Proteome analysis in Arabidopsis reveals shoot- and root-specific targets of cytokinin action and differential regulation of hormonal homeostasis.

The plant hormones cytokinins (CKs) regulate multiple developmental and physiological processes in Arabidopsis (Arabidopsis thaliana). Responses to CKs vary in different organs and tissues (e.g. the response to CKs has been shown to be opposite in shoot and root samples). However, the tissue-specific targets of CKs and the mechanisms underlying such specificity remain largely unclear. Here, we show that the Arabidopsis proteome responds with strong tissue and time specificity to the aromatic CK 6-benzylaminopurine (BAP) and that fast posttranscriptional and/or posttranslational regulation of protein abundance is involved in the contrasting shoot and root proteome responses to BAP. We demonstrate that BAP predominantly regulates proteins involved in carbohydrate and energy metabolism in the shoot as well as protein synthesis and destination in the root. Furthermore, we found that BAP treatment affects endogenous hormonal homeostasis, again with strong tissue specificity. In the shoot, BAP up-regulates the abundance of proteins involved in abscisic acid (ABA) biosynthesis and the ABA response, whereas in the root, BAP rapidly and strongly up-regulates the majority of proteins in the ethylene biosynthetic pathway. This was further corroborated by direct measurements of hormone metabolites, showing that BAP increases ABA levels in the shoot and 1-aminocyclopropane-1-carboxylic acid, the rate-limiting precursor of ethylene biosynthesis, in the root. In support of the physiological importance of these findings, we identified the role of proteins mediating BAP-induced ethylene production, METHIONINE SYNTHASE1 and ACC OXIDASE2, in the early root growth response to BAP.

The shoot and root of Arabidopsis (Arabidopsis thaliana) share the expression and function of molecules crucial for organ formation and maintenance (for review, see Dinneny and Benfey, 2008;Stahl and Simon, 2010). This has led to the hypothesis that similar regulatory networks underlie comparable developmental processes in the shoot and root of Arabidopsis (Stahl and Simon, 2010). However, additional studies have indicated that the regulatory networks associated with shoot and root development also exhibit important differences, particularly in the case of hormonal regulation. Identical hormones or gene families involved in hormonal regulation may interact with tissue-specific molecules and thus can have contrasting effects on the proliferationdifferentiation dynamics in each part of the plant (Kyozuka, 2007). This implies that the regulation of shoot and root developmental processes might differ more than previously thought. Thus, tissue-specific studies on hormonal action may help to clarify to what extent the respective regulatory networks in different organs and tissues are similar in terms of their components and regulatory interactions.
The plant hormones cytokinins (CKs) constitute a class of growth regulators involved in the stress response, senescence, photosynthesis, nutrient assimilation, and mobilization as well as modulation of a plant tissue's ability to act as a sink or source of metabolites (for review, see Werner and Schmülling, 2009). However, the molecular targets associated with the aforementioned processes are still largely unknown. Moreover, the CK-mediated regulation may operate differently in distinct parts of the plant. For example, whereas CKs stimulate shoot apical meristem activity and size (Kurakawa et al., 2007), they act as negative regulators of the proximal root apical meristem (RAM) size, mostly via CK-induced cell differentiation (Dello Ioio et al., 2007). Thus, it is important to identify the mechanisms behind the tissue-specific regulation of CK molecular targets.
CKs are recognized by a subset of sensor His kinases, and the signal is transmitted to the nucleus via a multistep phosphorelay, thus activating the expression of target genes (for review, see Hwang et al., 2012). Previous studies have identified a wide spectrum of CK-regulated genes. CKs have been shown to modulate the expression of genes involved in the control of meristem activity, hormonal cross talk, nutrient acquisition, and various stress responses . Recently, some level of shoot and root specificity was proven in the CK-mediated regulation of the transcript abundance of several genes. However, the vast majority of CK-regulated transcription was found to be similar in the shoot and root . In addition to the CK-mediated regulation of transcription, several recent studies have provided growing evidence for the role of CKs in the posttranscriptional and/or posttranslational regulation of various developmental and physiological processes in Arabidopsis. CKs and CK primary response genes have been shown to regulate the stability of transporters of another plant hormone, auxin (Pernisová et al., 2009;R u zi cka et al., 2009;Marhavý et al., 2011;Zhang et al., 2011). CKs mediate the feedback regulation of CK signaling via CK-mediated proteolysis of B-type ARABIDOPSIS RESPONSE REGULATOR (ARR), ARR2 (Kim et al., 2012), and changes in the proteome and phosphoproteome have been shown to be part of the fast CK response in Arabidopsis ( Cerný et al., 2011). These findings suggest that posttranscriptional regulatory activity may indeed be a mechanism involved in the contrasting behavior and response of the different plant tissues. Thus, in order to study the origin of CK tissue specificity and the nature of hormonal cross talk in different plant organs and tissues, it is important to identify proteins under CK control at a posttranscriptional and/or posttranslational level in a tissue-specific manner. Moreover, whole-genome transcriptomic data sets are now widely available (Rashotte et al., 2003;Brenner et al., 2005;, which can be complemented and compared with proteomic studies. This should help to elucidate the relative contributions of both types of regulation during plant development. In this study, we analyzed 6-benzylaminopurine (BAP)-mediated proteome regulation specifically in the shoot and root of Arabidopsis. We showed that the proteome response to BAP has a strong tissue and time specificity and identified targets of BAP-mediated regulation. In particular, the results revealed that the tissue-specific regulation of hormonal homeostasis is an intrinsic feature of the BAP response. These findings contribute to our understanding of how the regulatory interactions among similar molecules can be modified in a tissue-specific manner and provide a plausible explanation for one of the possible mechanisms behind the long known shoot and root specificity of the CK response in plants.

