SEARCH

SEARCH BY CITATION

Keywords:

  • CD4+ T-cell proliferation;
  • dissociation-enhanced lanthanide fluorescent immunoassay;
  • HLA-DR binding;
  • heat-shock proteins 70;
  • point mutations

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References
  10. Supporting Information

Heat-shock protein 70 (Hsp70)–peptide complexes are involved in MHC class I- and II-restricted antigen presentation, enabling enhanced activation of T cells. As shown previously, mammalian cytosolic Hsp70 (Hsc70) molecules interact specifically with HLA-DR molecules. This interaction might be of significance as Hsp70 molecules could transfer bound antigenic peptides in a ternary complex into the binding groove of HLA-DR molecules. The present study provides new insights into the distinct interaction of Hsp70 with HLA-DR molecules. Using a quantitative binding assay, it could be demonstrated that a point mutation of amino acids alanine 406 and valine 438 in the substrate binding pocket led to reduced peptide binding compared with the wild-type Hsp70 whereas HLA-DR binding remains unaffected. The removal of the C-terminal lid neither altered the substrate binding capacity nor the Hsp70 binding characteristics to HLA-DR. A truncated variant lacking the nucleotide binding domain showed no binding interactions with HLA-DR. Furthermore, the truncated ATPase subunit of constitutively expressed Hsc70 revealed similar binding affinities to HLA-DR compared with the complete Hsc70. Hence, it can be assumed that the Hsp70–HLA-DR interaction takes place outside the peptide binding groove and is attributed to the ATPase domain of HSP70 molecules. The Hsp70-chaperoned peptides might thereby be directly transferred into the binding groove of HLA-DR, so enabling enhanced presentation of the peptide on antigen-presenting cells and leading to an improved proliferation of responding T cells as shown previously.


Abbreviations
ATPase

adenosine triphosphatase

CD

cluster of differentiation

CLIP105–117

Class II-associated invariant chain peptide fragment SKMRMATPLLMQA

DELFIA

dissociation-enhanced lanthanide fluorescent immunoassay

HA307–319

haemagglutinin fragment PKYVKQNTLKLAT

HSP

heat-shock proteins

HSP70

heat-shock proteins of the 70 000 molecular weight family

HSP90

heat-shock proteins of the 90 000 molecular weight family

Hsp70

heat shock protein of 70 000 molecular weight

Hsp90

heat shock protein of 90 000 molecular weight

Hsc70

cytosolic heat shock protein of 70 000 molecular weight

TT947–966

tetanus toxin fragment FNNFTVSFWLRVPKVSASHL

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References
  10. Supporting Information

Heat-shock proteins (HSP) have been shown to play crucial roles in a variety of intracellular processes. In addition, important immunological functions in both innate and adaptive immunity have been reported for most HSP molecules (reviewed in ref. [1]). Heat-shock protein molecules may be released from cells by necrotic cell death, e. g. after infection, inflammation or in tumour cells.[2, 3] Released HSP are able to induce cytokine secretion and expression of co-stimulatory molecules and to enhance antigen presentation in antigen-presenting cells.[4, 5] Besides these innate HSP effects, mainly members of the 70 000 and 90 000 molecular weight (MW) HSP families (HSP70 and HSP90) additionally contribute to the adaptive immune response.[1, 3] For example, extracellular Hsp72 has been reported to induce antigen-specific immune responses in consequence of stress and is thought to function as a systemic alert signal.[6] These effects are attributed to the ability of HSP70 and HSP90 molecules to interact with a wide variety of other peptides and proteins.

The HSP70 molecules consist of a nucleotide binding domain with an ATPase active site A and a substrate binding domain that is subdivided in the substrate binding pocket S and the C-terminal lid C. The binding of substrate such as antigenic peptides is ATP-dependent and causes a multitude of conformational changes of all subunits.[7, 8] Extracellular Hsp70–peptide complexes can be taken up into antigen-presenting cells by macropinocytosis or receptor-mediated endocytosis via specific receptors (reviewed in ref. [9]). Former studies have shown that cluster of differentiation 91 (CD91), along with other scavenger receptors, is one of these specific receptors.[9-11] The uptake of Hsp70–peptide complexes results subsequently in an enhanced presentation of Hsp70-chaperoned peptides via MHC class I and II, so inducing improved antigen-specific cytotoxic T-cell and CD4+ T-cell activation.[12-16]

Although the HSP-enhanced cross-presentation of MHC class I-restricted epitopes is well investigated, the role in MHC class II-restricted antigen presentation remains less clear. Nevertheless it becomes increasingly obvious that peptides chaperoned by HSP, particularly of the 70 000-MW family, are involved in MHC class II-restricted presentation.[10, 12, 13, 17] As a pre-condition for facilitated MHC class II presentation of Hsp70-chaperoned peptides, Hsp–peptide complexes have to reach the MHC class II loading compartments to be correctly positioned for presentation of chaperoned peptides via MHC class II molecules. As Hsp70–peptide complexes could be detected after receptor-mediated uptake in MHC class-II-enriched compartments, this important requirement is given.[18] As a second mechanism, the antigen-presenting cells’ own cytosolic Hsp molecules may reach the MHC class II presentation pathway in complex with peptides from the cytosol via autophagic processes.[19]

In an earlier study, a direct and specific interaction between human Hsp70 and isolated intact HLA-DR molecules could be demonstrated.[20] This interaction was increased at lower pH, whereas Hsp70-bound peptides were found to be released under these conditions. These findings and the fact that the interaction of Hsp70 and HLA-DR in contrast to normal substrate binding was insensitive to nucleotide addition implicated that other structures than the substrate binding site of Hsp70 contribute to the specific binding of HLA-DR molecules.

