Novel Plant-specific Cyclin-dependent Kinase Inhibitors Induced by Biotic and Abiotic Stresses*

The EL2 gene of rice (Oryza sativa), previously classified as early response gene against the potent biotic elicitor N-acetylchitoheptaose and encoding a short polypeptide with unknown function, was identified as a novel cell cycle regulatory gene related to the recently reported SIAMESE (SIM) gene of Arabidopsis thaliana. Iterative two-hybrid screens, in vitro pull-down assays, and fluorescence resonance energy transfer analyses showed that Orysa; EL2 binds the cyclin-dependent kinase (CDK) CDKA1;1 and D-type cyclins. No interaction was observed with the plant-specific B-type CDKs. The amino acid motif ELERFL was identified to be essential for cyclin, but not for CDK binding. Orysa;EL2 impaired the ability of Orysa; CYCD5;3 to complement a budding yeast (Saccharomyces cerevisiae) triple CLN mutant, whereas recombinant protein inhibited CDK activity in vitro. Moreover, Orysa;EL2 was able to rescue the multicellular trichome phenotype of sim mutants of Arabidopsis, unequivocally demonstrating that Orysa;EL2 operates as a cell cycle inhibitor. Orysa;EL2 mRNA levels were induced by cold, drought, and propionic acid. Our data suggest that Orysa;EL2 encodes a new type of plant CDK inhibitor that links cell cycle progression with biotic and abiotic stress responses.

The model plant species Arabidopsis thaliana counts up to 12 CDKs and 49 cyclins that have been grouped into different types according to sequence similarity (2,3). A-type CDKs are most closely related to the mammalian CDK1 and CDK2, contain the characteristic PSTAIRE amino acid sequence in their cyclin-binding domain, and play a role at both the G 1 -to-S and G 2 -to-M transition points. By contrast, the plant-specific B-type CDKs hold a divergent cyclin-binding domain and control the G 2 -to-M transition only (1). The A-and B-type CDKs probably form complexes with cyclins of type A, B, and D. D-type cyclins respond to external and internal growth signals, such as hormones and carbohydrate levels (4,5). They operate at the G 1 -to-S transition point, although they might control the G 2 -to-M transition as well (6 -9). A-type cyclins are mainly produced from the onset of the S phase until the middle of the G 2 and B-type cyclins, specifically from G 2 until the end of mitosis (1,10).
Because of their sessile life style, it is plausible that plants have developed mechanisms that allow them to adjust their cell cycle in response to environmental cues. Both biotic and abiotic stress stimuli negatively affect plant growth through the inhibition of the cell cycle machinery. Salinity inhibits Arabidopsis root growth by reducing the pool of dividing cells in the meristem (11,12). Likewise, in leaves of wheat (Triticum aestivum) and maize (Zea mays), water stress induces a shortening of the meristem and prolongs cell cycle duration as a result of reduced CDK activity (13,14). On the biotic side, cell cycle activity is inhibited in parsley (Petroselium crispum) and tobacco (Nicotiana tabacum) cell cultures upon treatment with fungal elicitors (15,16).
