Cytokinin oxidase/dehydrogenase genes in barley and wheat

Cloning and heterologous expression


P. Galuszka, Division of Molecular Biology, Department of Biochemistry, Faculty of Science, Palacký University, Šlechtitelů 11, 783 71 Olomouc, Czech Republic. Fax: +420 58 5634933, Tel.: +420 58 5634929, E-mail:


The cloning of two novel genes that encode cytokinin oxidase/dehydrogenase (CKX) in barley is described in this work. Transformation of both genes into Arabidopsis and tobacco showed that at least one of the genes codes for a functional enzyme, as its expression caused a cytokinin-deficient phenotype in the heterologous host plants. Additional cloning of two gene fragments, and an in silico search in the public expressed sequence tag clone databases, revealed the presence of at least 13 more members of the CKX gene family in barley and wheat. The expression of three selected barley genes was analyzed by RT-PCR and found to be organ-specific with peak expression in mature kernels. One barley CKX (HvCKX2) was characterized in detail after heterologous expression in tobacco. Interestingly, this enzyme shows a pH optimum at 4.5 and a preference for cytokinin ribosides as substrates, which may indicate its vacuolar targeting. Different substrate specificities, and the pH profiles of cytokinin-degrading enzymes extracted from different barley tissues, are also presented.


cytokinin oxidase/dehydrogenase


expressed sequence tag


Murashige–Skoog medium



Cytokinins were initially viewed as factors promoting cell division and differentiation in plants. Since then, however, cytokinins have been shown to control other developmental events, such as the growth of lateral buds, the release of buds from apical dominance, leaf expansion, the delay of senescence, the promotion of seed germination, and chloroplast formation [1]. Naturally occurring cytokinins are mainly N6-substituted adenine derivatives that generally contain an isoprenoid or aromatic side-chain. Recently, considerable progress has been made in elucidating the regulation of cytokinin homeostasis during plant growth and development. New molecular biological techniques have allowed for the identification and characterization of genes encoding important enzymes participating in cytokinin metabolic pathways. Genetically engineered plants that overexpress some of these genes were prepared as a tool to study changes in physiological aspects caused by altered cytokinin levels. Seven genes for isopentenyltransferases – cytokinin de novo synthesizing enzymes – were identified in the Arabidopsis genome [2–4]. In addition, three novel genes, encoding cytokinin-specific glycosylation enzymes with different substrate specificities, have been described [5–7]. The principle of cytokinin catabolism has been studied for many years. Enzymes capable of degrading cytokinins with unsaturated side-chains have been found in many plant tissues [8], but the details of their features and the mechanism of their action remained unknown for a long time owing to their very low content in plant tissues. The ground-breaking cloning of the cytokinin oxidase maize gene ZmCKX1[9,10] opened up the possibility for more detailed study of cytokinin degradation, both at the molecular and at the biochemical levels. The recombinant maize enzyme is a glycoprotein containing a covalently bound FAD. The isoprenoid side-chain of the cytokinin molecule is most efficiently cleaved in the presence of an electron acceptor other than oxygen. Hence, the enzyme has been classified as a dehydrogenase with a new EC [11]. The detailed reaction mechanism of cytokinin oxidase/dehydrogenases (CKX) has recently been presented for the conversion of different types of cytokinin substrates [12]. Studies of reaction rates have revealed that oxygen is unlikely to be the physiological acceptor reoxidizing the FAD molecule of the enzyme in vivo. The exact characteristics of a naturally cooperating electron acceptor are still unknown, but experiments in vitro indicate that it might be p-quinone or a molecule with a similar structure [12].

The completed sequencing project of Arabidopsis and rice genomes allowed identification of the small CKX gene family of seven homologues in Arabidopsis (AtCKX1 to AtCKX7[13]) and 11 in rice [14]. Six AtCKX genes were individually overexpressed in tobacco or Arabidopsis plants, and a detailed phenotypic characterization was subsequently carried out. All transformants displayed reduced cytokinin content and showed distinct developmental alterations in the shoot and root [15,16], most of them in accordance with previous assumptions on cytokinin function. Two of the AtCKX proteins were found to be targeted to the vacuoles, while another accumulated in the reticulate structure, which may indicate its final extracellular localization [16]. One additional CKX gene has been identified in a Dendrobium orchid, and similar aspects of its overexpression in growth and development have been described in Arabidopsis plants [17]. In this work, we reveal the basic characterization of the CKX gene family in the cereal species Hordeum and Triticum, as well as report on the cloning of the first two CKX genes of barley and demonstrate functionality for one of them in transgenic tobacco and Arabidopsis plants.

Materials and methods

Plant materials

Commercial barley (H. vulgare cv. Luxor) and wheat (T. aestivum L. cv. Samantha) grains were soaked in tap water for 1 day to initiate germination. The soaked grains were then transferred to soil and grown in a greenhouse with a 15 h/9 h day/night cycle at 21 °C.

Isolation of poly(A+) RNA

All RNA was extracted from different plant tissues using TRIZOL Reagent (Gibco BRL, Grand Island, NY, USA). Polysaccharide contamination of the grain extract was removed with two additional centrifugations at 14 000 g and treatment with a high salt solution (0.8 m sodium acetate, 1.2 m NaCl) before precipitation with isopropyl alcohol/ethanol (20% isopropyl alcohol and 70% ethanol). Poly(A+) RNA was purified from the total amount of RNA using an Oligotex Suspension (Qiagen, Hilden, Germany), according to the manufacturer's instructions.

