INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Nucleoside diphosphate kinases (NDPKs) are ubiquitous enzymes that transfer phosphate groups from triphosphate nucleosides to nucleoside diphosphates (NDPs) (Parks and Agarwal, 1973). Besides this housekeeping function, recent reports have revealed the involvement of animal NDPKs in other vital processes such as control of cell proliferation (Cipollini et al., 1997), regulation of transcription (Postel et al., 1993; Ji et al., 1995), and protein phosphotransferase (Engel et al., 1998; Wagner and Vu, 2000). In plants, phytochrome B response (Choi et al., 1999), UV-B light signaling (Zimmermann et al., 1999), and hormone responses (Nato et al., 1997; Novikova et al., 1999) are among the processes in which NDPK isoforms have been shown to be involved.

The oligomeric structure of NDPKs has been shown to be important for function. With the exception of some prokaryotic NDPKs, which are tetrameric, most of the known NDPK isoforms are active as hexamers, including the chloroplastic NDPK II from spinach (Yang and Lamppa, 1996) and the human mitochondrial Nm23-H4 (Milon et al., 2000). However, the organellar isoforms lack the typical C-terminal motif that is believed to stabilize the hexameric conformation of cytosolic isoforms (Webb et al., 1995). A correlation between the stability of the NDPKs oligomeric structure and its function has been observed in vitro in Dictyostelium (Mesnildrey et al., 1997). Whereas the wild-type hexameric NDPK lacks DNA-binding activity, point mutations cause destabilization of the hexamer structure that lead to dimer formation and enable this mutant protein to bind DNA.

Interactions of NDPKs with different types of proteins have been reported in several cases. For example, in human erythrocytes dimers of nm23-H1 and of glyceraldehyde-3-phosphate dehydrogenase form a heterotetrameric complex that has Ser/Thr phosphotransferase activity (Engel et al., 1998). In Arabidopsis, interaction between phytochrome B and NDPK 2 has been shown to occur (Choi et al., 1999). Also commonly found but less studied is the interaction of NDPK with heat shock proteins. In Poeciliopsis lucida a 16-kD NDPK modulates the activity of Hsc70 (Leung and Hightower, 1997). Moreover, in Escherichia coli copurification of NDPK and Dna K, a Hsp70 protein, suggests a similar type of interaction (Barthel and Walker, 1999).

NDPK isoforms have been found in the matrix as well as in the inter-membrane space of mitochondria (Troll et al., 1993; Lambeth et al., 1997; Milon et al., 1997; Struglics and Håkansson, 1999). In animals, matrix NDPK isoforms have been suggested to catalyze transfer of the phosphoryl group from GTP, produced by the TCA cycle, to ATP (Herbert et al., 1955). However, in plants no GTP is directly generated by the TCA cycle (Palmer and Wedding, 1966). Early reports suggested that the function of an inter-membrane isoform would be generation of triphosphate nucleosides needed in the cytosol, using the ATP produced by the mitochondrion as substrate (Pedersen, 1973). However, the functions of the different isoforms remain under debate. Struglics and Håkansson (1999) purified the first plant mitochondrial NDPK isoform and suggested an inter-membrane space localization. This 17-kD isoform, purified from pea (Pisum sativum L. cv Oregon sugarpod) mitochondria, shows auto-phosphorylation on Ser residues, which is characteristic of the human NDPK isoforms involved in signal transduction (McDonald et al., 1993; Postel et al., 1993). Sequence and phylogenetic analysis of the cDNA encoding for this mitochondrial isoform (Escobar Galvis et al., 1999) revealed high similarity with NDPK3 from Arabidopsis and the chloroplastic NDK III from spinach. The similarity among these isoforms might reflect an analogous localization within the organelles: the mitochondrial inter-membrane space and the inter-envelope in chloroplasts. We believe that, due to its subcellular localization and biochemical features (Struglics and Håkansson, 1999), the pea mitochondrial NDPK (mtNDPK) possibly acts as a signaling element between mitochondria and cytosol.