BAP Uptake Is Comparable in the Shoot and Root
We analyzed the time and tissue specificity of the CK response in 6-d-old seedlings treated with exogenous 5 mM BAP separately in the shoot and root for two different time intervals: 30 min (hereafter referred to as the "early" response) and 120 min of incubation (hereafter referred to as the "delayed" BAP response; designations were according to Brenner et al. [2005]). To ensure that the applied BAP passed through the environmental barrier and conveyed to the examined tissues, we measured the levels of BAP in both tissues and for both time intervals. We did not detect any BAP in mock-treated controls, as expected, but observed similar BAP levels in the BAP-treated shoots and roots ( Fig. 1). Although surface-bound BAP might partially distort the measurements, a slight increase in BAP levels following the 120-min treatment, in contrast to the 30-min treatment, indicated that we predominantly measured internally accumulated BAP. Furthermore, levels of metabolized BAP (hydroxylated BAP derivatives, ortho-and para-topolin) showed a similar response in both tissues following the 30-min treatment, whereas after the 120-min treatment, we observed slightly higher levels in the shoot samples than in the Figure 1. Comparable BAP uptake in shoot and root samples. Levels of BAP and hydroxylated BAP derivatives (meta-, ortho-, and para-topolin, denoted mT, oT, and pT, respectively) were measured in the mocktreated (DMSO) and BAP-treated (5 mM) samples (denoted by -and +, respectively). The measurements were performed in three independent biological replicates. Error bars show SD. FW, Fresh weight. root samples (Fig. 1). Overall, these results indicated similar BAP uptake by the shoot and root, as suggested by the direct measurements of BAP and its metabolic products.

BAP Induces Time-and Tissue-Specific Proteome Changes in Arabidopsis
In order to understand the molecular mechanisms behind the BAP response at the protein level, we analyzed the proteome of BAP-treated seedlings. We used two-dimensional gel electrophoresis with fluorescent Sypro Ruby gel staining of both shoot and root samples treated with BAP for the above-mentioned time intervals. Image analysis of the two-dimensional gels (four biological replicas) revealed qualitative (presence or absence of the respective spot) and quantitative changes in the protein abundance in comparison with the mock-treated controls. Individual proteins with changed expression were identified by liquid chromatography-tandem mass spectrometry (for a detailed description of the assay, see "Materials and Methods"). On average, 1,900 protein spots were monitored on gels obtained from shoot isolates, whereas around 1,500 protein spots were detected on gels corresponding to root isolates. Overall, we identified 43/18 (early/delayed response) differentially regulated proteins in the shoot and 31/21 differentially regulated proteins in the root. We classified the proteins according to their type of response and time specificity (Fig. 2); for a complete list, see Supplemental Table S1.
Unexpectedly, we found only one protein regulated by BAP in both tissues (AT4G14880, AtCYS-3A/OASA1/ OLD3), which suggests the high tissue specificity of the Arabidopsis response to BAP. Considering the time specificity, only two proteins in the shoot, AT4G20360 (Rab GTPase homolog E1B) and ATCG00490 (large subunit of Rubisco [RBCL]), were identified as BAP regulated for both time intervals (both proteins were up-regulated at the early time point and downregulated at the delayed time point; Supplemental Table S1). In the root, we did not identify any proteins regulated by BAP in both time intervals.
Thus, following BAP treatment, we observed mostly shoot-and root-specific proteome changes that also depended on the time point.

BAP Regulates Carbohydrate and Energy Metabolism in the Shoot
To gain an insight into the biological role of the BAPregulated proteins in both tissues, we categorized the BAP-regulated proteins according to the classification introduced by Bevan et al. (1998) and mapped the proteins in the annotated pathway maps provided by the Kyoto Encyclopedia of Genes and Genomes (KEGG; http://www.genome.jp/kegg/). We were able to map 68 of the 104 proteins (65%) identified in this study (Supplemental Figs. S1 and S2; Supplemental Tables S1 and S2). Protein classification according to the functional categories of Bevan et al. (1998) revealed that the major portion of BAP-regulated proteins (both up-and downregulated) in both tissues were proteins involved in processing metabolites of different pathways, such as amino acid, nitrogen, nucleotide, sugar, and lipid metabolism (Fig. 2B).
In the shoot, the second major class of BAP-regulated proteins contained proteins involved in energy-associated (energy acquisition or storage) processes, such as glycolysis, gluconeogenesis, and photosynthesis. Accordingly, using the KEGG database, we mapped most of the BAP-regulated proteins in the shoot to carbohydrate metabolism and energy acquisition (Supplemental Fig.   Figure 2. BAP application induces specific proteome changes in the shoot and root. A, Classification of BAP-regulated proteins according to the type of response and its tissue and time specificity. Qualitative (presence or absence of the protein in comparison with the mocktreated control) and statistically significant changes in the protein abundance (Student's t test; four biological replicates) were evaluated separately for both tissues and assayed time intervals. The number inside the chart shows the total number of identified proteins in each category. B, Functional distribution of proteins (both up-and downregulated) by BAP in shoot and root samples. The color code represents the functional classification according to Bevan et al. (1998). Table S2). The roles of these proteins have been annotated in basic metabolic pathways, including starch and Suc metabolism, glycolysis, and gluconeogenesis (Supplemental Fig. S1B), pyruvate metabolism (Supplemental Fig. S1C), carbon fixation (Supplemental Fig. S1D), citrate cycle (Supplemental Fig. S1E), and oxidative phosphorylation, photosynthesis, and chlorophyll metabolism (Supplemental Table S2). Shoot-identified proteins that were not mapped but have previously been described in the literature also indicate CK control over protein import in chloroplasts (AT5G16620, PDE120/ATTIC40; Bédard et al., 2007) or control of meristem size via correct folding and/or complex formation of CLV proteins (AT4G24190, SHD/AtHSP90.7; Ishiguro et al., 2002).