Investigating the binding affinity of mutated bacterial Hsp70 homologue DnaK, Mayer et al.[21] proved that the amino acids methionine at position 404 and alanine at position 429 are essential for substrate binding. A mutation of valine at position 436 to phenylalanine caused a steric effect completely disabling enzyme-substrate interaction. Methionine at position 404 was proven to be responsible for enclosing the peptide backbone so influencing the substrate release. Via FASTA sequence comparison of bacterial DnaK and human Hsp70A1A the bacterial methionine 404 and valine 436 can be recognized as alanine 406 and valine 438 in human Hsp70A1A homologue. The C-terminal subdomain is an α-helical lid covering and opening the substrate binding domain in an ATP-dependent manner. In the absence of ATP, open and closed conformation are in equilibrium, whereas presence of ATP causes an open conformation and an increase in substrate binding and release rates with lowered substrate affinity.[7]

The present study investigates the binding characteristics of human purified HLA-DR with recombinantly expressed Hsp70 molecules. The Hsp70 molecules that had either mutations in the substrate binding cavity or truncated variants were evaluated. Considering previous findings,[21, 22] mutations in the substrate binding cavity of Hsp70 for alanine 406 to glycine (mutant A406G) and valine 438 to glycine (mutant V438G) were generated. A truncated Hsp70 molecule lacking the C terminus, hence consisting of the ATPase domain and the substrate binding subdomain (mutant AS-fragment) was produced. The isolated substrate binding domain (mutant SBD) was kindly provided by Lila Gierasch (University of Massachusetts). Binding assays of the recombinantly expressed Hsp70 molecules were performed with peptides (tetanus toxin TT947–966 and haemagglutinin HA307–319) and human HLA-DR molecules. To further investigate the physiological effects of the mutated Hsp70 molecules on the Hsp70-enhanced MHC class II-dependent antigen presentation in vitro, T-cell proliferation assays were performed. This study provides evidence that the HLA-DR binding is independent of the substrate binding cavity and the C-terminal lid of Hsp70 and might be referred to the ATPase domain. Hence, Hsp70-chaperoned peptides might be transferred into the binding groove of HLA-DR in a ternary complex, resulting in facilitated antigen presentation and improved proliferation of responding T cells as demonstrated.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References
  10. Supporting Information

Antibodies and proteins

The monoclonal antibodies to human stress-inducible Hsp70 (ADI-SPA-810) as well as recombinant human stress-inducible Hsp70 (ADI-ESP-555), bovine constitutive Hsc70 (ADI-SPP-751) and ATPase fragment of bovine Hsc70 (ADI-SPP-752) were purchased from Enzo Life Sciences Inc. (Farmingdale, NY). Goat anti-mouse IgG IRDye 800CW (926-32210) was purchased from Li-Cor Biosciences (Bad Homburg, Germany). Affinity-purified HLA-DRB1*0401 and *0402 molecules were isolated from HLA-DR homozygous Epstein–Barr virus-transformed B-lymhoblastoid cell lines. Cell culture, purification and labelling of HLA-DR molecules were performed as previously described.[20, 23]

Construction of mutant plasmids and expression of proteins

A pEQ plasmid containing human hsp 70-gene (Gene ID: 3303), especially designed for expression in Escherichia coli was purchased from Qiagen (Hilden, Germany). Mutagenesis was performed using the Stratagene QuikChange® II XL site-directed mutagenesis kit (Agilent Technologies, Waldbronn, Germany) according to the manufacturer's protocol with the primers 3′-tgggtctggaaacCGGaggtggtgttatgac-5′ and 3′-gtcataacaccacctCCGgtttccagaccca-5′, causing a mutation of amino acid alanine 406 to glycine (A406G) and the primers 3′-ccagatttttaccACCggctccgataatcagccgg-5′ and 3′-ccggctgattatcggagccggTGGtaaaaatctgg-5′ causing a mutation of amino acid valine 438 to glycine (V438G). A C-terminal shortened protein, containing only substrate binding subdomain and ATPase-domain (amino acids 1–541, AS fragment) but lacking the distal C-terminal domain of Hsp70 leading to a constitutively open conformation (described for the bacterial homologue DnaK in [21, 22, 24]) could be designed using primer pair 3′-ataaataCTCGAGcagggcattttttgcgctaacacgttcacgc-5′ and 3′-taatattaGGATCCatgaaacaggcaaaagcagcagcaattgg-5′ containing a BamHI and XhoI restriction site. After amplification in a PCR, DNA was digested, purified using a PCR purification kit (Qiagen) and ligated into a pET-28a+ vector (Merck KGaA, Darmstadt, Germany). All plasmids were sequenced before expression (Agowa, Berlin, Germany). The stable isolated substrate binding domain (amino acids 386–616), which lacks the nucleotide binding domain and the disordered C-terminus of the α-helical lid (compare DnaK truncation constructs described in ref. [25]) was kindly provided by Eugenia Clerico and Lila Gierasch, University of Massachusetts.