Perception of biotic and abiotic stress signals activates signaling cascades that trigger ion fluxes, kinase cascades, the generation of reactive oxygen species, and accumulation of antimitogenic hormones, such as abscisic acid (ABA) and jasmonic acid. These signaling molecules stimulate cell cycle checkpoints, resulting into an impaired G 1 -to-S transition, slowing down of DNA replication, and/or delayed entry into mitosis (16 -20). Still, insight is lacking into the molecular mechanisms that link the stress perception directly to the cell cycle machinery. Putative candidate proteins are CDK inhibitors (CKIs). In mammals, seven CKIs have been identified, which, based on structural and biochemical features, belong to two very distinct classes. Members of the INK4 family (p15 INK4b , p16 INK4a , p18 INK4c , and p19 INK4d ) are characterized by the presence of multiple ankyrin repeats and selectively inhibit G 1 -specific CDKs (CDK4 and CDK6). In contrast, inhibitors of the Kip/Cip family (p21 Cip1 , p27 Kip1 , and p57 Kip2 ) bind and inhibit a broader range of CDKs involved in the control of the G 1 -to-S transition (21). Until recently, only one class of plant CKIs has been described whose members were designated as Kip-related proteins (KRPs), because of their similarity with the mammalian Kip/Cip proteins, but some members are also known as Interactors of Cdc2 Kinases (ICKs) (21). Of late, a second putative CKI was found, distantly related to the ICK/KRPs, known as SIAMESE (SIM) (22). SIM interacts with A-type CDKs and D-type cyclins, and its overproduction results into a strong inhibition of cell division activity. However, biochemical proof of its inhibitory activity is still lacking. No plant homologs to the mammalian INK4 or yeast inhibitors have been discerned so far (2).
Here, we demonstrate that the Orysa;EL2 protein of rice (Oryza sativa), recognized as a novel D-type cyclin-interacting protein, interferes with the ability of plant D-type cyclins to complement a yeast strain deficient in its G 1 cyclins, suggesting that Orysa;EL2 acts as an novel CKI. In agreement, Orysa;EL2 expression was able to rescue the multicellular trichome phenotype of sim mutants of Arabidopsis, whereas the recombinant Orysa;EL2 protein efficiently inhibited the activity of purified CDKs. The Orysa;EL2 gene was transcriptionally induced by biotic and abiotic stress treatments, implying that Orysa;EL2 might coordinate stress perception and cell cycle progression.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Experiments-The full-length Orysa; CYCD5;3 open-reading frame was amplified by PCR with genespecific primers and subcloned into the pBD-GAL4 Cam vector (Stratagene, La Jolla, CA), resulting into the pBD-GAL4: CYCD5;3 bait plasmid. The Orysa;EL2 bait construct was obtained by cloning the Orysa;EL2 cDNA into a GATEWAYmodified pBD-GAL4 vector (Invitrogen, Carlsbad, CA). Orysa; EL2 cDNA was amplified from rice (Oryza sativa L. cv. Nipponbare) with gene-specific primers with attB adaptors, cloned into the pDONR TM 201 ENTRY vector (Invitrogen) by an attB ϫ attP (BP) recombination reaction, and subsequently mobilized into the pBD-GAL4 vector by an attL ϫ attR (LR) recombination reaction. All constructs were confirmed by sequencing. The cDNA library used for the two-hybrid screens was prepared with RNA extracted from rice cell suspension cultures harvested 0, 3, 6, 9, and 12 days after subculture in fresh medium. Equimolar amounts (100 g) of total RNA from each sample were used to purify poly(A)ϩ mRNA with the Poly(A) Quick mRNA Isolation kit (Stratagene). The cDNA was synthesized and subcloned into the HybriZAP-2.1 vector according to the manufacturer's instructions (Stratagene). Approximately 2 ϫ 10 6 independent plaque-forming units were produced, with an average insert size of 1 kb. The library was amplified once to yield 2 ϫ 10 9 plaque-forming units/ml. Screens were performed with the PJ69-4a (MATa; 23) yeast strain according to a protocol described previously (23). The cDNA inserts from interacting clones were amplified by PCR and sequenced. For pairwise two-hybrid interactions, different combinations between bait (pBD-GAL4:OSCYCD5) and prey constructs were transformed into yeast cells and assayed for their ability to grow on histidine-deficient minimal media at 30°C.
Sequence Analysis-Conserved protein domains were detected with MEME (24), whereas domain matches on the different protein sequences were identified with a combination of MEME and HMMer (25). Multiple sequence alignments were generated with ClustalW (26).
FRET Analysis-Onion epidermal cells were used for FRET analysis as described (22).
