Cytokinins are a class of mitogenic plant hormones that play an important role in most aspects of plant development, including shoot and root growth, vascular and photomorphogenic development and leaf senescence. A model for cytokinin perception and signaling has emerged that is similar to bacterial two-component phosphorelays. In this model, binding of cytokinin to the extracellular domain of the Arabidopsis histidine kinase (AHKs) receptors induces autophosphorylation within the intracellular histidine-kinase domain. The phosphoryl group is subsequently transferred to cytosolic Arabidopsis histidine phosphotransfer proteins (AHPs), which have been suggested to translocate to the nucleus in response to cytokinin treatment, where they then transfer the phosphoryl group to nuclear-localized response regulators (Type-A and Type-B ARRs). We examined the effects of cytokinin on AHP subcellular localization in Arabidopsis and, contrary to expectations, the AHPs maintained a constant nuclear/cytosolic distribution following cytokinin treatment. Furthermore, mutation of the conserved phosphoacceptor histidine residue of the AHP, as well as disruption of multiple cytokinin signaling elements, did not affect the subcellular localization of the AHP proteins. Finally, we present data indicating that AHPs maintain a nuclear/cytosolic distribution by balancing active transport into and out of the nucleus. Our findings suggest that the current models indicating relocalization of AHP protein into the nucleus in response to cytokinin are incorrect. Rather, AHPs actively maintain a consistent nuclear/cytosolic distribution regardless of the status of the cytokinin response pathway.
A key feature of many signal transduction pathways is the subcellular compartmentalization of pathway components. Receptor-mediated nuclear transport of regulatory proteins is particularly important for plant growth and development. Key steps in many plant signal transduction pathways involve the migration of regulatory proteins into and out of the nucleus to affect gene expression (reviewed in: Merkle, 2001; Lee et al., 2008). Examples include: the transcription factor BZR1, which localizes to the nucleus in response to brassinosteroid (Gendron and Wang, 2007); the phytochromes, which move from the cytosol to the nucleus in response to red light (Chen et al., 2005); and the cryptochrome CRY1, which moves from the nucleus to the cytosol in response to white light (Yang et al., 2000). In addition, components of the nuclear transport machinery have been implicated in diverse processes such as plant–microbe interactions, auxin responses, stress tolerance, and regulation of flowering time (Dong et al., 2006; Parry et al., 2006; Jacob et al., 2007).
An important aspect of the cytokinin response pathway has been postulated to be the nuclear translocation of the AHP proteins in response to cytokinin. The AHPs mediate the flow of phosphate from the plasma membrane-localized AHKs to the mainly nuclear-localized ARRs. Previous reports have indicated that cytokinin treatment, and by inference phosphorylation of the AHPs, results in the relocalization of the AHP proteins from the cytosol into the nucleus. For example, observations in Arabidopsis mesophyll protoplasts indicated that both AHP1 and AHP2 were localized to the cytosol prior to cytokinin treatment, but translocated into the nucleus in response to exogenous cytokinin (Hwang and Sheen, 2001). Likewise, Yamada et al. (2004) reported that several AHPs relocalized from the cytosol into the nucleus in response to exogenous cytokinin in T87 Arabidopsis tissue culture cells.
Despite broad acceptance for nuclear movement of the AHP proteins as a key aspect of cytokinin signal transduction, this has not been rigorously tested. First, while the observation that AHP1 and AHP2 move into the nucleus upon cytokinin treatment was reported in Arabidopsis mesophyll protoplasts, this has not been demonstrated in planta. Second, movement into the nucleus was not shown to require upstream components of the cytokinin signal transduction cascade, calling into question whether the observed movements were the result of cytokinin signaling or some other response. Finally, and perhaps most importantly, the data for nuclear movement of AHPs in protoplasts and cultured cells were not quantified. In contrast to what has been reported in Arabidopsis, the Saccharomyces cerevisiae homolog of the AHPs, YPD1, does not undergo nuclear transport in response to osmotic stress, suggesting that bulk movement of histidine phosphotransfer proteins into the nucleus is not necessary for relaying extracellular signals to the nucleus in eukaryotic two-component signal transduction cascades (Lu et al., 2003).