S1A; Supplemental
With regard to the time specificity of the shoot response, most of the aforementioned processes were found to be predominantly regulated at the earlyresponse time point, except for purine, porphyrin, and chlorophyll metabolism, which were found to be regulated during the delayed response (Supplemental Fig. S1).

BAP Targets Protein Synthesis and Destination in the Root
As in the shoot, the most abundant category of proteins regulated by BAP in the root related to metabolism. However, in contrast to the shoot response, the second most represented category in the root constituted proteins involved in protein destination and storage, such as protein folding, targeting, modification, and proteolysis. At the same frequency, BAP regulated proteins in the root associated with protein synthesis, like ribosomal proteins and tRNA synthases (Fig. 2B). With respect to KEGG mapped pathways, one of the most affected processes in the root was RNA transport (Supplemental Fig. S2A). Four of the BAP-regulated proteins have been annotated as proteasome components, three of which we identified in the root (Supplemental Fig. S2B). Several other BAP-regulated proteins in the root have been annotated as proteins involved in the proteasome-mediated regulation of protein stability (AT1G16190, AT5G22610), in protein folding (AT3G03960, AT3G18190), or in protein processing in the endoplasmic reticulum (ER; Supplemental  Fig. S2C). Regarding metabolic processes, proteins involved in amino sugar and nucleotide sugar metabolism (Supplemental Fig. S2D), as well as Cys and Met metabolism (Supplemental Fig. S2E), were also found to be predominantly regulated in the root. This suggests a role for BAP in regulating Met-related metabolism in the root (see below). Previously characterized nonmapped proteins connect BAP action in the root with, for example, the organization of cortical microtubule arrays and cellular patterning (AT3G55000, TONNEAU1; Azimzadeh et al., 2008) or the regulation of telomere-length homeostasis and salicylic acid-dependent disease resistance (AT1G14410, ATWHY1/PTAC1; Desveaux et al., 2004;Yoo et al., 2007). This is consistent with recently published data on CK cross talk with the salicylic acidmediated immune response Argueso et al., 2012).
Similar to the shoot, most of the BAP-regulated proteins in the root were found to be differentially regulated during the early response, except for proteasomemediated protein degradation or amino sugar and nucleotide sugar metabolism, which were mainly targeted during the delayed response (Supplemental Table S2).
In both shoot and root, we found a substantial proportion of BAP-regulated proteins to be stress related, including those involved in ascorbate metabolism (Supplemental Fig. S1F; Supplemental Table  S1) and others related to the cold or heat response (Supplemental Table S1).
Altogether, our results imply that the proteome response to BAP in Arabidopsis is tissue and time specific. The BAP-mediated proteome regulation appears to be important, particularly as an immediate regulatory mechanism, because of the higher number of differentially regulated proteins during the early response in both assayed tissues. Mapping the annotated activities of the identified proteins suggested a predominant role for BAP-mediated regulation in carbohydrate and energy metabolism within the shoot and in protein synthesis and protein processing within the root.

BAP-Mediated Proteome Regulation Occurs Dominantly at the Posttranscriptional and/or Posttranslational Level
CKs have been identified as effective regulators of gene transcriptional activity. Thus, we wanted to identify the contribution of CK-mediated transcriptional regulation to the observed proteome response. Therefore, we selected a subset of BAP-modulated proteins from our data set, in particular, proteins covering the two tissues and time points whose function had already been described. We analyzed BAP-mediated changes in the abundance of 21 mRNAs using quantitative realtime PCR. Two out of the 11 genes assayed in the shoot and two out of 10 in the root exhibited the same type of response as their protein products, whereas nine out of 11 genes in the shoot and eight out of 10 in the root displayed no statistically significant transcriptional response (Fig. 3, A and B). This agrees well with the results of previously performed transcriptional studies (Rashotte et al., 2003;Brenner et al., 2005;, demonstrating only minimal or no overlap with our proteome data set (data not shown). In the case of AtPHB3, one of the proteins exhibiting no transcriptional response, the posttranscriptional and/or posttranslational nature of BAPmediated regulation was confirmed by immunostaining ( Fig. 3C).
Overall, our results demonstrate only limited overlap between regulation at the transcript and protein abundance levels and suggest a substantial role for posttranscriptional and/or posttranslational regulation in the proteome response to BAP.

