A trans- zeatin riboside in root xylem sap negatively regulates adventitious root formation on cucumber hypocotyls

Shoot cultures of cucumber were used to analyse the roles of root-derived substances in adventitious root formation on hypocotyl tissues. Xylem sap collected from the roots of squash had a strong inhibitory effect on the formation of hypocotyl adventitious roots. Double-solvent extraction followed by fractionation with both normal and reverse phase column chromatographies and analysis by liquid chromato-graphy/tandem mass spectrometry identiﬁed trans-zeatin riboside (ZR) as the primary suppressor of adventitious root formation. ZR was the predominant cytokinin present in the xylem sap, occurring at a concentration of 2 3 10 –8 M. Application of ZR at concentrations from 3.16 3 10 –9 M effected inhibition of adventitious root formation. These results suggest that ZR transported from roots via xylem sap may act as an endogenous suppressor of hypocotyl adventitious root formation in planta.


Introduction
The formation of adventitious roots from the stems or hypocotyls of most plants is stimulated by the removal of root tissue, and appears to be controlled by various factors transported from both the roots and shoots. It is well established, for example, that basipetally transported auxin, which accumulates at the basal end of hypocotyl and stem cuttings in tomato and wild cherry, plays a stimulatory role during adventitious root formation (Maldiney et al., 1986;Label et al., 1989). Similarly, removal of the shoot apex decreases both the level of endogenous auxin in the basal part of a cutting and the number of adventitious roots in pea (Nordstro Èm and Eliasson, 1991), and auxin-overproducing mutants of Arabidopsis produce numerous adventitious roots from the hypocotyls of intact plants (Boerjan et al., 1995). Consequently, endogenous auxins that are produced in the shoot apex, transported basipetally and accumulate in the cut end of a stem are believed to induce adventitious root formation.
Ethylene is synthesized as a stress response to wounding in most plants, and promotes adventitious root development in cut tissues. Inhibitors of ethylene biosynthesis and perception reduce the number of adventitious roots on cuttings of sun¯ower and mung bean (Robbins et al., 1983(Robbins et al., , 1985, while ethylene-insensitive tomato and petunia plants produce fewer adventitious roots than wild-type cuttings (Clark et al., 1999).
Although it has been reported that root extracts have an inhibitory effect on adventitious root formation in bean (Libbert, 1956), and that removal of the root stimulates adventitious root formation (Soekarjo, 1965), the endogenous root-derived suppressor of adventitious root formation has not yet been identi®ed.
When exogenous chemical applications, rather than endogenous stimulators and suppressors, are considered, stimulatory effects on adventitious root formation have been reported (Haissig and Davis, 1994), especially by plant hormones such as the auxins (Went, 1939;Basu et al., 1969;Shibaoka, 1971) and ethylene (Zimmerman et al., 1933;Mensualisodi et al., 1995). There are fewer reports of adventitious root inhibition as a consequence of exogenous chemical treatment (Fellenberg, 1966;Mitsuhashi et al., 1969), but these too include effects of plant hormones such as gibberellin (Brian et al., 1960;Fabijan et al., 1981) and cytokinin (Humphries, 1960;Bollmark and Eliasson, 1986). Both the studies of endogenous and exogenous factors leading to the control of adventitious root development suggest that plant hormones probably play a critical role.
It is reasonable to suppose that a root-derived suppressor would be transported to the hypocotyl through the xylem via the transpiration stream. Squash plants, from which a large amount of xylem sap can be obtained (300±3000 ml plant ±1 in 2 d), were used as a source of xylem sap. Squash plants are widely used as a rootstock for commercial cucumber production, as they ensure protection from soilborne disease. Cucumbers graft well to squash rootstock, and produce many adventitious roots from their own hypocotyl tissues as well, providing an excellent study system.
When hypocotyls of cucumber seedlings were cultured in squash root xylem sap fractions, inhibition of adventitious root formation was observed (Satoh et al., 1998). A signi®cant concentration of a novel amino acid, methoxybenzylglutamine [N 5 -(4-methoxyphenyl)methyl-L-glutamine], was identi®ed in the inhibitory fraction of the root xylem sap (Satoh et al., 1998;Inouye et al., 1999). Chemically synthesized methoxybenzylglutamine inhibited the formation of adventitious roots at a concentration of 5Q10 ±3 M, but not 5Q10 ±4 M (Satoh et al., 1998), but the endogenous concentration of methoxybenzylglutamine was estimated to be 10 ±6 M in root xylem sap. The endogenous concentration is far too low to be effective in above-ground organs, even if concentrated in the shoots via transpiration by several hundred times. These results suggest that methoxybenzylglutamine is a poor candidate for being the sole suppressor of adventitious root formation, and that another suppressor is likely to be present in root xylem sap.