All proteins (wild-type Hsp70, A406G, V438G and AS-fragment) were expressed in E. coli BL21 at 22°. The expression was induced with 0·6 mm isopropyl β-D-1-thiogalactopyranoside (Peqlab, Erlangen, Germany). Cell pellets from a 200-ml culture were lysed by using BugBuster (Merck) according to the manufacturer's protocol and purified using Ni-nitrilotriacetic acid columns. Fractions containing the protein were pooled and dialysed using Amicon Ultra-15 centrifugal filter units with ultracell-50 membranes (cut-off < 50 000 MW; Merck). Correct folding of the recombinantly produced proteins was verified in a circular dichroism spectroscopy analysis using a Jasco J-720 spectropolarimeter (Jasco Corporation, Cremella, Italy). Proteins at a concentration of 3 mg/ml in Dulbecco's PBS were diluted fivefold with bi-demineralized water (R = 18·2 MΩ cm) before circular dichroism measurement. Proteins were stored at 4° and used in all experiments for no longer than 7 days. All proteins that were used in T-cell proliferation studies were endotoxin-purified with high-capacity endotoxin removal spin columns (Pierce, Rockford, IL) according to the manufacturer's protocol. Subsequently, the residual amount of endotoxin was measured using the LAL chromogenic endotoxin quantification kit (Pierce). Endotoxin content of protein stock solutions (0·5 mg/ml) was not detectable (< 0·2 EU/mg protein, which equals approximately 0·02 ng endotoxin per mg protein) for all protein preparations used in cell culture.

SDS–PAGE and Western blot

All recombinantly expressed proteins were analysed for size and purity by Coomassie staining in a gel electrophoretic separation. Protein concentration was determined by a BCA Protein Assay (Pierce) according to the manufacturer's protocol. Preparations of 200 ml yielded an average amount of 10–12 mg protein. Protein preparations were mixed with NuPAGE® LDS Sample buffer (Invitrogen, Karlsruhe, Germany). Five micrograms of each purified protein was loaded per lane and size-separated by SDS–PAGE in Tris–HEPES–SDS 4–20% protein gels (Pierce). PageRuler Prestained Protein Ladder (10 000–170 000 MW; Pierce) was used as size standard in protein electrophoresis. Size-separated proteins were blotted to an Immobilon FL low fluorescence PVDF-membrane (0·45 μm; Merck), incubated with anti-Hsp70 antibody (ADI-SPA-810; Enzo) and visualized with goat anti-mouse IRDye 800CW (926-32210; Li-Cor) using an Odyssey Fc dual-mode imaging system (Li-Cor). Coomassie gel staining was performed using Imperial Stain (Pierce) according to the manufacturer's protocol.

Peptides

Tetanus toxin TT947–966 (FNNFTVSFWLRVPKVSASHL), influenza haemagglutinin HA307–319 (PKYVKQNTLKLAT) and class II-associated invariant chain peptide CLIP105–117 (SKMRMATPLLMQA) peptide synthesis was performed on a modified SyRo multiple peptide synthesizer (MultiSynTech, Witten, Germany) as described previously.[20] Biotinylated peptides were specifically labelled at the N-terminus using succinimidyl-6-(biotinamido)hexanoate before deprotection and cleavage. Identity and purity of peptides were verified by reverse-phase HPLC and electrospray ionization-mass spectrometry (> 90%). Molecular weight was determined as 2718·196 g/mol for biotinylated TT947–966, 1871·59 g/mol for biotinylated HA307–319 and 1845·92 g/mol for biotinylated CLIP105–117.

ELISA anti-Hsp70 monoclonal antibody

To detect different binding affinities of anti-Hsp70 monoclonal antibody to either the wild-type Hsp70 or the mutated molecules, an indirect ELISA was performed. Wells of 96-well FluoroNunc Maxisorp microtitre plates (Thermo, Rochester, NY) were pre-coated overnight at 4° with Hsp70 molecules (20 μg/ml in PBS). After blocking with blocking reagent (Roti®-block; Carl Roth GmbH & Co. KG, Karlsruhe, Germany), proteins were incubated with anti-Hsp70 monoclonal antibody (2 μg/ml) for 1 hr at 25°. Wells were extensively washed with PBST (Hanks’ PBS, PAA Laboratories GmbH, Cölbe, Germany, with 0·1% Tween-20; Carl Roth) and treated with 0·2 μg/ml horseradish peroxidase-conjugated anti-mouse-IgG monoclonal antibody (BML-SA-204; Enzo) for another hour at room temperature. Antibody binding was quantified by using TMB substrate (TMB substrate kit; Pierce) according to the manufacturer's protocol. Absorbance was read out at 450 nm in a microplate reader (Milenia Kinetic Analyzer; DPC, Los Angeles, CA).

Heat-shock protein 70–peptide binding assays

Peptide and protein binding were assayed via dissociation-enhanced lanthanide fluorescent immunoassay (DELFIA)[23] using the same conditions as described previously.[20] For complex formation Hsp70 molecules (10 μg/ml) and biotinylated peptide (0·01–200 μg/ml) were incubated overnight at 37° in phosphate buffer pH 7·0. Two milimolar ATP (Sigma-Aldrich, St. Louis, MO) was added to this assay as indicated to investigate the impact of the ATPase domain on substrate binding. The 96-well FluoroNunc Maxisorp® (Nunc, Wiesbaden, Germany) plates were coated overnight at 4° with 5 μg/ml anti-Hsp70 monoclonal antibody (mAb; ADI-SPA-810, clone C92F3A-5). Subsequently plates were blocked with blocking reagent Roti®-block (Carl Roth) for 1 hr. Plates were washed with DELFIA washing buffer (Perkin Elmer, Boston, MA), the Hsp–peptide mixture was transferred to the microtitre plate and incubated for 1 hr at room temperature. After another washing step, the biotinylated peptides were incubated with europium-labelled streptavidin in DELFIA buffer (Perkin Elmer) for 1 hr. After adding DELFIA enhancement solution fluorescence was detected by time-resolved fluorescence measurement at 615 nm with a Viktor 1420 multilabel counter (Perkin Elmer).

