• Here, cytokinin-induced nitric oxide (NO) biosynthesis and cytokinin responses were investigated in Arabidopsis thaliana wild type and mutants defective in NO biosynthesis or cytokinin signaling components.
• NO release from seedlings was quantified by a fluorometric method and, by microscopy, observed NO biosynthesis as fluorescence increase of DAR-4M AM (diaminorhodamine 4M acetoxymethyl ester) in different tissues.
• Atnoa1 seedlings were indistinguishable in NO tissue distribution pattern and morphological responses, induced by zeatin, from wild-type seedlings. Wild-type and nia1,2 seedlings, lacking nitrate reductase (NR), responded to zeatin with an increase within 3 min in NO biosynthesis so that NR does not seem relevant for rapid NO induction, which was mediated by an unknown 2-(2-aminoethyl)2-thiopseudourea (AET)-sensitive enzyme and was quenched by 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-1-oxy-3-oxide (PTIO). Long-term morphological responses to zeatin were severely altered and NO biosynthesis was increased in nia1,2 seedlings. As cytokinin signaling mutants we used the single-receptor knockout cre1/ahk4, three double-receptor knockouts (ahk2,3, ahk2,4, ahk3,4) and triple-knockout ahp1,2,3 plants. All cytokinin-signaling mutants showed aberrant tissue patterns of NO accumulation in response to zeatin and altered morphological responses to zeatin.
• Because aberrant NO biosynthesis correlated with aberrant morphological responses to zeatin the hypothesis was put forward that NO is an intermediate in cytokinin signaling.
One model of cytokinin signal transduction in Arabidopsis proposes a direct transduction of the cytokinin signal from the receptors of the AHK gene family to the nuclear-localized type B response regulator (ARRs) transcription factors via histidine phosphotransfer proteins (AHPs) (Hwang & Sheen, 2001). However, other inputs into this signaling pathway are possible (Hutchison & Kieber 2002, Kakimoto, 2003; Grefen & Harter, 2004; Kiba et al., 2005), which could be switched on, for example, by phosphorylation of as yet unknown proteins by the kinase function of the receptors or the AHPs (Hutchison & Kieber, 2002). Here we provide data supporting the hypothesis of NO as a signaling intermediate in zeatin action. Nitric oxide biosynthesis was upregulated in Arabidopsis seedlings after 2 min of zeatin addition not only in wild type but also in nia1,2 seedlings deficient in NR (Wilkinson & Crawford, 1993), demonstrating that NR does not mediate rapid induction of NO by zeatin. Long-term responses of nia1,2 seedlings to zeatin, however, were changed as was the tissue distribution of NO biosynthesis. AtNOA1 also did not seem to be necessary for zeatin action. The receptor knockout mutant ahk4, which is allelic to WOL and CRE1 (Mähönen et al., 2000; Inoue et al., 2001; Yamada et al., 2001; Nishimura et al., 2004), and the triple knockout ahp1,2,3 mutant (Hutchison et al., 2006) showed aberrant patterns of NO tissue distribution in response to zeatin. This suggests that WOL/CRE1/AHK4 and AHP two-component signaling proteins are upstream of zeatin-activated NO biosynthesis and that NO is important for some cytokinin responses.
Materials and Methods
Arabidopsis thaliana (L.) Heynh. seedlings (Columbia ecotype, nia1,2 double mutant (Wilkinson & Crawford, 1993), ahk4 knockout, ahk2,3, ahk2,4 and ahk3,4 knockouts (Hutchison et al., 2006), Atnoa1 knockout (Guo et al., 2003), and ahp1,2,3 triple knockout (all in Columbia background) were used in the experiments. The ahk4 and ahp1,2,3 lines were generated in one partner laboratory (J. J. K.) and Atnoa1 was a gift from Dr N. Crawford. Arabidopsis seeds were surface-sterilized, maintained for 3 d at 4°C for vernalization. Arabidopsis seeds (20 per well) were grown in half-strength MS (Murashige & Skoog, 1962) medium on a rotating shaker (60 r.p.m.). Seedlings 7–8 d old were used in the experiments. For phenotypic responses and visualization of NO biosynthesis by 2.5 µm DAR-4M AM (diaminorhodamine 4M acetoxymethyl ester; or as otherwise stated) seedlings were grown for 10–14 d on half-strength MS agar plus 2% sucrose, containing different concentrations of zeatin. A 16-h photoperiod and a photon flux of 20–23 µmol m−2 s−1 and 22°C were used.