Here, we show that the AHP proteins are present in both the nucleus and the cytosol, in the presence or absence of cytokinin treatment, in both transiently transformed mesophyll protoplasts and intact plants. Treatment with cytokinin altered neither the proportion of cells showing nuclear-localized AHPs nor the ratio of nuclear-localized:cytosolic-localized AHP proteins. Genetic and molecular disruption of cytokinin signaling did not alter the localization of the AHPs, suggesting that their localization is independent of cytokinin signaling. Finally, despite being unresponsive to cytokinin treatment, the distribution of the AHP proteins between nucleus and cytosol was maintained by an active nuclear import and export mechanism.
AHPs are localized to both the nucleus and cytosol in planta; this localization is unchanged in response to cytokinin treatment
Prior studies in Arabidopsis mesophyll protoplasts suggested that AHP–GFP fusion proteins are mainly localized to the cytosol, but rapidly translocate into the nucleus in response to cytokinin treatment (Hwang and Sheen, 2001; Yamada et al., 2004). To determine the subcellular localization of these proteins in planta, we analyzed transgenic Arabidopsis plants harboring either a ProAHP2:AHP2–GFP (Figure 1) or a ProAHP5:AHP5–GFP transgene (Figure S1 in Supporting Information). The ProAHP2:AHP2–GFP construct complements an ahp1,2,3 loss-of-function mutant (Figure S2), and the ProAHP5:AHP5–GFP construct was previously shown to complement an ahp1,2,3,4,5 mutant (Hutchison et al., 2006), indicating that these fusion proteins are functional. We observed that AHP2–GFP fluorescence was distributed throughout the cells distal to the meristematic zone (early transition zone) of the root in plants harboring a ProAHP2:AHP2–GFP transgene (Figure 1a,b). A high concentration of GFP fluorescence was observed within the nuclei; however, GFP signal was also clearly visible in the cytosol (Figure 1k). Application of trans-zeatin did not modify this localization pattern (Figure 1e,f). Similar results were observed in the meristematic zone in the absence (Figure 1c,d) and presence of cytokinin (Figure 1g,h), root hair cells (data not shown), and epidermal cells of the leaf (Figure 1i,j) as well as with ProAHP5:AHP5–GFP (Figure S1). These observations suggest that in planta the AHP proteins are distributed in both the cytosol and nucleus, and that AHPs do not change their subcellular localization in response to exogenous cytokinin.
AHPs are localized to both the nucleus and cytosol in mesophyll protoplasts, and this localization is unchanged in response to cytokinin treatment
Because our in planta observations of AHP–GFP subcellular localization contradict the initial observations of cytokinin-induced AHP–GFP relocalization in protoplasts, we repeated these original experiments and quantified the results. To this end, we analyzed Arabidopsis mesophyll protoplasts transfected with various Pro35S:AHP–GFP constructs. We grew our plants in conditions as similar as possible to those used in the original experiments, used similar constructs and a similar transfection protocol (Hwang and Sheen, 2001; Yoo et al., 2007). Figure 2 shows that in the absence of exogenous cytokinin, AHP1–GFP, AHP2–GFP, and AHP5–GFP are located in both the nucleus and the cytosol of most cells. However, across the population of protoplasts, we did observe a subset of cells with primarily nuclear or primarily cytosolic AHP localization (Figure S3). Nuclear localization was confirmed by co-localization with 4′,6-diamidino-2-phenylindole (DAPI) as well as with confocal microscopy (Figure S4). Protoplasts were treated with two cytokinins, trans-zeatin (Figure 2g–i) or N6-benzyladenine (Figure S5). In both cases, cytokinin treatment had no obvious effect on the subcellular distribution of AHP–GFP across a population of protoplasts. Importantly, these treatments did induce the expression of primary cytokinin-response genes (Table S1), indicating that the protoplasts are physiologically normal and cytokinin-responsive, and that our treatments activated the endogenous cytokinin signal transduction pathway.