Expression of NDPKs varies between different tissues and developmental stages. In wheat grains, NDPK protein levels were found to be more abundant in the embryo than in pericarp-testa (Hurkman et al., 1998). The pea mtNDPK showed higher transcript levels in young leaves and reproductive tissues as compared with mature leaves and roots (Escobar Galvis et al., 1999). In addition, stress conditions, such as wounding in tomato has been shown to affect transcription of an NDPK isoform (Harris et al., 1994). Moreover, phosphorylation of a NDPK isoform was affected by heat stress in sugarcane (Moisyadi et al., 1994). Nevertheless, heat-stress response in plants has been shown to involve proteins such as NDPK (Moisyadi et al., 1994) and BiP (binding protein) (Hurkman et al., 1998). Studies on heat-stress response in plant mitochondria have been mainly focused on characterization of heat shock proteins (HSPs) (Neumann et al., 1993; Lund et al., 1998). Only one mitochondrial small heat shock protein, which provides thermotolerance to respiratory complex I (Downs and Heckathorn, 1998), has been functionally studied. Investigations of the possible involvement of other proteins such as NDPK in mitochondrial response to heat stress are therefore relevant.

The purpose of this work was to functionally characterize pea mtNDPK, investigating tissue specificity in expression as well as a possible role in response to different kinds of stress. In this paper, by using immunocytochemistry, we show a higher expression of the pea mtNDPK protein in young leaves and the reproductive parts of the flower bud as compared with mature leaves and vegetative parts of the flower. We also report a novel interaction of a 86-kD protein with the pea mtNDPK under heat stress, providing evidence for a role of this mitochondrial NDPK isoform in stress response in plants. In addition, studies on the oligomerization of the pea mtNDPK revealed that this protein could form protein complexes of different sizes.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Tissue-Specific Expression of the Pea mtNDPK

To avoid cross-reaction of the mtNDPK antibody with other NDPK isoforms, we selected a region at the C terminus of the protein for antibody production. This region has no significant similarities with other plant NDPKs (Escobar Galvis et al., 1999). When isolated mitochondria were probed with the antibody, only one band (16.5 kD) in the region corresponding to NDPK isoforms (approximately 17 kD) was detected (Fig. 1, lanes 1-4). The size of the detected protein corresponds to the size of the purified mitochondrial NDPK (lane 5, supplied by Struglics and Håkansson). No cross-reaction of the anti-mtNDPK with any of the chloroplast isoforms was observed (lane 6).

View larger version (85K): [in this window] [in a new window]   Figure 1.   Western analysis of the pea mtNDPK in various subcellular fractions. Lane 1, Twenty-five micrograms of flower mitochondria; lane 2, 25 µg of root mitochondria; lane 3, 25 µg of 7-d-old leaves mitochondria; lane 4, 25 µg of 9-d-old leaves mitochondria; lane 5, purified pea mtNDPK (according to Struglics and Håkansson, 1999); lane 6, 25 µg of purified chloroplasts.

We were able to detect the presence of the mtNDPK in isolated mitochondria from flowers, leaves (7 and 9 d old), and roots using western analysis. No significant differences were observed when mitochondria isolated from the different tissues were compared (Fig. 1, lanes 1-4). Based on these observations we concluded that the steady-state level of pea mtNDPK does not seem to vary among mitochondria of different origin.

On the other hand, when whole tissues probed with anti-mtNDPK were studied by immuno-light microscopy clear differences in signal were observed within each tissue as well as between leaves of different developmental stages. In flower buds (Fig. 2A), the signal specifically corresponding to anti-mtNDPK was higher in central and lateral parts of the bud than in the outer layers of tissue. When compared with the sample incubated only with the preimmune serum (Fig. 2B), it is clear that the anti-mtNDPK specifically reacts only with ovary, stamen, and petal tissues (Fig. 2A). When the anthers are observed in more detail (Fig. 2C) and compared with the preimmune serum sample (Fig. 2D), the labeling is found mainly in the cytoplasm (which contains the mitochondria) of the inner cells. No label is present in the vacuole. When sections of young leaves (nonexpanded leaves) were labeled we were able to see that the mtNDPK is preferentially localized in the lower mesophyll (Fig. 2E), a tissue that later contains large air space and lower chloroplast number. In the sample probed with the preimmune serum, no such specific localization of label is seen (Fig. 2F). However, in expanded leaves, a photosynthetically active tissue, mtNDPK seems to be found in lower amounts and with no preference for any cell type (results not shown).

