Arabidopsis immunophilins ROF1 (AtFKBP62) and ROF2 (AtFKBP65) exhibit tissue specificity, are heat-stress induced, and bind HSP90
- First Online:
- Cite this article as:
- Aviezer-Hagai, K., Skovorodnikova, J., Galigniana, M. et al. Plant Mol Biol (2007) 63: 237. doi:10.1007/s11103-006-9085-z
- 426 Views
The plant co-chaperones FK506-binding proteins (FKBPs) are peptidyl prolyl cis-trans isomerases that function in protein folding, signal transduction and chaperone activity. We report the characterization of the Arabidopsis large FKBPs ROF1 (AtFKBP62) and ROF2 (AtFKBP65) expression and protein accumulation patterns. Transgenic plants expressing ROF1 promoter fused to GUS reporter gene reveal that ROF1 expression is organ specific. High expression was observed in the vascular elements of roots, in hydathodes and trichomes of leaves and in stigma, sepals, and anthers. The tissue specificity and temporal expression of ROF1 and ROF2 show that they are developmentally regulated. Although ROF1 and ROF2 share 85% identity, their expression in response to heat stress is differentially regulated. Both genes are induced in plants exposed to 37 °C, but only ROF2 is a bonafide heat-stress protein, undetected when plants are grown at 22 °C. ROF1/ROF2 proteins accumulate at 37 °C, remain stable for at least 4 h upon recovery at 22 °C, whereas, their mRNA level is reduced after 1 h at 22 °C. By protein interaction assays, it was demonstrated, that ROF1 is a novel partner of HSP90. The five amino acids identified as essential for recognition and interaction between the mammalian chaperones and HSP90 are conserved in the plant ROF1-HSP90. We suggest that ROF/HSP90 complexes assemble in vivo. We propose that specific complexes formation between an HSP90 and ROF isoforms depends on their spatial and temporal expression. Such complexes might be regulated by environmental conditions such as heat stress or internal cues such as different hormones.
Peptidyl prolyl cis-trans isomerase
Tripartite tetratricopeptide repeats
HSP70 and HSP90 organizing protein
Calmodulin binding domain
Heat shock protein
Protein phosphatase 5
Heat shock factor
Heat shock element
Stress response element
The FK506-binding proteins (FKBPs) belong to the superfamily of peptidyl prolyl cis-trans isomerases (PPIases) (EC 22.214.171.124), which are present in all organisms and almost all subcellular compartments (Galat 2004; Romano et al. 2005). The catalytic activity of PPIase is the rotation of a peptide bond preceding proline in the proteins (Fischer et al. 1998; Fanghanel and Fischer 2004). The FKBPs are a distinct set of cellular receptors, which bind the immunosuppressive-drugs FK506 and rapamycin. When FKBPs bind these drugs their PPIase activity is inhibited (Schreiber 1991). The complexes formed by the FKBPs and their ligends are the functional modules for immunosuppression. Based on this activity, the FKBPs, together with the unrelated family of cyclophilins (CyPs) [which bind the immunosuppressive drug cyclosporine A (CsA)] have also been termed immunophilins. The drugs FK506 and CsA are involved in the mechanism of immunosuppression, and therefore the immunophilins are extensively studied as reviewed (Galat 2003, 2004; Dornan et al. 2003; He et al. 2004; Romano et al. 2005).
During the past decade, a growing number of immunophilins from mammalian, bacteria, and plants have been characterized. The high-level of conservation among the FKBP-protein family and ubiquitous distribution suggest that they participate in important cellular processes. However, only in few cases the cellular function of FKBPs has been elucidated (He et al. 2004; Romano et al. 2005).
The FKBP family consists of multiple members that are distinguished by molecular weight ranging from 12 kDa (hFKBP12, AtFKBP12) (Standaert et al. 1990; Faure et al. 1998a) to over 77 kDa (wFKBP77) (Kurek et al. 1999). They are found in almost all subcellular locations including endoplasmic reticulum (Luan et al. 1996; Zhang et al. 2004), chloroplast (Luan et al. 1994), cytosol (Standaert et al. 1990; Dwivedi et al. 2003), nucleus (Jin et al. 1992), and mitochondria (Breiman et al. 1992).
The human FKBP12, which comprises of a single FKBP domain, is the gene that provides the common denominator for the family. The proteins possessing sequence homology to this domain have been designated members of the FKBP family with FKBP12-like domains.
Multidomain FKBPs are structurally characterized by additional protein modules, typically a tripartite tetratricopeptide repeats (TPR) domain, which is involved in protein–protein interaction (Goebl and Yanagida 1991) and a calmodulin-binding domain (CaMBD) (He et al. 2004). In the large FKBPs studied, the first FKBP12-like domain exhibits PPIases activity in vitro (Chambraud et al. 1993; Blecher et al. 1996). The large FKBP group is exemplified by the mammalian FKBP52, which is the major immunophilin of the multiprotein glucocorticoid receptor complex (Smith et al. 1993).
In plants, high-molecular weight FKBP members have been identified from wheat (Blecher et al. 1996; Kurek et al. 1999), Arabidopsis (Vucich and Gasser 1996; Vittorioso et al. 1998; Geisler et al. 2004; Perez-Perez et al. 2004), maize (Hueros et al. 1998), and recently also from rice (Magiri et al. 2006).