We next quantified the distribution of AHP–GFP throughout a population of cells by scoring GFP fluorescence in individual cells as: (i) predominantly cytosolic; (ii) both nuclear and cytosolic; or (iii) predominantly nuclear (Figure S3). Across populations of cells, and for multiple AHP fusion proteins, the percentage of cells exhibiting these various GFP patterns did not significantly change in response to treatment with cytokinin (Figures 2 and S5). This is consistent with the in planta results, and suggests that the localization of AHP–GFP fusion proteins is unchanged in response to cytokinin. These data also suggest that the initial observations that AHP proteins are translocated into the nucleus after cytokinin treatment do not represent general features of the AHPs nor of cytokinin signal transduction.
Quantification of the subcellular distribution of AHP2–GFP shows that the fraction of AHP2–GFP localized to the nucleus did not change in response to cytokinin
Although the subcellular localization of AHP–GFP fusion proteins did not appear to change in response to cytokinin treatment across populations of cells, it is possible that, within a cell, the relative amount of nuclear-localized AHP increases upon cytokinin treatment. To address this question, we analyzed the percentage of AHP–GFP fluorescence that was localized to the nucleus before and after cytokinin treatment in cells that showed both nuclear- and cytosolic-localized AHP–GFP fluorescence (Figure S3). For each individual cell, we captured a z-series of images from multiple focal planes. These images were subjected to deconvolution to remove out-of-focus fluorescence. Regions of interest were generated that represented either the entire cell or the nucleus, and the amount of fluorescent signal from each of these regions was quantified. As shown in Figures 2(p) and S5(j), the relative level of nuclear-localized fluorescent signal from multiple AHP–GFP fusion proteins did not appreciably change in response to treatment with cytokinin. These data support our previous observations that AHPs do not change their subcellular localization in response to exogenous cytokinin.
AHP2 localization did not change in response to genetic and molecular disruption of the cytokinin signaling pathway
Our observations suggest that AHP–GFP fusion proteins are localized to both the nucleus and cytosol, irrespective of activation of cytokinin signal transduction via application of exogenous cytokinin. During cytokinin signaling, phosphoryl groups are transferred from an aspartate residue on the activated cytokinin receptors (AHKs) to a histidine residue on the AHP proteins (Mähönen et al., 2006). To test whether phosphorylation of AHP2 might be required for its proper localization, we examined the subcellular localization of a phospho-insensitive version of AHP2–GFP. We mutated the conserved histidine (H82) (Hutchison et al., 2006) residue to an alanine (A) residue in AHP2–GFP and examined the subcellular localization of the resulting fusion protein. This mutant fusion protein, AHP2H82A–GFP, exhibited a subcellular localization similar to that of the wild-type AHP2–GFP fusion protein (Figure 3a) and displayed a similar ratio of nuclear:cytosolic localized fluorescent signal as compared to the wild-type AHP2–GFP, even in the presence of cytokinin (Figure 2p). Similar results were obtained with a phospho-insensitive version of AHP1–GFP (Figure S6).