Hsp70:HLA-DR protein binding assays

To determine the affinity of Hsp70 to HLA-DR, wild-type Hsp70 and its mutated analogues were pre-coated overnight at 4° to microtitre plates (20 μg/ml in PBS). After blocking with blocking reagent Roti®-block, the proteins were incubated for 1 hr with biotinylated HLA-DRB1*0402 molecules at a concentration of 6 μg/ml PBS in a total volume of 50 μl/well. The bound HLA-DR molecules were detected using DELFIA as described above. To determine the interaction of the ATPase fragment with HLA-DR molecules, microtitre plates were pre-coated with 5 μg/ml Hsp70, Hsc70, ATPase fragment of Hsc70 (Hsc70–ATPf) or human serum albumin. Blocking was performed with bovine serum albumin. Binding of 10 nm biotin-labelled purified HLA-DRB1*0401 was analysed in 150 mm phosphate buffer pH 5 (0·05% Tween-20) and the resulting fluorescence was measured using DELFIA as described above.

T-cell purification and culture

A tetanus-pre-immunized donor with the HLA-DR haplotype HLA-DRB1*11 was chosen for T-cell proliferation experiments. HLA-DR typing was performed by oligonucleotide typing (Institute for Clinical and Experimental Transfusion Medicine, Tübingen, Germany). CD4+ T-cell purification and irradiation treatment of antigen-presenting cells was performed as described previously.[12] All cells were cultured in very low endotoxin RPMI-1640 medium (Biochrom AG, Berlin, Germany) supplemented with 10% heat inactivated pooled human serum, 2 mm l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin and 10 mm HEPES buffer (all Biochrom AG).

Staining and proliferation

Carboxy fluoroscein succinimidyl ester (CFSE) staining of CD4+ T cells was performed using CellTrace Cell Proliferation Kit (Invitrogen Ltd., Paisley, UK). Cells were incubated with 10 μm CFSE at room temperature for 6 min. Staining was quenched by adding a fivefold amount of cold pure fetal calf serum (Biochrom AG) and washing in culture medium. Irradiated antigen-presenting cells were incubated with either protein–peptide complexes (0·1 μg/ml Hsp70 molecules/0·01 μg/ml TT947–966), or protein or peptide alone, respectively. Then, 2 × 105 CFSE stained CD4+ T cells were co-cultured with 5 × 104 irradiated and pre-incubated antigen-presenting cells at 37° for 7 days. CD4+ T cells were labelled with anti-CD4-Peridinin chlorophyll protein fluorescent dye (BD Biosciences, Franklin Lakes, NJ) for 10 min at 4°. Washed cells were analysed on a FACSCalibur flow cytometer (BD Biosciences).

Statistical analysis

All resulting data were subjected to analyses using graphpad prism software version 5.04 for Windows (GraphPad Software, San Diego, CA). To determine significant differences (*P < 0·05, **P < 0·01, ***P < 0·001) between all test groups and control groups, analysis of variance and Bonferroni's post test were carried out. Outliers were detected and eliminated using Grubb's outlier test.

Gene bank accession number: 3303 (M11717) HSP A1A heat-shock 70 000 MW protein 1A (Homo sapiens)

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References
  10. Supporting Information

Mutation of the substrate binding cavity of Hsp70

As binding of Hsp70 to HLA-DR independent of the substrate binding is expected, a mutation in the substrate binding cavity supposedly causing a reduction of substrate binding was performed on His-tagged Hsp70A1A-DNA. Substrate binding amino acids in the binding cavity, alanine 406 and valine 438,[22] were replaced by glycine. The resulting mutated proteins were called A406G and V438G, respectively. To narrow down the possible binding site of HLA-DR to Hsp70, two different truncated Hsp70 variants were used: a C-terminal shortened protein AS consisting of amino acids 1–541 and the isolated SBD consisting of amino acids 386–616. The AS-fragment lacks the C-terminal lid but expresses the ATP binding domain as well as the substrate binding subdomain. The lack of the C-terminal lid results in an ATP-independent all-open conformation of the protein. The SBD lacks the nucleotide binding domain, and so remains unaffected by the addition of ATP. Circular dichroism spectra displayed in Fig. S1 (see Supporting information) show correct folding of the recombinant Hsp70 molecules.

To ensure high purity and correct expression of both wild-type Hsp70 and mutated Hsp70 proteins, pooled and washed elution fractions were separated on an SDS–PAGE and stained with Coomassie. Figure 1 demonstrates protein bands at a size of approximately 70 000 MW for wild-type Hsp70A1A and mutated A406G and V438G. A C-shortened AS-fragment with a calculated size of 55 000 MW shows a protein band at a size of about 60 000. The truncated SBD with a calculated size of 26 000 MW shows a protein band at a size of about 30 000.

image

Figure 1. Coomassie staining and Western blot analysis of heat-shock protein 70 (Hsp70) purification. (a) SDS–PAGE analysis of purified, recombinantly expressed Hsp70 molecules wild-type Hsp70 A1A, A406G and V438G with mutated substrate binding site as well as truncated C-fragment shortened AS and A-fragment shortened SBD. Coomassie staining was performed after gel electrophoresis. A1A, A406G and V438G show protein bands at approximately 70 000–72 000, AS is detectable at approximately 60 000 and SBD at approximately 30 000. (b) Western blot analysis of A1A, A406G, V438G, AS and SBD. One of three independent experiments with similar outcome is represented.