The completion of the Arabidopsis genome sequencing revealed 23 FKBP members, being one of the largest FKBP families identified. The Arabidopsis FKBPs can be classified into those bearing single-or multiple-domains, among which seven members contain multiple domains (He et al. 2004; Romano et al. 2005).
The role of two Arabidopsis large FKBPs has been recently explored. The PASTICCINO1 (PAS1) protein is a member of the large FKBP-family and the mutants, impaired in the AtFKBP72 coding region, display severe developmental abnormalities throughout the growth stages, altered response to cytokinins and down-regulation of primary auxin responsive genes, suggesting that this protein play a role in cytokinin and auxin response (Vittorioso et al. 1998; Harrar et al. 2003). The characterization of the twisted dwarf1 (twd1) (Geisler et al. 2003) and ultracurvata 2 (UCU2) (Perez-Perez et al. 2004) both knockouts of At FKBP42, revealed common pleiotropic effects leading to dwarfism, distorted roots, and stems as well as helical rotation of organs. The AtFKBP42 protein was shown to interact with ABC transporters, and when mutated, vacuolar transport is affected (Geisler et al. 2003, 2004). In addition TWD1 was shown to bind HSP90 (Kamphausen et al. 2002). Recently, the crystal structure of the AtFKBP42 was solved (Eckhoff et al. 2005; Weiergraber et al. 2006). The function of the other, structurally different, plant large FKBPs in the cell requires further investigation.
Among the large FKBPs, which have been identified in plants, the expression of the wheat wFKBP77 (Kurek et al. 1999) and rice rFKBP65 (Magiri et al. 2006) genes were shown to be elevated by heat stress. The Arabidopsis ROF1 gene expression was found to be increased by wounding or exposure to elevated NaCl levels (Vucich and Gasser 1996).
High molecular-weight immunophilins from mammals (FKBP51, FKBP52, and CyP40) have been identified by their abilities to bind HSP90 via TPR-domains, as components of unliganded steroid receptor heterocomplexes. This interaction occurs through the heat shock protein 90 (HSP90) (Sanchez et al. 1990; Pratt et al. 2004). HSP90 is a highly abundant cytosolic protein, which is important under both stress and physiological conditions. It acts as the central platform for distinct multichaperone complexes as the steroid receptor complex, with additional cofactors, including p23 (Pirkl and Buchner 2001; Riggs et al. 2003).
The crystal structures of the TPR-containing immunophilins: CyP40 (Ward et al. 2002), FKBP51 (Sinars et al. 2003) and FKBP52 (Wu et al. 2003, 2004) in complex with a peptide containing the C-terminal residues of HSP90 (MEEVD motif) enabled the identification of the five residues of the TPR, essential for HSP90 binding. These five key residues were also found to be conserved in the TPR-containing chaperones HOP (HSP70 and HSP90 organizing protein, also known as p60 or Sti1p) (Scheufler et al. 2000; Carrigan et al. 2006) and in protein phosphatase 5 (PP5) (Das et al. 1998). Recent experiments in plants have suggested that the multiprotein complex components of this chaperoning machinery are likely to be conserved between animal and plant kingdoms (Owens-Grillo et al. 1996; Reddy et al. 1998; Harrell et al. 2002).
From studies of the mammalian and plants FKBPs it appears that the common denominator of this protein family, namely the FKBP12-like domain, represents an evolutionary conserved domain relevant to taxonomic resemblance (Galat 2004). However, each member has gained an unique function in context to its intracellular localization and protein client (Geisler et al. 2004, Davies et al. 2005).
The FKBP superfamily contains pairs of proteins that share a high-sequence identity but were shown to have different functions. Such is the case for the mammalian homologues FKBP51 and FKBP52, which share high-sequence similarity, but associate with different steroid receptors (Davies et al. 2005). In plants, the wheat FKBP73 and FKBP77 were shown to be cognate and heat-stress induced counterparts (Kurek et al. 1999). The Arabidopsis large FKBPs, AtFKBP62 (ROF1), and AtFKBP65 (ROF2) share similar domain structure and high sequence identity.
In order to gain better understanding for the role of the two homologues, the Arabidopsis ROF1 and ROF2 have been characterized. A detailed glucuronidase (GUS) histochemical analysis of ROF1 promoter at various stages of plant development, revealed specific vascular staining pattern in young and mature vegetative tissues and in reproductive organs. Differential expression between the two ROF genes at RNA and protein levels was displayed. Furthermore, we show that the Arabidopsis ROF1 can specifically bind HSP90.
Materials and methods
Arabidopsis thaliana plants ecotype columbia (Col-0) were used for ROF1 and ROF2 promoter analysis, and ecotype wassilewskija (WS) for creation of rof1 and rof2 mutants and for RNA and protein expression analysis. Plants were grown at long-day conditions (16 h light, 8 h dark) at 22 °C, unless differently mentioned.