BAP Regulates Proteins Involved in Hormonal Metabolism
In our proteomic analysis, we detected several BAPregulated proteins that were mapped in diverse steps of plant hormone biosynthesis ( Fig. 4; Supplemental Fig. S4). In general, we found that modulation of the ethylene biosynthetic pathway was the most prominent. Ethylene is biosynthesized in four catalytic steps, and it has been shown previously that CKs stabilize ACC SYNTHASE (Chae et al., 2003;Hansen et al., 2009), catalyzing the third step in this pathway. Here, we identified that all three remaining enzymes of the ethylene biosynthetic pathway were up-regulated by BAP: MET SYNTHASE1 (AtMS1; AT5G17920), MET ADENOSYLTRANSFERASE3 (MAT3; AT2G36880), and Figure 3. Comparison of the transcriptional type of regulation with the proteome response. A and B, BAP responses of selected genes encoding proteins revealing BAP-mediated changes in the shoot (A) and root (B); note logarithmical scale of the y axes. The statistical significance of differences between mock-treated (2) and BAP-treated (+) samples (Student's t test) at a = 0.05 and 0.01 is denoted by single and double asterisks, respectively. Error bars indicate SD. C, A 1.2-fold increase in AtPHB3 abundance was detected in the independent isolation by immunodetection using AtPHB3 antibodies compared with a 1.5-fold increase identified by Sypro Ruby staining. The histone H3A antibody was used as an internal control for total protein amount normalization. Figure 4. BAP regulates endogenous hormone metabolism specifically in the shoot and root. A, Overview of BAP-regulated proteins participating in the metabolism of endogenous hormones. Proteins involved in the MVA pathway and ethylene biosynthesis are highlighted as blue and orange ovals, respectively. B, BAP tightly regulates ethylene biosynthesis in the root. In addition to the enzymes identified as up-regulated by BAP in our study, CK-mediated stabilization of a subset of ACS has been demonstrated previously (Chae et al., 2003;Hansen et al., 2009 4B). Importantly, we found that the enzymes were up-regulated specifically during the early response of the root, indicating that BAP-mediated up-regulation of ethylene biosynthesis is very fast and tissue specific. It is worth noting that we identified another MAT isoform, MAT4, that was down-regulated in the delayed shoot response, thus further contrasting the tissue-specific BAP regulation of the ethylene biosynthetic pathway (Supplemental Fig. S3; Supplemental Table S1). Furthermore, in the case of AtMS1, we found that the BAP-mediated up-regulation occurs at the transcriptional level (Fig. 3B).
Besides ethylene biosynthesis, we found that BAP regulates proteins belonging to the mevalonate acid (MVA) pathway, namely AT2G30200 (EMB3147) and AT4G11820 (MVA1; Fig. 4A). This pathway is generally important in hormone biosynthesis; it is crucial for brassinosteroid synthesis and is also involved in the biosynthesis of gibberellin (Kasahara et al., 2002), CKs (Kasahara et al., 2004), and possibly abscisic acid (ABA; Milborrow, 2001; Fig Table S1). Also, acetyl-CoA production, which is a key metabolite initiating the MVA pathway, seems to be under strong positive control by BAP during the early response of the shoot (Supplemental Figs. S1, B and E, and S3). In agreement with this, we found several ABA-related proteins to be CK regulated predominantly in the shoot. Among the nine ABA-related proteins found in our data set, seven were identified in the shoot, five of which were up-regulated in the early shoot response (Supplemental Table S3). Three (RBCL, FBA2, and CRU3/CRC) have been shown to be Tyr phosphorylated in response to ABA (Ghelis et al., 2008), and transcription of AtCYS-3A/OASA1/OLD3 is reportedly increased by ABA (Barroso et al., 1999). In agreement, we observed both up-regulation of AtCYS-3A/OASA1/ OLD3 in the shoots of CK-treated seedlings and also that CRU3 was under distinctive CK-mediated transcriptional control during the early shoot response (Supplemental Fig. S3A).
Besides ethylene-and ABA-related proteins, we also identified several BAP-regulated proteins that are potentially associated with CK biosynthesis (Supplemental Fig. S3). Specifically, AT1G58080, AT3G22960, and AT5G35170 regulate the production of AMP, ADP, and ATP (Supplemental Fig. S3; Supplemental Table  S2), which are substrates of adenosine phosphateisopentenyltransferases in the first step of CK biosynthesis. This is consistent with previously published data showing that CKs can regulate their own biosynthesis (Sakakibara et al., 2006, and refs. therein).
Taken together, our results show that BAP regulates several proteins involved in hormone biosynthesis or associated with hormone-mediated regulation and that BAP modulation of hormonal regulation exhibits significant tissue specificity. We identified strong, rapid up-regulation of proteins catalyzing ethylene biosynthesis in the root and up-regulation of the MVA pathway and several ABA-related proteins mainly in the shoot.

BAP Treatment Affects the Tissue-Specific Distribution of Endogenous Hormones
Results of our proteomic analysis suggested potential tissue-specific cross talk between CKs and the biosynthesis or action of other phytohormones. Therefore, we examined the BAP-mediated changes in endogenous levels of 1-aminocyclopropane-1-carboxylic acid (ACC), the rate-limiting precursor of ethylene biosynthesis, ABA, and endogenous CKs and their metabolites, again for both inspected tissues and time points.
ACC distribution and its CK-dependent regulation revealed remarkable tissue specificity. In untreated controls, levels of ACC in the root were below the detection limit but increased substantially upon CK application. By contrast, levels of ACC in the shoot were almost unaffected by CK treatment (Fig. 5A). The root-specific increase in ethylene production was further confirmed by the staining of the ethyleneresponsive reporter line carrying GUS under the control of four synthetic copies of the EIN3-binding site (EBS; Solano et al., 1998;Supplemental Fig. S7). This is notably in agreement with our results suggesting tight CK-mediated control over ethylene biosynthesis specifically in the root (Fig. 4B). In the case of ABA, we detected slightly higher levels in mock-treated roots than shoots. Following BAP treatment, we found significant up-regulation of ABA levels, particularly in the early response of the shoot (Fig. 5A). In the root, we identified a slight (statistically insignificant) decrease in the ABA amount for both time points. These ABA measurements corroborate with the positive BAP regulation of the MVA pathway and with acetyl-CoA biosynthesis in the shoot and negative regulation of the MVA pathway in the root (Supplemental Figs. S1E and S3).
Endogenous levels of the free CK bases trans-zeatin (tZ), isopentenyladenine (iP), and cis-zeatin (cZ) also displayed strong tissue specificity (Fig. 5B), which was most pronounced in the case of cZ. The amount of cZ in the root was approximately 10 times higher than the amount detected in the shoot. BAP treatment even enhanced the asymmetric distribution of cZ during the delayed response, leading to its down-regulation in the shoot and up-regulation in the root (Fig. 5B). Furthermore, we noticed that cZ was the dominant CK in the root, with levels six to 10 times higher than for tZ. This cZ dominance was even more apparent in the case of CK nucleosides, where we found about 15 to 185 and 15 to 40 times higher levels of cZ riboside in comparison with tZ riboside and iP riboside, respectively (Supplemental Fig. S4). Interestingly, as for cZ, an asymmetry favoring root accumulation was detected for dihydrozeatin riboside but not for tZ riboside. For iP and tZ, a consistent type of regulation was found in both tissues and time intervals assayed, whereas opposing types of responses were observed for both CK types: While the levels of tZ significantly decreased in both tissues upon the addition of BAP, iP was slightly upregulated or not affected by BAP in both root and shoot (Supplemental Fig. S2, B and C, respectively).
Collectively, these results agree with our proteomic data suggesting tissue-specific BAP-mediated hormonal cross talk. Our results provide experimental evidence for tight BAP-mediated control of ethylene biosynthesis in the root and ABA biosynthesis in the shoot. The significant tissue specificity is also apparent in the distribution of endogenous CKs as well as in the BAP-mediated regulation of endogenous CK metabolism.

