|
SUPPLEMENTARY INFORMATION TO MANUSCRIPT: The Disulfide Bonds in Glycoprotein E2 of Hepatitis C Virus Reveal the Tertiary Organization of the Molecule
Abstract
the divalent metal ions used for induction with the histidine tag used for purification of the secreted protein, we replaced the region including the V5 epitope and the 6-Histidine tag by a segment coding for a specific proteolytic cleavage site, followed by a tandem strep-tag (IBA, www.iba-go.com) with a linker region (GlyGlySer) 4 in between. The proteolytic cleavage site was added to allow the specific removal of the tag for structural studies. We avoided the use of cysteine proteases (like Prescission, TEV or 3C proteases) which, although are very specific, require a reducing agent for activity, which could also reduce some of the exposed disulfides of the glycoprotein. We instead engineered an enterokinase (EK) cleavage site, which is a serine-protease relatively specific for the sequence (Asp) 4 Lys↓X, cleaving at the site indicated by the arrow with a cleavage efficiency between 60 and 80 % (X being any amino acid) [4] . This resulted in the following amino acid sequence downstream of the ApaI and BstBI sites ...DDDDKAGWSHPQFEKGGGSGGGSGGGSWSHPQFEK-COOH. All synthetic HCV glycoprotein genes were purchased from GeneCust (Dudelange, Luxemburg) and amplified by PCR using strain specific 5'-oligonucleotides containing Bgl II, which allows insertion immediately downstream of the BiP secretion signal, and strain specific 3'-oligonucleotides containing Apa I.
A full list of oligonucleotides used in this study is available upon request
Transfection
For large scale production of E2e the cells were cultured in spinner flasks or in Wave Bioreactors (2/10, Wave Biotech, Somerset, USA) and induced with 4µM CdCl 2 at a density of approximately 7x10 6 cells per ml. After 8 days at 28°C cells were pelleted and E2e was purified by affinity chromatography from the supernatant using a StrepTactin Superflow column (IBA, Goettingen, Germany) followed by gel filtration chromatography using a Superdex200 column (GE Healthcare, Uppsala, Sweden). Pure protein was quantified using adsorption at UV 280nm and concentrated to approximately 1 mg/ml.
25µg of H77 E2e and 70µg of either mAb H53 or CBH-4D, respectively, were incubated as isolated proteins as well as in complex for 1h at 10°C followed by analysis on a Superdex200 Mini column (column volume 3 ml, Amersham).
25µg of E2e was bound to a StrepTactin Superflow mini column (column volume 0.2ml) and washed with 10 column volumes of washing buffer. Subsequently, 10µg
of CD81 large extracellular loop (produced as described before [5] ) or 50µg of conformation dependent antibodies CBH-4B, CBH-4D against HCV E2 (kindly provided by S. Foung, Stanford, USA) or a control antibody were added, followed by washing with 10 column volumes. Complexes were eluted in 4.5 column volumes elution buffer and concentrated 20-fold by ultrafiltration. This concentrate was analysed by SDS-PAGE and Coomassie Blue staining.
Huh7.5 cells plated on glass coverslips in 24-well plates (4. Secondary structure contents were estimated from the far-UV CD spectra using the CDSSTR routine [6] of the DICHROWEB server [7, 8] run on the SP175 reference dataset [9] , containing 72 proteins representing a large panel of secondary structures. Similar results were obtained on different datasets [10] or by using the CONTIN/LL routine [11] .
Soluble HCV E2 and Dengue virus 3 E protein in a concentration of 5-10mg/ml in After 8h, 0.25µg of trypsin were added and the digestion was continued for another 16h. The peptides were eluted from the gel using 200µl water and two times 100µl 60% Acetonitril.
As a control, the trypsin digestion experiment was carried out as described above, but in the presence of 5% DMSO, acting as oxidizing agent. The deglycosylated 8 protein was analysed by SDS-PAGE followed by transfer onto a Nitrocellulose membrane and staining with Poinceau-Red. Bands containing approximately 15µg
E2e were cut out of the membrane and saturated with 1ml of 0.2% PVP K30 for 15 minutes followed by six washes, four with water and two with 50mM TrisHCl pH 7.6.
Subsequently the protein was digested with 1µg Trypsin for 3h at 37°C in the absence or presence of 5mM NEM. The peptides were eluted from the membrane using 200µl of water.
The tryptic digest was divided in two half fractions. One half was submitted directly to reverse-phase HPLC using DEAE-C18 columns (1mm diameter) and a gradient
N-terminal sequencing was performed using a ABI 494 Protein Sequencer (Applied
ChipReader System 4000 using a H4 (reversed phase, Ciphergen Biosystems, Fremont, CA, USA) surface and a SPA matrix, which was prepared according to the manufacturer's instructions. Peak identification was carried out using ProteinChip Software 3.1 (Ciphergen). Molecular weight prediction of disulfide-connected peptides was performed using MS-BRIDGE [12] , while molecular weight of reduced peptides was predicted using PeptideMass [13] .
In order to identify the disulfide bridges of HCV E2e we first performed a tryptic digestion of E2e of the JFH-1 isolate. The HPLC chromatogram of the resulting digest revealed peaks 6-3, 12-3 and 16-3 to be TCEP sensitive and disappear upon reduction.
Peak 6-3 revealed a mixture of peptides, the N-terminal sequencing of which showed that only J1 and J2 (Table S1 and Fig. S3A ) contained a cysteine residue (position 452 and 459, respectively). In the respective mass spectrum a peak corresponding to the disulfide linked dipeptide could be identified (1471.71 Da), which disappeared upon reduction (Fig. S3A) , indicating a disulfide bridge between Cys452 and Cys459.