Quantification of NO release
The release of NO to the medium was determined by binding to the cell-impermeant DAR-4M in the medium in a fluorometric assay (Kojima et al., 2000, 2001). Twenty Arabidopsis seedlings (equivalent to 88.9 + 7.7 mg fresh weight) were used for each sample. The averages of relative units of fluorescence of NO release were based on seedling number (20 per each sample). When with 12 identical samples the average of units was set as 100% a standard deviation of ±9.2% was obtained. When, in the same experiment, units were based on fresh weight directly the average was 100 ± 13.8% or when based on protein extracted by 0.5 m KOH and measured by a Bradford (1976) assay 100 ± 14.0% was obtained so that variance of data was lowest within one experiment when based on seedling number. Arabidopsis (20 seedlings) were incubated in 2.5 µm or 1 µm DAR-4M in 0.5 mm KPO4 buffer pH 5.7, including different concentrations of zeatin or mock addition and with or without 2-phenyl-4,4,5,5-tetramethyl imidazoline- 1-oxyl-3-oxide (PTIO) or 2-(2-aminoethyl)2-thiopseudourea (AET) (dissolved in distilled water). Incubation condition was in the light at 22.5°C on a rotating shaker (60 r.p.m.). Supernatant was taken at each time-point and the relative fluorescence was measured by excitation at 560 nm and emission at 575 nm in a LS-5 Luminescence spectrometer (Perkin-Elmer, Überlingen, Germany). The values at t = 0 time-points were subtracted. Currently, relative fluorescence units obtained with DAR dyes cannot be converted to absolute (molar) values as no NO-derivatized DAR compounds are available. All experiments were repeated two to five times with similar results and the data presented are of a single representative experiment or averages of three replications.
Observation of NO by microscopy
Nitric oxide was visualized under fluorescence microscope by binding to the cell-permeable derivative DAR-4M AM. Twenty Arabidopsis seedlings were used for each treatment. For microscopy in Fig. 1, seedlings were loaded with 1 µm dye for 4 h, then washed twice with water, incubated with 0 µm or 5 µm zeatin and without or with 1 mm or 2 mm PTIO for 48 h and washed again twice with water. Incubation was in the light at 22.5°C on a rotary shaker (60 r.p.m.) and then kept at 4°C before microscopy. For other microscopy experiments, seedlings grown in half-strength MS agar supplemented with 2% sucrose and, if not stated otherwise, with 2.5 µm DAR-4M AM and various zeatin concentrations were harvested after 14 d of germination for microscopy. Plates were prepared and observed under a fluorescence Axioskop2 Mot Plus microscope (Zeiss), Filter set n°20 from Zeiss (excitation: BP 546/12; beam splitter: FT 560; emission: BP 575–640). Digital photos to be compared were taken at exactly the same camera settings of a digital camera and are were further processed. Several micrographs were made from each treatment and micrographs were selected from the experiment only for one-color plates shown.
Morphological responses to zeatin
Arabidopsis seedlings grown for 10–14 d in the same light conditions as for other experiments were scanned by Epson perfection 1250 at 1200 dpi resolution. From these digital pictures, quantifications of lengths and areas were measured by using the AxioVision Rel.4.3 (Zeiss, Göttingen, Germany) program. The basal medium was half-strength MS supplemented with 2% sucrose or Gamborg B5 medium diluted 1 : 50 supplemented with 0.5% sucrose and various zeatin concentrations (Gamborg et al., 1968; De Smet et al., 2003).
To test the NO specificity of cell-permeable DAR-4M AM, seedlings were kept for 2 h in distilled water containing 0.2 mm (±)-S-nitroso-N-acetylpenicillamine (SNAP), a chemical donor alone or additional 2 mm of the NO scavenger PTIO. The NO-induced fluorescence above the endogenous control without SNAP was evident after this time and both endogenous and NO generated by SNAP was quenched by PTIO (Fig. 1). Treating seedlings with 10 µm zeatin for 2 h led to no significant NO-dependent fluorescence (data not shown).