In the search to identify the endogenous suppressors of the formation of adventitious roots, double-solvent extraction was used, followed by normal and reverse phase column chromatographies and analysis by liquid chromatography/tandem mass spectrometry (LC/MS/MS). These techniques revealed a trans-zeatin riboside (ZR), one of the cytokinins, which acts as an inhibitory factor in root xylem sap. The negative regulation of adventitious root formation by ZR transported from roots to shoots in planta will be discussed.

Plant materials and chemicals
Seeds of interspeci®c hybrids of squash Cucurbita maxima DuchesneQC. moschata Duchesne cv. Shintosa-ichigou, commonly used as the commercial rootstock for cucumber, and seeds of cucumber, Cucumis sativus cv. Shimoshirazu-jibai were obtained from the Sakata Seed Co. (Kanagawa, Japan). For the identi®cation and quanti®cation of cytokinins using LC/MS/MS, ZR, cis-ZR, Z, cis-Z, and [ 2 H]Z were purchased from Apex Organics (Devon, UK). For the bioassay on the formation of adventitious roots, ZR, Z and BA were purchased from Sigma (St Louis).
Collection of root xylem sap Squash plants, cv. Shintosa-ichigou, were grown in the ®eld for 2±3 months (May to August), and the stems were cut off 15±30 cm above the soil level. After the ®rst drops of exudates were discarded, the cut surface was washed with distilled water; further exudates were collected in¯asks over ice and stored at ±30°C as root xylem sap.
Fractionation of the inhibitory factor in root xylem sap Root xylem sap (5.0 l) was ®ltered under reduced pressure with a glass ®lter (110 mm ID; Whatman, Maidstone, UK). The ®ltrate was subjected to reverse phase chromatography on a silica gel ODS-Q3 open column (80 ml volume; Wako, Osaka, Japan) ( Fig. 1). The bound substances were eluted sequentially with 200 ml of 20%, 50% and 100% (v/v) acetonitrile and the samples were evaporated to completely remove acetonitrile. Eluate (150 ml) of the 20% acetonitrile sample with inhibitory activity was partitioned with 150 ml of butanol (Fig. 1). This step was repeated three times and the results combined. The butanol and aqueous phases were separately evaporated at 40°C, and solubilized in 5 ml of water. The solution was centrifuged at 3000 g for 10 min. The supernatant was subjected to reverse phase chromatography on an ODS-80 Ts column (4.6 mm IDQ15 cm; Tosoh, Tokyo, Japan) in an HPLC system (Tosoh), and eluted with a linear gradient of acetonitrile (from 0% to 60% (v/v) in 30 min) at a¯ow rate of 0.5 ml min ±1 with monitoring absorbance at 280 nm (Fig. 1). The pooled fractions were evaporated at 40°C, to remove acetonitrile, and used for the bioassay.
The eluate from 35±39 min with inhibitory activity was dissolved in 80% (v/v) acetonitrile and applied to normal phase chromatography on a Spherisorb W silica column (4.6 mm I.D.Q25 cm; Waters, Tokyo) in an HPLC system, with a monitoring absorbance of 280 nm (Fig. 1). The mobile phase was isocratic and consisted of 80% (v/v) acetonitrile at a¯ow rate of 0.5 ml min ±1 . The pooled fractions were used for the bioassay following evaporation. The inhibitory fraction at the retention time from 10±11 min was again subjected to reverse phase chromatography on the ODS-80Ts column in the HPLC system (Fig. 1). The mobile phase was isocratic and consisted of 0.1% (v/v) acetic acid solution mixed with 10% acetonitrile (v/v) at a¯ow rate of 0.5 ml min ±1 . The inhibitory activity at the retention time from 18±22 min was recovered with a single UV-peak on HPLC analysis.
Quanti®cation of cytokinins in root xylem sap was performed by the LC/MS/MS system as follows. Each root xylem sap sample (0.5 l) was lyophilized and extracted with 80% (v/v) methanol, to which [ 2 H]Z was added as an internal standard. The methanol extract was applied to an OasisÔ HLB extraction cartridge (Waters, Tokyo) and eluted with 40% (v/v) methanol containing 0.1% acetic acid. The eluate was evaporated, the resulting residue was redissolved into an aliquot of methanol and applied to a reverse phase CapselPak MG 18 column under the same conditions as described above for the identi®cation of the inhibitory factor. Cytokinins were quanti®ed using the selected reaction monitoring mode (ZR and cis-ZR, m/ z=352/220; Z and cis-Z, m/z=220/136). The internal concentration of each cytokinin was calculated using the recovery value of the corresponding internal standard.