Download figure to PowerPoint

Anti-Hsp70 mAb binding to mutated Hsp70 molecules is not affected

Before Hsp70–peptide fluorescence assays, it was investigated whether the mutation of the substrate binding site of Hsp70 affects the binding affinity of mAb anti-Hsp70 to Hsp70 molecules. Figure 2 displays the results of an indirect ELISA with anti-Hsp70 mAb (ADI-SPA810) and wild-type Hsp70A1A, mutated A406G and V438G as well as AS-fragment and the isolated SBD. All the Hsp70 molecules tested showed no significantly different (P > 0·05) binding affinity to anti-Hsp70 mAb.

image

Figure 2. Binding affinity of anti-heat-shock protein 70 (Hsp70) monoclonal antibody (mAb) to Hsp70 molecules. Binding affinity of Hsp70 molecules to anti-Hsp70 mAb was detected in an indirect ELISA. Microtitre plates were pre-coated with Hsp70 molecules A1A, A406G, V438G, AS and SBD (20 μg/ml) and subsequently incubated with anti-Hsp70 mAb (2 μg/ml). Bound anti-Hsp70 mAb was detected with horseradish peroxidase-conjugated secondary antibody. Binding affinity of the mutated Hsp70 molecules A406G, V438G, AS and SBD to anti-Hsp70-antibody was not significantly different (P > 0·05) from wild-type Hsp70A1A. Sample size n = 4. Data show mean + standard deviation (SD). a, significantly different to blank (P < 0·001) but not significantly different (P > 0·05) to each other; anova, analysis of variance.

Download figure to PowerPoint

Substrate binding of mutated A406G and V438G is reduced

Binding capacity to HA307–319 and TT947–966 was tested for all Hsp70 molecules. For that purpose, anti-Hsp70 mAb was pre-coated to microtitre plates. Pre-incubated Hsp70–peptide complexes or Hsp70 molecules alone were added and incubated for 1 hr. Binding of biotin-labelled peptides was evaluated using DELFIA. As displayed in Fig. 3, substrate binding of 20 μg/ml biotinylated TT947–966 was significantly reduced for mutated A406G and V438G (P < 0·001) but not for AS-fragment and SBD (P > 0·05) compared with substrate binding of wild-type Hsp70A1A. Addition of 2 mm ATP during pre-incubation of Hsp70 molecules and peptide could significantly reduce the substrate binding of Hsp70 molecules A1A, A406G, V438G and AS (P < 0·001). Substrate binding of SBD to TT947–966 was not altered by the addition of 2 mm ATP. The binding assay with biotinylated HA307–319 showed reduced binding capacity for A406G, V438G and SBD compared with wild-type A1A. Binding capacity of the AS-fragment to HA307–319 was not reduced but even increased. Additionally Hsp70–TT947–966 binding affinity was analysed in saturation experiments. Figure 4 shows titration curves for A1A, A406G, V438G, AS and SBD with increasing amounts of TT947–966. Titration curves for all mutant Hsp70 molecules, except the AS-fragment were significantly different (P < 0·05) compared with the titration curve of TT947–966 peptide binding to wild-type Hsp70A1A. Figures S2 and S3 (see Supporting information) show titration curves of peptide binding of wild-type Hsp70A1A and the isolated SBD to biotinylated HA307–319 (Fig. S2) or biotinylated CLIP105–117 (Fig. S3). Peptide binding of the SBD is significantly reduced (P < 0·0001) compared with wild-type Hsp70A1A.

image

Figure 3. Peptide binding is reduced for A406G and V438G. Overnight pre-coating with 5 μg/ml anti-heat-shock protein 70 (Hsp70) antibody was followed by either incubation with Hsp70 molecules (10 μg/ml) alone or pre-incubated Hsp70–peptide (10 : 20 μg/ml) complexes. The Hsp binding of biotinylated peptides TT947–966 (a, b) or HA307–319 (c, d) was detected via DELFIA. Additionally, Hsp70–peptide binding was measured in the presence or absence of 2 mm ATP as indicated. Experiments were carried out in triplicate and represent one of four independent experiments with similar outcome. Data show mean + SEM of triplicates. ***P < 0·001, a: P < 0·001 compared with respective incubations without ATP; anova, analysis of variance.

Download figure to PowerPoint

image

Figure 4. Titration curves for heat-shock protein (Hsp70)–TT947–966. Overnight pre-coating with 5 μg/ml anti-Hsp70 antibody was followed by incubation with pre-incubated Hsp70–TT947–966 complexes. A1A, A406G, V438G, AS and SBD (10 μg/ml) were incubated with increasing amounts of biotinylated TT947–966 (0·0140 μg/ml ≡ 0·03109 μm). Bound TT947–966 was detected via DELFIA. Sample size n = 3. Experiments were carried out in triplicate. Data show non-linear regression. *P < 0·01; **P < 0·001; n.s., not significant.