Prior reports have shown that AHPs can homo- and hetero-dimerize in yeast two-hybrid assays (Dortay et al., 2006, 2008). This suggests that AHP2H82A–GFP may have entered the nucleus in the above assays by dimerizing with an endogenous, wild-type AHP2 protein. To confirm the yeast two-hybrid assays, we tested whether AHP2 can homo-dimerize using bimolecular fluorescence complementation. Both a wild-type and a phospho-insensitive version of AHP2 (AHP2 and AHP2H82A, respectively) were fused at their C-termini to both the N-terminal and C-terminal halves of the yellow fluorescent protein (YFP) coding region (YFPn and YFPc, respectively). Combinations of these proteins as well as either half of YFP alone were tested for interaction in Arabidopsis mesophyll protoplasts by assaying for YFP fluorescence using confocal microscopy. If the two proteins interact when co-expressed in the same cell, the two complementary halves of YFP are brought together, resulting in fluorescence. As shown in Figure 3(b), both AHP2–YFPn and AHP2H82A–YFPn interact with AHP2–YFPc in the nucleus and cytosol. In contrast, YFPn alone did not interact with AHP2–YFPc, suggesting that the interaction was not due to affinity between the two halves of YFP. Reciprocal experiments showed a similar pattern of interactions (data not shown). Combined with previous yeast two-hybrid assays, these results strongly suggest that AHP proteins are able to homo-dimerize in vivo.
Because dimerization of AHP2 may have confounded our analysis of phospho-insensitive AHP2 in protoplasts derived from wild-type plants, we analyzed the subcellular localization of AHP2H82A–GFP in protoplasts isolated from ahp1,2,3,4,5 mutant plants, which are null for four of the five AHP genes and have a strongly reduced level of expression for the remaining gene, AHP2 (Hutchison et al., 2006). As shown in Figure 3(a), both AHP2–GFP and AHP2H82A–GFP exhibit a wild-type subcellular distribution in cells largely lacking endogenous AHP proteins, being found in both nucleus and cytosol, consistent with a model in which the localization of AHP2 is independent of phosphorylation at the conserved histidine residue. Because we used the 35S promoter to drive expression of the AHP2–GFP and AHP2H82A–GFP fusion constructs in protoplasts, it is unlikely that the residual expression of endogenous AHP2 in the quintuple mutant was sufficient to result in substantial entry of AHP2–GFP or AHP2-H82A–GFP into the nucleus via dimerization.
Finally, it is possible that activation of the cytokinin signal transduction pathway results in phosphorylation of AHP2 at a non-conserved phospho-acceptor, which may in turn cause AHP2 to become localized to the nucleus. To address this, we analyzed the subcellular distribution of AHP2–GFP in mesophyll protoplasts isolated from ahk2,3 or ahk3,4 mutant plants. AHK2, AHK3, and AHK4 encode the cytokinin receptors of Arabidopsis; AHK2 and AHK3 are expressed in all tissues, whereas AHK4 is predominantly expressed in the root (Higuchi et al., 2004; Nishimura et al., 2004). Therefore, leaf mesophyll protoplasts isolated from ahk2,3 plants are strongly insensitive to cytokinin signaling through the canonical two-component signaling pathway (Higuchi et al., 2004; Nishimura et al., 2004). As in wild-type cells, AHP2–GFP is distributed in both the nucleus and cytosol of the ahk2,3 and ahk3,4 protoplasts (Figure 3a). Taken together, these results strongly suggest that AHP2 localization is not regulated by changes in the phosphorylation state of its conserved histidine or by activation of the cytokinin signaling pathway through the known cytokinin receptors.
AHP–GFP fusion proteins are actively shuttled between the nucleus and cytosol
The above experiments suggest that subcellular localization of AHP–GFP fusion proteins is not regulated by the cytokinin signaling pathway. The predicted size of the AHP2–GFP fusion protein monomer is ∼46 kDa, which is above the size exclusion limit of the nuclear pore (∼20–40 kDa) (Paine et al., 1975), suggesting that AHP–GFP fusion proteins are actively transported both into and out of the nucleus. To test if AHP2 enters the nucleus via active transport, we generated a series of AHP2 fusions in which we fused multiple copies of the GFP coding region to the AHP2 coding region. These additional fusion constructs are predicted to encode proteins that are well above the size exclusion limit of the nuclear pore (∼75 kDa in the case of AHP2–2x GFP or GFP–AHP2–GFP and ∼104 kDa in the case of AHP2–3x GFP). Protoplasts transfected with these constructs exhibit a distribution of GFP similar to AHP2–GFP, which is functional as determined by its ability to complement ahp1,2,3 (Figure S2). By contrast, both a 2x GFP and a 3x GFP fusion protein were predominantly localized to the cytosol (Figure 4g,h). These results suggest that AHP–GFP fusion proteins are actively transported into and out of the nucleus, even in the absence of exogenously applied cytokinin. To test this observation further, we attempted to identify sequences within the AHP2 protein that were required for proper active transport by generating a series of N- and C-terminal deletions of AHP2. These deletion proteins were fused with two GFP coding regions, one at the N-terminus and one at the C-terminus, to ensure that the resulting fusion protein was larger than the predicted size exclusion limit of the nuclear pore.
