DnaJ proteins are located in various compartments of the eukaryotic cell. As previously shown, peroxisomes and glyoxysomes possess a membrane-anchored form of DnaJ protein located on the cytosolic face. Hints as to how the membrane-bound co-chaperone interacts with cytosolic soluble chaperones were obtained by examining the affinity between the DnaJ protein and various potential partners of the Hsp70 family. Two genes encoding cytosolic Hsp70 isoforms were isolated and characterized from cucumber cotyledons. In addition, cDNAs encoding Hsp70 forms attributed to the cytosol, plastids and the lumen of the endoplasmic reticulum were prepared. His-tagged DnaJ proteins and glutathione S-transferase–Hsp70 fusion proteins were constructed. Using these tools, it was demonstrated that the soluble His-tagged form of DnaJ protein exclusively binds the cytosolic isoform 1 of Hsp70. This interaction was further analyzed by characterizing the interaction between the glyoxysome-bound form of the DnaJ protein and various isoforms of Hsp70. Specific binding to the glyoxysomal surface was only observed in the case of cytosolic isoform 1 of Hsp70. This interaction was strictly dependent on the presence of ADP. Glyoxysomes did not bind other cytosolic or plastidic isoforms or the BiP-related form of Hsp70. Analyzing the enzymatic properties of cytosolic Hsp70s, we showed that the ATPase-modulating activity of DnaJ was highest when isoform 1 was assayed. Collectively, the data indicate that the partner of the DnaJ protein anchored at the glyoxysomal membrane is the cytosolic isoform 1 of Hsp70. In addition to the chaperones located at the surface of glyoxysomes, two isoforms of Hsp70 and one soluble form of DnaJ protein were detected in the glyoxysomal matrix.
Chaperones of the Hsp70 type function as escorting partners [1,2] or as machinery to support the folding of protein substrates [3,4]. Hsp70s accompanying proteins to their targets may already have made contact with their partners at the ribosome during translation. They possibly maintain this contact until a particular receptor system takes over to guide the protein substrate to the target organelle or the final folding machinery .
In several intracellular transport pathways, Hsp70s act in concert with DnaJ proteins [6,7]. Accordingly, DnaJ proteins and Hsp70s accompany proteins to cytosolic T complexes or to Hsp90 complexes responsible for the maturation of transcription factors and other regulators [1,8,9]. Several investigations have shown that, in Saccharomyces cerevisiae, a DnaJ protein, YDJ1p, facilitates the import of protein into mitochondria and the endoplasmic reticulum [10–12]. Other examples of transport through membranes show that DnaJ proteins are part of the translocation machinery [13,14]. It has been argued that proteins with J domains may participate in this protein transport as components of a molecular ratchet .
Since the detection of a DnaJ protein located on the glyoxysomal membrane , there has been speculation about its function at the cytosol–peroxisome interphase. As later on another DnaJ protein was shown to be a peroxin , a component of the targeting apparatus required for peroxisome biosynthesis, questions about the function of DnaJ protein in the context of peroxisome biosynthesis became even more relevant.
As one of several hypotheses put forward, we followed the idea that the membrane-bound DnaJ protein co-operates with an Hsp70 in providing substrate proteins destined for translocation into peroxisomes/glyoxysomes. Although the identity and function of such protein complexes in vivo have remained elusive, we favor the idea that DnaJ either mediates the liberation of the transport protein from the accompanying Hsp70 or is required to keep the region that contains the targeting sequence in an accessible conformation.
Materials and methods
Preparation of genomic DNA and cDNAs
Seeds of cucumber (Cucumis sativus L., Chinesische Schlangengurken) were grown in continual darkness at 27 °C and > 90% humidity . Genomic DNA was obtained from cotyledons of 7-day-old seedlings homogenized under liquid nitrogen using the DNeasy-Plant-Mini-Kit (Qiagen) according to the manufacturer's instructions.