In most experiments, seedlings were grown in agar containing 1–5 µm water-soluble and membrane-permeable DAR-4M AM which was taken up by the transpiration stream. During this time, zeatin might have influenced the uptake of the dye. Therefore, to exclude dye transport effects as a reason for the observed tissue patterns, an experiment was done where the 4-h dye uptake and the hormone treatment for 2 d were kept separate. Using this method the suitability of PTIO could be demonstrated and potential long-term side-effects of PTIO could be avoided (Fig. 2). Shorter hormone treatments did not convincingly show the hormone or PTIO effects, especially because DAR-4M AM acts as a competitor for NO binding (Tun et al., 2006); therefore, the DAR-4M AM concentration needed to be lowered to 1 µm. Zeatin induced strong fluorescence increases in leaves (Fig. 2a,b), in trichomes (Fig. 2e), in cotyledons (Fig. 2g), in the root–shoot transition zones (Fig. 2i), in the elongation zones of root tips (Fig. 2k) and in all bundles (Fig. 2g,i; see also below, 3, 4 and 8a). At higher magnification it seems that vacuoles as acidic compartments accumulated the dye (Fig.Figs 2a,b) although the resolution of vacuoles, cytoplasm, and cell wall is incomplete. Probably also certain lipophilic substances, like cuticles of guard cells and trichomes accumulate dye, whereas during the 48-h treatment the guard cells themselves remained dark in the controls and increased in fluorescence in zeatin treatments. Controls without DAR-4M AM showed that chlorophyll fluorescence was effectively excluded (see below, Fig. 4). As DAR-4M AM and PTIO compete for NO binding (Tun et al., 2006) the decrease by PTIO was partial at 1 µm DAR-4M AM and may result in a slightly reddish hue (Fig. 2d,f), and was not observed at 2.5 µm DAR-4M AM (not shown). Yet, in cytokinin-treated seedlings PTIO clearly decreased fluorescence (Fig. 2d,f,h,j,l), thus showing that the zeatin-induced fluorescence increase was caused by NO accumulation. The endogenous NO level may result from other processes or endogenous cytokinin.
To date, two enzymes to generate NO have been identified in Arabidopsis: putatively AtNOA1 (Guo et al., 2003; Guo & Crawford, 2005) and NR (Harper, 1981). We tested whether the absence of these two enzymes affected the hormone response. Atnoa1 seedlings were grown in agar containing DAR-4M AM and with or without 5 µm zeatin. The strong increase in fluorescence in the leaves caused by zeatin was indistinguishable from the response of wild-type seedlings (Fig. 3). The responses in all other tissues (root tip, root–shoot transition zone, cotyledon, hypocotyls and shoot tips) was comparable to wild-type plants (Fig. 3e–l). Furthermore, the morphological responses to cytokinin were indistinguishable in wild-type and Atnoa1 plants (data not shown). Hence, we conclude that this particular protein was not regulated by zeatin.
