; sense primer 5′-CACTCGCATCCACTTCAA-3′ and antisense primer 5′-TCACCGTCTTTATCCTCGT-3′ for EuIAA2; sense primer 5′-GCGGAGCAAAGAAATGAT-3′ and antisense primer 5′-ACCCCGAGGAAGAAGAAGT-3′ for EuIAR3; sense primer 5′-AGATTCTGCTCCTTTGTTG-3′ and antisense primer 5′-CTTCCCATCCAATCCGTATC-3′ for EuSUT2; and sense primer 5′-GATATGTTTCATTGGCAAGC-3′ and antisense primer 5′-GGGCTCCACTGTTCTTTCA-3′ for EuUE2 were used. Cycle numbers for standard PCR amplification were chosen to stay below saturation (25 cycles). The PCR bands were separated on 1.5% agarose gels and visualized by ethidium bromide staining. The GeneBank accession numbers of EuABP1, EuIAA2, EuIAR3, EuSUT2, and EuUE2 are AY509875, EU170472, EU177878, AY946204, and EU170471, respectively.

Results

Phloem formation during bark regeneration after girdling in E. ulmoides

During studies on bark regeneration after girdling in E. ulmoides trees, it was noticed that the bark regeneration process was complete within 1 month by protecting the wounded trunk with transparent plastic sheets (Fig. 1a–c).

Fig. 1.

Bark regenerates after girdling in E. ulmoides trees. Bark regenerated after girdling for 1 m in living trees. (a) The trunk surface is smooth after girdlings; (b) the trunk surface become flocky at 12 DAG; (c) the trunk surface changes to hard and dry at 30 DAG.

To investigate the timing of wound phloem formation, time-course experiments were performed on the girdled region of E. ulmoides during bark regeneration. Over the 3 years of the study, SEs formed in more than half of the trees at 10 DAG, and in >90% of the trees at 12 DAG (Fig. 2).

Fig. 2.

The newly formed SEs appear during the time-course of detection after girdling in E. ulmoides trees during 2004–2006. The percentage of girdled areas of 4–5 trees forming new SEs 6, 8, 9, 10, 12, and 15 DAG. At 10 DAG, there are 60, 50, and 80% trees of total trees that formed new SEs in 2004, 2005, and 2006, respectively; at 12 DAG, there are 80, 100, and 100% trees of total trees that formed new SEs in 2004, 2005, and 2006, respectively.

Both fluorescence light microscopy and scanning electron microscopy revealed newly formed SEs, with typical sieve plates on the end walls (Fig. 3a, d) and sieve areas on the other walls (Fig. 3b, e). In some cases, companion cells accompanied the SEs (Fig. 3c). The SEs were present 1–3 cell layers under the isodiametric cells, arranged between the xylary ray cells (Fig. 4a, b), and connected longitudinally to each other (Fig. 4c, d). Therefore, the SEs typified sieve tube members and it was confirmed that wound phloem was formed at ∼12 DAG.

Fig. 3.

The newly formed SEs have structures typical of sieve tube members after girdling in E. ulmoides trees. (a–c) Hand-cut sections are stained with aniline blue 12 DAG; (a) transverse section of sieve plates (arrowhead); (b) radial longitudinal section of sieve areas (arrowhead); (c) radial longitudinal section showing companion cells beside SEs; the arrowhead indicates companion cells. (d, e) Scanning electron microscopy of sieve plates and sieve areas 12 DAG. Cal, callus; Xy, xylem. (a–c) Bars=50 μm; (d, e) bars=10 μm.

Fig. 4.

The newly formed SEs are localized under the callus and arranged between xylem ray cells after girdling in E. ulmoides trees. Hand-cut sections 15 DAG, co-stained with aniline blue and toluidine blue O (TBO). (a, b) The newly formed SEs are arranged between the xylem ray cells on the transverse sections. (a, b) The same section under UV light and white light. (c, d) The newly formed SEs are present 1–3 cell layers under the isodiametric cells and connected longitudinally to each other on the radial longitudinal sections. (c, d) The same section under UV light and white light. Cal, callus; Xy, xylem. Arrowheads show SEs. Bars=100 μm.

Sieve element transdifferentiation from immature xylem cells

To identify the phloem progenitor cell type, tissue anatomical changes in bark regeneration were observed. In the growth season, the vascular cambium produces secondary phloem outside and secondary xylem inside (Fig. 5a). Along with the bark girdling, most cambial cells have been removed with the bark (Fig. 5b), and the cells remaining on the trunk surface were largely immature xylem cells with a low frequency of split cambial cells (Figs 5c, 6a, 7a).