Design of primers

A collection of degenerate oligonucleotide primers (CKX01, 5′-GAYTTYGGXAAYATHAC-3′; CKX02, 5′-AADATRTCYTGXCCXGG-3′; CKX03, 5′-TTXARCCAXGGRTGXGG-3′; CKX04, 5′-CCXCAYCCXTGGYTXAA-3′; and CKX05 5′-TRXARRTARTCXGTCCA-3′) covering the entire assumed sequence was synthesized on the basis of highly conserved areas between the sequences of maize ZmCKX1 (AF044603) and Arabidopsis AtCKX2 (AC005917) genes. The previously determined N-terminal amino acid sequence of wheat CKX [11] was not suitable for use in the primer design.

Three gene-specific primers (CKX07, 5′-CGGGGCACGAGCACGTTGAGCCAGGGAT-3′; CKX08, 5′-AAGATGTCTTGGCCCGGGGAG-3′; and CKX09, 5′-GTTCTGCGCCTCCAGCCGCC-3′) were designed for amplification of a 5′-end region of barley HvCKX1, wheat TaCKX1 genes, and one antisense primer (CKX06, 5′-ATCCCTGGCTCAACGTGCTCGTGCCCCG-3′) for amplification of the 3′-end region in RACE-PCR.

Three specific primers (two sense: CKX11, 5′-GCAATGGACTTCGGCAACCTCTCTAGCTTC-3′; CKX14, 5′-GATTGTCATCAGAATGGAATCCCTTCGGAG-3′; and one antisense: CKX13, 5′-GCACCCTATCCAAGAACTCAATGTAAGTGA-3′) were designed to amplify fragments of HvCKX2 and HvCKX3 genes according to sequences from the barley cDNA library of top adult leaves (AV835311, AV836048) that show particular homology with the maize ZmCKX1 gene. A pair of primers was designed to amplify part of the gene predicted as HvCKX7 on the basis of the coding region of the genomic DNA fragment (AJ234763; CKX19, 5′-GACATGCTCACGCACCAAGACCCCGGA-3′; CKX20, 5′-TGCCCTGGTGATGATGCCAAACTGGCC-3′) showing high homology with other CKX genes.

To amplify full-length genes, and to distinguish between HvCKX2 and HvCKX3 genes, one sense primer (CKX25, 5′-CAGTGAACCACTACCCTGCTACACG-3′) and two antisense primers (HvCKX2 specific, CKX23, 5′-GCTGATCTTCATTGATCTCAGTGCT-3′; HvCKX3 specific, CKX24, 5′-CATATTGCTAACCACGTGACATATG-3′), covering the dissimilar region, were designed.


The first-strand cDNA was reverse transcribed from 0.1 to 1.0 µg of poly(A+) RNA using a reverse transcriptase RAV-2 (Takara Shuzo Co., Shiga, Japan) and oligo(dT) primer (Promega, Madison, WI, USA). Hot-start touchdown PCR [18] was carried out using 45 cycles of amplification, with the annealing temperature of the first five cycles scaled down 1 °C per cycle. The usual cycle consisted of melting at 94 °C for 30 s, annealing at 53–49 °C for 30 s and extension at 72 °C for 1 min. The PCR mixture was prepared using a Takara Taq polymerase, as recommended by the manufacturer, with aliquots of the RT reaction, diluted 1 : 10 (v/v), as a template.


Different RACE-PCR techniques were used to amplify the full-length cDNA strands of barley CKX genes. Positive results were obtained by using a Marathon™ cDNA Amplification Kit (Clontech Laboratories, Palo Alto, CA, USA). The 0.5 µg of isolated poly(A+) RNA was treated exactly as advised by the manufacturer to obtain an adaptor-ligated ds cDNA library. The final 3′- and 5′-end products of HvCKX2 and HvCKX3 genes were obtained after 35 cycles of amplification in the GeneAmp® High Fidelity PCR System (Applied Biosystems, Foster City, CA, USA) using primers CKX13 and CKX14. Full-length cDNA was constructed by PCR with the template from the ds cDNA library using specific primers from 5′- and 3′-product termini (HvCKX2r, 5′-GCTGATCTTCATTGATCTCAGTGCT-3′; HvCKX3r, 5′-CATATTGCTAACCACGTGACATATG-3′; CKX23f, 5′-CAGTGAACCACTACCCTGCTACACG-3′). The annealing temperatures and the concentration of dimethylsulfoxide (4–10%) in the PCR mixture were altered to permit amplification of the cDNA ends of barley and wheat CKX genes from poly(A+) RNA treated using two other RACE-PCR kits [the SMART™ RACE cDNA Amplification Kit (Clontech); and FirstChoice™ RLM-RACE Kit (Ambion, Austin, TX, USA)].

Amplified fragments were excised from polyacrylamide gels and eluted by water for 1 day at 37 °C. DNA was subsequently recovered by ethanol precipitation and ligated into a pDRIVE vector (Qiagen). Transformations of Escherichia coli TOP10F′ competent cells were made by electroporation (1.8 kV, 5 ms). Positive transformants were selected by a β-galactosidase blue/white screening test. Inserted DNA was completely sequenced on both strands after amplification with internal or universal vector primers using a BigDye-terminator Cycle sequencing kit (Applied Biosystems) and an ABI PRISM 310 DNA sequencer (Applied Biosystems).

Search and analysis for novel gene sequences

DNA sequences encoding putative CKX proteins in cereals were searched using a wu-blast 2.0 program [19] in the expressed sequence tag (EST) clone database of the Institute for Genomic Research (TIGR: All Arabidopsis CKX protein sequences [14] were searched, one by one, against EST database subsets for wheat and barley using the blosum62 comparative matrix. The search produced gene indices that were constructed by assembling related ESTs after filtering for possible sequence contaminants. The resulting tentative consensus sequence was numbered and listed by relevant GenBank accession numbers representing the most overlapping sequences. Alignment of all sequences was performed with bioedit software [20] using the clustal w multiple sequence alignment program.