View larger version (27K): [in this window] [in a new window]   Figure 2.   Immunolocalization of the pea mtNDPK in flower bud and young pea leaf. Positive fluorescent immunolabelling of the mtNDPK is seen as bright green spots. A through D, Flower bud; E and F, nonexpanded pea leaf. The pictures represent the following: A, longitudinal flower bud section incubated with anti-mtNDPK; B, longitudinal flower bud section incubated with preimmune serum; C, transversal anthers section incubated with anti-mtNDPK; D, transversal anthers section incubated with preimmune serum; E, transversal young pea leaf section incubated with anti-mtNDPK; and F, transversal young pea leaf section incubated with preimmune serum. A, Anthers; O, ovary; ST, stamen; P, petals; SE, sepals; UM, upper mesophyll; LM, lower mesophyll; V, vein. Scale bars = 100 µm.

An 86-kD Protein Is Newly Synthesized upon Heat Stress and Coprecipitates with the Pea mtNDPK

Previous studies have shown involvement of NDPK isoforms in stress responses (Harris et al., 1994; Zimmermann et al., 1999). We therefore investigated mtNDPK expression after exposure to stress conditions such as high salt (400 mM NaCl), oxidative stress (2% [v/v] H₂O₂), cold (4°C), and heat (42°C). Western analyses confirmed the presence of mtNDPK in the different samples (Fig. 3), showing slightly decreased levels at 42°C (lane 4) as compared with the control (lane 1). Decreased levels of mtNDPK were also found in the presence of high salt (lane 2) or H₂O₂ (lane 3), probably a result of increased unspecific proteolytic activity upon exposure to all these types of stress. Thus, neither the steady-state levels of mtNDPK mRNA (Escobar Galvis et al., 1999) or protein levels are regulated by the stress conditions used here.

View larger version (37K): [in this window] [in a new window]   Figure 3.   Western analysis of the pea mtNDPK in crude mitochondria prepared from pea leaves exposed to various stress conditions for 4 h. Lane 1, control; lane 2, high salt stress (400 mM NaCl); lane 3, oxidative stress (2% [v/v] H2O2); lane 4, heat stress (42°C); lane 4, cold stress (4°C).

To detect whether the mtNDPK was synthesized de novo under stress conditions, pea seedlings were supplied with [³⁵S]Met during cold and heat stress. Immunoprecipitation assays of solubilized proteins isolated from crude mitochondria, with the mtNDPK antibody, were carried out to detect differential expression of this protein upon stress. No band corresponding to the 17-kD pea mtNDPK was ³⁵S-labeled (Fig. 4A, lanes 1-3). The mature mtNDPK contains only one Met and one Cys, making labeled protein difficult to detect. However, a single labeled 86-kD band (lane 2) was observed in seedlings that had been exposed to heat stress (4 h at 42°C). Incorporation of [³⁵S]Met into the 86-kD protein was observed after 2 h of heat treatment, increasing in a time-dependent manner for 8 h (Fig. 4B, lanes 1-3). The 86-kD protein coprecipitates with the mtNDPK, suggesting a specific interaction between the two proteins. In addition, the results indicate involvement of the pea mtNDPK in mitochondrial heat-stress response.

View larger version (25K): [in this window] [in a new window]   Figure 4.   Analysis of de novo synthesized protein in pea upon heat and cold stress. A, PhosphorImage of immunoprecipitation of [35S]Met-labeled crude mitochondrial proteins using the pea mtNDPK antibody. Lane 1, Control; lane 2, heat stress (42°C); lane 3, cold stress (4°C). B, PhosphorImage of the time course of incorporation of [35S]Met into the 86-kD heat stress up-regulated protein, immunoprecipitated using the pea mtNDPK antibody. Lane 1, 2 h; lane 2, 4 h; lane 3, 8 h. C, Western blot of immunoprecipitations of crude mitochondrial proteins prepared from pea leaves exposed to various stresses probed with anti-mtNDPK. Lane 1, Control; lane 2, high salt stress (400 mM NaCl); lane 3, oxidative stress (2% [v/v] H2O2); lane 4, heat stress (42°C); lane 4, cold stress (4°C).