Fig. 5.

The cambium has been removed with bark by girdling in E. ulmoides trees. Transverse sections show the removal of cambium within bark by girdling 0 DAG. (a) The tissues before bark girdling including phloem, cambium, immature xylem, and xylem; (b) the cambium and a few immature xylem cells have been removed with the bark by girdling; (c) immature xylem cells and split cambial cells on the trunk surface after girdling. Plastic sections stained with TBO. Ph, phloem; C, cambium; Xy, xylem. The arrowhead shows a split cambial cell. Bars=100 μm.

Fig. 6.

The immature xylem ray cells divide and form callus during the first week after girdling in E. ulmoides trees. Tangential surfaces by SEM show cell divisions of immature xylem ray cells 1, 3, and 6 DAG. (a) Many immature xylem ray cells broken on the trunk surface after girdling; (b) some immature xylem ray cells expand outwards at 1 DAG; (c) the proliferation of ray cells form cell groups on the surface at 3 DAG; (d) callus forms from the proliferation of ray cells on the tangential surface 6 DAG. Bars=100 μm.

Fig. 7.

The immature xylem axial cells divide during the first 2 DAG in E. ulmoides trees. Radial longitudinal sections show cell divisions of immature xylem axial cells 2 DAG. (a) Immature xylem cells with a low frequency of split cambial cells are left on the trunk after girdling; (b) the nucleoli with DNA replication are labelled by IdU; (c, d) division phases of immature xylem axial cells. Plastic sections stained with TBO. Arrowheads show dividing cells. Bars=100 μm.

Immature xylem cells showed different proliferation patterns. Ray cells radiated, divided rapidly in irregular directions at 1–3 DAG (Fig. 6b, c), and spread diffusely to form a soft callus covering the wounded surface at 6 DAG (Fig. 6d). Callus cell division continued for >1 week. Simultaneously, axial immature xylem cells (comprised of differentiating tracheary elements and fibres) underwent periclinal and infrequent transverse division, most of which occurred during the first 2 DAG (Fig. 7b). Dividing cells were observed on the surface layer (Fig. 7c), but a high frequency of different division phases was present, even 8–10 layers under the wounded surface (Fig. 7d).

Immature xylem axial cells were observed undergoing only one or two cell divisions, resulting in the maintenance of an elongate shape in the cell axis. Interestingly, newly formed SEs and companion cells developed from these xylem cells rather than from callus at 12 DAG, and no cambium could be found in the regenerated tissues at that time (Fig. 8a, d). Furthermore, the appearance of newly formed SEs developed almost simultaneously on the top and base of the wounded trunks (data not shown). Two to 3 weeks after girdling, the wound cambium dedifferentiated from deeper immature axial xylem cells. At 15 DAG, discontinuous cambial cells appeared under newly formed SEs (Fig. 8b, e). There was intact cambium at 21 DAG (Fig. 8c, f). Thus, the newly formed SEs could not be derived from the wound cambium, but rather from the immature xylem cells. These immature xylem axial cells were likely to undergo transdifferentiation, directly ‘changing’ the cells into phloem mother cells. Subsequently, the cells differentiated into sieve tube members and companion cells.

Fig. 8.

The newly formed SEs appear earlier than wound cambium recovery during bark regeneration in E. ulmoides trees. Transverse sections (a–c) and radial longitudinal sections (d–f) show the already formed SEs and wound cambium forming 12, 15, and 21 DAG. (a, d) The new SEs already forms while no cambium is found 12 DAG; (b, e) discontinuous cambial cells appear under newly formed SEs 15 DAG; (c, f) intact cambium forms 21 DAG. Cal, callus; Xy, xylem. The arrowhead indicates SEs. Arrows show wound cambium. Bars=100 μm.