Construction of recombinant DNA for transformation and expression

A 10 µL aliquot of a heat-treated (7 min, 100 °C) commercial barley genomic library (partial Sau3AI DNA digest cloned into the Lambda FIX II vector; Stratagene, La Jolla, CA, USA) was used as a template to amplify genomic sequences of HvCKX genes with HvCKX2r, HvCKX3r, and HvCKX23f primers. Amplified DNA was cloned into the pDRIVE vector and sequenced. The same primers, with Asp718 and XbaI overhangs, were used to reamplify both genes using PCR with Pwo DNA Polymerase (Roche Applied Science, Mannheim, Germany) for direct sense subcloning into a binary pBINHygTx vector downstream of the cauliflower mosaic virus 35S promoter [21].

Full-length cDNAs were subcloned into the pYES2 (Invitrogen, Groningen, the Netherlands) and pDR197 binary vectors, with constitutive or inducible expression, respectively. The pDR197 plasmid was constructed from pDR195 [22] by introducing an additional cloning site (donated by D. Rentsch, ZMBP, University of Tübingen, Tübingen, Germany). Cells of Saccharomyces cerevisae strain 23344c ura were transformed by electroporation [23], and positive transformants were selected on the basis of the acquired uracil autotrophy. CKX activity was measured in the media and cell lysates within 48 h of growth, or within 48 h after induction with galactose when an inducible system was used.

Plant transformation

Agrobacterium tumefaciens strain GUS3101, harboring the binary vector pBINHygTx with different transgenes, was used to transform the A. thaliana ecotype Col0 via vacuum infiltration [24]. A standard protocol [25], using leaf discs of 8-week-old Nicotiana tabacum L. cv. Samsun NN plants, was employed to generate transgenic tobacco plants. The selection of all transformants was performed by adding hygromycin (15 mg·L−1) to the selection and rooting medium.

Transformed Arabidopsis plants were grown in a greenhouse until seed production. T1 progeny seeds of Arabidopsis transformants were surface sterilized and germinated on Murashige–Skoog medium (MS-medium) [26] in a controlled-environment chamber. Resistant seedlings were transferred to soil and placed in the greenhouse.

Immediately after transformation, tobacco leaf discs were placed on MS-medium supplemented with selection antibiotics and an appropriate growth regulator ratio for shoot regeneration (0.7 mg·L−1 of benzylaminopurine, 0.1 mg·L−1 of b-naphthoxyacetic acid). After 2 days, the discs were transferred to the same medium supplemented with claforam (0.5 mg·L−1; Ratiopharm, Ulm, Germany), to inhibit Agrobacterium growth. Developing shoots were transferred to MS-medium (without growth regulators) for root induction. Young plants with several leaves were then transferred to the soil and grown in the greenhouse under the conditions described above.

CKX activity assay

Plant samples for activity measurements were cut into pieces, powdered with liquid nitrogen using a hand mortar, and extracted with a 1.5-fold excess (v/w) of 0.2 m Tris/HCl buffer, pH 8.0, containing 1 mm phenylmethanesulfonyl fluoride and 1% Triton X-100. Cell debris was removed by centrifugation at 12 000 g for 10 min. The extract was loaded onto a Sephadex G-25 (50 × 2.5 cm) column equilibrated with 0.1 m Tris/HCl, pH 8.0, to remove the low molecular mass fraction. The protein fraction was then concentrated to a minimum volume by ultrafiltration and used to assay CKX activity.

The assay was performed according to a method described previously [27]. Samples were incubated in a reaction mixture (total volume 0.6 mL in an Eppendorf tube) of 100 mm reaction buffer, 0.5 mm electron acceptor [2,6-dichloroindophenol or 2,3-dimethoxy-5-methyl-1,4-benzoquinone (Q0)] and 0.5 mm substrate, for 0.5–12 h at 37 °C. The following buffers (and pH ranges) were used for determining the pH profile: Tris/HCl buffer (pH 7.5–9.5), imidazole/HCl buffer (pH 6.0–7.0), Mes/NaOH buffer (pH 5.0–5.5), and Na2HPO4/citric acid buffer (pH 3.0–4.5).

For determination of specific activities, the protein content of the samples was assayed according to Bradford [28], with BSA as the standard.

Extraction and analysis of cytokinins

Two grams of frozen plant material (barley kernels, 7- and 14-day-old barley seedlings) was ground in liquid nitrogen and extracted in 20 mL of 70% ethanol containing diethyldithiocarbamate (400 µg·g−1 of tissue) for 3 h at 4 °C. After centrifugation (20 min, 14 000 g), the pellet was re-extracted for 1 h in the same extraction mixture. The supernatants were combined and applied to a C18 cartridge (Waters, Milford, MA, USA), prewashed with 80% methanol to retain pigments. The pass-through fraction was collected and combined with a second fraction obtained by elution with 8 mL of 80% methanol. The resulting sample containing cytokinins was dried on a vacuum rotary evaporator. Cytokinins were then separated by reverse-phase HPLC, and individual HPLC fractions were analyzed by ELISA, according to a previously described protocol [29].


Isolation of HvCKX genes

RT-PCR with degenerate primers designed on the basis of two conserved motifs found among CKX proteins corresponding to amino acid sequences PHPWLN and PGQdIF, starting at positions 389 and 528 of the ZmCKX1 protein, revealed a 413 bp 3′-end fragment of a potential barley CKX gene. The gene transcript was most abundant in the poly A+ RNA fraction isolated from mature barley seeds. Thus, the putative gene was named HvCKX1 (AF362472; Hordeum vulgarecytokinin oxidase/dehydrogenase). A fragment of the same length was also isolated from the poly(A+) RNA of mature wheat grains and was named TaCKX1 (AF362471; Triticum aestivumcytokinin oxidase/dehydrogenase).