Total RNA was isolated from cotyledons of seedlings grown in the dark for 2 days  basically following the procedure of Chomczynski & Sacchi . Poly(A)-rich RNA was obtained from the total RNA preparation by affinity chromatography using the Oligotex mRNA Maxi Kit (Qiagen). cDNA was synthesized by utilizing the Superscript™ II reverse transcriptase (Life Technologies) and the SMART™ PCR cDNA Library Construction Kit (Clontech) according to the manufacturer's instructions. cDNAs were cloned into vector pT7T3-18 U (Pharmacia) using Escherichia coli XL2-blue (Stratagene) as host. Standard hybridization techniques were applied for isolation of particular cDNAs .
Preparation of two Hsp70 genes by PCR amplification of genomic DNA
Generation of PCR fragments of genomic DNA was performed as described by Siebert et al. . The adaptor was created by incubating equal amounts of primer A1 (5′-CTAATACGACTCACTATAGGGCTCGAGCGGCCGCCCGGGCAGGT-3′) and primer A2 (5′-PO4-ACCTGCCC-NH2-3′) for 3 min at 95 °C. The solution was allowed to cool to room temperature. After ligation of the adaptor to the digested genomic DNA, the ligation reaction mixture was diluted 10-fold by addition of 180 µL 10 mm Tris/HCl, pH 7.5, containing 1 mm EDTA.
Primary PCRs were conducted in 50-µL volumes containing 5 µL ligated and diluted DNA, 10 mm Tris/HCl, pH 8.8, 50 mm NaCl, 0.08% Nonidet P40, 3 mm MgCl2, 0.2 mm each dNTP, 0.4 µm adaptor primer 1 (5′-CTAATACGACTCACTATAGGGC-3′), 0.4 µm gene-specific primer (cytHsp70 isoform 1: 5′-CGATGCATCGCTAAATCTTCGG-3′, cytHsp70 isoform 2: 5′-CGCCGGTAAAACAGTGGCTTTCAC-3′), and 1 U Taq polymerase (1 U·µL−1; Molecular Biotechnology Institute-Fermentas).
The cycle parameters (Perkin-Elmer Gene Amp PCR System 2400) were as follows: denaturation at 94 °C for 2 s and annealing/extension at 72 °C for 4 min; seven cycles of PCR were used; 32 cycles of PCR followed by denaturation at 94 °C for 2 s and annealing/extension at 67 °C for 4 min and a final annealing/extension time of 8 min.
A secondary PCR was conducted with 1 µL of the primary PCR mixture using adaptor primer 2 (5′-TATAGGGCTCGAGCGGC-3′) and the nested gene-specific primers (cytHsp70 isoform 1: 5′-AAGGATCCTAAATCTTCGGCCAATTAAACGC-3′) and (cytHsp70 isoform 2: 5′-AAACGCGTCAGCTCCGGCACCTTGGTAC-3′). The same reaction composition and cycle parameters were used, except that only five cycles of PCR were performed with annealing/extension at 72 °C and 25 cycles of PCR with annealing/extension at 67 °C.
The PCR product generated from genomic DNA using the nested gene-specific primer for cytHsp70 isoform 1 was digested with NotI and BamHI and inserted into the complementary sites in pcDNA3 (Invitrogen). The PCR product generated from genomic DNA using the nested gene-specific primer for cytHsp70 isoform 2 was digested with NotI and MluI and unidirectionally cloned into pCI (Promega) digested with the same restriction enzymes.
Full-length cDNA of both isoforms of cytHsp70 were generated by PCR using 1 µL PCR-amplified cDNA as template. PCRs were conducted in 50-µL volumes containing 20 mm Tris/HCl, pH 8.8, 10 mm KCl, 10 mm (NH4)2SO4, 0.1% Triton X-100, 2 mm MgCl2, 0.2 mm each dNTP, 0.5 µmXhoI-primer (cytHsp70 isoform 1: 5′-ACGCTCGAGATGGCCGGTAAAGG-3′; cyt Hsp70 isoform 2: 5′-TCTCTCGAGTCCACTACCTCCATTATGG-3′), 0.5 µmMluI-primer (cytHsp70 isoform 1: 5′-AACACGCGTTAATCAACTTCTTCAATC-3′; cytHsp70 isoform 2: 5′-AACCGAAACGCGTAAAATTGGGAACTTAGTCG-3′) and 0.6 U VentR® DNAPolymerase (2 U·µL−1, New England Biolabs). Both PCR products were digested with XhoI and MluI and each undirectionally cloned into pCI (Promega) digested with the same restriction enzymes.