Twig girdling assays

To investigate the factors responsible for immature xylem cell transdifferentiation, a series of water culture assays was conducted. Twigs of E. ulmoides were girdled at the mid-portion, treated with IAA and/or sucrose (applied to the excised tops), and cultured in water for 12 d. The results demonstrated that all treated twigs formed new SEs, even those treated with lanolin alone (Fig. 9c). Compared with the development of SEs in the trunks of trees, the SEs in the treated twigs appeared earlier, while the position of the SEs was the same (Fig. 9a, b). Furthermore, the girdling experiments were conducted in late September, the quiescence-1 phase of E. ulmoides (Mwange et al., 2003b, 2005), with IAA treatment. Entire girdled twigs (control) or IAA-treated entire girdled twigs (dipped in IAA solution after having been girdled, d-IAA) were cultured in water (Fig. 10a). The percentage of girdled areas forming new SEs at 6, 9, and 12 DAG showed that exogenous IAA treatment accelerates SE differentiation (Fig. 10b).

Fig. 9.

Eucommia ulmoides twigs undergoing mid-portion girdling in excised culture form SEs. Mid-portion girdled twigs were treated with lanolin alone; lanolin and 1 mg g⁻¹ IAA; lanolin and 1 mg g⁻¹ IAA plus 25 mM sucrose; or lanolin and 25 mM sucrose, then cultured in water. (a, b) The same hand-cut sections from lanolin-treated twigs (12 DAG). The sections co-stained with aniline blue and TBO show the same position of newly formed SEs to that in trunks of girdled trees. Cal, callus; Xy, xylem. Arrowheads show SEs. Bars=100 μm. (c) All treated groups of twigs show newly formed SEs from 8 DAG.

Fig. 10.

The Eucommia ulmoides twigs undergoing complete girdling in excised culture form SEs. (a) Completely girdled twigs (control) are cultured in water; (b) the percentage of girdled areas forming new SEs at 6, 9, and 12 DAG. The d-IAA group comprises twigs that were completely girdled, treated with IAA (dipped in IAA solution after girdling), and cultured in water (same as a), Hand-cut sections were stained for SEs with aniline blue.

[¹⁴C]Sucrose transport of E. ulmoides twigs

Using [¹⁴C]sucrose as a tracer, 24 water-cultured twig experiments were conducted and demonstrated temporal synchronization between sucrose transportation and SE formation (Fig. 11a). Radioactive signals were detected in the 10 cm (Cx) and 15 cm (Dx) girdled areas below the [¹⁴C]sucrose application site (A) with formation of SEs (Fig. 11b). Autoradiography revealed the pathway of [¹⁴C]sucrose transport, which was near the twig surface, spatially co-localized with the new SEs (Fig. 11c–e). These results indicated that the sucrose transport was blocked by bark girdling, while the newly formed SEs ensured transport during the recovery process, and the formation of SEs does not require the longitudinal transportation of sucrose and other substances during the process.

Fig. 11.

Isotope tracing of twigs with [¹⁴C]sucrose shows that newly formed SEs undergo sucrose transportation. (a) Schematic drawing of [¹⁴C]sucrose treatment and sampling positions of twigs; deep grey areas are reserved bark, grey areas girdled bark; A, [¹⁴C]sucrose treatment site, B, sampling for bark (Bp) and xylem (Bx), separately; C and D sampling for xylem (Cx, Cx). (b) [¹⁴C]Sucrose contents at three positions on 2, 6, 10, and 14 DAG; Cx-SEs and Dx-SEs are samples showing newly formed SEs; Cx-N and Dx-N are the samples lacking newly formed SEs. (d) Autoradiograph of the C area specimen (c). Bar=1 cm. (e) Hand-cut section from a part of c, stained with aniline blue. Bar=100 μm. Cal, callus; Xy, xylem. The arrowhead shows the SEs.

IAA- and sucrose-related gene expression during bark regeneration

The expression of several genes was detected during bark regeneration. These genes were involved in IAA binding (EuABP1, EuIAA2) or hydrolysis of bound IAA (EuIAR3), and sucrose transport (EuSUT2). RT-PCR assays revealed that all the gene transcripts were expressed in a similar manner (Fig. 12). At 4 DAG, they were weakly detectable and then they increased 9 DAG and 12 DAG. At 18 DAG, the expression of the four genes reached a peak, and neared the 0 DAG level at 28 DAG. The change in the expression of these genes was consistent with the tissue recovery during the process. At 4 DAG, most cells were dividing and showed few differentiating characteristics. However, 9 DAG and 12 DAG, SEs were forming, which need the participation of IAA and sucrose. The highest expression of genes involved IAA at 18 DAG, when cambium is regenerating, corresponding to the IAA concentration peak in cambium across the vascular tissue. The results indicated that IAA and sucrose may contribute to the differentiation of SEs during bark regeneration.