Download figure to PowerPoint

HLA-DR binding is not affected by mutations of the substrate binding site but by the removal of the nucleotide binding domain

To investigate the impact of the mutation of the substrate binding site, the removal of the C-terminal lid or the nucleotide binding domain on the binding affinity of Hsp70 to HLA-DR molecules, an indirect ELISA was performed. Therefore, Hsp70 molecules were pre-coated to microtitre plates and subsequently incubated with biotinylated HLA-DRB1*0402. HLA-DR binding was detected using DELFIA. Figure 5 demonstrates equal HLA-DR binding characteristics for wild-type Hsp70 and mutated A406G, V438G and AS-fragment (P > 0·05). Binding affinity of the isolated SBD is significantly different to all other Hsp70 molecules (P < 0·05) and not different to the blank (P > 0·05). These experiments indicate an interaction of HLA-DR and Hsp70 independent of conformational changes in the substrate binding cavity but dependent of the nucleotide binding domain.

image

Figure 5. Binding of heat-shock protein 70 (Hsp70) A1A, A406G, V438G and AS to HLA-DRB1*0402 is not reduced. Microtitre plates were pre-coated with Hsp70 molecules (20 μg/ml). Biotinylated HLA-DRB1*0402 (6 μg/ml) was added and incubated for 1 hr. Bound HLA-DRB1*0402 was detected by DELFIA. Binding affinity of the mutated Hsp70 molecules A406G, V438G and AS to HLA-DRB1*0402 was not significantly different (P > 0·05) from wild-type Hsp70A1A. Binding affinity of SBD to HLA-DRB1*0402 was similar to blank (P > 0·05). Sample size n = 4. Data show mean + SD. a, significantly different (P < 0·001) compared with blank; b, not significantly different (P > 0·05) compared with blank; n.s., not significant (P > 0·05) to each other; anova, analysis of variance.

Download figure to PowerPoint

Hsp70 binding to HLA-DR molecules is mediated by the ATPase domain of Hsp70

So far, the HLA-DR–Hsp70 interaction could be attributed outside the substrate binding cavity of Hsp70 and independent of the C-terminal lid as the HLA-DR binding was not influenced by a lack in substrate binding or a split-off of the C-terminal lid. Further, it could be demonstrated that the removal of the nucleotide binding domain leads to a complete loss of the binding affinity of Hsp70 to HLA-DR. With DELFIA a specific interaction between human Hsp70 molecules and different allelic variants of purified and recombinant HLA-DR molecules could be demonstrated.[20] This specific interaction was shown for both the constitutively expressed (Hsc70) and the stress-inducible (Hsp70) form of cytosolic Hsp70 molecules. The interaction was saturable and could be competed by unlabelled molecules. KD values for the Hsp70–HLA-DR interaction were found to be in the low nm range and the interaction was not sensitive to the addition of ATP. Together with the previous findings, it can be assumed that regions other than the peptide binding pocket of the Hsp70 molecules are involved in the Hsp70–HLA-DR interaction. To further clarify which domain of the Hsp70 molecule mediates the HLA-DR binding, binding experiments using both the N-terminal ATPase fragment of Hsc70 protein and complete Hsc70 were performed. Figure 6 shows the interaction of purified HLA-DRB1*0401 with Hsp70, Hsc70 and the ATPase-subunit of Hsc70 (Hsc70-ATPf). Similar results were obtained with other allelic variants of purified HLA-DR molecules (HLA-DRB1*0801 and HLA-DRB1*0402, data not shown). In accordance with our previous findings, a slightly higher HLA-DR affinity was found for Hsp70 compared with Hsc70. Interestingly, a significant difference (P < 0·05) in signal strength between complete Hsc70 and the truncated Hsc70 ATPase fragment lacking the peptide binding domain with any of the allelic HLA-DR variants could not be detected. These findings demonstrate that HLA-DR binding of Hsc70 is attributed to the ATPase fragment. As bovine Hsc70 and human Hsp70 are approximately 98% homologues, and because the removal of the ATPase domain leads to a complete loss of the HLA-DR binding affinity it can be assumed that the interaction of Hsp70 with HLA-DR is mediated by the ATPase domain.

image

Figure 6. ATPase domain of mammalian cytosolic Hsp70 (Hsc70) takes part in the HLA-DR–Hsc70 interaction. Microtitre plates were pre-coated with 5 μg/ml Hsp70, Hsc70, ATPase fragment of Hsc70 (Hsc70-ATPf) or human serum albumin (HSA). Binding of 10 nm biotin-labelled purified HLA-DRB1*0401 was analysed in 150 mm phosphate buffer pH 5 (0·05% Tween-20) as described in the Materials and methods section. Experiments were carried out in triplicate and represent one of at least three independent experiments with similar outcome. Data show mean + SEM of triplicates. n.s., not significant; a, significantly different from both HSA and non-specific binding, P < 0·01; anova, analysis of variance.