Proteins Catalyzing BAP-Induced Ethylene Biosynthesis Mediate the Root Response to BAP
The roles of both CK and ethylene in mediating root shortening have been described (Cary et al., 1995;R u zi cka et al., 2007). Our proteomic analysis identified that three enzymes catalyzing the ethylene biosynthetic pathway are promptly up-regulated by BAP specifically in the root. This implies that BAP-induced ethylene may contribute to the immediate root growth response via tight control over ethylene production. To test this, we examined BAP-induced root shortening of mutants in the ATMS1 and ACO2 genes, encoding the first and last proteins of the ethylene biosynthetic pathway (Fig. 4B). To approximate the physiological assay to the early CK response, as studied in our proteome analysis, we employed the "seedling transfer" assay, where 6-d-old seedlings were grown on pure Murashige and Skoog (MS) medium and then transferred to MS medium supplemented with BAP and cultivated for a further 2 d. In close agreement with our model, both atms1 and aco2 mutants exhibited higher resistance to BAP-induced root shortening than the wild type. This effect was apparent by only day 1 after transplanting the seedlings from pure MS medium to medium supplemented with BAP (Fig. 6A). Furthermore, we compared these results with those for a mutation in the gene encoding ACS9, which has previously been shown to be stabilized by CKs (Hansen et al., 2009). Interestingly, acs9 did not display any significant changes in root growth during day 1 of BAP incubation, and differences were only apparent after 2 d of growth in the presence of BAP (Fig. 6A). This agrees well with the lack of ASC9 stabilization by BAP in our experimental system.
CKs have been shown to control root growth via regulation of the RAM size (Dello Ioio et al., 2007;R u zi cka et al., 2009), whereas ethylene is reported to act specifically on the regulation of cell elongation (R u zi cka et al., 2007). Thus, to elucidate the mechanism of the observed influence of ethylene biosynthesis on the CK-mediated root shortening, we examined the effects of BAP on the RAM size and cell elongation in the wild type and ethylene biosynthetic mutants. Surprisingly, we found that both atms1 and aco2 lines exhibited resistance to the BAP-mediated reduction of RAM. However, this was not the case in the acs9 line (Fig. 6B). In our seedling transfer assay, we were unable to detect any reduction in the elongation of cells leaving the RAM in the BAP-treated wild type and acs9 mutants. In contrast, we observed a large reduction in root cell elongation in atms1 and a smaller but still distinct reduction in the aco2 line (Fig. 6B). Notably, the mock-treated atms1 mutant had longer cells leaving the RAM than the wild type (Supplemental Fig. S6).
Altogether, these findings correlate well with our proteomic data, suggesting that BAP rapidly up-regulates Figure 5. Exogenous BAP application affects levels of endogenous hormones and their metabolites differentially in the shoot and root. A, Root-and shoot-specific changes of ACC and ABA levels, respectively. B and C, Endogenous CKs and their metabolites exhibit tissue-specific distributions and responses to exogenous BAP. The statistical significance of differences between mock-treated (2) and BAP-treated (+) samples (Student's t test) at a = 0.05, 0.01, and 0.001 is denoted by single, double, and triple asterisks, respectively. Error bars indicate SD. cZOG, Cis-zeatin-O-glucoside; FW, fresh weight; iP9G, isopentenyladenide-N9-glucoside; tZOG, trans-zeatin-O-glucoside. several members of the ethylene biosynthetic pathway that cause root-specific ACC production and interfere with the BAP effect on root growth (Fig. 7). Our data also provide evidence for the unexpected role of ethylene biosynthesis in the BAP-mediated reduction of the RAM size.