For preparation of glutathione S-transferase (GST)-fusion proteins the coding region of both Hsp70 isoforms was subcloned into pGEX-5x-1 (Pharmacia). Full-length cDNA of both isoforms in pCI were digested with XhoI and NotI and inserted in frame into the complementary sites of pGEX-5x-1. Both constructs were transformed into E. coli M15, grown at 37 °C to D600 = 1.0 and induced with 0.2 mm isopropyl thio-β-d-galactopyranoside at 20 °C for 6 h.
Purification of recombinant proteins
To purify the GST-fusion proteins, induced cells were harvested, frozen, thawed and lysed by sonication in buffer A (NaCl/Pi, 2 mm MgCl2, 2 mm 2-mercaptoethanol). The cleared lysate was mixed with GSH–agarose (Fluka) equilibrated with buffer A and washed consecutively with 40 bed vol. buffer A, 20 bed vol. buffer A supplemented with 2 mm ATP, 40 bed vol. buffer B (50 mm Na2HPO4, pH 7.8, 300 mm NaCl, 2 mm 2-mercaptoethanol) and 20 bed vol. buffer A. GST-fusion proteins were eluted with 5 bed vol. buffer A containing 50 mm GSH and transferred into buffer C (10 mm Mops/KOH, pH 7.2, 150 mm KCl, 3 mm MgCl2) for binding assays.
His-tagged DnaJ protein from cucumber  and Sis1 protein from Apiotrichum curvatum and Phytophthora megasperma were purified as described.
Antiserum raised against the C-terminal region of Hsp70
A cDNA fragment of Hsp70 isoform 2 encoding 116 amino acid residues of the C-terminus was inserted into pQE-9 (Qiagen). The protein was overexpressed in E. coli M15 [pREP4] and purified by chromatography on Ni2+/nitriloacetate/agarose. The antiserum raised against this protein in rabbits was tested for its specificity, and was shown to recognize only cytosolic isoforms of Hsp70.
Purification and subfractionation of glyoxysomes
Glyoxysomes were prepared from 4-day-old etiolated cucumber cotyledons as described previously  and purified by density-gradient centrifugation. Matrix proteins were obtained in the fraction of soluble proteins after osmotic shock of glyoxysomes. After collection of the membrane fraction by centrifugation, the peripheral membrane proteins were solubilized using 150 mm MgCl2. The remaining membranes were purified on a sucrose gradient and used as source of the integral or anchored membrane proteins.
Comparative binding assays with glyoxysomes in the presence of other organelles
Both cytosolic Hsp70species, the plastid Hsp70 and the BiP-related Hsp70, were prepared as radioactively labeled proteins by translation in vitro in the presence of l-[35S]methionine using the TNT-T7-coupled reticulocyte lysate system (Promega). Proteins were transferred into buffer C using NickSpin Columns (Pharmacia). Before the binding assay, each protein preparation was assayed for the amount of radioactivity incorporated into the protein during the in vitro translation. Emphasis was put on the fact that, for each Hsp70 form tested, equal amounts of Hsp70 radioactivity were applied to the organelle preparation. A pellet obtained by centrifugation at 12 000 g, containing mitochondria and plasma membrane besides glyoxysomes, was resuspended in buffer C and incubated with the radioactive proteins in the presence of 2 mm ADP or 2 mm ATP. Subsequently, the glyoxysomes were isolated by density-gradient centrifugation in the presence of 2 mm ADP (or 2 mm ATP). Their protein components were analyzed by SDS/PAGE and the amount of radioactivity confined to Hsp70 was visualized by photoimaging.
