Dissection of SARS Coronavirus Spike Protein into Discrete Folded Fragments.

The spike protein of the severe acute respiratory syndrome coronavirus (SARS-CoV) mediates cell fusion by binding to target cell surface receptors. This paper reports a simple method for dissecting the viral protein and for searching for foldable fragments in a random but systematic manner. The method involves digestion by DNase I to generate a pool of short DNA segments, followed by an additional step of reassembly of these segments to produce a library of DNA fragments with random ends but controllable lengths. To rapidly screen for discrete folded polypeptide fragments, the reassembled gene fragments were further cloned into a vector as N-terminal fusions to a folding reporter gene which was a variant of green fluorescent protein. Two foldable fragments were identified for the SARS-CoV spike protein, which coincide with various anti-SARS peptides derived from the hepated repeat (HR) region 2 of the spike protein. The method should be applicable to other viral proteins to isolate antigen or vaccine candidates, thus providing an alternative to the full-length proteins (subunits) or linear short peptides.

Introduction

Severe acute respiratory syndrome coronavirus (SARS-CoV), which is the causative agent of the atypical pneumonia, was first identified in the fall of 2002 to be a previously unknown member of the family of coronaviruses. The rapid transmission by means of aerosols and the high mortality rate (up to 10%) make SARS a potential global threat. An attractive approach to interfere with SARS disease progression focuses on one of the earliest infection processes by blocking the fusion process that mediates the delivery of the viral genome into the host cell. The spike protein (S protein) is a 180- to 200-kDa type-I transmembrane glycoprotein that is responsible for the initiation and propagation of infection by interacting with a cellular receptor to induce cell-to-cell fusion. Binding of the S protein to a specific soluble or cell surface glycoprotein receptor induces global changes in the conformation of S protein that displays a previously hidden hydrophobic surface area which allows the virions to interact with the host cell membrane2, 3. Earlier attempted expressions of this S protein failed to obtain soluble full-length polypeptides in Escherichia coli (E. coli). Subsequent work aimed at identifying smaller but folded SARS-CoV spike fragments for use as possible antigen or vaccine candidates. The approach involved digestion and reassembly of the target gene to generate a pool with smaller DNA fragments of random ends but controllable lengths which were screened for foldable fragments using a green fluorescent protein as a folding reporter4, 5. Two foldable fragments were identified, which coincide with various SARS peptides reported to have SARS neutralization activity6, 7, 8. This dissection approach has the potential to be a generally applicable tool for producing foldable fragments of viral surface proteins that may provide discontinuous epitopes. These fragments should be easier to express in E. coli or other recombinant hosts.

Materials and Methods

pET30a-linker-GFP construction

The green fluorescent protein (GFP) gene was amplified from an in-house GFP-containing vector pET30a-hydA (Wang and Lin, unpublished result), which was in turn constructed from pQB-2, then ligated into the pET30a(+) (Novagen) to yield the pET30a-linker-GFP.

Fragment library construction

The SARS-CoV spike gene was obtained from the Huada Beijing Genomics Institute. Fragmentation and re-assembly of the target gene were performed as described by Lorimer and Pastan. The reassembled DNA sample was then purified and phosphorylated with T4 polynucleotide kinase at 37°C for 30 min. The backbone vector pET30a-linker-GFP was digested with EcoR I, blunt-ended with T4 DNA polymerase in the presence of 0.1 mmol/L each dNTP, and purified with a QIAgen® gel purification kit to remove residual enzyme activity. The linearized and blunt-ended vector was then dephosphorylated with shrimp alkaline phosphatase (SAP) followed by heat denaturation to deactivate the enzyme. The gene fragments and the backbone vector were ligated at 12°C overnight in the presence of 5% PEG8000 and then transformed into E. coli BL21(DE3) (Novagen) competent cells by electroporation.

Screening of fragments

Transformed E. coli BL21(DE3) cells were plated on Luria-Bertani (LB) medium supplemented with 50 μg/mL kanamycin and grown overnight at 37°C, then grown further on a bench for about 20 h. The fluorescent colonies were picked and tested with colony PCR by using primers flanking the fragment inserts, and sequenced. No isopropylthio-β-D-galactoside (IPTG) was used in these experiments, as it would inhibit the formation of fluorescent colonies.

Expression analysis of fusion proteins

Saturated overnight cultures were diluted 100-fold into LB medium containing 50 μg/mL kanamycin and grown at 37°C for about 2 h to reach an optical density at 600 nm (OD600) of 0.5-0.6. Protein expression was initiated with 0.2 mmol/L of IPTG, and continued for 4 h at 23°C. Cells were then collected and lysed for soluble protein extraction. The supernatant fractions (soluble protein) and cell pellets (insoluble protein) were resolved by SDS-PAGE using a 12% acrylamide gel.

Results and Discussion

This work sought to identify smaller but folded SARS-CoV spike fragments for use as possible antigen or vaccine candidates. Compared with linear short peptides derived from the protein, folded fragments may be advantageous as they have the potential to provide discontinuous epitopes. The SARS-CoV spike gene was digested by DNase I to generate a pool of short DNA segments, followed by an additional step of reassembly of these segments to produce a library of DNA fragments with random ends10, 11. This is in part analogous to the DNA shuffling protocol12, 13, but the purpose here is not to produce full-length hybrids from a group of different parental genes, but to generate various smaller DNA fragments from a single template gene. The reassembly step following the DNase I treatment is necessary to prepare a large number of DNA sequences with controlled lengths, which was achieved by tailoring the number of PCR cycles used in the reassembly (see Methods). To screen for discrete folded polypeptide fragments, the reassembled gene fragments were further cloned into a vector as N-terminal fusions to a folding reporter gene which was a variant of the green fluorescent protein that exhibits strong fluorescence upon UV excitation. GFP has been shown to be an effective indicator for the foldability of the upstream polypeptide partner4, 5. The vector construction is shown in Fig. 1 .