All of the C-terminal deletions resulted in fusion proteins that were predominantly localized to the cytosol (Figure 5). This suggests that there are residues in the C-terminus of AHP2 that are necessary for active transport into the nucleus. Likewise, an N-terminal deletion of 38 amino acids resulted in a fusion protein that also predominantly localized to the cytosol (Figure 5f). Conversely, deletion of an additional 30 amino acids from the N-terminus (NΔ68, Figure 5g) resulted in a fusion protein that was predominantly nuclear, indicating that residues required for nuclear export were removed. Additional deletions from the N-terminus resulted in subcellular localization patterns that were mixed, but significantly different from the wild type. These results suggest that active transport of AHP2 is complex, and may involve multiple protein–protein interaction domains.
The cytokinin signal transduction pathway uses a two-component phosphorelay signaling system to convey extracellular information to the nucleus and ultimately affect gene expression. This signal transduction pathway shares many features with the two-component pathways of prokaryotes. A striking difference between these signaling pathways is the requirement to transduce the signal into the nucleus in eukaryotes. The generally accepted model for cytokinin signaling suggests that the subcellular location of the Arabidopsis histidine phosphotransfer proteins (AHPs) is regulated by cytokinin (Aoyama and Oka, 2003; Heyl and Schmulling, 2003; Ferreira and Kieber, 2005; Lohrmann and Harter, 2002; Muller and Sheen, 2007; To and Kieber, 2008). This model posits that AHP proteins undergo bulk translocation to the nucleus in response to cytokinin and implies that non-phosphorylated AHPs are found primarily in the cytosol, while phosphorylated AHPs are found primarily in the nucleus. Our results call this aspect of the model into question.
The model that AHPs translocate into the nucleus in response to cytokinin is based on observations of the movement of AHP made in protoplasts and cultured cells; however, this process has not been examined in intact plants. We analyzed the subcellular localization of both AHP2 and AHP5 in Arabidopsis roots and observed that these proteins are distributed throughout the cytosol and nucleus. We were unable to detect marked differences in subcellular localization in response to treatment with cytokinin. We next analyzed the localization of AHP1, AHP2, and AHP5 in protoplasts and observed that these proteins were again distributed throughout the cytosol and nucleus. Surprisingly, the proteins maintained this distribution pattern irrespective of treatment with trans-zeatin or N6-benzyladenine. Furthermore, genetic and molecular disruption of the cytokinin signaling pathway did not affect the subcellular distribution of AHPs in protoplasts. We cannot exclude the possibility that for a particular cell type or small subset of cells under certain conditions, the AHP proteins do undergo bulk relocalization to the nucleus in response to cytokinin, or that AHP movement is below the level of detection in our system. Nevertheless, our observations directly contradict prior evidence that AHP proteins relocalize into the nucleus in response to cytokinin and suggest that bulk nuclear transport is not a general feature of cytokinin signal transduction.