Sim Complementation Analysis-The full-length Orysa; EL2-coding region was introduced into the GATEWAY compatible vector pLEELA-pGL2 (a gift from Dr. Arp Schnittger) by an LR recombination reaction, placing the Orysa;EL2 open-reading frame under direct control of the GLABRA2 promoter. The resulting plasmid was introduced into Agrobacterium tumefaciens by electroporation, and subsequently into sim-1 plants via the floral dip method (29). Transgenic plants were planted in soil, then sprayed with 1 mM phosphinothricin after the first leaves had established, and resistant plants were inspected for complementation of the sim phenotype.

Glutathione S-Transferase (GST) Pull-down Assays-
The open-reading frames of Orysa;EL2 and Orysa;el2 ELERLF were cloned into pGEX4T-1 (GE-Healthcare, Little Chalfont, UK). The obtained plasmids were transformed in the Escherichia coli BL21-codonPlus(DE3)-RIL strain (Novagen, Madison, WI). E. coli cells were grown in a Luria-Broth medium to an A 600 of 1 at 37°C. Production of GST fusion proteins was induced by the addition of 1 mM isopropyl-D-galactoside for 6 h. The GST fusion proteins were purified with glutathione-Sepharose 4B (GE-Healthcare) according to the manufacturer's protocol. Subsequently, a total of 500 g of total protein extracted from a 2-day-old dividing Arabidopsis cell culture was mixed with 3 l of beads from a dilution series (100ϫ, 10ϫ, and 1ϫ). The pulldown was performed in a total volume of 200 l of homogenization buffer (HB; 25 mM Tris-Cl, pH 7.6, 75 mM NaCl, 15 mM MgCl 2 , 15 mM ethylene glycol-bis(␤-aminoethyl ether) N,N,NЈ,NЈ-tetraacetic acid, 15 mM p-nitrophenylphosphate, 60 mM ␤-glycerophosphate, 1 mM dithiothreitol, 0.1% Nonidet P-40, 0.1 mM Na 3 VO 4 , 1 mM NaF, and protease inhibitor mixture P9599 (Sigma-Aldrich) and incubated on a rotating wheel for 2 h at 4°C. The bead-bound fractions were washed three times with HB. The beads were divided in two parts: one part for kinase assay and one fraction was resuspended in 1ϫ sodium dodecyl sulfate loading buffer and boiled. Proteins were separated on 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and blotted onto Immobilon-P membranes (Millipore, Bedford, MA). Filters were blocked overnight at 4°C in 3% (v/v) milk powder in 25 mM Tris-HCl, pH 8, 150 mM NaCl, and 0.05% Tween 20, and incubated for 1 h at 4°C with a cdc2(anti-PSTAIRE):sc-53 (1/1000) (Santa Cruz Biotechnology, Santa Cruz, CA) or anti-GST (1/2000) (GE-Healthcare) antibodies in blocking buffer. Antigen-antibody complexes were detected with horseradish peroxidase-conjugated IgG diluted 1/10000 (GE-Healthcare) with a chemiluminescence system (Perkin-Elmer Life Sciences, Norwalk, CT).