Next the nia1,2 seedlings which lack NR were examined and compared with wild-type seedlings (Wilkinson & Crawford, 1993) (Figs 4 and 5). In both seedling genotypes, in the micrographs of leaves from seedlings grown on agar containing 0, 0.5, 1 or 5 µm zeatin, the effects of zeatin on fluorescence and on morphology could be observed. The cellular and subcellular distribution of zeatin-induced fluorescence was very similar to the patterns shown in Fig.Fig. 2 (data not shown) in that trichomes and the cuticles of guard cells were highlighted as an effect of zeatin treatment and fluorescence increased in hydathodes at leaf tips and also in intercostals fields (Fig. 4a–h). When DAR-4M AM was omitted, no fluorescence was observed in leaves and the other tissues (Fig. 4a–f) so that chlorophyll fluorescence did not obscure the results. The zeatin effect was already strongly apparent at 0.5 µm zeatin and increased with the zeatin concentration in leaves of both genotypes (Fig. 4a–h). Both genotypes exhibited the pointed tips at the leaves when treated by cytokinin but stronger leaf growth inhibition was found in the nia1,2 leaves at higher zeatin concentration (Fig. 4d,h). Cotyledons of nia1,2 seedlings always showed brighter fluorescence than wild-type seedlings and the increase induced by hormone was highest at 0.5 µm zeatin (Fig. 4i–p). Hypocotyls of nia1,2 seedlings showed stronger fluorescence than that of wild type even without zeatin (Fig. 4q–x). The morphological responses to zeatin of wild type and nia1,2 seedlings were studied by growing them on horizontal agar plates where the roots could grow into the agar (Fig. 5a–h). Growth of nia1,2 seedlings was different in the absence of exogenous hormone in that the roots were more branched and shorter (Fig. 5l) and the leaf areas of primary leaves were larger (Fig. 5i). Moreover, in response to zeatin the cotyledons became larger and fleshier than in wild type (Fig. 5g,h,j), and the expansion of primary leaves was delayed (Fig. 5f–h). In addition, hypocotyls became thicker and longer than in wild type (Fig. 5g,h,k). Anthocyanin accumulated typically more strongly in nia1,2 seedlings but this also depended on the light intensity.
Because nia1,2 seedlings exhibited a morphological zeatin-response phenotype and an altered tissue pattern of NO accumulation in response to zeatin, the total NO release in both seedling types was compared with the fluorometric quantification method. It was found that both genotypes exhibited endogenous NO release which was stronger in nia1,2 seedlings (Fig. 6a). The response was dependent on the concentration of zeatin in both types of seedlings (Fig. 6b). Endogenous NO release was rapidly enhanced by 20 µm zeatin after a 2–3 min lag phase, showing that NR is not the zeatin-regulated enzyme (Fig. 6c). In the short initial kinetics differences between the two genotypes were not apparent but in experiments of longer duration the amount of NO released by nia1,2 seedlings was slightly higher during the first hours and clearly higher after 24 h in nia1,2 seedlings (Fig. 6a). The NO scavenger PTIO inhibited the endogenous and the zeatin-induced fluorescence increase, verifying that the effects were NO-specific (Fig. 7a,b). Zeatin-induced NO release was inhibited in both genotypes by AET (Fig. 7c,d), an inhibitor of animal NO synthase and of zeatin- and polyamine-induced NO biosynthesis (Tun et al., 2001, 2006). As the inhibition by AET in nia1,2 plants was less effective this may indicate the presence of more than one NO-generating enzyme. However, it supports the notion that the zeatin-regulated enzyme was not NR since arginine analogues such as AET did not inhibit NR (Rockel et al., 2002).
Next, the response to zeatin of two cytokinin-signaling mutants was investigated with respect to their NO tissue pattern and their morphological responses. In leaves, cotyledons, and hypocotyls of cre1/ahk4 knockout seedlings no difference in the fluorescence tissue distribution was found (not shown) compared with wild-type seedlings. However, a remarkable difference was found in the mature central zone of cre1/ahk4 seedlings (Fig. 8A). Even though the fluorescence increase in 5 µm zeatin in the basal hypocotyl zone was as strong as in the wild type, the fluorescence in the bundle in the mature part of the cre1/ahk4 root did not increase when treated by zeatin. This tissue of the root is the site of strongest expression of the CRE1/AHK4 receptor (Mähönen et al., 2000; Nishimura et al., 2004). Wild-type and cre1/ahk4 seedlings responded very similarly to zeatin with respect to leaf size, cotyledon size, and hypocotyls length and thickness (data not shown). However, a transient inhibition of root growth was observed in cre1/ahk4 seedlings approx. 10 d after germination at high zeatin concentration (Fig. 8B,C), which was no longer apparent at 14 d or more after germination (data not shown). Thus, a correlation of the transient morphological response phenotype and the NO fluorescence phenotype is apparent.