As preliminary studies showed that a stable interaction between cytosolic Hsp70 isoform 1 and glyoxysomes was absolutely dependent on the presence of ADP, all following standard protocols included 2 mm ADP during the incubation of the resuspended 12 000 g pellet with the radioactive Hsp70 as well as in the gradient buffers.
Binding assays using Ni2+/nitriloacetate/agarose
Affinity-purified His-tagged fusion proteins were incubated in buffer C containing the combination of chaperones and nucleotides indicated for each case. The incubation was for 5 min at 20 °C and then for 5 min at 37 °C. After removal of traces of large aggregates by centrifugation, supernatants were mixed with 20 µL Ni2+/nitrilotriacetate/agarose beads for 20 min on ice. The beads were washed with 6 mL buffer D (50 mm Hepes/KOH, pH 7.5, 300 mm NaCl, 3 mm MgCl2). As indicated, the respective nucleotides were added to the washing buffer. Eventually, the proteins retained on the beads were eluted with 500 µL buffer B containing 200 mm imidazole. After precipitation with trichloroacetic acid, the proteins were analyzed by SDS/PAGE and Western blotting, or by exposure to an imaging plate in the case of radioactive products.
Single turnover rates of ATP hydrolysis were determined for the two different cytosolic Hsp70 isoforms at 37 °C in buffer C as described previously . [α-32P]ATP (4.3 µCi; ≈ 40 nm) and 1 µm Hsp70 were incubated in either the absence or presence of 2 µm His-tagged DnaJ protein. Aliquots of the reaction mixture were analyzed by TLC on polyethyleneimine–cellulose, and the radioactivity was visualized by autoradiography and quantitated by photoimaging.
To test the hypothesis that the localization of the DnaJ protein at the cytosolic face of peroxisomes may point to a function of DnaJ protein as partner of cytosolic proteins, various forms of DnaJ protein were used as a bait in selecting the partner. Although a cytosolic Hsp70 is the most likely partner of the membrane-bound DnaJ protein, we included in this assay several other forms of Hsp70 to demonstrate the specificity of the interaction with the respective DnaJ protein.
Preparation and comparison of a collection of Hsp70 isoforms
To investigate which cytosolic Hsp70 acts as the partner of the peroxisomal membrane-bound DnaJ protein extending into the cytosol, we first searched for cytosolic Hsp70s by screening several cDNA libraries prepared from diverse developmental stages of cucumber cotyledons. After the isolation of full-length clones coding for the endoplasmic reticulum-located BiP-related Hsp70 and for the chloroplast Hsp70 (cpHsp70) attributed to etioplasts of the dark-grown cotyledons (R. Preisig-Müller, unpublished results), we detected cDNAs encoding the C-terminal half of cytosolic Hsp70s. The sequence information of two different cDNAs was used to isolate fragments of genomic DNA by PCR-based walking in uncloned genomic DNA. Both fragments obtained in this way contained part of the promoter region and the part of the coding sequence missing from the cDNA fragments. Two genes, Hsp70-1 and Hsp70-2, were thus identified. As outlined in Fig. 1A, both genes were characterized by heat shock elements (Hsp70-1: 282 bp upstream of the translation start; Hsp70-2: 160 bp upstream of the translation start) and the presence of one intron (423 bp in gene 1 and 652 bp in gene 2) at the position 72 of the amino acid sequence. The findings of heat shock promoters are in agreement with our hypothesis that storage mobilization and germination development is somehow correlated with a slight heat shock situation.
All 10 cDNA clones subsequently characterized and found to encode cytosolic Hsp70 forms could be grouped into two classes corresponding to the two genomic structures described above. The full-length cDNAs representing the developmental stage of cotyledons during maximal lipid mobilization and corresponding to gene 1 and gene 2 code for proteins of molecular mass 71 408 Da and 70 748 Da, respectively. To facilitate a survey of the cucumber Hsp70 forms used in this study and to show the correct assignment of the isoforms, an alignment is shown of the N-terminal and C-terminal parts of the amino acid sequences. Their targeting sequences and C-termini allow their intracellular location to be defined (Fig. 1B).