Expression and Purification of Orysa;EL2-The Orysa;EL2 gene was cloned in an expression vector for E. coli (30), containing a His-tagged NusA fusion partner. After induction at 20°C with isopropyl-␤-D-thiogalactopyranoside for 16 h, the cells were lysed and loaded on an IMAC column (Ni-Sepharose HP; GE-Healthcare). The fusion part was removed enzymatically with His-tagged caspase-3 at an engineered DEVD site immediately preceding the Orysa;EL2 mature sequence. The protein was further purified on a mono-Q ion exchange column (GE-Healthcare) and a Ni-Sepharose IMAC column. The monomeric Orysa;EL2 was polished on a gel filtration column while the buffer was changed to phosphate-buffered saline, concentrated, and stored at Ϫ80°C. For kinase assays, Arabidopsis cells were harvested from actively dividing cell cultures and either used immediately or snap-frozen in liquid nitrogen and stored at Ϫ70°C. The cells were ground in liquid nitrogen and proteins were extracted in HB. Equal amounts of total protein were incubated with 100 l of 25% (v/v) p10 CKS1At -Sepharose beads overnight at 4°C on a rotary shaker. For immunoprecipitations, 300 g of total protein in HB were precleared with 30 l of 50% (v/v) protein A-Sepharose beads (GE-Heathcare) for 1 h at 4°C. After a short centrifugation, the precleared supernatants were transferred to new Eppendorf tubes containing CDKA;1 (1/250) or CDKB1;1 (1/100) antibodies (31,32) and incubated at 4°C for 2 h. In the following step, 30 l of 50% (v/v) protein A-Sepharose was added, and the tubes were incubated on a rotating wheel for 1 h at 4°C. Thereafter, beads were washed three times with 20 mM Tris-HCl, pH 7.4, 5 mM ethylenediaminetetraacetic acid, 2 mM ethylene glycol-bis(␤-aminoethyl ether) N,N,NЈ,NЈ-tetraacetic acid, 100 mM NaCl, 2 mM NaF, 0.2% Nonidet P-40, 300 M phenylmethylsulfonyl fluoride, and 10 g/ml aprotinin and pepstatin, and used for CDK activity reactions. Kinase assays were performed as described (33).
Stress Treatment, RNA Preparation, and Quantitative PCR-Seeds of rice were germinated in the dark in sand saturated with water. After 5 days, the germinating seedlings were transferred to 24-multiwell tissue culture plates (Falcon, BD Biosciences, San Jose, CA) filled with tap water, and grown at 55% relative humidity, a temperature regime of 28°C day/22°C night, with an 11-h day length at 3,000 Lux light intensity for 2 more days. Seven-day-old seedlings at leaf stage 2 were exposed to heat (37°C), cold (4°C), drought, ABA (100 M), or propionic acid (1.5 mM). Control plants were kept in tap water. After stress treatment, the seedlings were cut at the crown region with a clean sterile blade. The seeds, roots, and first leaves were discarded. Each, biological sample containing four seedlings was collected in triplicate. Total RNA and first-strand cDNA were prepared with the RNeasy plant mini kit (Qiagen, Hilden, Germany) and Superscript III first-strand cDNA synthesis kit (Invitrogen), respectively, according to the manufacturer's instructions. Multiplex PCR reactions were carried out on an LC480 Q-RT-PCR machine (Roche Diagnostics, Brussels, Belgium) in 384-well plates. Specific primers used were: 5Ј-CTCAGCAGCAGGGGAAGG-3Ј (forward primer) and 5Ј-AGGTGAAGGTTGGCATTATTGG-3Ј (reverse primer) for Orysa;EL2 and 5Ј-TTGTGTTGGACTCTGGT-GATGG-3Ј (forward primer) and 5Ј-CCGTCAGGATCT-TCATGAGGTAAT-3Ј (reverse primer) for Orysa;ACTIN1 (Os03g50890). The transcript level of the Orysa;ACTIN1 gene was used as an internal control. cDNA was amplified for multiplex PCR with SYBR Green I master mix (Roche Diagnostics) according to the manufacturer's instructions. Relative expression levels of Orysa;EL2 were determined by the comparative ⌬Ct method (34)  Meta-analysis of Arabidopsis EL2-like Gene Expression-With the "Response Viewer" tool of GENEVESTIGATOR, the expression profiles of genes to different stimuli were analyzed (35). Only biotic and abiotic stress treatments with a more that 2-fold change in transcription level for at least one of the EL2like genes were taken into account. Fold-change values were hierarchically clustered for genes and experiments by average linkage in "Multiple Experiment Viewer" of The Institute for Genome Research.