BAP-Mediated Proteome Changes Reveal Root-and Shoot-Specific Targets of CK Action
Here, we identified a set of proteins that respond to BAP application in a tissue-and time-specific manner. In the shoot, the identification of BAP-mediated control over proteins that are potentially involved in the regulation of basic metabolic processes, such as carbon fixation and chlorophyll, Suc, and starch metabolism, provides a molecular link to the previously shown CKmediated regulation of photosynthesis (Wareing et al., 1968), chloroplast biogenesis (Reski, 1994), chlorophyll metabolism (Mok, 1994), and sink and source strength of plant tissues (Werner et al., 2008).
In the root, the processes mainly affected by BAP include the regulation of protein synthesis and destination, protein processing in the ER, and proteasomemediated degradation. This supports recent evidence suggesting that CKs have an important role in protein targeting and proteasome-mediated protein stability. CK-induced internalization of the auxin transporter PIN1 during de novo organogenesis has been reported (Pernisová et al., 2009). CKs have been suggested to regulate the stability of diverse PIN proteins (R u zi cka et al., 2009;Zhang et al., 2011) and vacuolar targeting of PIN1 (Marhavý et al., 2011). The 26S proteasome subunit RPN12 has been implicated in sensitivity to CKs (Smalle et al., 2002), potentially regulating CK signaling via the stabilization of ARR5 (Ryu et al., 2009). The type-B response regulator ARR2 has also recently been found to be under the control of the 26S proteasome (Kim et al., 2012). In addition, the 26S proteasome regulatory subunit RPN10 was proposed to mediate the degradation of CK and/or auxin response repressors in Physcomitrella patens (Fu et al., 1999). Indeed, we found both core (20S) and regulatory subunits of the 26S proteasome to be under CK control (Supplemental Fig. S2B), suggesting potential feedback in the CK signaling via CK-mediated regulation of the 26S proteasome activity. Finally, recently identified localization of CK receptors in Figure 6. Relative root lengths of mutants in ethylene biosynthesis in response to BAP. Col-0 (always grown on the same plate) is shown in comparison with all mutants to demonstrate the SD used in Student's t test calculations. A value of 100% corresponds to the root length of the respective mock-treated controls (mean 6 SD). The relative root lengths are given as follows: 4.33 6 0.58, 4.27 6 0.81, and 4.77 6 0.65 mm (day 1; Col-0 controls to atms1, aco2, and acs9-1, respectively) and 10.82 6 1.10, 9.68 6 1.45, and 10.38 6 1.05 mm (day 2; Col-0 controls to atms1, aco2, and acs9-1, respectively); for atms1, 4.95 6 1.17 mm (day 1) and 11.64 6 1.83 mm (day 2); for aco2, 4.67 6 0.54 mm (day 1) and 10.34 6 0.89 mm (day 2); for acs9-1, 5.93 6 0.93 mm (day 1) and 13.11 6 1.42 mm (day 2). B, Relative root meristem lengths and lengths of cells that just left the transition zone of 100 nM BAP-treated plants in comparison with mock-treated plants (2 d of incubation); the absolute values are shown in Supplemental Figure S6. The statistical significance of differences between Col-0 and the mutant lines following 100 nM BAP application (Student's t test) at a = 0.05, 0.01, and 0.001 is denoted by single, double, and triple asterisks, respectively. the ER (Caesar et al., 2011;Wulfetange et al., 2011) supports our results suggesting CK control over protein processing in the ER.
Taken together, our data reveal several novel regulators influenced by BAP in a tissue-specific way that may be potential molecular targets of the CK-mediated regulation of several processes previously identified to be under CK control.

Nontranscriptional Regulations Are Involved in the Proteome Response to BAP
The proteins identified in this study as BAP regulated showed little or no overlap with transcripts that have been previously identified as CK regulated (Rashotte et al., 2003;Brenner et al., 2005), including a recently published study on the shoot-and root-specific transcriptional profiling of CK action in Arabidopsis . These results are consistent with several reports showing little or no correlation between transcript and protein levels at a given time point in several model organisms and experimental setups (Foss et al., 2007;Fu et al., 2009;Taniguchi et al., 2010;Ghazalpour et al., 2011;Ning et al., 2012). Differences in mRNA and protein lifetimes and/or extrinsic translational noise or regulatory networks may account for the observed differences in regulation at the transcript and protein levels (Fu et al., 2009;Taniguchi et al., 2010).
Accordingly, in the subset of proteins that we found to be regulated by BAP, we observed corresponding changes of the respective transcripts in only a minor (approximately 20%) fraction (Fig. 3). This clearly suggests the involvement of posttranscriptional and/or posttranslational regulation in the tissue-specific BAP regulation of the Arabidopsis proteome and implies that the transcriptional and nontranscriptional regulation of protein abundance might be rather complementary. Nevertheless, further studies are required to perform a statistical comparison of the transcript and protein abundance for different treatments as well as to search for long-term spatiotemporal correlations, both at the cell and tissue levels. The observed BAPmediated regulation of proteins involved in RNA transport, initiation of translation, or protein degradation, as evident in our data set (Supplemental Fig.  S2, A and B), revealed potential BAP targets for this type of regulation. However, identification of the molecular mechanisms underlying the BAP-mediated control over protein abundance remains a considerable challenge for future work.