strains, vectors, and cosmids
MJL12 (Arabidopsis thaliana genomic DNA, chromosome 3)
Amplification of ROF1 promoter
Cloning ROF1 and ROF2 promoters upstream β-GUS gene
Christou and Ford (1995)
Binary vector for Arabidopsis transformation
Cloning RNAi constructs for ROF2 silencing
Wesley et al. (2001)
Binary vector for cloning NotI fragment from pKANNIBAL
Expression vector for ROF1 mutated forms
Primers used for PCR in this study
ROF1 promoter F
ROF1 promoter amplification
ROF1 promoter R
ROF2 promoter F
ROF2 promoter R
GUS transgenic lines verification by PCR
ROF2 1197–1418 F
ROF2 1197–1418 R
ROF1 3′UTR F
ROF1 3′UTR R
ROF2 3′UTR F
ROF2 3′UTR R
β-actin I F
β-actin I R (An et al. 1996)
5′ ROF1 F
ROF1 coding sequence amplification
3′ ROF1 R
ROF1 K404A & N408A F
Mutation of Lys 404 to Ala, and Asn 408 to Ala
ROF1 N454D F
Mutation of Asn 454 to Asp
ROF1 K484A F
Mutation of lys 484 to Ala
ROF1 R488A F
Mutation of Arg 488 to Ala
Genomic DNA from transgenic plants (shoots) was isolated by the method described by Clark (1997). PCR was performed to verify the T-DNA insertion by using the primers of ROF1 promoter F or ROF2 promoter F with the GUS primer (Table 2). The pCAMBIA1300 containing the promoter fragments were used as a positive-template DNA. The number of incorporated copies of the GUS gene was analyzed by Southern blot. This was carried out by digesting 4 μg of genomic DNA isolated from the 4 ROF1 and ROF2 transgenic T3 lines (expressing the GUS reporter gene) with NcoI and HindIII, respectively. Following enzymatic digestion the DNA was separated using 0.8% agarose gels, transferred onto nylon membrane (Hybond-N+ Amersham Pharmacia Biotech) according to manufacturer instructions, and hybridized with a fragment of the GUS gene and NOS terminator (EcoRV-KpnI from pWRG2920) as a probe. The analysis showed that two ROF1 transgenic lines (ROF1.7.1 and ROF1.10.1) and two ROF2 lines (ROF2.2.1 and ROF2.15.1) contained two copies of the insert while the other lines contained more than two copies except ROF2.8.1 containing one copy. These lines were subjected to GUS expression assays.
For isolation of the Arabidopsisrof1 knockout mutant, a tagged T-DNA library available in versailles was screened (this work was done in collaboration with Drs. Bellini and Faure in INRA, Versailles, France). The rof1 knockout mutant was found to contain a T-DNA insertion in its third intron; consequently the protein is not produced.
For isolation of rof2 mutant we used the SiRNA strategy (Fire et al. 1998). Specific primers according to ROF2 nucleotide sequence were designed (nucleotide 1,197–1,418) (Table 2). This region was cloned into pKANNIBAL transient vector (Table 1). A pKANNIBAL NotI fragment containing the CaMV35S promoter, the ROF2 regions separated by the intron, and the OCS terminator, was cloned into pART27 (Table 1) for transformation mediated by Agrobacterium into Arabidopsis plants (as detailed above). Transgenic plantlets were selected on antibiotics and than transferred to soil for three generations. Subsequently, homozygous T3 2-week-old seedlings from eight independent lines were analyzed by Western blot for ROF1 and ROF2 expression. In three independent lines the ROF2 protein was not detected, while in the other lines the ROF2 expression level was reduced. Under normal growth conditions no obvious phenotypes were found, although the ROF1 or ROF2 proteins were not detected by Western blot (Fig. 8A). The generation and characterization of the rof1/rof2 double mutant is in progress.
Growth and heat-stress conditions
Arabidopsis thaliana plants were grown on soil (as detailed above) for about 30 days. Seedlings were germinated on MS-medium plates (Murashige and Skoog including vitamins, Duchefa) containing 0.2% MS (w/v), 2.5 mM MES, 1% sucrose, and 0.5% plant agar (Duchefa) for 7–14 days. After 2–3 days at 4 °C in the dark for stratification, seeds grown on soil or plates were grown at 22 °C in long-day conditions.
The mature Arabidopsis plants and the seedlings were subjected to heat-stress treatments. For histochemical GUS expression analysis, 2-weeks old ROF2–GUS transgenic seedlings were incubated for 3 h at 37 °C. For reverse transcription polymerase chain reaction (RT-PCR) analysis, mature wild-type (WT) plants were exposed to 37 °C for 1 h in an incubator with high humidity to avoid drought stress. Then tissues (open flowers, buds, siliques, stems, cauline leaves, and rosette leaves) were collected and immediately frozen in liquid nitrogen, followed by RNA isolation. For protein analysis 2-week old seedlings were subjected to various temperatures and times (see legends for figures).
To obtain an estimate of the relative levels of ROF1 and ROF2 mRNAs at various tissues and in response to heat treatments, locus IDs were used to query the genevestigator A. thaliana microarray database and analysis toolbox (https://www.genevestigator.ethz.ch) (Zimmerman et al. 2004, 2005).
Chosen experiments used for ROF1 and ROF2 tissue and heat-stress profiles
# of chips
Development baseline III
Seed and silique development
Silique senescence in Arabidopsis thaliana
Thomas Yang/Jerry Roberts
Young (green) seed
Seed and silique development
Gene expression of Col-0 seeds at 0, 1 and 3 h post imbibition
Role of COV in vascular patterning
Development baseline II
Development baseline II
Development baseline II
Rosette development (phase change)
Root & shoot development
Plant organs: comparison with CAGE data
The digital northern tool was used to retrieve signal values of ROF1 (At3g25230), ROF2 (At5g48570), and for the heat stress series also HSP 17.6 class I (At1g59860), in selected set of chips. Data average was carried between chips of different experiments only if the biological material was congruent and the standard deviation value was no greater then 20% of the average value.