As more severe cytokinin receptor mutants double-knockout lines of the three AHK genes, ahk2,3, ahk2,4 and ahk3,4 were investigated (Fig. 9). The tissue-dependent NO accumulation revealed a more complex pattern than in the single-knockout cre1/ahk4. When gene AHK2 was inactivated (ahk2,3 and ahk2,4), the root tips (Fig. 9b,c,f,g) produced more NO without exogenous zeatin and roots and root tips responded to zeatin with a further increase (Fig. 9j,k,n,o). When AHK3 was inactivated (ahk2,3 and ahk3,4), leaves produced more NO in response to zeatin (Fig. 9v,x) but without zeatin were not different compared to wild type (Fig. 9r,t). When gene AHK4 was inactivated (ahk2,4 and ahk3,4), cotyledons accumulated more NO and responded to exogenous zeatin with a further increase (Fig. 9α,β,ɛ,ζ). Wild-type cotyledons did not respond to 5 µm zeatin, only to a lower concentration of 0.5 µm zeatin (Fig. 4i,j). Even though overproduction of NO seems to be associated with loss of one gene and response in one organ (AHK2, root; AHK3,, leaf; AHK4, cotyledon), this seems not to provide an easy correlation with the growth responses to zeatin which we conducted on minimal nutrient agar (Gamborg B5 medium, diluted 1:50; Gamborg et al., 1968). All these receptor mutants showed cytokinin insensitivity, although slightly differently in different organs (Fig. 10). Without exogenous zeatin no difference was apparent compared with wild type in all three mutants. Wild type responded to zeatin (0.2 µm, 1 µm, 5 µm) at 0.2 µm with strong reduction of root growth and side roots and inhibition of leaf growth. The mutant ahk2,3 had decreased root growth at 1 µm zeatin or higher concentrations and decreased leaf growth at 5 µm. In ahk2,4 root growth was decreased at 0.2 µm and higher concentrations but leaf growth was decreased at 1 µm zeatin. The mutant ahk3,4 was not sensitive with regard to leaf growth up to 1 µm but root growth sensitivity was less than in wild type and, similar to ahk2,3, apparent at 0.2 µm and higher zeatin concentrations (see also Riefler et al., 2006).
The cytokinin receptor kinases phosphorylate the AHPs which migrate into the nucleus to regulate by phosphorylation the B-type ARR transcriptional cofactors. This, in turn, regulates the transcription of the A-type ARR genes. Upregulation of NO biosynthesis by zeatin could be a function downstream of the AHPs. To test this the tissue distribution and response to zeatin of NO-dependent fluorescence were studied in a triple-knockout line ahp1,2,3 (Hutchison et al., 2006). The Arabidopsis genome contains five AHP genes so that only AHP4 and AHP5 were active in this mutant. When grown without zeatin, the tissue distribution of fluorescence in ahp1,2,3 was very similar to wild-type seedlings (Fig. 11). However, when grown in the presence of 5 µm zeatin in the agar, fluorescence in the primary leaves of the triple mutant accumulated more strongly than in wild type. We observed no zeatin response in the roots, hypocotyls, and the cotyledons of the ahp1,2,3, in contrast to wild type where a clear zeatin-induced fluorescence response was observed. When the NO release from seedlings was quantified in the fluorometer, both the endogenous NO biosynthesis and NO biosynthesis in response to 5 µm zeatin was very similar in wild-type and ahp1,2,3 seedlings (not shown), probably because the overall balance of NO release in all tissues was still similar in wild-type and mutant seedlings. In half-strength MS medium the ahp1,2,3 seedlings did not show a phenotype or a response to zeatin that was different from wild type (not shown). When grown on minimal medium in the presence of increasing zeatin concentrations (Gamborg B5, diluted 1:50; Gamborg et al., 1968), the roots of ahp1,2,3 seedlings showed a response that was clearly different from wild-type roots in that roots of ahp1,2,3 were shorter and more branched than wild-type roots without hormone (Fig. 11B,C). Inhibition of main root growth by increasing zeatin concentrations was less affected in ahp1,2,3 than in wild-type seedlings, which is consistent with a lower zeatin sensitivity. The inhibition of side root formation was also less affected in ahp1,2,3 by zeatin compared with the wild type seedlings. The shoot, cotyledons, and leaves did not differ (compare also Hutchison et al., 2006). Gene AHP1 is strongly expressed in the roots, AHP2 in flowers, roots and stems, and AHP3 preferentially in roots. AHP4 is strongly expressed in the leaves and AHP5 in all organs (Tanaka et al., 2004) so that the observed responses in the ahp1,2,3 triple-knockout is also consistent with the preferential expression of the three inactivated genes in the roots.