In support of our observation that the AHP proteins do not respond to cytokinin by translocating to the nucleus, previous reports for AHP movement into the nucleus are not consistent with the genetic analysis of the AHPs. Different AHP proteins have been reported to display distinct nuclear movement kinetics in response to cytokinin treatment: AHP1 showed transient nuclear localization, AHP2 showed sustained nuclear localization, and AHP5 was found in the nucleus prior to cytokinin treatment (Hwang and Sheen, 2001). Subsequent mutant analysis revealed functional redundancy among the AHPs with respect to cytokinin perception and responsiveness (Hutchison et al., 2006). Such redundancy would suggest that the kinetics of translocation into the nucleus would be largely conserved amongst the encoded proteins if it were an integral part of the signal transduction pathway. Our data are consistent with genetic analysis of the AHPs and with the behavior of YPD1, the histidine phosphotransfer protein from S. cerevisiae, which is found in both the cytosol and nucleus, and does not change localization in conditions that activate its phosphorelay (Lu et al., 2003).
The data presented here indicate that the overall distribution of AHPs remains nucleo-cytosolic irrespective of the status of the cytokinin signal transduction pathway. We further showed that the subcellular distribution of AHPs is maintained by active nuclear-cytosolic transport mediated by cis-acting residues. Taken together, these observations suggest that localization of the endogenous AHPs is not static but constantly cycling between the nucleus and the cytosol. The cycling of AHPs mediates the relay of phosphoryl groups from the membrane-localized receptors to the mainly nuclear-localized response regulators. However, this raises the possibility that AHPs could also carry phosphoryl groups back out of the nucleus. AHPs could then donate these phosphoryl groups to either cytosolic-localized response regulators or could interact in a phospho-dependent manner with other cytosolic-localized proteins, regulating their activity. Furthermore, AHPs could be de-phosphorylated by the cytokinin receptors, a subset of which has been shown to possess phosphatase activity on AHPs (Mähönen et al., 2006). This backflow of phosphoryl groups to the receptors could serve to attenuate cytokinin signaling.
Our observations suggest that the AHP proteins are actively transported into and out of the nucleus in a cytokinin-independent manner. Interestingly, this active movement was not responsive to Leptomycin B (LMB) treatment (Figure S7), suggesting that nuclear export of AHP2 is not mediated by LMB-sensitive exportins such as AtXPO1 (Haasen et al., 1999; Kudo et al., 1998). We also confirmed that AHP2 could form homo-dimers in protoplasts. Consistent with the model that the endogenous AHPs are actively transported into and out of the nucleus, the sizes of endogenous AHP dimers (∼35 kDa) are near the size exclusion limit of the nuclear pore (∼20–40 kDa) (Paine et al., 1975). The ramifications of the active transport of AHPs are potentially twofold. First, different cells or tissues of the plant may modulate the mechanism of active transport to alter the ratio of nuclear:cytosolic AHPs. This could result in cell-type-specific cytokinin sensitivity. Second, it is also possible that other signaling or hormone response pathways modulate cytokinin signaling by affecting AHP localization via manipulation of the AHP active transport system.
Here, we examined the subcellular localization of the AHPs in response to cytokinin as well as to genetic and molecular perturbations. Our observations suggest that the current model for cytokinin signal transduction should be modified to indicate that bulk nuclear translocation of the AHPs is not a general feature of the cytokinin signal transduction pathway. Instead, we propose that the AHP proteins undergo constant, active, bidirectional nuclear–cytosolic transport regardless of their phosphorylation status and that this movement serves to shuttle phosphoryl groups into and out of the nucleus. The mechanism regulating the active transport of AHPs remains to be elucidated, which may add another level of complexity to the regulation of the cytokinin signaling pathway in plants.
Plant materials and growth conditions
Arabidopsis thaliana seeds were surface-sterilized with chlorine gas and germinated on plates containing 1 × Murashige and Skoog (MS) salts supplemented with 0.05% 2-(N-Morpholino)-ethanesulfonic acid (MES) (Research Products International Corporation, http://www.rpicorp.com), 1.0% sucrose, and 0.7% agar. Ten-day-old seedlings were transferred to soil and grown under 24-h light at 24°C. Protoplasts were obtained from plants grown as above, with the exception that they were grown under a 12-h light regime at 22°C.