Signal intensity value below 200 means no or very-low expression, signal intensity value between 200 and 800 means the noise area where results are ambiguous, and signal intensity value higher than 1,000 means expression.
Sequences alignment was carried out using ClustalW (http://www.ebi.ac.uk/clustalw/) and BoxShade (http://www.ch.embnet.org/software/BOX_form.html) softwares.
The Swiss pdb program (www.expasy.org) was used for the modeling of ROF1 TPR domain-structure based on HOP crystal structure, together with its bound HSP90-peptide (HSP90-MEEVD) (Scheufler et al. 2000). The mutational analysis of the HSP90 interacting domain of Cyp40 (Ward et al. 2002) was used to produce mutated ROF1 in order to identify the amino acids that contribute to the HSP90 binding.
GUS expression analysis
Histochemical GUS staining
The histochemical localization of GUS in transformed Arabidopsis plants was preformed as described by Jefferson (1987). Arabidopsis seedlings or tissues (roots and different stages of leaves, flowers, and siliques) expressing ROF1 or ROF2 promoter–GUS fusion were freshly collected and immersed in a histochemical reaction mixture containing 1.1 mM X-Gluc (5-bromo-4-chloro-3-indolyl-β-D-Glucuronide; Duchefa) in 100 mM potassium phosphate buffer pH 7.0, 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide and 0.1% Triton X-100. Plant material was vacuum infiltrated from 5 min (seedlings) to 15 min (mature tissues). The histochemical reaction was preformed in the dark at 37 °C for 20–360 min. Stained seedlings and tissues were incubated in a clearing solution (2-chloral hydrate: 1-lactic acid v/v) for up to 2 weeks and then placed in a 90% lactic acid medium on a slide and examined with a stereomicroscope (Zeiss SV11) and documented by a Nikon coolpix camera.
Histological sections and microscopy
For localization of GUS expression in various stages of flower development (Smyth et al. 1990; Bowman 1994) the stained tissues were cleared in 70% ethanol overnight, fixed in FAA solution (50% ethanol, 5% acetic acid, and 3.7% formaldehyde) and dehydrated using the tertiary butyl alcohol method. The plant material was embedded in paraffin and cut into 8–12 μm cross thick sections using Microtome 820 (Spencer). After removing the paraffin by solvent, the sections were mounted in glycerol and examined by light microscopy. For analysis of GUS expression in transverse sections of transgenic roots, 2-week old stained tissue were washed with water and cut into 1–1.5 cm long sections. The sections were dipped in petri dish containing 3% agar (in parallel to the agar surface). Pieces of agar containing the root sections were taken out and cut into 60–80 μm sections in ice water surrounding using cold vibrotome Leica VT1000S (Leica, Germany). The histochemical sections were observed using Zeiss Axioplan2 microscope and documented using a Nikon digital camera as indicated above.
RNA isolation and reverse transcription polymerase chain reaction analysis
Total RNA was isolated from various tissues (50–100 mg) before and after heat-stress treatment using SV total RNA Isolation Kit (Promega, USA) according to the manufacturer’s instructions. All RNA samples were quantified spectrophotometrically at 260 nm using a nano drop-ND-1000 spectrophotometer (Rockland, DE). First-strand cDNA was produced using 2 μg of total RNA, oligo (dT)15 primer and M-MLV reverse transcriptase (Promega, USA). The resulting cDNA was diluted 10 times and 2 μl were used as a template in 25-μl PCR reaction using gene specific primers designed from the non-homologous regions (3′UTR) of the ROF1 and ROF2 genes (Table 1). The primers were analyzed using Vectors NTI software program. β-actin I was chosen as an internal standard for determining the RT-PCR amplification efficiency among different reactions. PCR amplification was carried out using the following program: 94 °C for 4 min followed by five cycles at 94 °C for 45 s, 55 °C for 45 s and 72 °C for 1 min and then 18 (for β-actin) or 22 cycles (for ROF1 and ROF2) at 94 °C for 30 s, 55 °C for 45 s and 72 °C for 2 min. The reaction was terminated by 72 °C for 5 min. A total of 5 μl of each PCR-product were analyzed by standard 1.5% agarose gel electrophoresis and visualized with ethidium bromide staining. The PCR reactions were performed using the PTC-200-PCR Peltier Thermal Cycler (MJ Research, USA).
Preparation of protein extract for SDS-PAGE
Arabidopsis seedlings (300 mg fresh weight) were homogenized with small amount of sand and 50 μl of extraction buffer [50 mM tris-HCl pH 7.6, 300 mM sucrose, 100 mM NaCl, 10 mM EDTA pH 8.0, 1 mM DTT and 1X Protein inhibitors cocktail (Roche Diagnostics GmbH Germany)]. After adding 600-μl-extraction buffer, the homogenate was vortexed rapidly and immediately cooled in ice. The crude extract was cleared by two successive centrifugations at 14,000 rpm for 10 min. Protein concentration in the supernatant was measured by the Bradford assay (Bradford 1976), with bovine serum albumin as standard.
Thirty micrograms of total protein extract were separated on 7.5% SDS-polyacrylamide gels and transferred to a nitrocellulose membrane. The western blot was preformed as described above (Blecher et al. 1996).