Induction of the formation of adventitious roots
Cucumber seedlings cv. Shimoshirazu-jibai were grown in an arti®cial soil (Kurehakagaku, Tokyo, Japan) for 6 d under whitē uorescent light (15 mmol m ±2 s ±1 ) with a 16 h photoperiod at 28°C. The hypocotyl of each seedling was cut off 4 cm below the cotyledon using a razor blade; seedlings were cultured with 1 ml of the solution to be tested with 1/20 strength Murashige and Skoog's medium (MS medium; Murashige and Skoog, 1962) without sucrose in plastic microcentrifuge tubes (2 ml; Sorenson, Salt Lake City, UT, USA) under white¯uorescent light (15 mmol m ±2 s ±1 ). After 5 d in culture, the adventitious roots that had emerged from epidermal cells of the hypocotyl were counted using a loupe. From ®ve to nine plants were used for each bioassay and all experiments were repeated at least twice.

Results
Fractionation of the inhibitory factor in root xylem sap In a previous study, inhibitory activity in squash root xylem sap was observed only in the 20% acetonitrile eluate on an ODS-80Ts column of the butanol phase fraction (Satoh et al., 1998). Therefore, the inhibitory fraction was prepared using double-solvent extraction coupled with reverse phase chromatography; the butanol phase partitioned from the 20% acetonitrile eluate (using open column chromatography on a silica gel ODS-Q3) was subjected to an ODS-80Ts column in the HPLC system (Fig. 1). Inhibition of the formation of adventitious roots on hypocotyl was observed in fractions No. 4 (35±39 min) and No. 11 (51±57 min) on an ODS-80Ts column by reverse phase chromatography ( Fig. 2A, B). The inhibitory activity of fraction No. 4 was signi®cantly stronger than that of fraction No.11, indicating that fraction No. 4 included the main active compounds in xylem sap. Therefore, fraction No. 4 was further subjected to normal phase chromatography on a Spherisorb W silica column in the HPLC system (Fig. 1). The inhibitory activity of fraction No. 7 (10±11 min) was similarly detected (Fig. 3A,  B).
Identi®cation and quanti®cation of ZR as an inhibitory factor To identify the substance in the inhibitory fraction of root xylem sap, mass spectrometry (MS) and tandem mass spectrometry (MS/MS) was performed. Fraction No. 7 (10±11 min) from a Spherisorb W silica column, which showed the inhibitory activity (Fig. 3A, B) was applied to a reverse phase CapselPak MG 18 column and analysed by MS. The molecular mass of the major component of the fraction was found to be 351, indicating a [M+H + ] ion at m/z=352. An ion peak at m/z=220 was also detected as a minor component (data not shown). After the fragmentation of the peak at m/z=352 by MS/MS, this peak produced the component ion at m/z=136. These results and the standard m/z of ZR (352/220) and Z (220/136) indicated that zeatin riboside was a major component of this fraction.
This fraction was further analysed by liquid chromatography/tandem mass spectrometry (LC/MS/MS). In the total ion chromatograms using normal mode, two major peaks were detected at the retention time (around 8 and 9 min, Fig. 4, upper pro®le of total). In the ion chromatograms using selected reaction monitoring mode for ZR (m/ z=352/220) and Z (m/z=220/136), two major peaks were detected at the same retention time as two peaks in the total ion chromatograms (Fig. 4, upper pro®les of ZR and Z). The retention times of these major peaks were identical to those of authentic trans-ZR and cis-ZR (Fig. 4, lower pro®les). In contrast, ion peaks of iPMP, ZMP, ZOG, and ZOGR were not detected in the ion chromatograms using the selected reaction monitoring mode for each cytokinin (data not shown).
To determine which peak in the chromatogram on LC/ MS/MS had the inhibitory activity, the fraction from a Spherisorb W silica column was subjected to reverse phase chromatography on an ODS-80Ts column of an HPLC system under the same conditions as reverse phase chromatography on the CapselPak MG 18 column (Fig. 1), except for the¯ow rate and the concentration of acetonitrile in the mobile phase (see Materials and methods); the pooled fractions were used for the bioassay. Inhibitory activity was detected in fraction No. 2 (18±22 min) (Fig. 5). This peak was identical with that of trans-ZR in the HPLC chromatograph on LC/MS/MS. Weak inhibition was also observed in fraction No. 4 (24±40 min), but the activity was not signi®cant (Fig. 5).