Protoplast isolation and transformation
Protoplasts were isolated and transformed according to the method described in Yoo et al., 2007. Briefly, Arabidopsis leaf tissue was cut into thin strips and digested for 3–5 h in enzyme solution (1.5% Cellulase R-10 and 0.4% Macerozyme R-10, Yakult Pharmaceutical, http://www.yakult.co.jp/ypi/en/) to obtain protoplasts. Isolated protoplasts were washed and counted. For each transfection reaction, 20 000 cells were incubated with 20 μg of plasmid DNA and in 20% poly(ethylene glycol) (PEG; Fluka, http://www.sigmaaldrich.com) for 5 min. Cells were washed and incubated overnight in the dark to allow for gene expression. Protoplasts were analyzed 12–18 h after transfection.
For treatment of roots, seedlings were first grown horizontally on agar plates. Five days after germination, seedlings were carefully removed from plates and were placed in a solution of 1 × MS salts, 0.05% MES and 1.0% sucrose supplemented with either 1 μmtrans-zeatin (Sigma, http://www.sigmaaldrich.com/) or 1 μmN6-benzyladenine (Gibco BRL, http://www.invitrogen.com/site/us/en/home/brands/Gibco.html) for 1 h (water or DMSO was added to control samples). Following treatment, seedlings were counter-stained in propidium iodide to reveal cell walls and were immediately analyzed using confocal scanning laser microscopy.
For treatment of protoplasts, aliquots of transformed protoplasts were pelleted at 100 g and then resuspended in a solution containing 2 mm MES, 154 mm NaCl, 125 mm CaCl2, and 5 mm KCl at pH 5.7 supplemented with either 1 μmtrans-zeatin or 1 μmN6-benzyladenine (water or DMSO was added to control samples). Protoplasts were directly analyzed following cytokinin treatment.
The T-DNA constructs were introduced into Agrobacterium tumefaciens strain GV3101 by electroporation. Arabidopsis plants (accession Col-0 or ahp1,2,3) were transformed using a modified floral dip procedure (Clough and Bent, 1998). Transformed progeny were selected by germinating surface-sterilized T1 seeds on growth medium containing antibiotics (30 μg ml−1 kanamycin sulfate) supplemented with 15 μg ml−1 cefeotaxime. Resistant seedlings were transplanted to soil 10 days after germination. Transgene identity was verified by PCR using a gene-specific primer and a T-DNA-specific primer. Fusion protein sizes were verified by western blotting using the Pierce West Pico Chemiluminescence substrate (Thermo Scientific, http://www.thermofisher.com/) with an anti-GFP antibody (Clontech, http://www.clontech.com/) and a horseradish peroxidase (HRP)-conjugated secondary antibody (Santa Cruz Biotechnology, http://www.scbt.com/).
Microscopy and analysis of fluorescence
Intact plants and protoplasts containing GFP and bimolecular fluorescence complementation (BiFC) (YFP) constructs were analyzed using either confocal scanning laser microscopy or epifluorescence microscopy. In the case of confocal microscopy, a Zeiss LSM 510 META scanning confocal microscope (http://www.zeiss.com/) was used to analyze and capture images. The following excitation lines and emission ranges were used: GFP (excited using a 488-nm line, detected between 505 nm and 530 nm), propidium iodide (excited using a 488-nm line, detected with a 560-nm long-pass filter), YFP (excited using a 514-nm line, detected between 530 nm and 560 nm), and chlorophyll autofluorescence (excited using a 488-nm line, detected with a 560-nm long-pass filter). In the case of epifluorescence microscopy, a Nikon Eclipse 80i equipped with a Nikon DS-Qi1Mc high sensitivity cooled monochromatic camera (http://www.nikon.com/) was used to analyze tissue and capture images. Tissues were excited using a Hg-lamp (Nikon Intensilight C-HGF) and signal was detected with the ETGFP filter set for GFP, the DAPIHYB filter set for DAPI, and the ETYFP filter set for YFP. Image analysis was performed with Nikon NIS-Elements Ar and Br version 2.3. Deconvolution was performed by using a real-time iterative 2-D blind algorithm from AutoQuant (Media Cybernetics, http://www.mediacy.com/) to determine the point spread function and subsequently de-blur the images (Biggs, 2004; Brittain et al., 2009; Taguchi et al., 2009).