Orysa;EL2 Is a Novel D-type Cyclin-interacting Protein-To
identify novel rice proteins that control the G 1 -to-S transition of the plant cell cycle, a yeast two-hybrid screen was performed with a D-type cyclin of rice Orysa;CYCD5;3 (Os03g10650) as bait. For the screening, a galactose 4 (GAL4) activation domain cDNA fusion library was used, constructed with RNA isolated from an actively dividing rice cell suspension culture. A total of 10 6 independent yeast cotransformants were screened, resulting into 31 positive colonies. Most cDNA clones (77%) encoded a small protein of 106 amino acids (predicted molecular mass of 11.6 kDa), identical to the Orysa;EL2 protein (Os03g01740) whose gene had originally been identified as an early-response gene upon elicitation with N-acetylchitoheptaose, a potent biotic elicitor for phytoalexin biosynthesis genes (36). Orysa; EL2 has no sequence homology with any characterized func-tional protein but shares a small peptide domain of six amino acids with the ICK/KRPs (Fig. 1) Sequence similarity searches with this domain as a query identified one related protein from rice (Fig. 1). Also six Arabidopsis proteins were found, among which the recently described SIM protein (22). The EIEDFF domain resides into the mapped cyclinbinding domain of KRPs (37). To test whether the ELERLF motif is required for cyclin binding, it was mutated into a stretch of alanine residues, resulting into the mutant el2 ELERLF protein that did not interact with Orysa;CYCD5;3 (Fig. 2).
The protein-protein interactions observed with the two-hybrid system were confirmed by FRET experiments. As a positive-control FRET protein pair, the Arabidopsis transcription factor TGA (At5g06960), whose self-interaction in plants had previously been detected by FRET analysis (38), was used, and as a negative control, the noninteracting LexA-nuclear localization signal (NLS) and TGA5 proteins (39). FRET efficiency was evaluated by the increase in donor fluorescence after photobleaching of the acceptor molecule, in cells that produced YFP:CYCD5;3 (as acceptor) together with either CFP:EL2 or CFP:el2 ELERLF (as donors). When measuring the fluorescence intensity of CFP:EL2 before and after YFP photobleaching, an average increase in fluorescence of 11.4% was seen for CFP:EL2, demonstrating that Orysa;EL2 interacted in vivo with Orysa;CYCD5;3 ( Fig. 3; Table 1). No FRET was observed for CFP:el2 ELERLF . By contrast, an efficient FRET signal was noticed for both CFP:EL2 and CFP: el2 ELERLF in the presence of YFP:CDKA1;1. (Table 1), again indicating that the interaction between Orysa;EL2 and Orysa;CDKA1;1 is independent of the ELERLF motif. No significant association was observed between Orysa;EL2 and Orysa;CDKB;1 or Orysa;CDKD;1. Interestingly, both Orysa; EL2 and Orysa;CYCD4;1 localized specifically into the nucleus (Fig. 3).
To confirm the interaction of Orysa;EL2 with CDKA;1, an in vitro pull-down experiment was done with a GST-EL2 fusion protein. In addition, a GST protein fused to the Orysa;el2 ELERLF protein with the mutated cyclin-binding domain was used. After extensive washing, bound proteins were eluted and used for protein gel blot analyses with an anti-PSTAIRE antibody, which specifically recognizes A-type CDKs. Whereas no CDKs were found to associate with the mock-treated sample, a clear CDK signal was seen for the GST-EL2 beads (Fig. 4A). Also the  , together with the AD-EL2 or AD-el2 ELERLF plasmids, encoding a fusion protein between the GAL4 activation and the wild-type (AD-EL2) or mutant (AD-el2 ELERLF ) Orysa;EL2 protein, respectively. Transformants were streaked on plates with (ϩHis) or without (-His) histidine. Reconstitution of GAL4 activity restored the ability to grow on His-deficient medium. mutant Orysa;el2 ELERLF associated with CDKs, albeit with reduced affinity. A dilution series of bead-bound recombinant proteins was prepared to obtain pull-downs with similar amounts of CDKs. The CDK activity associated with the purified complexes was measured. More kinase activity was detected with the GST-el2 than the GST-EL2 fraction (Fig. 4B), indicating that the ability of Orysa;EL2 to interact with cyclins is essential to exert its role as a CDK inhibitor.