Antibodies used for Western blot analysis
For ROF1 and ROF2 detection, polyclonal antibodies that have been rised against the recombinant wFKBP73 (α-73) (Blecher et al. 1996; Kurek et al. 1999) diluted 1:3,000 were used. For specific detection of ROF2, immunoaffinity purified specific antibodies against ROF2 were used (detailed below).
A specific ROF2 peptide (amino acid 562–575) was synthesized from the 3′-UTR unconserved region, containing an extra cysteine in its amino-terminus (Genemed Synthesis Inc USA). Two milligram of the ROF2 specific peptide were coupled to a 2 ml SulfoLink gel column (Pierce) according to the manufacturer procedure.
IgG enriched ROF2 peptide antisera (7 ml) was prepared by 50% ammonium sulfate precipitation followed by dialysis against PBS. The ROF2-enriched antiserum was coupled (1 h at room temperature) to the SulfoLink gel column containing the ROF2 peptide. After washing the SulfoLink gel column with PBS, the specific affinity bound antibodies were eluted with glycine buffer (100 mM pH 2.5) and neutralized by adding 100 μl of 1 M tris-pH 7.5. The specific ROF2 antibodies were dialyzed against PBS and used for Western blot analysis at a 1:2,000 dilution. For detection of wheat HSP90, the rabbit R2 antiserum against Brassica HSP90 (Krishna et al. 1997) was used.
Labeling with second antibody was preformed by anti rabbit IgG horseradish peroxidase-linked whole antibodies (Pharmacia) diluted 1:15,000. Detection was carried out using ECL Western Blotting Detection Reagents and Analysis System (Amersham Biosciences, UK).
ROF1-wheat HSP90 heterocomplex formation
The ROF1 coding sequence was amplified from 2-week old seedlings cDNA library with specific primers (Table 2), designed with BamHI restriction enzyme for further cloning in expression vectors. The clone was sequenced to confirm lack of mutations.
In order to construct mutated ROF1, a site-directed mutagenesis strategy was applied. Single ROF1 mutation was generated by a PCR reaction using a sense primer, carrying the mutation K484A, with the antisense 3′ ROF1 R primer (Table 2) to amplify a 210 bp fragment at the c-terminal. This fragment was then used as a megaprimer with the 5′ ROF1 F primer to obtain ROF1 K484A. In order to construct a ROF1 TPR domain harboring five mutations, we have further used the “megaprime” technique. A primer carrying R488A mutation was used with 3′ ROF1 R primer and the single ROF1 mutation (K484A) clone as a template. The amplified fragment carrying two mutations was then used as reverse megaprimer with forward primer carrying the mutation N454D (Table 2). The amplified fragment was then used as a megaprimer with 5′ ROF1 F primer to obtain full ROF1 harboring three mutations. Five mutations were achieved with forward primer carrying two mutations K404A & N408D (Table 2) and 3′ ROF1 R primer using the ROF1 three mutations clone as a template. The amplified fragment carrying the five mutations was then used as a megaprimer with 5′ ROF1 F primer to obtain full ROF1 harboring the five mutations. The ROF1 containing single and five mutations were sequenced to confirm lack of additional mutations.
For expression of ROF1, ROF1 K484A (single mutation), and ROF1 five mutations (K484A, R488A, N454D, K404A, N408D), were cloned into pET28 vector (Novagen) with an amino–terminal His-tagged (Table 1). Escherichia coli strain BL21 cells were transformed with the various pET28–ROF1 constructs. For protein expression, single transformed colonies were grown at 37 °C to an OD600 and transferred overnight to 4 °C, then diluted 1:25 and grown at 37 °C to an OD600. Induction was performed with 1 mM IPTG for 4 h at 37 °C. Purification conditions of the ROF1 recombinant proteins were according to the manufacturer instructions (Novagen). For adsorption of the purified mutated proteins to Ni-NTA (Novagen) 300 μl ROF1, 400 μl ROF1 K484A mutant and 550 μl ROF1 five mutations were airfuged at 100,000 x g for 15 min with 0.05% Tween-20. The supernatant was rotated for 1 h at 4 °C with 60 μl Ni-NTA resin (50%), washed four times with 1.5 ml of washing buffer (50 mM Pi, 0.3 M NaCl, 20 mM imidazole, 0.05% Tween-20, pH 8.0) and twice with 1.5 ml of incubation buffer (0.1 M Hepes, 10 mM imidazole, 50 mM KCl, pH 7.6). The pellet was incubated for 30 min at 30 °C with 70-μl wheat germ lysate (Promega) supplemented with 10 μg of recombinant human p23 (Harrell et al. 2002) and protease inhibitor cocktail (Complete-Mini, Roche) in the presence or absence of 100 μg pure flag-TPR from rat protein PP5 purified as described in Silverstein et al. (1997). The tubes were shaken every 2 min during the incubation and washed four times with 1.5 ml of washing buffer (as detailed above). The proteins were resolved by a 10% SDS-PAGE followed by Western blotting using the R2 anti-Hsp90 (Krishna et al. 1997) and anti-wFKBP73 (Blecher et al., 1996).