Fig. 1

Expression construct (pET30a-linker-GFP). The sequence is flanked by the Nde I and Hind III sites which otherwise is identical to pET30a(+) (Novagen, Madison, WI). It contains a linker sequence GNSAGSSAAGSGS (boxed) upstream of the GFP gene, and an internal EcoR I site (underlined) used for insertion of gene fragments.

Among about 4300 clones screened, 230 clones were found to be fluorescent (see Fig. 2a ). These clones were then subjected to rapid colony PCR analysis. Many of the fluorescent clones were found to contain vectors with SARS spike gene fragments smaller than 100 base pairs (bp). In addition, some others (a total of 20) contained vectors with inserts in the reverse orientation or not in frame as indicated by sequencing. The SDS-PAGE results showed that the peptides encoded by these gene inserts were degraded in the corresponding fusion proteins (data not shown). Finally, the two inserts larger than 150 nucleotides or 50 deduced amino acid (aa) residues that were identified were ssPtu-15 (residues 1118-1175 of the original protein) and ssPtu-16 (residues 1129-1186). The expression of these fragments (in the GFP-fusion form) was further examined by SDS-PAGE. As shown in Figs. 2b and 2c, the fragments were partially soluble when expressed at 23°C. Higher temperatures significantly reduced the amount of soluble protein (data not shown).

Fig. 2

Expression of fusion proteins (a) Colonies obtained from inserting and expression of SARS-CoV spike gene fragments in pET30a-linker-GFP using E. coli BL21(DE3) as the host. (b) Images of the supernatant fractions for the clones containing ssPtu-15, ssPtu-16, and GFP itself, with E. coli BL21(DE3) as control (denoted as “C”). All the pictures were taken under UV irradiation. (c) Coomassie brilliant blue-stained 12% acrylamide SDS-PAGE using E. coli BL21(DE3) as control (denoted as “C”). Calculated molecular masses for GFP, ssPtu-15, and ssPtu-16 were 29.9 kDa, 36.3 kDa, and 36.3 kDa. Corresponding band positions are indicated by arrows. “s” indicates supernatants of lysates and “in” denotes insoluble pellets of the lysates. M: protein marker, broad range (NEB), whose bands were 175, 83, 62, 48, 33, 25, and 17 kDa, respectively.

Widely disparate virus families have been shown to contain two hepated repeat (HR) regions, which play a critical role in viral fusion with the target cell14, 15. Often, one N-terminal HR region (HR1) is adjacent to the cell fusion peptide while a C-terminal HR region (HR2) is close to the transmembrane anchor. The SARS fragment ssPtu-15 isolated in this work overlaps with the HR2 (residues 1147-1185) of the SARS-CoV spike protein, while fragment ssPtu-16 contains the whole SARS HR2 (Fig. 3a ). A hydrophobic cluster analysis of these two fragments showed two significant hydrophobic clusters (Figs. 3b, 3c, and 3d). Presumably, these clusters play a role in the stability and oligomeric specificity of the HR2 structure. In addition, compared with the wild-type sequence, both of the fragments contain a mutation at 1163 (K replaced by E). ssPtu-16 also contains a second mutation at 1151 (from I to T), while ssPtu-15 contains a second mutation at 1157 (S substituted by Y). These mutations are likely a result from the fragment reassembly process. The mutation at 1163 seems to increase the helicity of the fragments. Interestingly, our more recent dissection studies with other proteins rarely resulted in mutations.

Fig. 3

Sequence analysis (a) CLUSTALW alignment of ssPtu-15 and ssPtu-16 with HR2 derived peptides which interfere with SARS-CoV S-mediated fusion to host cells: peptide CP-1, peptides HR2, GST-HR2-38, GST-HR2-44, peptides SHR2-1, SHR2-2, SHR2-8, and SHR2-9. (b) Secondary structure prediction for ssPtu-15 and ssPtu-16 by 3D-Jury (http://bioinfo.pl/Meta/). E, β-strand; H, α-helix. (c) and (d) Hydrophobic clusters analysis (HCA) plot for ssPtu-15 and ssPtu-16 drawn using DRAWHCA (http://bioserv.rpbs.jussieu.fr). Protein sequences are displayed on a duplicated helix using one-letter codes for the amino acids except for prolines (⋆), glycines (♦), threonine (□), and serine (⊡). Hydrophobic residues are automatically contoured.

Several studies6, 7, 8 have reported that SARS-CoV S-mediated fusion can be inhibited by HR2 but not HR1-derived peptides, most likely by interfering with the six-helix bundle formation, a process essential to drive the membrane fusion reaction and to initiate infection. For the majority of these peptides, micromolar concentrations were required for efficient inhibition of the viral infection, indicating that although these peptides are effective, further optimization is required to achieve efficient inhibition of SARS-CoV in infected individuals. Given the high similarity of ssPtu-15 and ssPtu-16 with these peptides derived from the HR2 region6, 7, 8, ssPtu-15 and ssPtu-16 may both have potential as therapeutic agents for the direct inhibition of SARS-CoV cell entry, as an anti-SARS vaccine, and as a high throughput assay for screening for small molecule inhibitors of SARS envelope-mediated cell fusion.

In summary, the dissection approach described in this study has the potential to produce foldable fragments of viral surface proteins that may be useful for the design of antiviral compounds and provide alternative antigen or vaccine candidates. The method is target protein independent and thus can be applied to various viral proteins. The process is also simple and rapid. The method should be applicable for dissecting and understanding other non-viral proteins, for example, to identify smaller polypeptide units that are structurally, functionally, or evolutionally relevant.