Orysa;EL2 Abolishes Rescue of the CLN Budding Yeast Mutant by Plant D-type Cyclins-A functional interaction between Orysa;EL2 and Orysa;CYCD5;3 was demonstrated with the mutant budding yeast strain BF305-15d-21 that lacks two out of its three G 1 -specific cyclins (CLN1 and CLN2) and is conditionally deficient for the third one (CLN3). In the presence of galactose, the CLN3 gene is expressed and cells are able to divide. When transferred to glucose-containing media, CLN3 expression is repressed, and cells become arrested in G 1 (40). When BF305-15d-21 cells were transformed with an expression vector containing the Orysa;CYCD5;3 gene (pTH-CYCD5), cells resumed growth (Fig. 5), indicating that Orysa; CYCD5;3 encodes a functional cyclin able to complement a CLN mutant yeast strain. By contrast, when the Orysa;CYCD5; 3-expressing yeast strain was transformed with an Orysa;EL2expressing construct, containing the Orysa;EL2 cDNA sequence under the control of the alcohol dehydrogenase 1 (ADH1) promoter, Orysa;EL2 hampered the ability of Orysa; CYCD5;3 to complement the BF305-15d-21 strain (Fig. 5). Cell proliferation was completely inhibited, illustrating that Orysa; EL2 blocked the function of Orysa;CYCD5;3 in yeast.
Orysa;EL2 Is a Functional CKI in Vitro-The observed interaction of Orysa;EL2 with CDKs and cyclins and its interference with the capability of the Orysa;CYCD5;3 cyclin to complement the yeast CLN mutant suggested that Orysa;EL2 could operate as an inhibitor of CDK-cyclin complexes. To test this hypothesis, CDK-cyclin complexes were purified from cell suspension culture extracts either by affinity chromatography with p10 CKS1At -Sepharose beads or by immunoprecipitation with specific antibodies for CDKA;1 and CDKB1;1. Recombinant Orysa;EL2 was added to the affinity-purified CDKs and kinase activity was measured with histone H1 as substrate. CDK activity did not decrease significantly upon addition of Orysa; EL2 to p10 CKS1At -Sepharose-purified complexes (Fig. 4, C and D). By contrast, Orysa;EL2 was found to be a potent inhibitor of immunopurified CDKA;1 complexes (Fig. 4, C  and D). CDKA;1 activity was only partially inhibited. Addition of a higher quantity of Orysa;EL2 did not result in a stronger inhibition of CDK activity (data not shown). Orysa; EL2 inhibited only slightly B-type CDK activity at the highest dose applied (Fig. 4, C and D).
Orysa;EL2 Complements the Arabidopsis Sim Phenotype-Orysa;EL2 shows sequence similarity to the recently reported SIM gene (22). In Arabidopsis, trichome cells are unicellular. By contrast, recessive mutations in the SIM gene gives rise to multicellular trichomes, indicative that SIM is required to suppress cell division during hair development, according to its anticipated role as CKI. To test whether Orysa;EL2 was able to complement the sim trichome phenotype, sim mutant plants were transformed with a construct harboring the Orysa;EL2 construct under the control of the trichome-specific GLABRA2 promoter. In 12 out of the 14 transgenic lines obtained, Orysa; EL2 was able to completely rescue the multicellular trichome  phenotype (Fig. 6), demonstrating that Orysa;EL2 and Arath; SIM are functionally related. Orysa;EL2 Is Transcriptionally Induced upon Stress Treatment-Because of its observed elicitation with N-acetylchitoheptaose, the transcriptional response of Orysa;EL2 toward different stress treatments was tested in 7-day-old rice seedlings. Stresses applied included heat (37°C), cold (4°C), and drought. Steady-state mRNA levels were slightly reduced by heat, but remarkably strongly induced by 4 h of cold treatment (Fig. 7A). This induction was even more pronounced after 24 h of cold (Fig. 7B). At this time point, Orysa;EL2 was also induced by drought. Transcript levels did not change upon treatment with the antimitogenic hormone ABA (Fig. 7, A and  B), suggesting that induction of Orysa;EL2 by cold and drought occurred in an ABA-independent manner. Orysa;EL2 was also up-regulated by propionic acid that triggers cytoplasmic acidi-fication, a downstream effect in N-acetylchitoheptaose elicitation (41).