Genome organization and amino acid alignment of ROF1 and ROF2
Comparison between ROF1 and ROF2 protein sequences with that of the wheat FKBP73, FKBP77, the human FKBP52, and Arabidopsis FKBP42 (TWD1) reveals high similarity to the wheat homologues in their amino acid sequences and domain organization (Fig. 1B). ROF1 and ROF2 possess three FKBP12-like domains, whereas, the mammalian homologue hFKBP52 harbors two FKBP12-like domains and the Arabidopsis TWD1 only one (Fig. 1B). The first FKBP12-like domain (FKBD1) spans amino acid G-39–D-145 and G-46–R-153 of ROF1 and ROF2, respectively, and possesses the 10 amino acids reported to be essential for binding and maintaining the hydrophobic core of FK506 (van Duyne et al. 1993). The second and third FKBP12-like domains (FKBD2 and FKBD3) span amino acids G-155–K-262 of ROF1 and G-163–K-271 of ROF2 and K-272–K-384 of ROF1, and K-281–K-394 of ROF2, respectively. Sequence comparison between ROF1/ROF2 (85% identity), to the wheat and human FKBPs reveals the highest identity to the wFKBP73 (ROF1 87% and ROF2 84%), lower identity to the wFKBP77 (ROF1 78% and ROF2 80%), 58% with hFKBP52 and only 34% to the TWD1 (Fig. 1B).
ROF1 harbors 3-TPR repeats, which span from K-403 to F-436, N-452 to A-485, and L-486 to K-520. The ROF2-TPR repeats are located between K-413 and F-446, N-462 and A-495, and between M-496 and K-530. The CaMBD (calmodulin binding domain) spans from K-529 to F-545 and from K-539 to L-555 of ROF1 and ROF2, respectively. The amino acids identified as essential for HSP90 binding in the mammalian FKBP52 (Wu et al. 2004) are conserved in the Arabidopsis homologues ROF1/ROF2 and TWD1, except for N-408 (Fig. 1B, see asterisks).
Histochemical localization of ROF1 promoter-GUS expression in Arabidopsis plants
ROF1-promoter expression in 7-day-old plants
To examine ROF1 expression pattern in planta, we generated Arabidopsis transgenic lines expressing the β-glucoronidase gene under the control of ROF1 putative promoter (1,848 bp upstream the ATG codon) and analyzed its promoter-GUS activity during developmental stages.
ROF1-promoter expression in 14-and 28-day-old plants
To determine the expression pattern of ROF1 in root-specific cell layers, transverse sections of 14-and 28–day-old roots were prepared. Strong activity was detected only in the vascular bundle (including the internal stele layer and the pericycle) of the young and mature roots (Fig. 3D, E).
Furthermore, we examined the expression pattern of the ROF1-promoter-GUS cassette during leaf development. In young rosette leaves (second to third true-leaf stage) of 14-day-old seedlings, a basipetal gradient of GUS activity was observed (Fig. 3F). A strong expression was detected in the distal part of the leaf, in hydathodes located in the leaf lobes (Fig. 3F, see arrow), in the stomata (Fig. 3J) and in the basal cells of the trichomes (Fig. 3F see arrow, and 3I). In rosette leaves of 28-day-old plants, ROF1-promoter GUS activity was mainly observed along the vascular tissue and in the hydathodes (Fig. 3G, K). The ROF1 promoter was strongly expressed in the leaf tip and hydathodes (Fig. 3H, see arrows), whereas, GUS expression was barely detected in the center and leaf base of young cauline leaf (from 28-day-old plants). Cauline leaves of secondary young inflorescences showed GUS activity in the leaf tip and in the trichomes basal cells distributed along the leaf (Fig. 3L).
ROF1-promoter expression during flower development
Taken together, ROF1-promoter activity was shown to be developmental and organ specific with highest expression in the vascular elements that are common to all organs analyzed.
ROF1 and ROF2 expression in response to heat stress
ROF1 and ROF2 mRNA expression
The overall ROF1 GUS-expression results fit well with the microarray data collected from the AtGenExpress project (supplemental Fig. S2). There are few discrepancies, for example, the decline in ROF1 transcript level during flower maturation in the microarray data was not demonstrated in the GUS assays (Fig. 4). This finding could be explained by the fact that data obtained from the promoter analysis (which enables a more detailed examination involving localization of the activity) was compared to the analysis of mRNA expression at steady-state level.
Plants exposed for 1 h to 37 °C, show increased expression of ROF1 mainly in flowers, siliques, and roots, whereas, under the same conditions ROF2 was significantly increased in all organs tested (Fig. 6).
ROF2-promoter expression in Arabidopsis
After we found that ROF2 transcript is induced by heat-stress conditions (Fig. 6), a detailed ROF2 promoter-GUS analysis was performed on plants after exposure to heat stress (supplemental Fig. S3). The histochemical analysis revealed a pattern similar to that described for ROF1 promoter-GUS activity in plants grown under normal growth conditions. The similarity in expressions is maintained throughout plant development: in 7-day-old seedlings, 14-day-old roots, and in flowers at the various developmental stages (supplemental Fig. S3>, Figs. 2A, 3A, and 4G, respectively). The comparison of promoter expression analysis of ROF1 and ROF2 GUS fusions did not reveal a difference in the target tissues between the two genes.
Does ROF1 and ROF2 protein accumulation correlate with RNA expression pattern?