Fig. 12.

The changes in gene expression are similar during the period of SE formation after girdling in E. ulmoides trees. Gene expression detected by RT-PCR 0, 4, 9, 12, 18, and 28 DAG. These genes are involved in IAA binding (EuABP1, EuIAA2) or hydrolysis of bound IAA (EuIAR3), and sucrose transport (EuSUT2). EuUbiquitin E2 (EuUE2) was used as the internal standard gene.

Discussion

Sieve element transdifferentiation during bark regeneration after girdling

It was observed that the process of bark regeneration after girdling on a large scale was quite different from that following a partial stripping of the bark (Grunwald et al., 2002; Stobbe et al., 2002). During bark regeneration after girdling, the newly formed functional SEs could not be derived from the wound cambium, but from immature xylem cells. Some immature xylem cells were expected to undergo transdifferentiation, directly ‘changing’ the cells into phloem mother cells, and subsequently differentiated into sieve tube members and companion cells.

Previous studies revealed that cells can ‘change’ their tissue types under certain conditions. Mesophyll cells were transformed to tracheary elements in a suspended cell culture of Zinnia elegans (Fukuda and Komamine, 1980) and Arabidopsis (Oda et al., 2005). An APL mutation resulted in the formation of cells with xylem characteristics in locations characteristic of phloem cells (Bonke et al., 2003). However, there are few reports where cells were shown to ‘change’ into phloem cells. The present results showed that immature xylem cells transdifferentiated into phloem cells under certain conditions. Xylem and phloem cells might act similarly early in differentiation, where immature xylem cells could enter the phloem pathway under some conditions, such as wounding. Therefore, this bark regeneration system could provide a novel approach to study xylem and phloem cell differentiation.

Sieve element transdifferentiation affected by IAA and sucrose

The polarized auxin flow results in the formation of a vascular pattern (Fukuda, 2004). In the secondary vascular tissue, the radial distribution pattern of endogenous IAA across the cambial region exhibits a peak level in the cambial zone, steeply decreasing toward the mature xylem and phloem (Uggla et al., 1996, 1998; Mwange et al., 2005). In a previous study, there was an obvious change in the content and distribution of endogenous IAA within the girdled areas during the bark regeneration process (Mwange et al., 2003a). It was interesting to note that the regeneration of the functional SEs was not dependent on the substance provided by the axial system outside the girdled areas, and the exogenous IAA applied on the wound surface accelerates SE differentiation revealed by the assays of water-cultured twigs.

Sucrose is another important factor in the formation of phloem cells. Experiments of sucrose injections on silver birch trunk tissues showed that higher sucrose solution concentrations increased the number of phloem layers (Novitskaya et al., 2006). According to the vascular tissue gradient induction hypothesis, the position of phloem sucrose is integral in stress responses, such as wounding or oxidative stress (Ahn and Lee, 2003; Couée et al., 2006). Here it is suggested that the induction of phloem transdifferentiation was associated with the relative changes of some endogenous substances, such as IAA and sucrose, across their radial distribution. In addition, following girdling and the subsequent regeneration process, wound periderm, phloem, and cambium were formed from callus and immature xylem at different positions and at different times. This suggested that the position of immature xylem cells might control their differentiation pathways. The phloem induction factors could be tied to an auxin:sucrose concentration ratio at the position of phloem formation.

Furthermore, gene expression was analysed in E. ulmoides girdled areas during bark regeneration by RT-PCR, including genes involved in IAA binding (EuABP1, EuIAA2) and hydrolysis of bound IAA (EuIAR3), and sucrose transport (EuSUT2). The expression of these genes showed similar fluctuations from 0 to 28 DAG. In a previous study on protein expression patterns, a proteomics approach was employed during bark regeneration in Populus tomentosa (Du et al., 2006). The results of the work demonstrated that >200 proteins changed during the regeneration process. Together with other transcriptome analyses of secondary xylem and phloem in Arabidopsis (Zhao et al., 2005) and Populus (Schrader et al., 2004), it is suggested that bark regeneration after girdling is a complex process and involves regulation of a diverse array of genes.

We thank the Institute for Application of Atomic Energy, Chinese Academy of Agricultural Sciences, for use of its facilities. Special thanks are addressed to Professor Hong-Wei Guo (Peking University) for his critical reading of the manuscript and suggestions. This study was supported by the National Natural Science Foundation of China (30530620; 30670120).