Parallel assays were carried out in the absence of the isotope to check the efficiency of the immunoprecipitation assay. Gel blots of the immunoprecipitation assays probed with the mtNDPK anti-serum confirmed the presence of NDPK in all samples (Fig. 4C, lanes 1-5). These results show that the absence of a ³⁵S-labeled band corresponding to the mtNDPK is not due to failure of the immunoprecipitation assay but is probably a reflection of the low synthesis rate and the low specific labeling. No up-regulation of mtNDPK expression level during the stress conditions tested was high enough to bring the amount of mtNDPK above the detection limit. Rather, we can conclude that mtNDPK is not newly synthesized upon cold or heat stress conditions, in agreement with the results from the western analysis (Fig. 1).

The 86-kD Protein Is a Novel Protein

The [³⁵S]Met-labeled 86-kD protein was also detectable by Coomassie staining of the SDS-PAGE gels, allowing sequencing of this heat up-regulated protein. Unfortunately, N-terminal sequencing of the 86-kD protein was inconclusive due to N-terminal blockage of this protein. Using mass spectrometry, sequences of trypsin digested peptides were obtained. The sequences obtained were TWFM(L/I), ATGTVT(L/I) V, and (L/I) SVPTS(L/I). Leu and Ile have the same molecular mass and can hence not be distinguished using this method. However, analysis of the sequences revealed no similarity with other proteins found in the databases, making identification of the 86-kD heat up-regulated protein impossible. We conclude that the 86-kD protein is an as yet uncharacterized novel protein.

The mtNDPK Can Be Found in Various Oligomeric States

Taking into account the reports of correlation between function and structure of the NDPKs in animals (Mesnildrey et al., 1997; Mesnildrey et al., 1998), we wanted to determine the oligomeric structure of this mitochondrial isoform. After gel filtration of mitochondrial soluble proteins and immunodetection, the pea mtNDPK was detected in several fractions with peaks corresponding to approximately 100, 80, 60, 45, 30, and 15 kD (Fig. 5).

View larger version (44K): [in this window] [in a new window]   Figure 5.   Gel filtration of pea mitochondria soluble fraction. Fractions of 1 mL were collected and analyzed using the pea mtNDPK antibody. On the y axis, bottom, quantification of the signal corresponding to the pea mtNDPK (analyzed with ImageQuant 1.2 software, Molecular Dynamics). At the top, the calibration curve showing the used standards.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Western analysis revealed no differences in the amounts of mtNDPK compared with total mitochondrial protein in the various tissues studied in this work. On the other hand, immunolocalization studies revealed high levels of mtNDPK in flower and young leaves, which are tissues that require high mitochondrial activity to cover their energy requirements (Moneger et al., 1994; Smart et al., 1994; Thompson et al., 1998). The higher mitochondrial activity in these tissues could be explained by a higher number of mitochondria (Huang et al., 1994). The results obtained by immunolocalization of the pea mtNDPK are in agreement with previous studies using northern analysis (Escobar Galvis et al., 1999), where higher levels of mtNDPK mRNA were found in young and reproductive tissues. Taking our new observations into account we can conclude that the differences in pea mtNDPK expression previously reported are likely due to differences in mitochondrial number among the studied tissues.

In young nonexpanded pea leaves, the pea mtNDPK was preferentially found in the lower mesophyll, a tissue that is known to contain fewer chloroplasts than its upper counterpart. Since the lower mesophyll is a less photosynthetically active tissue, respiration must be an important source of metabolites and ATP. Furthermore, when mature leaves were studied, the amounts of mtNDPK were found to be very low, probably an indication of lower mitochondrial activity or mitochondrial number. These results are in agreement with previous work showing that the levels of respiratory capacity and efficiency of oxidative phosphorylation decreases with aging in pea leaves (Azcón-Bieto et al., 1983).