To test the temperature dependent ROF1 and ROF2 protein accumulation, seedlings were exposed for 3 h to the temperatures of 22–42 °C. ROF1 was expressed at all temperatures tested except for 42 °C (Fig. 8B, upper panel), whereas, ROF2 was only detectable at 35 °C and 37 °C (Fig. 8B). These results indicate that, in contrast to ROF1, ROF2 protein is strictly heat-stress induced, expressed at a narrow range of elevated temperatures in 14-day-old seedlings. In order to study the stability of the protein after its induction at 37 °C, the seedlings were allowed to recover at 22 °C. Both ROF1 and ROF2 were detectable for at least 4 h of recovery from heat stress (Fig. 8C).
In order understand the inexistence of ROF2 at 42°C in plants exposed directly to this temperature, we have exposed the seedling first to 37 °C and only then exposed them to 42 °C (Fig. 8D). The ROF2 protein, which appeared at 37 °C, remained stable at 42 °C for at least 1.5 h, as compared to plants, which were exposed only to 42 °C for 1.5 h, where no ROF2 could be detected (Fig. 8D). These results indicate that the absence of ROF2 at 42 °C is caused by arrest of its synthesis and is not due to its instability at 42 °C (Fig. 8B, D). The stability of ROF1 depends on the period of exposure to 42 °C: the protein is detectable after 1.5 h (Fig. 8D), whereas, it disappears after 3 h (Fig. 8B, D).
ROF1 binds HSP90: amino acids of the TPR domain, found to be essential for HSP90 binding, are conserved in ROF1
The initial hypothesis that ROF1 may bind HSP90 was reached from the high sequence identity between the TPR region of plant and mammalian large FKBPs (Fig. 1B). In order to find the amino acids relevant for the interaction between ROF1 and HSP90, we took advantage of the crystal structure of mammalian co chaperones possessing TPR such as HOP.
The mammalian HOP protein (Hsp70 and Hsp90 organizing protein) functions as an adaptor protein that provides specific binding sites for Hsp70 and Hsp90 (Chang et al. 1997; Chen and Smith 1998). This protein possesses the TPR-domain and its crystal structure together with its target peptide the HSP90-MEEVD has revealed the amino acids essential for this interaction (K229, N233, N264, K301, and R305) (Scheufler et al. 2000).
To characterize the interaction between ROF1 and HSP90, we performed a HOP-modeled ROF1 comparison, by superimposing the ROF1-TPR domain on the crystal structure of HOP C-terminal TPR domain (TPR2A domain), in complex with the five C-terminal residues of HSP90 (MEEVD peptide) (by using www.expasy.org). ROF1 putative residues involved in binding HSP90 were identified as being: Lys 404; Asn 408; Asn 454;Lys 484; Arg 488 corresponding to the HOP amino acids which were shown to be essential for HSP90 binding (Fig. 1B, see asterisks).
To examine the contribution of the amino acids defined as being directly involved in binding HSP90, we have produced site directed mutagenised ROF1 protein in the five amino acids.
To further validate the interaction between ROF1 and HSP90, competition experiments using the TPR containing rat protein phosphatase 5 (PP5) were performed. The PP5 excluded the binding of the ROF1 to HSP90 as observed in Fig. 9 lanes 6–8.
These results clearly prove the concept that the TPR, and specifically the conserved amino acids found by crystallography of HOP to interact with the HSP90–MEEVD peptide, are conserved in plants, and that the HOP-modeled ROF1-HSP90-MEEVD can serve as an efficient model to predict additional interactions between large FKBPs and HSP90.
To get first insights into the function of ROF1 in Arabidopsis development we have characterized its temporal and spatial expression pattern by promoter-driven GUS histochemical analysis. ROF1 promoter expression was pronounced in the vascular tissues of roots and leaves of young seedlings and mature plants, in vascular elements of reproductive tissues, such as stigma, the inner wall of the gynoecium’s and in the stamens filament (Figs. 2, 3, 4 and 5). Furthermore, strong ROF1 promoter activity was observed in restricted cell types of leaves, namely the basal cells of the trichomes, hydathodes, and stomata (Figs. 2, 3).
A common concept that FKBPs are constitutively expressed has been revised in this study, based on the observation that ROF1and ROF2 expression is developmentally regulated at the organ and tissue levels. Promoter analysis, semi quantitative RT-PCR and data analysis from microarray experiments all strengthen this observation. Previously published data describing a comprehensive analysis of all the immunophilins in A. thaliana (He et al. 2004) by semi quantitative RT-PCR indicated that AtFKBP62 (ROF1) and AtFKBP65 (ROF2) transcripts are constitutively expressed (except for roots, expressing low levels of ROF1, or below detection of ROF2). A possibility to explain these discrepancies may be due to technical differences such as usage of more cycles for the PCR reaction by He et al., which we found to minimize the differences between the two genes ROF1 and ROF2.
The pattern of ROF1 and ROF2 expression may point to a potential interaction with known chaperones with a similar expression profile such as HSP90 (Haralampidis al. 2002; Prasinos et al. 2005). The Arabidopsis large FKBP TWD1 (Geisler et al. 2003; Perez-Perez et al. 2004) located at the plasma membrane, has been also found to interact with HSP90 via the TPR domain (Kamphausen et al. 2002). However, the other Arabidopsis large FKBP, PAS1 (Vittorioso et al. 1998) although possessing a TPR-domain was not shown to interact with HSP90, but recently shown to interact with a member of the plant-specific family of NAC transcription factors via its C-terminal domain (Smyczynski et al. 2006). PAS1 was found to be expressed mainly in zones with high levels of mitoic activity such as shoots and root apical meristems (Vittorioso et al., 1998).