Attempts to amplify the 5′ cDNA end sequence by different RACE-PCR techniques did not yield any product for either of the genes. This failure may have occurred for several reasons, such as decreased quality of the isolated poly(A+) RNA owing to starch contamination, an amplified GC-rich sequence, or possibly the short half-life of the target CKX transcripts and their rapid degradation from the 5′-end.

In addition to this PCR-based strategy, a GenBank database search revealed several barley and wheat ESTs displaying homology to the Arabidopsis CKX gene family. Sets of gene-specific primers were designed to amplify the 3′ and 5′ cDNA ends of potential genes using the Marathon RACE-PCR kit (Clontech Laboratories). cDNA libraries generated from different barley tissues were used as templates for amplification. Two 3′-RACE and two 5′-RACE reaction products of a similar size were obtained when overlapping primers corresponding to EST-AV835311 (a barley cDNA library fragment generated from top adult leaves) were used for amplification. Both RACE products were cloned. Sequence analyses of several clones revealed the presence of two nearly identical gene sequences (94% homology between coding regions at the nucleotide level). Full-length gene sequences were recovered from independently amplified PCRs with primers flanking the predicted ORF regions where the reverse primer was designed on the basis of dissimilarity at the 3′-end of the noncoding region. The new gene of the 1578 bp coding sequence, fully corresponding to the above mentioned EST, was designated HvCKX2 (AF540382), and its 1560 bp close homologue was named HvCKX3 (AY209184).

Wheat and barley CKX ESTs

High homology between cereal gene fragments (HvCKX1 and TaCKX1 share 94% identity on the 130 amino acid fragment that includes the C-terminal region) may indicate the same evolutionary origin and possibly similar functions of both predicted genes. Both fragments show the highest degree of homology to ZmCKX1 (76%) and AtCKX2 (49%) proteins, CKX family members belonging to an evolutionary group with a predicted secretory pathway targeting.

The HvCKX2 gene encodes a protein of 526 amino acids with a predicted molecular mass of 58.8 kDa and a predicted pI value of 6.3. There is a very high identity between the HvCKX2 and the HvCKX3 gene products (92% at the amino acid level, Fig. 1). The latter is shorter (58.1 kDa) with one in-frame deletion within the sequence and its predicted pI value is shifted to 7.1. Both gene sequences contain an FAD-binding motif and other conserved regions typical of the CKX gene family. An N-terminal signal peptide for targeting to the secretory pathway was predicted by the cellular localization program, targetp[30], for both barley genes. However, predicted results were classified as medium-reliable using the ipsort program [31], the HvCKX3 protein classified as a mitochondrial protein. Encoded CKX proteins are predicted to be glycosylated at five potential N-glycosylation sites (calculated by NetNGly; distributed along the entire amino acid sequence.

Figure 1.

Alignment of barley and wheat cytokinin oxidase/dehydrogenase gene families compiled from TIGR EST clone databases. Amino acid residues conserved in more than half of the protein fragments are shown in white on a black background. Putative consensus sequences for N-glycosylation sites of HvCKX2 and HvCKX3 proteins are shaded grey. Signal peptides predicted by the targetp program [30] for both full-length genes are underlined. For detailed identification of gene indices see Table 1.

The genomic structure of HvCKX2 was determined by PCR using gene-specific primers flanking the cDNA and a barley genomic library cloned into bacteriophage λ as a template. Comparison of the genomic DNA sequence and the cDNA sequence showed the presence of four small introns, which corresponds well to the evolutionary conserved intron/exon pattern of most higher plant CKX genes [32].

A search for novel CKX genes in wheat and barley DNA databases revealed 24 EST clones showing significant homology to some members of the CKX gene family. Correct ORFs of partial sequences were compiled in an alignment and numbered according to the homology of the 11 rice gene family members [14]. For translated protein sequences of the genes and gene fragments, see Fig. 1. Sequences without mutual overlapping regions showing considerable homology to only one rice template (see Table 1) were assigned the same number. The compilation shown in Table 1 provided evidence for at least four additional barley (HvCKX4 to HvCKX7) and seven wheat (TaCKX2 to TaCKX8) gene homologues.

Table 1. Cytokinin oxidase/dehydrogenase (CKX) gene families in cereals. Sequences without mutual overlapping regions, showing considerable homology to only one rice template, are marked by the same number but with a different lowercase letter.
Closest rice
Homology to
rice protein
Tissue description
BQ462284  Callus
HvCKX2AF540382OsCKX784%7-day-old leaves
AV835311  Top three adult leaves
HvCKX3AY209184OsCKX784%7-day-old leaves
HvCKX4aBJ479455OsCKX489%Top three adult leaves
HvCKX4bBJ479606OsCKX494%Top three adult leaves
HvCKX5aBF264028OsCKX572%Seedling green leaves
HvCKX5bCB877904OsCKX577%Epidermis (seedlings)
HvCKX6CA031729OsCKX675%Seedling apex
HvCKX7AJ234763OsCKX384%Genomic DNA
AL825717  Drought-stressed seedlings
AL822297  Drought-stressed seedlings
TaCKX2aBG905097OsCKX666%Puccinia-infected leaf
TaCKX4BM138354OsCKX489%Fusarium-infected spikes
BJ306089  Spikelet at late flowering
TaCKX5aBM137409OsCKX587%Fusarium-infected spikes
TaCKX5bBQ161648OsCKX579%Fusarium-infected spikes
TaCKX6aCA705202OsCKX275%Developing kernels
BQ903062  Fusarium-infected spikes
BQ235927  Developing seeds
TaCKX6bBQ238832OsCKX252%Developing seeds
TaCKX7aCA603337OsCKX350%7-day-old roots
TaCKX7bBJ316444OsCKX392%Spikelet at early flowering
TaCKX8BJ322935OsCKX377%Spikelet at early flowering