Comparison of GST–Hsp70 fusion proteins by their affinity to DnaJ proteins
Both cytosolic proteins Hsp70 isoform 1 and Hsp70 isoform 2 with their N-terminus fused to the C-terminus of GST were isolated from transformed bacteria. The proteins were assayed in vitro for their capacity to bind to DnaJ protein. As DnaJ protein, we used the soluble form corresponding to the peroxisomal membrane-anchored DnaJ species. For technical reasons, the DnaJ protein was applied with an N-terminal His tag which allowed the specific selection of the fusion protein by Ni2+/nitriloacetate affinity material.
By this means, the GST–Hsp70 isoform 1 fusion protein was characterized by its binding to the affinity matrix via the DnaJ protein whereas isoform 2 showed no interaction with the His-tagged DnaJ (Fig. 2). Analysis of the proteins bound to and eluted from the Ni2+/nitrilotriacetate matrix demonstrated that the binding of Hsp70 isoform 1 is dependent on the presence of DnaJ protein. Controls analyzing the presence of DnaJ protein demonstrate that the GST–Hsp70 isoform 1 was bound only when DnaJ protein was also detectable in the mixture of eluted proteins (Fig. 2). Controls with GST alone did not indicate any affinity of the GST towards the Ni2+/nitrilotriacetate matrix. Furthermore, experiments with heterologous His-tagged Sis1 protein  instead of DnaJ protein resulted in selective binding of Hsp70 isoform 1, indicating that other proteins with a J domain also interact with Hsp70 isoform 1 (data not shown).
The formation of the binary complex between DnaJ protein and Hsp70 isoform 1 was further characterized by demonstrating that the presence of ATP prevented both the formation of the complex and its maintenance.
Glyoxysome-bound DnaJ protein as partner of Hsp70 isoforms
To extend our studies of DnaJ protein–Hsp70 interactions previously performed with soluble His-tagged DnaJ protein, efforts were made towards studying the interaction based on the membrane-bound form of DnaJ protein. Furthermore, to perform a comparative study on the tendency of the membrane-anchored DnaJ protein to form complexes with various isoforms of Hsp70, we added radioactively labeled Hsp70 isoforms to a suspension of a mixture of organelles including glyoxysomes. This procedure provides a competitive situation in which the labeled protein ligand can choose between alternatives. After separation of the glyoxysomes from mitochondria, plasma membrane, and unbound proteins, the distribution of the respective Hsp70 within these fractions was determined. This analysis characterized the above mentioned interaction in two ways: (a) Hsp70 isoform 1 only but not isoform 2 or the plastid or the endoplasmic reticulum isoforms was bound to the organelle fraction (Fig. 3B); (b) the binding of Hsp70 isoform 1 was specific to the glyoxysomal fraction (at 55% sucrose). Neither the mitochondrial nor the plasma-membrane-containing subfractions (at 44% or 32% sucrose) contained radioactive protein (Fig. 4B).
To examine in more detail the conditions under which the formation or dissociation of the Hsp70 isoform 1–DnaJ protein complex takes place, the effect of the nucleotide present in the medium was tested. For complex formation, the presence of ADP was required. Replacement of ADP with ATP markedly reduced the interaction of glyoxysomes with Hsp70 isoform 1 (Fig. 3C, lane 2). Equally, ADP was essential also for all further steps encountered by the complex, including purification in sucrose gradients. Leaving out ADP during the density-gradient centrifugation resulted in a dramatic decrease in the amount of Hsp70 isoform 1 in the glyoxysomal fractions (data not shown).