Flower buds showed high amounts of mtNDPK localized to the central and lateral parts of the bud, whereas in the peripheral parts the signal was comparable with the sample incubated with the preimmune serum. Development of pea flowers has been shown to be a complex process that involves the presence of common primordia to petals and stamens (Ferrandiz et al., 1999). This peculiarity would explain that tissues like petals and stamens showed similar levels of mtNDPK at this early stage of development (Fig. 2A). In sunflower (Smart et al., 1994), mitochondrial gene expression has been found to be correlated with flower development. Smart and co-workers showed that the mitochondrial -subunit of the F₁-ATP synthase was most abundant in young meiotic cells in anthers, a tissue responsible for the development of haploid microspore cells. This process has a high-energy demand that can only be covered by high mitochondrial activity and/or increased mitochondrial number.

The in vivo labeling assays showed that upon heat stress the pea mtNDPK interacts with a newly synthesized 86-kD protein. The results indicate involvement of the mtNDPK in heat-stress response. Previous work has shown that modulation of protein activity by interaction with NDPK isoforms occurs. In the fish P. lucida NDPK controls oligomerization of Hsc70, thereby affecting its activity (Leung and Hightower, 1997). The nature of the interaction between mtNDPK and the 86-kD protein as well as the possibility that either of the two interacting proteins is involved in regulation of the other remains to be investigated. Besides control of oligomerization, the phospho-transferase activity reported for some NDPK isoforms (Engel et al., 1995, 1998; Wagner et al., 1997; Wagner and Vu, 2000) could regulate activity of other proteins.

Studies of the oligomeric state of the mtNDPK revealed the presence of this mitochondrial isoform in several fractions corresponding to a wide range of molecular masses (100, 80, 60, 45, 30, and 15 kD). It is possible that these fractions represent a step-wise oligomerization from monomer to hexamer. However, if hexamerization is the result of trimerization of dimers that needs the formation of a tetramer intermediate, as has been reported for the Dictyostelium NDPK (Mesnildrey et al., 1998), then the peaks at 45 and 80 kD would involve mtNDPK interactions with other proteins. It was further proposed that interaction of NDPK with other proteins could occur at the tetramer intermediate state, where the structure would allow interaction with a larger substrate. One cannot exclude that such interactions of the pea mtNDPK with other soluble proteins could be detected through gel filtration. For example, in E. coli Dna K and NDPK have been shown to copurify (Barthel and Walker, 1999). An analogous interaction between the mtNDPK and the mitochondrial Hsp70 could explain the peak observed around 80 kD. The peak observed at 100 kD could include a complex of mtNDPK interacting with the 86-kD protein observed in this paper.

In conclusion, we have shown that the pea mtNDPK is likely to be involved in heat-stress response through its interaction with a novel heat shock inducible 86-kD protein.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material

Garden peas (Pisum sativum L. cv Oregon sugarpod) were grown on vermiculite in a growth chamber at 20°C with a 12-h day (25 µE m² s¹).

Mitochondria Isolation

Six grams of pea leaves from 12- to 13-d-old seedlings was incubated for 4 h at 4°C or 42°C. The seedlings were ground with a Polytron (9,500 rpm) in 40 mL of ice-cold homogenization buffer [0.4 M Suc, 50 mM Tris, 1 mM ethylene glycol-bis(oxyethylenenitrilo) tetraacetic acid, 10 mM KH₂PO4, 5 mM 2-mercaptoethanol, 1% (w/v) bovine serum albumine (BSA), 0.1% (w/v) polyvynilpirrolidone-44, pH 7.6]. The resulting supernatant, after centrifugation at 3,100g for 2 min, was spun at 15,300g for 15 min. The obtained pellet was used as crude mitochondria. Percoll (Pharmacia-Biotech, Uppsala) purified mitochondria were isolated as in Boutry et al. (1984) with modifications according to Håkansson et al. (1988). As a control, mitochondria were isolated from seedlings kept at room temperature (22°C) for 4 h.

Chloroplast Isolation

Chloroplasts were prepared from 12- to 13-d-old pea leaves according to Walker (1971).