With the GENEVESTIGATOR toolbox, the expression pattern of the six Arabidopsis SIM/EL2-like genes in response to different biotic and abiotic stress treatments was examined in more than 2,500 publicly available microarray experiments (35). The different family members were induced under various stress conditions, albeit with different specificity (Fig. 7C). Arath; SMR5 (At1g07500) and Arath; SRM3 (At5g02420) were activated by all or most stress conditions, respectively, whereas other EL2-like genes responded more specifically. Interestingly, distinct family members reacted often in a complementary manner; for example, Arath; SIM (At5g04470) and Arath;SMR1 (At3g10525) were induced by the nematode Heterodera schachtii, but down-regulated upon infection with the pathogen Pseudomonas syringae, whereas Arath; SMR2 (At1g08180) and Arath; SMR3 (At5g02420) had the opposite behavior.

DISCUSSION
We identified the Orysa;EL2 protein as a potent key regulator in the series of events that might transduce the early response of biotic and abiotic stresses directly to the cell cycle. Orysa;EL2, previously characterized as a gene induced within minutes after addition of the elicitator N-acetylchitoheptaose or purified flagellin protein of the pathogen P. avenae (36,42), was demonstrated here to encode a D-type cyclin-binding protein that inhibits CDK activity. EL2-like proteins can be found in mono-and dicotyledonous plants, but no homologs are known in non-plant species, indicating that this family of CKIs is plant-specific.
Recombinant Orysa;EL2 protein specifically inhibited CDKA;1, but not CDKB1;1 activity, a feature shared with the previously described ICK/KRPs (43). The interaction with CDKA;1 must probably be direct, because both the wild-type Orysa;EL2 and its mutant isoform with deleted cyclin-binding domain pulled-down the CDK subunit. However, compared with the wild-type isoform, the mutant Orysa;el2 was less efficient in pulling down CDKs, indicating that, besides a direct interaction, Orysa;EL2 also partially interacts indirectly with CDKs, probably through the binding of cyclins. Remarkably, FIGURE 4. In vitro binding and inhibition of A-type CDKs by Orysa;EL2. A, total plant extracts incubated with GST-EL2, mutated GST-el2 ELERLF , or mock-treated. Associated proteins were pulled-down with glutathione-Sepharose beads. Bound proteins were used for a protein gel blot with a PSTAIRE-specific antibody. A protein gel blot with an anti-GST antibody was used to measure the amount of pulled-down GST-EL2 or GST-el2 ELERLF . B, activity measurement of CDKs associated with GST-EL2 and GST-el2 ELERLF histone H1 as a substrate. C, CDKs purified from protein extracts with either p10 CKS1At -Sepharose beads or specific anti-CDKA;1 and anti-CDKB1;1 antibodies. Pulled-down kinase activity was measured in increasing levels of recombinant Orysa;EL2 protein with histone H1 as substrate. D, relative quantification of two independent kinase activity measurements as depicted in C. Kinase activity in the absence of Orysa;EL2 arbitrarily set at 100%. Bars indicate S.E.