Download figure to PowerPoint

Mutation of the substrate binding site influences the Hsp70–peptide-mediated MHC class II-dependent antigen presentation

As shown previously,[12] formation of the complex of peptide with Hsp70 enhances the proliferation of human CD4+ T cells compared with antigenic peptide alone. As demonstrated in this study, mutations of recombinantly expressed Hsp70 in the substrate binding site reduced peptide binding, but did not affect the interaction with HLA-DR molecules. The influence of these mutations on the Hsp70-facilitated antigen presentation of the antigenic tetanus peptide TT947–966 was therefore further investigated in a human antigen-specific in vitro CD4+ T-cell stimulation assay. As the recombinantly expressed Hsp70 molecules are of bacterial origin, the protein preparations were purified as described in the Materials and Methods section and cell cultures were incubated in very low endotoxin medium. CD4+ T cells from a healthy HLA-DR*11 donor pre-immunized with tetanus vaccine were stimulated with TT947–966 or Hsp70 molecules alone or Hsp70–TT947–966 complexes, respectively. HLA-DR*11/TT947–966-specific CD4+ T cells become activated and enter the cell cycle if they recognize TT947–966 peptide presented via HLA-DR*11 molecules on antigen-presenting cells. Proliferation of these activated CD4+ T cells was analysed by dilution of the fluorescent dye CFSE in CD4+ T cells (Fig. 7). Stimulation with wild-type and mutant Hsp70 molecules alone resulted in a low number of proliferated CD4+ T cells (Fig. 7b, usually < 3%). T-cell stimulation with the very low dose of 0·01 μg/ml antigenic TT947–966 peptide alone also yielded low numbers of proliferated T cells (Fig. 7c). However, proliferation of CD4+ T cells could be significantly enhanced by stimulation with TT947–966 complexed to wild-type Hsp70 (Fig. 7a; and as expected from our previous findings shown in ref. [12]) The modified Hsp70 variants A406G, V438G and AS fragment showed significantly lower enhancement of antigen presentation and subsequent CD4+ T-cell proliferation compared with wild-type Hsp70 protein (Fig. 7a). Data from several experiments are summarized in Fig. 7(d).

image

Figure 7. T-cell proliferation with heat-shock protein 70 (Hsp70) molecules and TT947–966. Purified CD4+ T cells were incubated for 7 days with irradiated antigen-presenting cells (APC). Before co-incubation, the APC were incubated with (a) Hsp70–TT947–966 complexes (0·1 : 0·01 μg/ml), (b) Hsp70 molecules (0·1 μg/ml) or (c) TT947–966 (0·01 μg/ml). CD4+ T cells were stained with anti-CD4 antibody and CFSE dilution in CD4+ T cells was measured in a flow cytometric analysis. (d) Proliferation of CFSElow CD4+high cells after treatment with Hsp70–peptide complexes was compared with control groups TT947–966 and Hsp70 molecules alone. Sample size n = 4. Data show mean + SD. ***P < 0·001; **P < 0·01; a, significantly different (P < 0·001) compared with peptide alone; b, significantly different (P < 0·001) compared with respective Hsp70 molecule without peptide; anova, analysis of variance.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References
  10. Supporting Information

In this study Hsp70 interactions were investigated by point mutations in the substrate binding site allowing insights into the Hsp70–peptide and the Hsp70–HLA-DR interplay. In earlier studies an interaction of Hsp70 and HLA-DR independent of peptide binding could be demonstrated.[20]

Members of the Hsp70 family consist of an N-terminal ATPase unit of 45 000 MW and a C-terminal substrate binding domain of 25 000 MW, which is further subdivided into a β-sandwich subunit of 15 000 and a C-terminal α-helical subdomain. The ATPase subunit and the substrate binding domain are linked by a short linker sequence (hinge).[22, 26] The Hsp70 molecules undergo conformational changes during the binding and release of substrate.[7, 8, 21, 22, 26-31] Substrate exchange at the substrate binding cleft on the β-sandwich subunit is thereby facilitated or handicapped by the α-helical subdomain of the substrate binding domain. This subdomain forms a lid that covers the peptide binding site and opens and closes in an ATP-dependent manner.

The hypothesis that the interaction of HLA-DR and Hsp70 is independent of the Hsp70 substrate binding site of the substrate binding domain could be supported by the findings that the point mutations alanine 406 to glycine and valine 438 to glycine resulted in a significant reduction of the substrate binding affinity whereas HLA-DR binding remained unaffected. In contrast to the results of Mayer et al.[21] with altered bacterial Hsp70 homologue DnaK, a complete lack of peptide binding could not be observed for mutant Hsp70 V438G, which might be due to the fact that a steric effect cannot be achieved by replacement of valine to the much smaller glycine. For the C-terminal shortened AS-fragment, the substrate binding domain remains unaffected but because it lacks a covering lid it presents itself in an all-open conformation in contrast to wild-type Hsp70 with a randomly distributed (1 : 1) open or closed covering lid. A similar or even slightly higher substrate binding affinity could be observed for AS compared with wild-type Hsp70. This might be a consequence of the all-open conformation allowing a statistically facilitated enzyme–substrate interaction. Peptide binding assays of the isolated SBD revealed binding of the TT947–966 peptide although with lower affinity compared with the wild type Hsp70, as demonstrated in a titration assay. Interestingly, binding of HA307–319 to the substrate binding domain could not be observed (compare with Fig. S2). As the binding characteristics of HA307–319 to the AS-fragment are similar to those of the wild-type Hsp70, it can be hypothesized that the nucleotide binding domain of Hsp70 has an effect on the substrate binding of HA307–319. However, substrate binding of the SBD to CLIP105–117 could be achieved although with lower substrate affinity (Fig. S3), pointing towards a binding specificity of the HA307–319 peptide. Substrate affinity of human Hsp70 to peptide fragments both TT947–966 and HA307–319 has already been approved in previously performed assays.[12] The ATPase domain of Hsp70 proteins is involved in the binding and release of substrates. Addition of ATP not only leads to an all-open conformation of the covering lid of the substrate binding domain but also increases substrate binding and release rates due to diverse conformational changes.[7, 26] It could be demonstrated that ATP leads to a distinct reduction of peptide binding. Interestingly, the C-terminal shortened AS-fragment, which lacks the covering lid, also revealed a reduced binding of both HA307–319 and TT947–966 in the presence of ATP. Therefore we conclude that the C-terminal lid does not primarily regulate the substrate access to the substrate binding cavity but probably exerts influence on substrate release. As would be expected, addition of ATP to the substrate binding domain did not have any influence on the SBD–peptide interaction, which is a result of the removal of the nucleotide binding domain.