Construction of ProAHP5:AHP5–GFP was described in Hutchison et al., 2006. For ProAHP2:AHP2–GFP, approximately 2 kb of the AHP2 promoter and the entire AHP2 coding region were amplified with primers that added unique restriction enzyme sites: 1–62 (GATTACGCCCTGCAGGTGATGTTCATTGCTGACTCTTTCG) and 1–34 (TTACGCCGGATCCGTTAATATCCACTTGAGGAACT). The resultant 3-kb PCR product was digested with SbfI and BamHI and cloned into pBI1GFP. The remainder of the constructs were made using the Gateway Recombination System (Invitrogen, http://www.invitrogen.com/) according to the manufacturer’s protocols. Table S2 describes the primers used to generate entry clones used in this study. Table S3 summarizes the Gateway reactions and the resultant expression clones.
To make several of the constructs, existing plasmids were converted to Gateway destination clones. Briefly, the multiple cloning sites of pHBT-sGFP(S65T)-NOS (Yoo et al., 2007), pUC-SPYNE, and pUC-SPYCE (Walter et al., 2004) were replaced with a Gateway cassette that contained attR sites flanking a ccdB gene and a chloramphenicol resistance gene. The resultant destination vectors gw-GFP, pUC-gw-SPYNE, and pUC-gw-SPYCE were used to fuse wild-type and mutant AHP coding regions with either a GFP coding region, a coding region encompassing the N-terminal half of YFP, or a coding region encompassing the C-terminal half of YFP. The original non-Gateway-compatible vectors were used to express unfused GFP and each unfused half of YFP in protoplasts. All constructs were sequence verified.
Real-time PCR analysis of cytokinin-induced gene expression
Protoplasts were treated with water or 1 μmtrans-zeatin for 1 h. RNA was extracted using the Qiagen RNeasy Micro Plus kit following the manufacturer’s protocol (http://www.qiagen.com/). Aliquots of RNA (2 μg) were reverse transcribed using the Invitrogen SuperScript III kit (Invitrogen) following the manufacturer’s protocol. For real-time RT-PCR, PCR was performed using SYBR Premix Ex Taq (Perfect Real Time) (http://www.takara-bio.com/) in a volume of 20 μl on a DNA Engine Opticon 2 Real-Time PCR machine (http://www.bio-rad.com/), analysis was performed using Opticon Monitor 3. The PCR reaction mixture consisted of 2 μl cDNA, 0.5 μm primers, and 1 × Master Mix. For each real-time RT-PCR run, ACTIN2 (At3g18780) was used as an internal control to normalize for pipetting error of the cDNA template. The PCR program consisted of 94°C for 2 min followed by 40 cycles of: 94°C for 15 sec, 55°C for 15 sec, and 72°C for 15 sec. To determine the specificity of the PCR, all amplified products were subjected to melting curve analysis using the machine’s standard methods and one product per primer set was run on a gel. The reported CT values are averages of four independent trials (three biological replicates and one technical replicate). Real-time RT-PCR primers are listed in Table S4.
We thank Jaimie Van Norman and Cristiana Argueso for critical review of this manuscript. We thank Ramin Yadegari for providing the vector pBI1GFP, Plant Systems Biology, Ghent University, Belgium for providing the vector pK7FWGF2, and the ABRC for providing the vector pHBT-sGFP(S65T)-NOS. This work was supported by a DOE grant (no. DE-FG02-05ER15669) and an NSF grant (no. IOS 0618286) to JJK and GES.