DnaJ protein selectively modulates the ATPase activity of Hsp70 isoform 1
ATPase activities of cytosolic Hsp70 isoform 1 and isoform 2 were determined using the corresponding GST-fusion proteins. Both proteins were purified by affinity chromatography and then subjected to the enzymatic test at a concentration of 1 µm. To demonstrate the time courses of ATP hydrolysis, the chromatographic analyses of the reaction mixtures are shown; they indicate that 50% of the ATP was cleaved after 10 min under the conditions used (Fig. 5A). When 2 mm DnaJ protein was added to Hsp70 isoform 1, the 50% conversion was already reached after 2 min. Unlike isoform 1, the ATPase activity of which was significantly changed by the co-chaperone, the ATPase activity of isoform 2 was not affected. After quantitation by photoimaging of the radioactivities attributed to the ADP, the results of the ATPase test could be summarized as shown in Fig. 5B.
Survey of Hsp70 forms located within glyoxysomes
The data presented here show that a cytosolic Hsp70 was recruited to the cytosolic surface of glyoxysomes, but only when ADP was present. In the absence of ADP, all exogenously added Hsp70 isoform 1 was removed from the glyoxysomes. However, glyoxysomes purified and treated in this way were still found to contain Hsp70 forms.
To study the localization of Hsp70 within glyoxysomes in more detail, we used glyoxysomes that had been highly purified by sucrose-density-gradient centrifugation. Catalytic activities of marker enzymes or amounts of marker proteins (on Western blots) for potential contamination were tested: vanadate-insensitive ATPase (plasma membrane), fumarase (mitochondrial matrix), ribulose bisphosphate carboxylase/oxygenase and triosephosphate isomerase (plastid stroma). Within the glyoxysome-containing fraction exhibiting the highest equilibrium density [equivalent to 54.5% (w/w) sucrose], no contamination by other organelles was observed. We then subfractionated the proteins of glyoxysomes into a matrix fraction, a fraction of membrane-associated proteins sensitive to high-salt washing, and integral membrane proteins not removed from the membrane by high-salt treatment. As a control for the quality of the subfractionation carried out in this particular preparation, we assayed in the subfractions a number of marker proteins (Fig. 6B).
Figure 6 shows that (a) the protein stains revealed all predominant enzyme proteins expected to be present in the matrix or the fraction of peripheral membrane proteins, (b) the distribution of the marker proteins detected by immunodecoration was in full agreement with our present knowledge of intraorganelle enzyme localization in glyoxysomes, and (c) two different forms of Hsp70 were present in the matrix fraction. The two Hsp70 forms exhibited different apparent molecular masses (71 and 78 kDa) and were distinguishable by their isoelectric points (pH 4.9 for the larger form and pH 5.4 for the 71-kDa form). In addition, the two isoforms could be differentiated by the two antisera available to us.
To demonstrate the specificity of the antisera, we carried out a survey of soluble proteins, either localized in the cytosol or originating in this fraction because of organelle breakage. First, an antiserum raised against the C-terminal region of cytosolic Hsp70 was applied to the blot of the two-dimensional electrophoresis. Spots corresponding to ≈ 71 kDa were observed (Fig. 6D, left). Subsequently, we decorated the same blot with the antiserum raised against the recombinant protein derived from the cpHsp70 cDNA. This yielded an additional spot at 78 kDa with a pI of 4.9 (Fig. 6D, right). In an independent experiment, the apparent molecular masses of the two forms in the glyoxysomal matrix fraction were determined more accurately (Fig. 6C).
Many questions about the function of DnaJ proteins have arisen from the findings that a DnaJ protein is anchored to the membrane of plant glyoxysomes  and leaf peroxisomes , and that the DnaJ protein Djp1p plays some role during the import of proteins into fungal peroxisomes . As a contribution to understanding why a DnaJ protein is present on the membrane site facing the cytosol, we have demonstrated that this protein finds a partner in the cytosol and that this interaction is specific. The glyoxysomal DnaJ protein, a YDJ1 homolog, stably interacts only with the cytosolic Hsp70 isoform 1 in the ADP-bound conformation. This specific binding of Hsp70 isoform 1 was demonstrated using a soluble fusion protein containing the glyoxysomal DnaJ structure as well as with purified glyoxysomes as a platform for the DnaJ protein. Interestingly, another protein containing a J domain but not the zinc finger region, i.e. Sis1 protein, also showed a binding preference for Hsp70 isoform 1.