Substrate binding of Hsp70 does not influence the binding of HLA-DR.[20] Hence we hypothesized that the interaction of HLA-DR and Hsp70 takes place outside the peptide binding cavity. The findings demonstrated in the present study affirm this hypothesis. Point mutations of the substrate binding cavity A406G and V438G did not influence HLA-DR binding, showing that amino acids 406 and 438 are not involved in HLA-DR binding. It could be demonstrated that the removal of the C-terminal lid did not influence the HLA-DR interaction either. In contrast, removal of the ATPase domain led to a complete loss of the HLA-DR binding properties. These findings indicate that HLA-DR binding is neither located within the substrate binding domain nor the C-terminal lid but can be attributed to the ATPase domain.

A truncated recombinant ATPase subunit of bovine constitutively expressed Hsc70, which is a 99% homologue to the human Hsp70 ATPase fragment, was used to further confine the binding area of HLA-DR. The ATPase-fragment is 388 amino acids long. Only six amino acids are changed within human and bovine Hsp70 ATPase fragment. According to the manufacturer, this fragment maintains its ATPase activity although lacking the entire substrate binding domain. In binding assays, comparable binding affinities of HLA-DR molecules to this ATPase fragment and the complete Hsc70 molecule could be detected. Unfortunately, isolated human Hsp70 ATPase fragment is neither commercially available nor was it possible for us to express it in a correct folding manner. The results of bovine ATPase fragment endorse that the HLA-DR binding site can be ascribed to structures on the N-terminal ATPase domain of the Hsp70 molecules.

In T-cell proliferation assays Hsp70-enhanced stimulation of CD4+ T cells with the tetanus toxin fragment TT947–966 could be confirmed using wild-type Hsp70–TT947–966 complexes. In contrast, stimulation with peptide alone, Hsp70 molecules alone or pre-incubations of TT947–966 with mutated Hsp70 molecules induced similarly low proliferation of CD4+ T cells. Considering the amount of bound peptide as a pivotal factor we would have expected that the stimulation with AS–TT947–966 would result in a similar proliferation effect as complexed wild-type Hsp70A1A–TT947–966. Unexpectedly, the impact of AS–TT947–966 on T-cell proliferation was not significantly different from the control groups or the incubations with peptide complexed to A406G and V438G. This effect might be due to the physiological dependencies on Hsp70 uptake into antigen-presenting cells, implying that C-fragment-shortened AS interferes differently with Hsp70-specific receptors such as CD91 and diverse scavenger receptors.[2, 9, 11]

Taken together, the results of this study affirm HLA-DR binding to Hsp70 outside the substrate binding groove and the C-terminal covering lid but dependent on the ATPase domain. The mechanism proposed from our data for the Hsp70–HLA-DR interaction might allow that Hsp70-chaperoned peptides are transferred in a ternary complex directly into the binding groove of HLA-DR, resulting in enhanced HLA-DR presentation of the Hsp70-bound peptide.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References
  10. Supporting Information

The experiments shown were performed by Karin Rohrer and Markus Haug. Daniela Schwörer executed the mutations and expression of the Hsp70 proteins. Hubert Kalbacher synthesized the tetanus and influenza peptides and performed the biotinylation of the used peptides and HLA-DR proteins. The study was designed by Ursula Holzer, Markus Haug, Daniela Schwörer and Karin Rohrer and was supervised by Ursula Holzer. The isolated substrate binding subdomain was constructed by Julia Leu, University of Pennsylvania and kindly provided by Eugenia Clerico and Lila Gierasch, University of Massachusetts. Circular dichroism spectroscopy was performed with kind support from Luisa Ströh, Interfaculty Institute of Biochemistry, Tuebingen. This work was supported by the Deutsche Forschungsgemeinschaft (DFG HO 2340/3-1) and SFB 685. The authors would like to thank Lila Gierasch and Eugenia Clerico for critical reading of the manuscript and Barbara Schmid-Horch for HLA-DR typing and providing buffy coats and leukaphereses of blood donors.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
imm12249-sup-0001-FigS1.tifimage/tif7869KFigure S1. Circular dichroism spectra of heat-shock protein 70 (Hsp70) molecules.
imm12249-sup-0002-FigS2.tifimage/tif4295KFigure S2. Titration curves for heat-shock protein 70 (Hsp70)–haemagglutinin (HA)307–319.
imm12249-sup-0003-FigS3.tifimage/tif4517KFigure S3. Titration curves for heat-shock protein 70 (Hsp70)– class II-associated invariant chain peptide (CLIP)105–117.
imm12249-sup-0004-FigS1-S3legends.docxWord document12K 

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.