Expression of CKX genes during barley plant development

To examine the expression of CKX genes in barley plants, a series of RT-PCR experiments were carried out using poly(A+) RNA prepared from representative plant organs during development, including roots, leaves, and kernels. As shown in Fig. 2, transcripts of HvCKX1 were found in all organs tested, such as mature kernels, roots and different developmental stages of leaves. HvCKX2 transcripts were detected in the leaves of 7-day-old seedlings, and the signal was also observed in kernels and roots. Interestingly, the expression of HvCKX3 transcripts was only observed in mature kernels and the leaves of young seedlings. Importantly, the presence of the HvCKX3 gene was not detected in the commercial barley genomic library. However, no signal was detected when primers designed for the amplification of the coding sequence of genomic DNA fragment (AJ234763, HvCKX7) were used for RT-PCR (data not shown).

Figure 2.

Expression patterns of HvCKX1, HvCKX2 and HvCKX3 genes during plant development. cDNA aliquots corresponding to 100 ng of mRNA were used as templates for PCR with gene-specific primers. Control reactions were set up with commercial barley genomic and cDNA libraries to distinguish between cDNA and genomic gene fragments. To eliminate reciprocal cross-reactivity between primers, plasmids with other cloned genes were used as templates (lane cross-reactivity). No template reaction contained water instead of mRNA. (A) An HvCKX1 gene cDNA fragment with a predicted size of 332 bp. (B) HvCKX2 gene cDNA with a predicted size of 1830 bp. (C) HvCKX3 gene cDNA with a predicted size of 1740 bp. (D) Time-dependence of the total specific cytokinin oxidase/dehydrogenase (CKX) activity (▪) and protein content (•) in the whole developing barley seedlings. Inset graph shows distribution of the CKX activity between shoots and roots of developing seedlings. The activity was determined with tissue extracts in imidazole/HCl buffer, pH 6.5, containing 5 mm CuCl2 and isopentenyladenosine as a substrate. All values represent mean values of data obtained from two parallel extractions, each measured in at least two replications.

An overview of cDNA lifetimes of in silico-derived genes suggests that cereal CKX enzymes are also expressed in additional tissues. Partially characterized novel barley genes HvCKX4 to HvCKX6 were found to be expressed in leaves, while the wheat genes were found in different tissues. Similarly, like HvCKX1, TaCKX1 is also expressed in both mature grains and developed seedlings. Transcripts of four TaCKX genes (TaCKX2, TaCKX4, TaCKX5 and TaCKX6) were observed in an mRNA pool collected after the infection of leaves and spikes by the cereal pathogens Fusarium and Puccinia. Like TaCKX6, TaCKX2 is also expressed in grains after pollination. Interestingly, a fragment of the gene coding for the TaCKX7 protein was present in a cDNA library generated from the mRNA of developing roots and also from spikelets at early flowering, where the fragment of TaCKX8 was also found. However, these descriptions are limited by the fact that only data presented in incomplete databases were used.

Transformants overexpressing HvCKX genes

To investigate whether the cloned genes code for active CKX enzymes, we overexpressed HvCKX2 and HvCKX3 in Arabidopsis and tobacco. The cDNAs and, for HvCKX2, also the genomic clone, were placed under the control of a constitutively expressed 35S promoter. At least 30 independent tobacco transformants were regenerated for each construct. Several regenerated plants transformed with the genomic version of HvCKX2 showed a very strong phenotype that was consistent with a cytokinin deficiency [15]Fig. 3. These plants had significantly shorter internodes, leading to a dwarf growth habit. On the contrary, the root system was noticeably enlarged in comparison with wild-type plants, similarly to the transgenic tobacco plants overexpressing the Arabidopsis AtCKX1 and AtCKX3 genes [15]. All of these plants were sterile and died without producing seeds. Other regenerated transformants overexpressing gHvCKX2, and also most of the HvCKX2 cDNA overexpressers, showed a milder phenotype. These plants were also characterized by shorter shoots, narrow leaves and a more branched and higher root mass than the wild type. Interestingly, T1 primary transformants overexpressing the HvCKX3 gene did not show any alteration of the phenotype. However, RT-PCR with specific primers for the HvCKX3 gene revealed the presence of HvCKX3 transcripts in tobacco leaves (data not shown). Increased CKX activity was detected in the leaves of several selected transgenic plants. As expected, the activity was elevated 10- to 50-fold in gHvCKX2 transformants (Fig. 4A) with a strong phenotype. Only a two- to fourfold increase was found in plants expressing the cDNA of the same gene. In the case of HvCKX3 overexpressers, no increase in activity was found.

Figure 3.

Shoot and root phenotype of gHvCKX2-expressing tobacco plants. (A) Tobacco overexpressers with mild (gHvCKX2-M) and strong (gHvCKX2-S) phenotypes, and wild-type (WT) plants, at the flowering stage. (B) Comparison of the root systems of phenotypically mild transgenic tobacco plants with those of wild-type plants.

Figure 4.

Substrate specificity of cytokinin oxidase/dehydrogenase(CKX) enzymes. Activity was measured in an Na2HPO4/citric acid buffer, pH 4.5, with 2,3-dimethoxy-5-methyl-1,4-benzoquinone (Q0) as the electron acceptor (dark bars) and in Tris/HCl buffer, pH 7.5, with 2,6-dichloroindophenol as the electron acceptor (light bars). (A) Activity of the HvCKX2 enzyme extracted from transgenic tobacco leaves. (B) CKX activity extracted from mature barley grains. (C) CKX activity extracted from 7-day-old barley roots. (D) CKX activity extracted from 7-day-old barley leaves.