If we hypothesize that DnaJ protein acts by binding to Hsp70, thus mediating the ‘holding’ of the complex with the cargo protein, we must consider primarily the influence of DnaJ protein on the ATPase-affecting conformation changes of the cytosolic Hsp70. DnaJ protein would then be a stabilizing factor located at the site at which the substrate protein waits for further partners. This would be necessary if a folding intermediate has to be released from one protein before it binds to another at a specific site on the peroxisome. If, however, the DnaJ protein is more directly involved in the import process at the peroxisomal membrane, we could envisage it as being a component of a multiprotein complex formed before the import and required for the translocation into the peroxisome.
To differentiate ultimately between the possible ways by which the membrane-bound form of DnaJ protein functions in peroxisome biosynthesis, we need data on whether the sidedness of the anchored DnaJ protein changes and if a dual localization outside and inside the organelle is possible. The role of glyoxysomal DnaJ protein resembles that of two other J proteins: the yeast DnaJ protein YDJ1p  and the human DnaJ homolog dj2  both facilitate protein import into mitochondria.
Unlike the Hsp70–DnaJ protein complex which is located before the translocation step as discussed here, BiP molecules associate with a translocation substrate subsequent to the interaction of the J domain of the Sec63p  and therefore acts within the target organelle. In both cases, the specific interaction of Hsp70 with the modulator DnaJ protein takes place at the site of Hsp70 function and at a stage when Hsp70 function requires modulation by cofactors.
In E. coli, the DnaJ-triggered conversion of DnaK–ATP to DnaK–ADP–Pi occurs simultaneously with ATP hydrolysis . In the absence of GrpE, DnaJ forms a stable and tightly binding ternary complex with peptide–DnaK–ADP–Pi. In bacteria, GrpE appears to control the chaperone cycle by transient interactions with DnaK and by promoting the dissociation of the ADP complex . In the case of complexes with eukaryotic cytosol-oriented DnaJ protein, however, the complex with Hsp70 is less tight and has to be stabilized by Hip or other partners [7,9,33].
Our analysis of glyoxysome subfractions (Fig. 6) indicated that, in addition to the membrane-anchored DnaJ protein behaving in the solubilization tests like an integral membrane protein, a soluble matrix form of DnaJ protein exists. This form is distinct from the previously reported membrane-bound species  and the DnaJ protein attributed to a membrane protein complex .
Notably, Hsp70 isoforms were also detected in subfractions obtained from purified glyoxysomes. Using selective antisera against either the plastid 78-kDa form or the 71-kDa cytosolic forms characterized by a C-terminal region containing the EEVD motif, we found that the main portion of both Hsp70 isoforms is confined to the matrix compartment (Fig. 6B). In glyoxysomes and plastids of watermelon cotyledons, two forms of Hsp70 have been identified by immunocytochemistry which are encoded by a single gene  but are targeted alternatively into two organelles. A 73-kDa peroxisomal Hsp70 form has been described as a component of a protein complex integrated into the boundary membrane . At present, these data taken together do not provide a consistent picture of the distribution and function of Hsp70s and DnaJ proteins in peroxisomes. Probably, a rather complex picture, as found for mitochondria, will eventually be found.
This work was supported by the Deutsche Forschungsmeinschaft (SFB 286). The contribution of Dr R. Preisig-Müller during the initiation of this work is thankfully acknowledged.
Enzymes: ATPase (EC 184.108.40.206); fumarase (fumarate hydratase, EC 220.127.116.11); glutathione S-transferase (EC 18.104.22.168); malate synthase (EC 22.214.171.124); ribulose bisphosphate carboxylase/oxygenase (EC 126.96.36.199); thiolase (acetyl-CoA C-acyltransferase, EC 188.8.131.52); triose phosphate isomerase (EC 184.108.40.206).Note: a web site is available at http://www.chemie.uni-marburg.de/