In contrast to the dramatic developmental phenotypes of twd1 and PAS1 knockout mutants, the Arabidopsisrof1 and rof2 mutants (Fig. 8A) have not shown clear phenotypes under normal growth conditions. We suggest that although ROF1 is present at 22 °C, its function may appear mainly after exposure to heat stress. At this condition, its close homologue ROF2 could compensate and hence no phenotype was observed. Therefore the recently obtained rof1/rof2 double-mutants and the single rof1, rof2 mutants are currently tested for phenotype after exposure to various heat-stress treatments. It should be mentioned that while ROF1 and ROF2 are close homologues, which may compensate for each other, twd1 and PAS1 are unique in their structure and therefore their phenotypes were possibly easier to detect.
The ROF2 characterized in this study was shown to be a bona fide heat-stress gene. At regular growth condition of 22 °C, ROF2 transcript level was below detection limit, whereas, it was highly expressed in response to heat (37 °C) in all plant tissues tested. The kinetics of the induction and decay, as well as the amounts of ROF2 transcript, were similar to those observed during heat induction for the small heat-shock protein HSP17.6 Class I (HSP17.6-CI) (Wehmeyer et al. 1996; Lohmann et al. 2004).
ROF2 expression was down regulated in mutants defective in heat shock transcription factors (AthsfA1a/AthsfA1b) similarly to major heat shock proteins such as HSP17.6-CI, HSP101, HSP70, and HSP83.1. In these mutants, the ROF2 displayed the most dramatic markdown (Busch et al. 2005). These results point to the possibility that one or both of the Arabidopsis heat-stress transcription factors HsfA1a and Hsf A1b may regulate the expression of ROF2.
Comparison between ROF1 and ROF2 protein accumulation patterns shows that only ROF1 is detected at 22 °C. Differential accumulation between the two ROF proteins was also observed at 42 °C, a temperature in which the ROF1 protein is accumulated, while the ROF2 protein is absent. We interpret this result by the arrest of ROF2 synthesis rather than instability of the protein at this temperature (Fig. 8D).
We have shown that ROF1 binds HSP90, the interaction is specific, and thus ROF1 is a novel HSP90 client protein. The ROF1-HSP90 interaction occurs via the TPR domain in which 5-amino acid-key residues were found to be essential for this interaction (Fig. 9). Using the mammalian HOP (HSP70 and HSP90 organizing protein) crystal structure we have modeled the ROF1 TPR structure. We have identified the amino acids responsible for interaction with HSP90. These amino acids were site directed mutagenized followed by binding assay analysis, which validated our concept of conservation of these amino acids, which are responsible for the interaction between the large FKBPs and HSP90 (Ward et al. 2002; Scheufler et al. 2000; Wu et al. 2004; Sinars et al. 2003).
The expression pattern of ROF1 and ROF2 described in this study shows high resemblance to the pattern of expression of the Arabidopsis HSP90 (Haralampidis et al. 2002; Prasinos et al. 2005). When comparing the reported promoter expression profile of Arabidopsis AtHSP90-3 (At5g56030.1) and AtHSP90-1 (At5g52640.1), we observed a similar pattern between AtHSP90-3 with that of ROF1. For example, both ROF1 and HSP90-3 are highly expressed in bundles of the siliques, roots meristems and vascular system. The expression pattern of ROF2 resembles that of AtHSP90-1, both being detected only after heat stress (Haralampidis et al. 2002; Prasinos et al. 2005).
The AtHSP90-3 and ROF1 promoters harbor common sequences known to mediate heat shock and other-stress stimuli. For example both contain heat-stress elements (HSEs, elements for binding heat shock transcription factor) (Wu 1995) at about -200 bp and CCAAT sequences essential for binding the C/EBP transcription factors (Akira et al. 1990), which acts cooperatively with HSEs (Prasinos et al. 2005). The ROF2, as the AtHSP90-1 promoter contains HSEs and STREs (Stress Response elements), involved in general stress response (Haralampidis et al. 2002).
In conclusion, our biochemical assays, showing interaction between the ROF1 and HSP90, together with the spatial expression pattern and induction by similar stresses of the ROF1/2 and AtHSP90-3/HSP90-1 lead us to the hypothesis that these proteins complexes are formed and function in vivo. For example, the ROF1-AtHSP90-3 may function under regular growth conditions as well as after heat stress, whereas the ROF2-AtHSP90-1 under heat-stress conditions.
HSP90 is known to have a diverse set of client proteins and participate in various developmentally, hormonally, and morphogenetically regulated processes (Queitsch et al. 2002; Sangster and Queitsch 2005). We suggest that part of the multiple functions of HSP90 can be attributed to the specific interactions between HSP90 and either ROF1 or ROF2.
In spite of the similarity between ROF1 and ROF2, their expression patterns emerge to be differentially regulated as shown during heat-stress response.
We propose that the Arabidopsis ROF1 and ROF2 large FKBPs function in concert with specific HSP90 partners and contribute to plant growth and response to environmental factors.
This work was supported by a grant from the Israeli National Academy of Science to A. Breiman; by CONICET–PIP6167 and ANPCyT PICT2003/01–14123 to M. Galigniana; by FEBS short-term fellowship to K. Aviezer-Hagai. We thank Dedi Meiri for his generous contribution and fruitful discussions. We also thank Neomi Dafni for her technical help.