Inhibitory effect of cytokinins on the formation of adventitious roots To con®rm that ZR had an inhibitory effect on the formation of adventitious roots, cytokinins (ZR, Z, BA) were tested by bioassay. All cytokinins clearly showed concentration-dependent inhibitory activities for the adventitious root formation on the cucumber hypocotyls (Fig. 6). ZR and Z had strong inhibitory activities at concentrations between 3.16Q10 ±9 M and 10 ±5 M (Fig. 6), but BA showed weaker inhibitory activity at 3.16Q10 ±9 M (Fig. 6).

Quanti®cation of cytokinins in root xylem sap
To evaluate whether the concentration of endogenous ZR in root xylem sap is suf®cient to inhibit adventitious root formation, relative quantities of endogenous cytokinins were determined. LC/MS/MS analysis showed that the concentration of endogenous ZR in root xylem sap was estimated to be 2.0Q10 ±8 M (SD=0.1Q10 ±8 M) ( Table 1). The concentration of cis-ZR was also estimated to be 6.6Q10 ±9 M (SD=0.5Q10 ±9 M), approximately three times lower than trans-ZR (Table 1). The concentrations of trans-Z and cis-Z were less than 1/10th of those of trans-ZR and cis-ZR (Table 1).

Discussion
In the present study, ZR has been identi®ed from squash root xylem sap as the main suppressor of the formation of hypocotyl adventitious roots in cucumber. Cytokinin is a phytohormone that is involved in such processes as cell division, photosynthesis, chloroplast differentiation, senescence, and nutrient metabolism (Mok, 1994). Naturally occurring cytokinins are adenine derivatives with a side chain at the N 6 -position. Isopentenyladenine (iP), Z and  Fig. 3, was applied to a reverse phase CapselPak MG 18 column. The ion peaks were detected using normal mode (Total) and selected reaction monitoring mode (ZR, m/z=352/220; Z, m/z=220/136) (upper pro®les). The conditions for MS detection of cytokinins were previously determined by infusion of authentic ZR and Z; the ion peaks of each authentic cytokinin (trans-ZR, cis-ZR, trans-Z, and cis-Z) are shown in the lower pro®les.  dihydrozeatin (DZ) are the predominant cytokinins found in the tissues and calli of higher plants. ZR is a riboside of Z, which is thought to be one of the most biologically active compounds, along with isopentenyladenosine (iPA) and dihydrozeatin riboside (DZR), ribosides of iP and DZ.
ZR showed signi®cant inhibitory effects on adventitious root formation at concentrations from 3.16Q10 ±9 M to 10 ±4 M (Fig. 6). In previous reports, cytokinins such as BA and kinetin were shown to inhibit the formation of adventitious roots in bean (Humphries, 1960), Pinus radiata (Smith and Thorpe, 1975) and rose (De Vries and Dubois, 1988). Z, the predominant active cytokinin in plants, inhibits the formation of adventitious roots in pea, as does BA (Bollmark and Eliasson, 1986). As far as is known, the inhibitory effect of ZR has not been reported, despite the current evidence showing that ZR and Z exhibit the same strong inhibition of adventitious root formation (Fig. 6). It is possible that ZR was converted to Z in the hypocotyl, which then affected the inhibition; in intact plants, the ZR conversion mechanism following translocation from roots to shoots is unknown. BA, which is not dominant as an endogenous cytokinin, had a weaker inhibitory activity on the formation of adventitious roots than ZR and Z (Fig. 6), perhaps as a consequence of different metabolism and/or transport within the plant.
As with species such as Urtica dioica (Beck and Wagner, 1994) and pea (Beveridge et al., 1997), ZR was the most abundant cytokinin in squash root xylem sap, at an endogenous concentration estimated to be 2.0Q10 ±8 M (SD=0.1Q10 ±8 M) (Table 1). This concentration is suf®cient to inhibit the formation of adventitious roots by cucumber hypocotyls (Fig. 6).