Antibody

A peptide corresponding to the last 15 amino acids of the C terminus of the pea mitochondrial NDPK (accession no. AF191098) was synthesized and conjugated to keyhole limpet hemocyanid using maleimide crosslinker. Innovagen (Lund, Sweden) produced a rabbit polyclonal antiserum raised against this peptide. Anti-mtNDPK was used in a 1:5,000 dilution for western analysis and 1:50 for immunolocalization experiments.

Western Blotting

Proteins were separated by SDS-PAGE (Laemmli, 1970) using a mini-gel system (Bio-Rad Laboratories, Hercules, CA) and transferred onto nitrocellulose membranes (Hybond ECL, Amersham, Buckinghamshire, UK). Transfer was performed in a Multiphore II NovoBlot unit (Pharmacia) using the transfer buffer of Bjerrum and Schafer-Nielsen (1986) (48 mM Tris, 39 mM Gly, 20% [v/v] methanol) at 5 mA cm² for 30 min. Membranes were blocked in 3% (w/v) BSA (fraction V, Sigma, St. Louis) in Tris-buffered saline (TBS: 100 mM Tris, 150 mM NaCl) at 4°C, overnight. Incubation with the primary antibody (diluted in 3% [w/v] BSA in TBST [TBS containing 0.1% {v/v} Tween 20]) was carried out for at least 4 h at room temperature. The secondary antibody, an alkaline phosphatase-conjugated goat anti-rabbit (Bio-Rad Laboratories) was used at 1:10,000 dilution, and the protein was detected using the Immun-Star Chemiluminiscent Protein Detection Systems (Bio-Rad Laboratories) following manufacturer's instructions.

Immunocytochemistry

Semithin sections were prepared from purified pea mitochondria, pea flower buds, young (6 d old), and mature leaves (fully expanded, 11 d old) as in Marttila et al., (1996). To be brief, the material was fixed in 4% (w/v) paraformaldehyde and 0.25% (v/v) glutaraldehyde in phosphate-buffered saline (PBS) (10 mM Na-phosphate buffer, pH 7.4, 150 mM NaCl) for 2 h at room temperature, dehydrated, infiltrated with medium-grade London Resin White (London Resin Company Ltd., Reading, UK; distributed by TAAB, Aldermaston, UK), and polymerized for 24 h at 58°C. Sections of 1 µm were cut on Superfrost Plus slides (Menzel Gläser, Braunschweig, Germany) and blocked in 5% (v/v) goat normal serum and 1% (w/v) BSA in PBS for 30 min at room temperature. Primary antibody incubation with anti-mtNDPK (1:50 in 1% [w/v] BSA in PBS) was carried out at 4°C overnight. After 4 washes in PBS for 20 min, the sections were incubated with a goat-anti-rabbit IgG conjugated to fluorescein isothiocyanate (Sigma) for 1 h at 37°C. On control slides, either preimmune serum was used instead of the primary antibody, or both primary and secondary antibody were replaced with dilution buffer. Samples were washed with PBS and water and mounted in 20% (v/v) Mowiol 4-88 (Calbiochem-Novabiochem Corporation, La Jolla, CA) in PBS, pH 8.6, containing 0.1% (w/v) phenylenediamine. Immunofluorescence was investigated under UV light (excitation filter 495 nm, barrier filter 520 nm) in a Leica microscope.

In Vivo Labeling under Stress Conditions

Six to eight (12- or 13-d-old) seedlings were cut off and incubated for 4 h at 4°C or 42°C in water containing 0.2 mCi of [³⁵S]Met (Amersham, SJ1015). The control sample was kept at room temperature (22°C) for 4 h. After incubation the leaves were detached and ground in 5 mL of ice-cold homogenization buffer together with 1 g of glass beads. All the following steps were carried out between 4°C to 8°C. The homogenate was filtered through four layers of Miracloth (Calbiochem-Novabiochem Corporation) and spun for 3 min at 2,500g in a microcentrifuge. The supernatant was then spun for 15 min at 16,200g. The resulting pellet was resuspended in 400 µL of solubilization buffer (1% [v/v] Nonidet P-40, 150 mM NaCl, 50 mM Tris, 1 mM phenylmethylsulfonyl fluoride). A soluble fraction was obtained after centrifugation for 5 min at 16,200g, and the pellet was discarded. A time course at 42°C was carried out under the same conditions, collecting samples every second hour during an 8-h period.