The same three constructs of HvCKX genes in the binary vector, pBINHygTx, were used to transform Arabidopsis plants via vacuum infiltration. Regeneration of fertile Arabidopsis transformants was successful only from the seed progeny collected from plants transformed with constructs containing HvCKX cDNAs. In contrast, the growth of gHvCKX2 transformants was characterized by an enhanced root system and very slow shoot development. All seedlings had died by the formation of the third pair of rosette leaves, ≈ 3–4 weeks after germination. Thus, several Arabidopsis plants transformed with a construct carrying HvCKX2 cDNA showed similar phenotypical alterations to those recently described for strong Arabidopsis expressers of 35S:AtCKX1 and 35S:AtCKX3[16]. Plants were distinctive in having delayed formation of rosette leaves, smaller leaf size, and delayed onset of flowering with a reduced number of flowers. After flowering, approximately half of the plants did not produce siliques, or produced only one or two siliques with a very small amount of seeds and afterwards died.

CKX activity and cytokinin content during barley plant development

The CKX activity was monitored in barley seedlings and young plants. The specific activity was highest in the extracts of coleoptiles collected 1 day after germination and declined continuously thereafter, reaching about 10% of the initial activity by day 30. A twofold increase in the enzyme activity was observed around day 9 of barley growth (Fig. 2D). About 95% of the total activity in seedlings was located in the roots, while the activity in the leaves increased 7 days after germination to only slightly above the detection limit of the assay method.

The content of endogenous cytokinins with unsaturated side-chains, including bases, ribosides, nucleotides, and N- and O-glucosides, was measured in three developmental stages, i.e. grains, and 7- and 14-day-old barley seedlings. The measured values are summarized in Table 2. The total cytokinin content in the grains was approximately threefold lower than in young seedlings. The increase was mainly observed in the content of free bases and riboside types of cytokinins, which are the preferred substrates of CKX. The level of nucleotides remained nearly constant throughout the entire period. A significant increase was also visible in the content of zeatin O-glucoside during seedling development. There were no major differences in the cytokinin content of 7- and 14-day-old plants.

Table 2. Endogenous isoprenoid cytokinin levels in Hordeum vulgare during early development. Values are expressed as pmol of cytokinin-equivalents per gram of fresh weight (FW). All values represent the mean of two independent measurements. Standard errors were in the range of 4–20%.
Cytokinin compoundCytokinin content (pmol·g−1 FW)
Grain7-day seedling14-day seedling
Isopentenyladenosine monophosphate0.110.150.09
Zeatin riboside0.592.132.65
Zeatin ribotide0.210.270.29
Zeatin O-glucoside0.685.045.78
Zeatin riboside O-glucoside0.430.310.83

pH optimum and substrate specificity of barley CKX

The effect of pH on the activity of recombinant HvCKX2 and CKXs from grain, root and leaf extracts of barley was measured under standard assay conditions across the pH range from 3.0 to 9.5, with Q0 in the acidic range and 2,6-dichloroindophenol in the basic range as electron acceptors (Fig. 5). Overlapping pH ranges were measured in two buffer systems to exclude salt effects. These varied by only up to 5% of the total value. Protein extract from tobacco with a strong phenotype overexpressing  gHvCKX2 was used as a source of the recombinant protein. The same pH-dependence experiment was carried out with the extract of wild-type tobacco to eliminate the contribution of naturally present tobacco CKX to the recombinant activity. Activity found within wild-type tobacco was more than 20-fold lower than activity found in the extract of gHvCKX2-expressing plants. Surprisingly, the maximum value of HvCKX2 activity with isopentenyl adenosine was observed at pH 4.5 and then the activity slowly declined through neutral to alkaline pH. A similar activity profile was observed when isopentenyl adenine was used as the substrate. This behavior contrasts with the previously described pH-dependence of CKX enzymes [8], but supports new results on the subcellular targeting of two AtCKX-green fluorescence protein fused proteins to the vacuoles [16], where the pH generally ranges from 3.0 to 5.0. This contention emphasizes the fact that one of the enzymes, AtCKX1, is the closest homologue to the HvCKX2 enzyme. At low pH, the turnover of cytokinin ribosides is significantly higher than that of free bases (Fig. 4). This is in agreement with the possible existence of vacuolar-targeted CKX [16] and the observation of glycosylated forms of cytokinins occurring in acidic content of lytic vacuoles [33].

Figure 5.

pH dependence of HvCKX activity with 2,3-dimethoxy-5-methyl-1,4-benzoquinone(Q0)(▪) and 2,6-dichloroindophenol(•) as the electron acceptor. (A) Activity of HvCKX2 enzyme extracted from transgenic tobacco leaves. (B) Cytokinin oxidase/dehydrogenase (CKX) activity extracted from mature barley grains. (C) CKX activity extracted from 7-day-old barley roots. (D) CKX activity extracted from 7-day-old barley leaves. See Materials and methods for details on the buffer and reaction mixture composition.

With barley grain, root and leaf extracts, the pH profiles varied with the type of tissue from which the extract was prepared. In this case, the total activity is, however, contributed by all CKX isoenzymes expressed in the particular tissue. CKX activity from grain extract showed two maxima, one at pH 4.5 and the other, more significant one at pH 7–7.5, while the pH profile of leaf enzymes more or less corresponds to the HvCKX2 profile. This may indicate a predominant expression of HvCKX2 or a similar type of CKX in barley leaves and the expression of other CKX forms having an optimum at pH 7.5 in grains and roots. These conclusions are in agreement with the RT-PCR expression pattern of two evolutionarily distant HvCKX1 and HvCKX2 genes studied in this work.