Cytokinin-induced NO biosynthesis and the potentially NO contributing enzymes
The cell-permeable dye DAR-4M AM was used for visualization of NO in tissue and the cell-impermeable dye DAR-4M was used for quantification of NO released into the medium by the seedlings (Kojima et al., 2000, 2001). Unlike the previously used 4,5-diaminofluoresceine acetoxymethylester (DAF-4M) (Tun et al., 2001), the DAR compounds bind NO even below pH 4, which is important in plant media and acidic plant compartments. The DAR derivatives are not light-activated whereas DAF derivatives are (Broillet et al., 2001; Gould et al., 2003). Investigations on the chemical specificity of NO-specific dyes showed that fluorescence increase can be caused by peroxynitrite originating from NO and H2O2 (Roychowdhury et al., 2002) but not from H2O2 alone (Roychowdhury et al., 2002; Tun et al., 2006). Even though nitrite and hydrogen peroxide can also generate peroxynitrite, nitrite cannot be suspected to be present in nia1,2 seedlings lacking NR, and nia1,2 seedlings actually produce more NO than the wild type. So, even though peroxynitrite cannot be excluded as a reactant with DAR dyes it will not originate without NO also originating. The fluorometric quantification is very sensitive and allows quantification within short time-spans (Tun et al., 2001, 2006; Zeidler et al., 2004) but fluorescence microscopy allows visualization of the spatial pattern of NO accumulation. As microscopy is less sensitive and requires much longer hormone treatments, a combination of both methods and different mutants is best. Using the NO-specific scavenger PTIO and the arginine analogue inhibitor AET (Tun et al., 2001, 2006) lead to inhibition of the zeatin-inducible activity and this further supports our interpretation of fluorescence data being not nitrite-dependent. This proves that, with our precautions, fluorescence changes correlate closely with NO biosynthesis and accumulation. Hence, the observed tissue patterns must also be regarded as one pattern of cytokinin action, not excluding other cytokinin actions.
Our results exclude the AtNOA1 as a potential enzyme responsible for NO induction by zeatin since Atnoa1 seedlings showed the same tissue fluorescence response to zeatin as wild-type seedlings. Nitrate reductase is obviously necessary as a long-term contributor or regulator in the NO balance in plants because its absence caused an aberrant response to zeatin of nia1,2 plants (see also below) but in short kinetics of NO induction by zeatin the response was very similar in both genotypes, indicating that some other enzyme was regulated by zeatin. Polyamines induced a tissue pattern of NO accumulation very similar to the one induced by zeatin (Tun et al., 2006). For both zeatin and polyamine-induced NO biosynthesis, the relevant enzyme was AET sensitive but not identified. It is not improbable that the rapidly polyamine-regulated enzyme(s) described could be similar or identical to the cytokinin-regulated enzyme(s) but their identity remains to be shown. Alternatively, arginine-dependent NO synthase activity could be regulated by zeatin (Cueto et al., 1996; Barroso et al., 1999; Corpas et al., 2006; Jasid et al., 2006; Valderrama et al., 2007; Zhao et al., 2007).