BAP Affects Endogenous Hormone Levels Specifically in the Shoot and Root
Importantly, we also obtained experimental evidence that differential proteome responses underlie the shootand root-specific BAP regulation of hormone biosynthesis, the clearest example being the regulation of ethylene production. Our data show that BAP mediates very fast up-regulation of the ethylene biosynthetic pathway, resulting in ACC accumulation specifically in the root. It is noteworthy that we did not observe the previously reported CK-mediated stabilization of ACC synthases, neither ACS5 nor ACS9 (Chae et al., 2003;Hansen et al., 2009), possibly due to a lower detection limit or the different experimental design of our system. This is in a good agreement with our physiological analysis, suggesting that in comparison with ATMS1 and ACO2, ACS9 appears to play a relatively minor role in the early root growth response to BAP (see below). In our study, however, all three remaining enzymes of the ethylene biosynthetic pathway, AtMS1, MAT3, and ACO2, were up-regulated by BAP in the root, indicating their crucial role in the early BAP response (Fig. 7). These results are complemented and supported by a recent study on ethylene-mediated proteome changes. Among eight proteins identified to be regulated by both BAP (this study) and ethylene (Chen et al., 2011), five were identified as BAP regulated in the root (Supplemental Table S4). Nevertheless, despite the fact that the root-specific response at the level of ACC biosynthesis predominated under our experimental conditions, the involvement of a cell type-specific, ethyleneassociated CK response in the delayed developmental regulation of the shoot cannot be excluded, as reported previously (Cary et al., 1995;Tanaka et al., 2006).
The observed up-regulation of ABA-related proteins in the shoot might be a consequence of the CK-mediated up-regulation of the MVA pathway and increase of ABA levels in the early shoot response (Fig. 7). Although it has been reported that the methyl erythritol phosphate pathway in chloroplasts constitutes the main branch of the ABA biosynthetic pathway, the chloroplast import of cytoplasmic isopenthenyl diphosphate produced via the MVA pathway still seems to contribute to ABA production (for review, see Milborrow, 2001). Thus, in addition to the recently reported positive role of CKs in the regulation of ABA biosynthesis at the transcriptional level (Nishiyama et al., 2011), proteome regulation by CKs could represent another mechanism of CK and ABA cross talk. The BAP-mediated up-regulation of proteins that are Tyr phosphorylated in response to ABA suggests that this type of posttranslational modification might have a stabilizing effect. This, however, remains to be clarified.
We also identified a strong asymmetry in the distribution and response of endogenous CKs and their metabolites. The most pronounced asymmetry was observed in the case of cZ and its ribosides, which we found to be the predominant CK in Arabidopsis root. Indeed, slightly higher concentrations of cZ than tZ have previously been found in the roots of hop (Humulus lupulus; Watanabe et al., 1982) and maize (Zea mays; Veach et al., 2003;Saleem et al., 2010). This tissuespecific distribution probably did not play an important role under our experimental conditions due to the relatively high concentration of exogenously applied BAP (compare CK levels in Figs. 1 and 5). However, the recently identified specificity of the CK-degrading enzymes cytokinin oxidases/dehydrogenases (CKX; Gajdošová et al., 2011) and the diverse ability among CK types, including cZ, to bind individual CK receptors suggest that this asymmetrical distribution might indicate a potential functional importance in CK-mediated developmental control. This is further supported by our findings demonstrating the higher root accumulation of dihydrozeatin riboside. Dihydrozeatin and its derivatives are not substrates for CK degradation by CKX (Sakakibara, 2010) but are still able to activate CK signaling via AHK3 (Spíchal et al., 2004), providing further evidence for the potential tissue specificity of CKXmediated developmental regulation, in agreement with previous observations (Werner et al., 2003).

Ethylene Biosynthesis Is Involved in the Early Root Growth Response to BAP
Root growth is regulated at the level of mitotic activity of the cells in the RAM, cell differentiation of the meristematic cells leaving the RAM in the transition zone, and elongation of cells that undergo differentiation outside the RAM. These events are integrated, with the balance between cell division and cell differentiation determining the RAM size (Dello Ioio et al., 2007). In particular, CKs have been shown to reduce RAM size via the activation of cell differentiation in the transition zone (Dello Ioio et al., 2007). Using ethylene signaling mutants, this CK effect has been suggested to be ethylene independent (R u zi cka et al., 2009). However, this was observed in Arabidopsis seedlings germinated and grown on medium supplemented with BAP. In our short-term experimental system via transplanting the seedlings from pure MS medium to medium supplemented with BAP, we have shown that both atms1 and aco2 mutants exhibit resistance to the BAP-mediated shortening of the RAM. This rather surprising finding suggests that both ATMS1 and ACO2 are necessary for the immediate root growth response at the level of BAPmediated regulation of RAM size. Nevertheless, further detailed analysis will be necessary to uncover the underlying molecular mechanisms of observed cross talk between CK action and ethylene biosynthesis in the regulation of RAM size.
In contrast to CKs, ethylene has been demonstrated to inhibit cell elongation in the root (R u zi cka et al., 2007). However, similar to the aforementioned CKmediated regulation of RAM size, this was observed during long-term cultivation of Arabidopsis seedlings in the presence of excessive amounts of exogenous ACC. Under the conditions of our assay, we were unable to detect any BAP effects on cell elongation in the wild type and the acs9 line, whereas aco2 and particularly atms1 exhibited large reductions in cell length. These results suggest that ethylene may interfere with immediate CK effects on root growth even at the level of cell elongation. The increased length of cells in mock-treated atms1 is consistent with the negative role of ethylene in root elongation and suggests that a physiologically effective amount of ethylene was still present in the root. This is in agreement with our analysis of the ethylene-responsive reporter line (EBS:GUS), revealing GUS activity even in the mock-treated root (Supplemental Fig. S7). Consistently, the presence of 2-aminoethoxyvinylglycine, an inhibitor of ethylene biosynthesis, has also been shown to increase root cell length (R u zi cka et al., 2007).
In summary, our findings suggest that CKs and ethylene are tightly interconnected in their regulatory roles on root growth and imply that endogenous ethylene homeostasis is important for the CK-mediated regulation of individual RAM parameters (Fig. 7).