The study of substrate specificity agrees with previously published data [11]. Cytokinins with isoprenoid side-chains are the preferred substrates for all tested enzyme samples. Isopentenyl adenosine is evidently the best substrate for HvCKX2 when measured under acidic conditions and with Q0 as an electron acceptor. This preference for riboside is less significant at basic pH and with 2,6-dichloroindophenol as an acceptor, while CKX enzymes generally prefer free bases when the pH of a reaction mixture is neutral or shifted to the alkaline region [11,13]. Riboside forms of cytokinins were found to be degraded better under acidic conditions (Fig. 4).

A newly described method for the detection of degradation products of aromatic cytokinins [27] was used to test them as potential substrates for barley CKXs. A low turnover of kinetin and its riboside was detected with HvCKX2 and the enzyme extract from grains. Activity with other aromatic cytokinins was probably under the threshold of method sensitivity for the quantities of enzyme used. Turnover of these substrates was described for maize recombinant CKX as being 200- to 1000-fold lower than that of isopentenyl adenine [12]. A newly estimated value of molar absorption coefficient for 4-(-4-hydroxyphenylimino)-3-methyl-2-buten-1-ol [26], the conjugated degradation product of zeatin-type cytokinins, results in a 4.5-fold increase in detected activities for these cytokinins than was previously assumed for purified CKX from barley grains [11]. Cleavage of cis-zeatin seems to be catalyzed only by some forms of CKX. Relatively high turnover rates were detected only with enzymes present in grains and roots at pH 7.5. This zeatin isomer does not serve as a substrate for HvCKX2 (Fig. 4).


In recent years, genomics and reverse genetics have developed tools and techniques that are crucial for a better understanding of the activity and function of cytokinins. Complete sequencing of the Arabidopsis genome revealed the presence of a small gene family encoding CKX. Detailed characterization of six out of the seven AtCKX gene family members demonstrated differential subcellular compartmentalization and their expression predominantly in meristematic tissues where the main pool of cytokinins is located [16]. Characterization of the CKX gene families in other species seems to be more difficult to assess, especially in monocot crop plants with large genomes in which complete sequences are unlikely to be obtained in the near future. Large genomes of cereals, with a high content of repetitive DNA sequences and their polyploid nature, make the study of gene organization more difficult. To date, one gene encoding a functional CKX enzyme has been described in maize [9,10], and two other full-length homologues have recently been deposited in the gene database.

In this work, we present the cloning of the first CKX genes of barley and their functional expression in tobacco and Arabidopsis plants. We describe two novel members of the CKX gene family with a typical FAD-binding domain and predicted glycosylation sites. Surprisingly, HvCKX cDNAs share 89% homology at the nucleotide level that leads to 37 changes in the protein sequence, and sequences noticeably differ only in the length of a 3′-end noncoding sequence. This high homology may indicate a rather recent evolutionary duplication of the HvCKX2 and HvCKX3 genes. A similar duplication event probably took place in the rice genome, where two paralogs of the CKX gene with 88% homology lie on neighboring loci on chromosome 2 (AP004996) [14]. Three recently annotated Zea mays mRNAs for CKX also show features of a recent duplication event. While two almost-identical isolated mRNAs (ZmCKX2: AJ606943, AJ606944) are obviously allelic versions of the same gene, the sequence annotated as ZmCKX3 (AJ606942) is probably their close paralog (sharing 93% homology at the amino acid level). Preliminary comparative mapping of selected gene regions in barley, wheat and maize has shown that gene duplication plays a significant role in the evolution of gene families within large cereal genomes [34]. However, it is still questionable whether all of these paralogs encode functional proteins. Whereas transformation of the HvCKX2 gene into the tobacco genome unambiguously elevates the level of the endogenous CKX activity and causes phenotypic alterations typical for cytokinin-deficient plants, no enhancement of the CKX level and no cytokinin-deficiency syndrome were found when the HvCKX3 paralog was overexpressed. Following heterologous expression of the HvCKX3 gene in the yeast S. cerevisiae, active CKX was not demonstrably present either in yeast media or in the cell extract (data not shown). Effectiveness in expression of the genomic and cDNA versions of the HvCKX2 transgene, respectively, in tobacco and Arabidopsis plants was significantly different. While transformation of model plants by the genomic version of the transgene led to strong cytokinin-deficiency phenotypes, the cDNA overexpresser showed just mild phenotypic alterations with only a moderately increased CKX level. A similar phenomenon was observed when expressing genomic and cDNA versions of ZmCKX1 gene in tobacco (K. Bilyeu, personal communication). It has been demonstrated many times that incorporating introns into transgenes has an enhancing effect on gene expression. This phenomenon was observed predominantly in GC-rich monocot genomes [35], but the mechanisms underlying the enhancement of gene expression are not entirely clear, especially when introducing monocot introns into dicot plants [36].

The great number of ESTs in public databases helped us to assemble at least a partial picture of CKX gene families in Hordeum and Triticum species. Gene indices were constructed for a minimum of seven barley and eight wheat CKX genes, respectively. However, background noise was observed within the constructed consensus sequences, which could be attributed to the limited fidelity of the reverse transcription step of cDNA library construction and sequence artifacts caused by the biochemistry of sequencing reactions. In addition, wheat is a hexaploid organism in which sequence diversity could be attributed to the origin of genomes inherited from different ancestors. Therefore, a partial sequence of several highly homologous ESTs did not allow us to distinguish whether they belonged to allelic variations or to two independent genes. Owing to the relationship of cereal plants, we were able to assign the closest rice CKX orthologs to all predicted HvCKX and TaCKX genes (50–94% homology). Hence, the rice genome revealed 11 CKX homologues, some of which had more than two corresponding copies in barley or wheat. Only detection and characterization of full-length genes, and their localization on cereal genomes, can give us a better understanding of the character and evolutionary relationships within this multiple gene family.