Peak 12-3 contained peptides J6 and J7, each of them with one cysteine (position 607 and 644, respectively). While peptides J6 and J7 were found as single peptides in the mass spectrum, indicating partial reduction, we also observed a peak at the predicted molecular weight of the two peptides linked by a disulfide bond (Fig. S3A, 2045 .37 Da). This peak disappeared, as expected, upon reduction. This clearly suggested a disulfide bridge between Cys607 and Cys644.
Peak 16-3 contained a mixture of peptides with one dominant sequence corresponding to peptide J4, containing two cysteines (position 503 and 508, respectively) and a proline residue in between. This peptide was unambiguously identified in the mass spectrum of peak 16-3 (Fig. 5B, Fig. S3B and Table S1 , 2341.37 Da). Reduction with TCEP resulted in a molecular weight shift by 2 Da, which was interpreted as two hydrogen atoms added upon reduction of the cysteines, demonstrating an intrapeptidic disulfide bridge between Cys503 and Cys508. Two more peptides, which could not be observed by mass spectrometry, were identified by N-terminal sequencing in peak 16-3: peptide J3 and peptide J5, containing Cys486 and Cys494 as well as Cys581 and Cys585, respectively.
Subsequently we subjected the E2e from isolate UKN2b_2.8 to trypsin digestion.
HPLC separation of the resulting peptides revealed that peaks 13-1, 20-1, 29-3, 42-observed by N-terminal sequencing in all three control experiments (data not shown), strongly suggesting that it is also present in the native protein.
Peak 20-1 contained exclusively peptide U3, which corresponds to peptide J4 in JFH-1 E2e, thereby confirming the presence of a disulfide bridge between Cys503
and Cys508 (Table S1 and Fig. S3C, 2194 .94 Da).
Analysis of peak 29-3 revealed two TCEP sensitive peptides, U2 and U3. We had identified U3 previously to carry an internal disulfide bridge, thus suggesting an additional internal disulfide bond between Cys486 and Cys494 in peptide U2. Cys552. Although the disulfide linked peptides could not be identified by mass spectrometry, upon reduction a peak corresponding to the reduced peptide U1 was observed (Fig. S3D , 2308 Da). Likely the high molecular weight of the disulfide linked dipeptide (U1 + U4 -6890.68 Da) prevented its appearance in the spectrum.
One peak (19-1) was found to contain a mixture of sequences, with one dominant sequence corresponding to peptide U5, in which two cysteines (position 581 and 585) are present. We observed a peak corresponding to the peptide harboring an intrapeptidic disulfide bridge in the mass spectrum (Table S1 and Fig. S3D, 1849.64 Da). Reduction resulted as expected in an increase of the molecular weight by 2 Da.
In addition, peptide U2 was found in the same peak, which has previously been shown to carry an intrapeptidic disulfide bond.
Finally, we performed a tryptic digestion of the of E2e of H77 followed by HPLC of the resulting peptides, which revealed that peaks 15-2, 6-2, 26-2, 32-2, 43-2 and 33-2 disappeared upon reduction.
Peptides H5 and H6, which correspond to J6/ J7 and U6/ U7 were identified in peak 15-2. For both E2 of JFH-1 and UKN2b_2.8 a disulfide bridge between the respective cysteines (position 607 and 644) was shown in this study. Mass spectrometry clearly demonstrated the presence of a disulfide bridge between Cys607 and Cys644 in the ectodomain of H77 E2 as well (Table S1 and Fig. S3E ,
Peptide H1 was found in two different peaks. Together with peptide H2, which corresponds to peptide J2, it was observed in peak 6-2, suggesting the presence of a disulfide bridge between Cys452 and Cys459, which has already been identified in E2e from strain JFH-1. A peak in the mass spectrum corresponding to this peptide confirmed the presence of this disulfide bond (Fig. S3E, 1544 .29 Da).
However, peptide H1 was also found together with peptide H6 in peak 26-2, which clearly suggested a disulfide rearrangement for these cysteines (Fig. S3F) Peak 43-2 consisted of the peptides H7 and H8, each containing one cysteine residue (position 652 and 677, respectively). In the mass spectrum we observed a peak corresponding to the disulfide linked dipeptide (Fig. S3G and Table S1 , 6849.91 Da), unambiguously identifying a disulfide bridge between Cys652 and Cys677 in the ectodomain of H77 E2.
Comparing the sequence alignment of E2 in the region between Cys569 and Cys581 we noticed that while UKN2b_2.8 and JFH-1 E2 contain three trypsin cleavage sites, H77 E2 has no cleavage sites in this region (Fig. S2 ). Thus trypsin cleavage prediction in this region resulted in one peptide containing 4 cysteines, aligned sequentially in a way that the first two cysteines and the last two each have a proline residue in between. Analysis of peak 33-2 revealed only peptide H4, which corresponds to the predicted peptide containing 4 cysteines (positions 564, 569, 581 and 585, respectively). Mass spectrometry revealed a peak matching the predicted mass of this peptide containing two intrapeptidic disulfide bridges (Fig. S3G , 2504.50 Da). Under non-reducing conditions two minor peaks could be observed, which are shifted by exactly 2 Da and thus likely correspond to partially reduced peptides in the original HPLC peak. Since we had already identified the disulfide bond between Cys581 and Cys585 in UKN2b_2.8 E2, this result strongly indicates the presence of a disulfide bridge between Cys564 and Cys569.
|