By contrast, a cis-isomer of ZR, cis-ZR, was also detected at an estimated concentration of 6.6Q10 ±9 M (SD=0.5Q10 ±9 M) (Table 1). Since no isomerization of the internal standard [ 2 H]Z was observed in root xylem sap (data not shown), extraction artefacts such as tRNA breakdown and isomerization during extraction, resulting in the formation of the cis-isomer of cytokinins, are unlikely to have in¯uenced the results. Earlier work has shown cis-ZR in xylem sap from chickpea (Emery et al., 1998), wheat and oats (Parker et al., 1989). Emery et al. (1998) suggested that because cis-isomers of cytokinin are predominant in chickpea, activity would be expected in this species. The present results failed to show activity in the cis-ZR fraction (Fig. 5), although considerable cis-ZR was present in root xylem sap from squash (Table 1), supporting a report of lower biological activity of the cisisomer than the trans-isomer in bioassays (Kaminek, 1982).
While Z was observed in root xylem sap at a concentration of 2.4Q10 ±10 M (SD=0.1Q10 ±10 M) (Table 1), and has a strong inhibitory capability, the concentration was insuf®cient to inhibit the formation of adventitious roots (Fig. 6). These results lend further support to the hypothesis that ZR is the only appropriate candidate as a rootderived endogenous suppressor of adventitious root formation.
Root tips are the major sites of cytokinin biosynthesis and the source of cytokinins translocated via the xylem to control shoot development (Letham, 1994). This current work would suggest that much of the ZR carried by the transpiration stream from the roots is taken up by hypocotyl tissues to inhibit the formation of adventitious roots, and never reaches the uppermost regions of the shoot. Other reports support the hypothesis about the distribution of ZR from root to shoot. When radioactive ZR was supplied to root-pruned plants from the cut end, the largest proportion of extractable radioactivity was detected in the stem, with more than 50% in lupin stems (Jameson et al., 1987) and 40% in tobacco stems (Singh et al., 1992). Of course, root-pruning is expected to lead to a decrease in the ZR level in the stem, which has been shown to occur in both tomato (Maldiney et al., 1986) and pea (Bollmark et al., 1988) 1 d after cutting.
The mechanism for the negative regulation of adventitious root formation by ZR remains unclear. The inhibition of adventitious root formation by exogenously applied cytokinins occurs during the induction phase of root cell division (Eriksen, 1974;Fabijan et al., 1981;Bollmark and Eliasson, 1986;De Klerk et al., 1995). During the induction phase, auxins induce adventitious root formation in apple cuttings (De Klerk et al., 1995) and seem to interact antagonistically with cytokinins to control the initiation of adventitious roots. Similarly, auxins and cytokinins have antagonistic effects during lateral root formation (Bo Èttger, 1974), and in controlling the gene expression of a cdc2-like protein, which regulates the G 2 -M transition of the cell cycle as mitotic cyclin (John et al., 1993). The control of the expression of a cdc2-like protein might be involved in the inhibitory effects of ZR in the formation of adventitious roots, particularly if cell cycling were inhibited.
These results imply that ZR prevents the formation of adventitious roots on hypocotyls or stems in the normal, Table 1. Cytokinin levels in the xylem sap collected from squash roots Cytokinins were quanti®ed with LC/MS using selected reaction monitoring mode (trans-Z, cis-Z, m/z=220/136; trans-ZR, cis-ZR, m/z=352/220). All results were calculated using the recovery value of the internal standard in the given sample. Results shown represent the means of ®ve replications under the same conditions.

Compound
Concentration (nM TSE) trans-ZR 20T1 cis-ZR 6.6T0.5 trans-Z 0.24T0.01 cis-Z 0.38T0.01 non-wounded condition. When the roots are cut off, a rapid decrease in the ZR level would occur, along with an accumulation of auxin and the synthesis of ethylene in the basal part of the cut hypocotyls or stems, resulting in the formation of adventitious roots from the hypocotyls or stems. Further support for this study's hypothesis comes from a report that transgenic tobacco plants with reduced cytokinin content, due to the overexpression of a cytokinin oxidase gene, showed an increase in lateral and adventitious roots (Werner et al., 2001). Moreover, transgenic tobacco plants overexpressing a zeatin O-glucosyltransferase gene showed increased adventitious roots on their lower stems, suggesting that the reduction of active cytokinin by O-glucosylation leads to a lower cytokinin/ auxin ratio (Martin et al., 2001).
Because weaker inhibitory activities were also observed in the other fraction than the ZR fraction (Fig. 2), it is possible that substances other than ZR have auxiliary roles in the negative regulation of adventitious root formation; these remain unidenti®ed.