CDK complexes purified on p10 CKS1At beads were not inhibited, suggesting that either CKS1At and Orysa;EL2 compete for the same CDK-binding site or that p10 CKS1At lacks affinity for the Orysa;EL2 target complexes. Human p9 CKShs1 and yeast p13 SUC1 do not bind the mammalian G 1 -specific CDK4 and CDK6 complexes (44,45). Analogously, fission yeast (Schizosaccharomyces pombe) cdc2 associates only to p13 SUC1 during mitosis (46). Selective CDK binding by p13 SUC1 has been demonstrated for maize as well, where only M-but not S-associated CDK complexes were found to bind p13 SUC1 (47). Therefore, the lack of inhibition of p10 CKS1At -associated CDKs complexes by Orysa;EL2 might indicate that the main targets of Orysa;EL2 can be found among G 1 /S-phase-associated CDK complexes. In support for a role during G 1 -to-S, Orysa;EL2 interacts only with D-type cyclins and not with A-or B-type cyclins. Analogously as for mammals, a model has been put forward for plants in which D-type cyclins act as primary sensors linking external and internal growth signals with cell division activity, as illus-    (n ϭ 3). C, hierarchical average linkage clustering of Arabidopsis EL2-like genes in publicly available microarray experiments. Data were obtained with GENEVESTIGATOR, a data base and Web browser data-mining interface for Affymetrix GeneChip data (Santa Clara, CA). Code: blue, down-regulation; yellow, up-regulation; black, no change.
trated by the cytokinin-independent growth of Orysa;CYCD3; 1-overexpressing calli or the delayed activation of the cell cycle machinery in seeds of D-type cyclin knock-out lines when sown under favorable conditions (48,49). The specific interaction of Orysa;EL2 with D-type cyclins suggests that the signaling cascades that arrest the cell cycle in response to biotic and abiotic stress signals directly impinge on the CYCD/CDK complexes. In such a model, the balance between CYCD/CDK complexes and EL2-like proteins might be a critical parameter to decide whether cells divide or not.
In mammals, interaction between Kip/Cip proteins and cyclins is controlled by the presence in the CKIs of one or more ZRXL amino acid motifs (with Z basic or cysteine and X preferentially basic). This ZRXL motif is found in other cyclin-interacting proteins, such as E2F1, p107, and p130 (50,51) and is present in most Orysa;EL2-related sequences, but only in some of the ICK/KRPs. Moreover, mutation of the ZRXL motif does not impair the binding of ICK2/KRP2 with D-type cyclins, 5 suggesting that a motif different from ZRXL controls the interaction of ICK/KRPs and EL2-like proteins with D-type cyclins. By mutational analysis, we identified the EIEDFF domain that is shared by all EL2-like proteins and ICK/KRPs as a cyclin-binding motif. This particular domain resides into the presumed cyclin-binding domain of ICK/KRPs (37). Moreover, its mutation hampers the association between Orysa;EL2 and Orysa; CYCD5;3 and is probably important for CDK inhibition. The EIEDFF motif can be found also in other potential plant cyclininteracting proteins, such as retinoblastoma-binding protein 2-like proteins and the DNA mismatch repair protein PMS2. EL2-like and ICK/KRPs share some other small peptide sequence proteins, but, because they are only shared between a few members of each protein class, they might play a role in fine-tuning the activity of specific family members.
Orysa;EL2 was able to complement the sim trichome phenotype, illustrating that both genes are functionally related. Interestingly, in sim mutants, the trichomes do not only undergo mitotic cell division, but display as well an inhibition of their endoreduplication program (52). These data point toward a role for the SIM protein at the mitosis-to-endocycle switch during trichome development. Our results on Orysa;EL2 suggest that this transition might be achieved through inhibition of specific CDK/cyclin complexes. Therefore, the EL2/SIM genes might not only operate as sensors of antimitogenic stimuli, but also play important roles during developmental processes.