Plant Growth Conditions
For proteome analysis, quantitative real-time PCR, and hormone measurements (CK, ABA, and ACC levels), seedlings were cultivated on square plates with nylon mesh (Uhelon 120T; Silk & Progress) positioned vertically in growth chambers (Percival Scientific; CLF Plant Climatics) under a 16/8-h light/dark photoperiod at 150 mmol m 22 s 21 as described previously (Pernisová et al., 2009). Six-day-old seedlings were treated with liquid MS medium containing 5 mM BAP (Duchefa) for 30 min or 2 h. For mock-treated samples, we used liquid MS medium supplemented with 0.1% dimethyl sulfoxide (DMSO). The treatment was carried out under gentle stirring in the growth chambers (Percival Scientific) under the same conditions as seedling cultivation. After the treatment, roots and shoots were detached with a scalpel, collected, immediately frozen in liquid nitrogen, and analyzed separately. For the proteome analysis, four independent biological replicates were performed, whereas for the quantitative real-time PCR and hormone measurements, three independent biological replicates were used.
In the root elongation assay and GUS staining assay, 6-d-old seedlings grown on vertical petri dishes and 13 MS medium were replanted into 13 MS medium supplemented with 100 nM BAP (Duchefa; dissolved in DMSO) and 0.01% DMSO as a control. After 2 d of incubation, root elongation was measured (ImageJ; National Institutes of Health; http://rsb.info.nih.gov/ij) and compared with the wild type (Col-0). At least 20 seedlings were evaluated in at least three independent experiments. An Olympus BX61 microscope using differential interference contrast microscopy was used for measurements of RAM lengths and root cell lengths. Root cell lengths were measured in the first fully elongated cortex cells following the transient zone. GUS staining was performed as described previously (Hejátko et al., 2003).

Proteomic Analysis
For two-dimensional gel electrophoresis, proteins were isolated separately from approximately 0.5 g fresh weight of the shoots and roots using the 10% TCA/acetate extraction method (Tsugita and Kamo, 1999). For isoelectric focusing (IEF; the first dimension of two-dimensional electrophoresis), vacuum-dried protein isolates were dissolved in IPG buffer (7 M urea, 2 M thiourea, 2% [w/v] CHAPS, 60 mM dithiothreitol [DTT], 0.8% Biolyte 3/10 Ampholyte [Bio-Rad], and 0.003% bromphenol blue) and the protein concentration was determined using a RC/DC kit (Bio-Rad). Solubilized protein samples were subjected to centrifugation at 20,000 relative centrifugal force for 60 min at 10°C before application onto IPG strips. A total of 150 mg of whole protein sample in 315 mL of IPG buffer per IPG strip was applied to 18-cm ReadyStrip IPG strips, pH 3 to 10 nonlinear (Bio-Rad), by passive rehydration. IEF was performed in a Protean IEF Cell (Bio-Rad) for 80,000 Vh. Prior to the second dimension, the IPG strips were equilibrated in buffer (6 M urea, 0.375 M Tris-HCl, pH 8.8, 2% SDS, 20% glycerol, and 2% DTT) for 10 min, followed by 10 min in a second equilibration buffer containing 2.5% iodoacetamide instead of DTT.
For SDS-PAGE (the second dimension of two-dimensional electrophoresis), 12% vertical polyacrylamide gels were employed in a Protean Plus Dodeca Cell (Bio-Rad). The Precision Plus Protein Standard (Bio-Rad) was applied to each gel. Gels were stained with Sypro Ruby (Invitrogen) fluorescent dye according to the manufacturer's recommendations. Gels were scanned by FLA-7000 Fluorescent Image Analyzer (Fuji Film).
Image analysis of protein maps was performed using PDQuest 8.0.1 Advanced software (Bio-Rad). The Spot Detection Wizard was used to select the parameters for spot detection; a faint spot (the sensitivity and minimum peak value parameter) and a large spot cluster (the radius of the background subtraction) were selected. Gel warping was carried out prior to spot matching. The results of automated spot detection and matching were checked and, if necessary, manually corrected. On average, 1,900 protein spots were monitored on gels of shoot isolates, whereas around 1,500 protein spots were detected on gels of root isolates. Protein spots with an intensity at least 10-fold higher than the gel background on at least three out of four gel replicates were considered for further evaluation. A local regression model (Loess) was used for spot intensity normalization. For quantitative differentiation, a 1.5-fold change or higher in the average spot intensity between compared samples was regarded as significant. Statistical significance of differences was assessed using Student's t test at a significance level of 0.05 in four biological replicates. Individual protein spots selected on the basis of image-analysis output as those showing significant changes were excised using a spot cutter (Bio-Rad), in-gel digested with trypsin, and identified by liquid chromatography-tandem mass spectrometry (for details, see Supplemental Materials and Methods S1). All identified proteins in qualitatively different spots were considered (Supplemental Table S1A). Only spots with a single identified protein were considered for quantitative evaluation (Supplemental Table S1B).

Hormonal Analysis
For endogenous CK analysis, extraction and purification of approximately 0.2 g fresh weight was performed according to the method described previously, and levels of CKs were quantified by ultra-performance liquid chromatography-electrospray tandem mass spectrometry. For ABA and ACC measurements, samples were purified by solid-phase extraction and then derivatized and analyzed by gas chromatography-tandem mass spectrometry. For details see Supplemental Materials and Methods S1.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Mapping of BAP-regulated proteins in the processes dominantly affected in the shoot.
Supplemental Figure S2. Mapping of BAP-regulated proteins in the processes dominantly affected in the root.
Supplemental Figure S4. Exogenous BAP application modulates levels of endogenous CKs and their metabolites.
Supplemental Figure S5. Typical two-dimensional electrophoresis maps showing the quality of gel separation.
Supplemental Figure S6. RAM lengths and root cell lengths of mutants in ethylene biosynthesis in response to BAP.
Supplemental Figure S7. BAP-mediated ethylene production in ethyleneresponsive reporter line EBS:GUS in the shoot (A) and root (B).
Supplemental Table S2. Overview of mapping BAP-regulated proteins in the KEGG database.
Supplemental Table S3. ABA-related proteins identified as BAP regulated in the shoot and root.
Supplemental Table S4. Overlay of proteins identified in our study with the ethylene-mediated proteome response in the whole seedling as identified by Chen et al. (2011).
Supplemental Materials and Methods S1.