Tissue distribution of zeatin-induced NO-dependent fluorescence and comparison to other cytokinin-related activities
Owing to the diffusion of NO, visualization of NO is influenced by subcellular distribution of dye and the NO sources. At the cellular or subcellular level a permeable cationic dye such as DAR-4 AM is concentrated in acidic compartments. Our extended dye exposure made it possible to clearly demonstrate zeatin-induced NO biosynthesis at relatively low cytokinin concentrations in growing plants and to describe characteristic patterns. In addition to accumulation in acidic compartments such as vacuoles, we saw fluorescence accumulation in cuticles of stomata and the sculptured cuticle of trichomes and in tracheidal cell walls. Highlighting of such structures might indicate lipophilic binding after derivatizing of DAR by NO. The tissues of veins (xylem and phloem parenchyma) were found to be a source of endogenous NO also by others (Corpas et al., 2004, 2006; Gabaldón et al., 2005; Valderrama et al., 2007). We assume that fluorescence increases in vacuoles, cuticles and specialized cell walls such as tracheids and trichomes was not the result of enzymatic activity in these compartments but rather the result of NO diffusion from the sources close by. Others found NOS-like activity in peroxisomes (Barroso et al., 1999) and NO-induced DAF-2-dependent (4,5-diaminofluorescein) fluorescence in the cytoplasm (Foissner et al., 2000; Zeidler et al., 2004; Zeier et al. 2004) and cell walls (Gould et al., 2003). At high cytokinin concentration, programmed cell death is induced (Carimi et al., 2005) which may be attributed to the release of NO by cytokinin as NO-induced cell death was shown for elicitors and pathogens (Delledone et al., 1998; Durner et al., 1998; Lamotte et al., 2004; Zeidler et al., 2004).
Aberrant tissue patterns of zeatin-induced NO-dependent fluorescence correlate with aberrant zeatin responses in cre1/ahk4, nia1,2, and ahp1,2,3 seedlings
Neither zeatin-induced NO-dependent fluorescence nor the physiological response to zeatin was changed in the Atnoa1 knockout seedlings so that we conclude that this gene product is not engaged in cytokinin signaling. This also shows that general decreases in the NO levels in this plant (Guo et al., 2003; Zeidler et al., 2004; Guo & Crawford, 2005) do not have specific consequences for zeatin-dependent responses. In contrast, we applied NO donors to generate a general increase in NO levels and the responses were only moderate (Scherer & Holk, 2000; Scherer, 2006), indicating that a more specific link between source of NO and downstream effect of NO exists which cannot be easily mimicked pharmacologically.
This contrasted with the consequences observed when another source of NO was absent, NR in the nia1,2 seedlings. In NR-deficient seedlings major disturbances of NO-dependent and zeatin-dependent fluorescence and of the morphological responses to zeatin were observed. The absence of one NO-generating enzyme, NR, apparently upregulated NO-producing enzyme(s) in mature roots, hypocotyls and cotyledons. As this nia1,2 line is a double-recessive mutant it would have not be possible to find it in any typical cytokinin mutant screen. The altered tissue response in those same organs to exogenous zeatin suggests NO to be a critical factor in the response to zeatin. These NO-producing enzyme(s), however, are not identical with NR since rapid NO release from nia1,2 seedlings was induced by zeatin just as in wild-type seedlings (and was actually slightly higher) and inhibited by the arginine analogue AET: therefore, NR is not the candidate enzyme of rapid regulation by zeatin (Rockel et al., 2002). Clearly, however, NR as a missing NO source changed the response to zeatin, identifying NO as a mediator of cytokinin action, generated by an unknown zeatin upregulated enzyme. This notion corresponds to and is supported by observations on the response phenotype of nia1,2 seedlings showing clear signs of hypersensitivity to zeatin. Alternatively, the observed long-term physiological responses in NR-deficient plants here may also be linked to the known mutual interplay of cytokinin, NR and nitrogen metabolism. Increase in nitrogen sources upregulates elements of cytokinin signal transduction (Takei et al., 2002) and cytokinin upregulates NR (Lu et al., 1990).
The NO tissue distribution in double-knockout mutants ahk2,3, ahk2,4 and ahk3,4 without and with zeatin addition was another example of a more complex response. The absence of receptor genes was expected to be associated with an increase in NO in certain tissues rather than with a lack of NO induction by zeatin. Even without exogenous zeatin, NO accumulation was greater in roots in plants without functional AHK2 than in cotyledons of plants without functional AHK4 and, in the presence of zeatin, more NO accumulated in leaves in plants without functional AHK3. Even though relief from repression of NO accumulation by these genes or protein activities may provide an explanation, it remains unclear how this regulation is achieved as this could be slow developmental steps within the 2-wk of growth of the seedlings. As in the nia1,2 plants, where we expected rather a lack of NO biosynthesis, again negative regulation of unknown genes/enzymes for NO biosynthesis by the AHK genes is indicated. It should be noted in this context, that knockout seedlings lacking AHK2 have very high cytokinin contents indicating repression of cytokinin biosynthesis of cytokinin by AHK2 and a further complexity in the coupling of responses to individual receptors (Riefler et al., 2006).