The expression patterns of the four CKX genes characterized during barley plant development suggest that the expression of particular genes is organ-specific. Three of the four genes studied are expressed in mature kernels, two of them in the roots and leaves of young seedlings, whereas expression of the fourth gene was not detected at all. Expression studies of maize CKX homologues have resulted in similar conclusions. The ZmCKX1 gene was found to be expressed in the vascular bundles of developing kernels, roots, and coleoptiles [37]. Transcripts of the other two maize genes were detected at two different time-points during kernel development, but they were not tested for in other organs [38]. These results indicate that different CKX genes are active mainly during kernel development. Furthermore, wide variations in the pH optimum of CKXs suggest that subcellular compartmentalization can be also a specific parameter among particular members of CKX gene families. Localization of two AtCKX proteins in the vacuoles and one in the endoplasmic reticulum, or possibly in the extracellular space, has been recently described [16].

The rapid loss of CKX activity during the first few days of plant growth, to the residual levels in the roots, is in agreement with the Western blot analysis presented for maize tissues [13]. The levels of cytokinins that are good substrates of CKX increase antagonistically with the enzyme level during the early days of the plant's development. This indicates that the CKXs detected in barley regulate the active cytokinin level principally in developing grains, and that the biologically active cytokinins are probably deactivated by other metabolic pathways, such as glycosylation, in the later stages of development when increased levels of both O- and N9-glucosides are observed.

The addition of the detergent Triton X-100 to the extraction buffer noticeably increased the total extracted CKX activity in all tested tissues. A relatively high concentration (1%) of detergent effectively destroys cellular membranes, probably allowing all forms of CKX proteins to be released without their substantial denaturation. Detailed characterization of the CKX reaction mechanism, carried out with recombinant ZmCKX1 protein, also helped us to determine conditions for the assay of CKX activities in vitro[12]. The mechanism of cytokinin cleavage involving a ternary enzyme–substrate–electron acceptor complex formation indicates that previously used concentrations of substrate and electron acceptors in the reaction mixture may limit the assays used for determination of the total activity [39]. The high stability of CKX proteins allowed us to prolong incubation times for the activity assay. A linear increase of product formation was determined within 7 h of incubation. Hence, the upgraded method utilizing p-aminophenol seems to be both sensitive enough and more effective than the previously and more commonly used radioisotope method.

Differences in CKX activity over the whole physiological range of pH were presented for partially purified enzymes from different sources [8]. Variations are also significant in one plant species. This heterogeneity may reflect differences in subcellular localization and/or glycosylation of enzymes. Indeed, differently glycosylated forms of CKX detected in conditionally cytokinin-overproducing tobacco cell suspensions have recently been studied [40]. While a glycosylated CKX with a pH optimum of 6.0 is secreted to the culture medium, the nonglycosylated form remains in the cells and shows a pH optimum of 8.5. On the contrary, an extract from barley grains, where the HvCKX1 form is predominant, shows a pH optimum of 7.0–7.5. In accordance, the two extracellular CKXs – AtCKX2 and ZmCKX1 – possibly the closest orthologs to HvCKX1, were also found to be most active under mildly alkaline conditions [13,27]. The surprisingly low pH optimum of the recombinant HvCKX2 indicates its potential targeting to the plant vacuole, despite the computer prediction showing possible secretion out of the cell. This form of enzyme is probably the main one in the leaf CKX pool. Preliminary characterization of Arabidopsis CKX proteins shows that enzymes with enhanced activity at low pH might also exist in other species [16].

The barley CKXs studied here have broad substrate specificities. Until recently, kinetin and aromatic cytokinins were not thought to be substrates for N6 side-chain cleavage by CKX. It was demonstrated recently, with the recombinant maize enzyme, that these compounds can also be irreversibly converted to adenine and a corresponding aldehyde by CKX. However, the degradation of aromatic cytokinins is not significantly affected by the cooperation of the enzyme with an electron acceptor. The reaction rates of these cytokinins are still several hundred fold lower then those of isoprenoid ones [12]. The barley enzymes studied also convert kinetin and its riboside with two-order lower velocity, and this conversion seems to be nonspecific for this particular isoenzyme. In contrast, the conversion of cis-zeatin was detectable only in grain and root extracts at pH 7.5. This zeatin isomer is obviously preferably cleaved by only a specific isoform of CKX that is apparently visible from a substrate preference study of the recombinant maize enzyme [13]. Cleavage of cis-zeatin by this isoenzyme was detected only with a large quantity of enzyme, and was determined to have a 30-fold higher Km and two orders lower kcat/Km values for cis-zeatin than that estimated for isopentenyladenine. This indicates that cis-zeatin has a much lower affinity for this enzyme than other isoprenoid cytokinins.

Progress made during recent years in the transformation of monocot plants, including barley [41], has promoted interest in the further investigation of barley and wheat CKX genes. Characterization of direct HvCKX barley overexpressers, and especially the possibility of preparing knockout mutants, will help us to elucidate the precise function of cytokinin-degrading enzymes in these crop plants. Genetic manipulation of CKX activity also holds the promise of improving agricultural traits, such as yield attributes or adaptation to environmental stress, in barley and other cereals.


This work was supported by grants 204/03/P103 and 522/03/0979 from the Grant Agency and MSM 153100008 and KONTAKT CZE01/023 from the Ministry of Education, Youth and Physical Education, Czech Republic and Bundesministerium für Bildung und Forschung, Germany.