Immunoprecipitation

Three-hundred-fifty microliters of the soluble fraction was incubated with the mitochondrial NDPK antibody (1:100 dilution) overnight at 4°C. Proteins were precipitated by addition of Protein G Sepharose 4 Fast Flow (Pharmacia-Biotech). After 1 h of incubation the beads were washed three times in solubilization buffer (0.25% [v/v] Nonidet P-40, 150 mM NaCl, 50 mM Tris, 1 mM phenylmethylsulfonyl fluoride) and once in washing buffer (50 mM Tris-HCl, pH 8.0) following manufacturer's instructions. The proteins were denaturated in reducing sample buffer (100 mM dithiothreitol, 1% [w/v] SDS, 50 mM Tris, pH 7.5) by heating at 95°C for 3 min. Proteins were separated by SDS-PAGE (Laemmli, 1970), and gels were vacuum-dried and exposed to PhosphorImager plates (Molecular Dynamics, Sunnyvale, CA). Data were analyzed using ImageQuant 1.2 software (Molecular Dynamics).

Protein Sequencing

For N-terminal sequencing, immunoprecipitation of crude mitochondrial soluble proteins prepared from pea leaves exposed to heat stress were separated by SDS-PAGE. After electroblotting onto PVDF membrane (Pall Europe Limited, Portsmouth, UK) proteins were detected via Coomassie staining according to the manufacturer's instructions. For internal sequencing, proteins prepared as above were separated by SDS-PAGE and briefly Coomassie stained and destained. Sequence analysis (N-terminal sequencing, in-gel digestion, peptide extraction, mass mapping, and tandem mass spectrometry) were performed by the Protein Analysis Center, Karolinska Institutet (Stockholm). The excised protein was applied to a Procise cLC sequencer (PE-Applied Biosystems, Foster City, CA) for N-terminal Edman degradation.

Internal sequencing was carried out as follows: The Coomassie-stained protein band was cut from the gel, and the piece was placed in an Eppendorf tube for in-gel digestion. To be brief, washing was carried out in 0.2 M ammonium bicarbonate containing 50% (v/v) acetonitrile. The protein was reduced (dithiothreitol) and alkylated (iodoacetamide) followed by in-gel digestion with 0.5 to 3 µg of trypsin (Promega, Madison, WI; modified) in 0.2 M ammonium bicarbonate overnight at 37°C. The tryptic peptides were extracted using acetonitrile in 0.1% (w/v) trifluoroacetic acid, first at 60% then at 40%. Aliquots of the peptide extract were desalted (ZipTip C18, Millipore, Bedford, MA) and analyzed by MALDI mass spectrometry (Voyager DE-PRO, Applied Biosystems). For protein identification, the resulting mass map was analyzed with computer algorithms (MS Fit, Pep Sea, and Pro Found) in screens of internet-accessible sequence databases. For tandem mass spectrometry analysis (Q-TOF, Micromass, Manchester, UK), aliquots were applied using a nano-electrospray ion source and argon as the collision gas. Amino acid sequence interpretation from peptide fragment data was aided by software supplied by Micromass.

Gel Filtration

Purified mitochondria were osmotically ruptured in ice-cold 5 mM EDTA, in the presence of 200 µM ATP. The resulting suspension was spun at 16,200g in a microcentrifuge for 15 min at 4°C, separating the membrane from the soluble fraction. Soluble mitochondrial proteins were separated in Sephacryl S-200 HR column (Pharmacia) using PBS buffer (pH 7.3). The column was operated using the FPLC system (Pharmacia) at a flow of 0.3 mL min¹. The collected (1 mL) fractions were blotted onto nitrocellulose membranes (Hybond-ECL, Amersham), using a BIO-DOT SF slot-blot apparatus (Bio-Rad Laboratories) according to the manufacturer's instructions and probed with the anti-mtNDPK as in the western-blotting analysis. Phosphorylase B (97.16 kD), BSA (66 kD), carbonic anhydrase (30 kD), and cytochrome c (12.4 kD) were used as molecular mass standards.