A simpler aberrant tissue pattern of zeatin-induced fluorescence corresponding to an aberrant morphological response of these plants to zeatin was found in the mutants cre1/ahk4 and ahp1,2,3. The cre1/ahk4 root base showed a diminished zeatin-induced fluorescence compared with wild type and, c. 10 d after germination, a transiently stronger inhibition of root growth by zeatin. Since this region is identical with a strong expression of AHK4 in the young seedling (Mähönen et al., 2000; Nishimura et al., 2004) this suggests that the lack of AHK4 expression and the lack of NO accumulation correlated with the change in the morphological response. Since two other cytokinin receptors, AHK2 and AHK3, are also expressed along the root, they may replace functionally the absent AHK4 gene product in most developmental stages (Nishimura et al., 2004; Higuchi et al., 2004; Riefler et al., 2006) so that redundancy may explain this rather subtle phenotype.
The ahp1,2,3 seedlings showed a quite differently aberrant pattern of NO accumulation than the nia1,2 seedlings. Without exogenous zeatin the mutant seedlings the NO tissue accumulation was indistinguishable from wild-type plants. However, the NO response to exogenous 5 µm zeatin was missing in roots, hypocotyls and cotyledons, and opposite to wild type in leaves, suggesting that a correct tissue response requires at least one of the three inactivated genes. Moreover, the morphological response of ahp1,2,3 seedlings to zeatin exhibited a decrease in hormone sensitivity in the root response where AHP1, AHP2 and AHP3 are expressed most strongly (Tanaka et al., 2004). Taken together, the results here suggest a close correlation of defects in NO biosynthesis and cytokinin signaling defects.
Signal transduction models for cytokinin action and NO as signaling intermediate
Most current signal transduction models for cytokinin suggest a biochemical signal transfer chain consisting of histidine kinase receptors, histidine phosphotransfer proteins, and A-type regulator response proteins acting as transcriptional cofactors to regulate the expression of the B-type response regulators (Hwang & Sheen, 2001; Hutchison & Kieber, 2002; Kakimoto, 2003). However, most authors do not exclude other mediators or second messengers in cytokinin action besides this phosphorylation cascade (Hutchison & Kieber, 2002, Kakimoto, 2003; Grefen & Harter, 2004; Kiba et al., 2005). For example, proteins other than B-type ARR proteins could be also phosphorylated. Proposed models for the role of NO in cellular signaling in plants so far are similar to animal models since cGMP (Durner et al., 1998; Wendehenne et al., 2001) and calcium-activated channels could be involved (Gould et al., 2003; Lamotte et al., 2004; Lecourieux et al., 2006). More recent models of animal NO-mediated signal transduction stress that certain plasma membrane calcium channels are regulated by reversible nitrosylation at defined cysteines by a NOS activity in direct vicinity of the channel without participation of cGMP (Stamler et al., 2001; Martinez-Moreno et al., 2005). These authors stress that strictly local NO sources regulate nearby proteins in their activity by reversible nitrosylation and probably not by a general change in NO level. Also, nitrosylation of cysteines seems not to be random but probably occurs only at cysteines surrounded by suitable neighboring amino acids. An enzyme capable of removing nitrosylation at cysteines, to make nitrosylation reversible, was also identified in plants (Liu et al., 2001; Sakamoto et al., 2002; Feechan et al., 2005) and nitrosylation of plant proteins has been shown (Lindermayr et al., 2005, 2006).
We cannot yet identify the NO-synthesizing enzyme(s) regulated by zeatin. We could only eliminate NR and AtNOA1 from the list of potentially directly zeatin-regulated NO-generating enzymes. Hence, positive identification of zeatin-regulated NO-generating enzymes remains a major challenge.
Financial support was provided by a Graduiertenstipendium from the University of Hannover to N. T. and by the DFG Sche207/15 to G. S and a grant to J. K (NSF nr. IOB0227669). We thank Dr N. Crawford for the Atnoa1 knockout seeds.