Membrane insertion and assembly of epitope-tagged gp9 at the tip of the M13 phage
© Ploss and Kuhn; licensee BioMed Central Ltd. 2011
Received: 28 April 2011
Accepted: 26 September 2011
Published: 26 September 2011
Skip to main content
© Ploss and Kuhn; licensee BioMed Central Ltd. 2011
Received: 28 April 2011
Accepted: 26 September 2011
Published: 26 September 2011
Filamentous M13 phage extrude from infected Escherichia coli with a tip structure composed of gp7 and gp9. This tip structure is extended by the assembly of the filament composed of the major coat protein gp8. Finally, gp3 and gp6 terminate the phage structure at the proximal end. Up to now, gp3 has been the primary tool for phage display technology. However, gp7, gp8 and gp9 could also be used for phage display and these phage particles should bind to two different or more surfaces when the modified coat proteins are combined. Therefore, we tested here if the amino-terminal end of gp9 can be modified and whether the modified portion is exposed and detectable on the M13 phage particles.
The amino-terminal region of gp9 was modified by inserting short sequences that encode antigenic epitopes. We show here that the modified gp9 proteins correctly integrate into the membrane using the membrane insertase YidC exposing the modified epitope into the periplasm. The proteins are then efficiently assembled onto the phage particles. Also extensions up to 36 amino acid residues at the amino-terminal end of gp9 did not interfere with membrane integration and phage assembly. The exposure of the antigenic tags on the phage was visualised with immunogold labelling by electron microscopy and verified by dot blotting with antibodies to the tags.
Our results suggest that gp9 at the phage tip is suitable for the phage display technology. The modified gp9 can be supplied in trans from a plasmid and fully complements M13 phage with an amber mutation in gene 9. The modified phage tip is very well accessible to antibodies.
The bacteriophage M13 is assembled during a secretion process in the cytoplasmic membrane of Escherichia coli. Membrane inserted phage proteins contact the single stranded phage DNA in an helical array and pass through the outer membrane by a porin-like structure composed of gp4 . In the inner membrane a protein complex probably consisting of gp1, gp11 and thioredoxin catalyses the assembly process . First, membrane inserted gp7 and gp9 proteins form a tip structure  that is extended by a multiple array of gp8 proteins, the major coat protein. Gp8 is synthesised as a precursor protein, termed procoat, that is inserted into the inner membrane by the YidC protein [4, 5] and is then processed by leader peptidase . After processing, the transmembrane coat proteins assemble into oligomers and bind to the viral DNA forming the nascent phage filament [7, 8]. This filament traverses the outer membrane through the gp4 complex . Finally, the membrane inserted gp3 and gp6 proteins are assembled onto the extruding phage at the proximal end of the virion terminating phage assembly.
The gp3 protein has been extensively used for the phage display technology. Since gp3 is engaged in the adsorption of the phage onto the host cell certain restrictions on the infectivity of the modified phage have to be encountered . This might be different for gp9 modifications since this protein is localized at the distal end of the filamentous phage particle. Previously, it has been shown that gp9 is accessible in the phage particle . Therefore, gp9 might be a good target for phage display technology . In addition, an attractive idea is to have both ends supplied with functional peptide moieties applicable as molecular measures or bifunctional binders.
Gp9 is a 32 amino acid long protein that is synthesised without a signal sequence. It is thought that the membrane-inserted protein displays its N-terminus into the periplasm. However, the first amino-terminal 17 residues are hydrophobic and it is questionable whether the protein spans the entire bilayer. One possibility to explore this is to fuse hydrophilic peptides onto the N-terminus. When these modified gp9 proteins are inserted into the membrane their amino-terminal region can be analysed whether they are exposed in the periplasm. Therefore, we have fused short antigenic peptides to the N-terminus of gp9 between the residues 2 and 3. They extend the protein by 17 to 36 amino acid residues. The proteins are inserted into the membrane and efficiently assemble onto phage progeny particles since they can substitute for the wild-type protein. Also, the antigenic epitopes are detectable with gold-labelled antibodies by electron microscopy.
The insertion of gp9-T7 into the membrane was then investigated in E. coli JS7131. In these cells, the membrane insertase YidC can be depleted when the cells are grown in the presence of glucose . After 2 h growth under glucose conditions the cells were pulse-labelled with 35S-methionine for 10 min and converted to spheroplasts. The protease mapping (Figure 5B) shows that the YidC depleted cells did not allow the digestion of the T7-epitope at the N-terminus of gp9 (lane 2). These results suggest that the membrane insertion of gp9-T7 is YidC-dependent. In both cases, the integrity of the spheroplasts was verified by the protection of GroEL (lane 4) and the proteolytic activity was corroborated by the accessibility of the OmpA protein (lane 6).
The minor coat protein gp9 of the filamentous phage M13 is exposed from the phage particle and can be modified with short peptides. Here we have shown that peptides of 17, 18, 32 and 36 amino acid residues can be incorporated into the amino-terminal region of the protein without interfering with membrane insertion and assembly of the phage. The epitopes of these peptides are accessible by antibodies and allow binding of gold nanoparticles that can be visualised by electron microscopy. This implicates that gp9 could be used for the phage display methodology allowing a combination with the well-established display of modified gp3. Previous experiments have shown that gp9 of the closely related fd phage is localised at the distal end, together with gp7 . In that study, it was also shown, that the gp9 protein is exposed to the surface in contrast to gp7. However, a fusion with glutathione-S-transferase (GST) did not support phage production suggesting that membrane insertion or phage assembly is inhibited. Since GST is a folded structure of about 35 kDa we tested smaller fusion proteins that may be tolerated for membrane insertion and phage assembly. By introducing short antigenic sequences between the amino acid residues 2 and 3 of gp9 on a plasmid membrane insertion and phage assembly was followed. Also, longer fusions consisting of 32 and 36 additional residues that code for two tandem tags were constructed. Intriguingly, all gp9 fusion proteins complement an amber-9 phage infection and lead to progeny production up to wild-type levels. When the phage progeny particles were analysed for the presentation of their antigenic epitopes we observed by dot-blot analysis (Figure 6) and immunogold labelling (Figure 7) a clearly positive response. We conclude that the amino-terminal end of gp9 is capable to accept modifications and provides a new possibility for phage display.
The extended amino-terminal region with an antigenic tag allowed the investigation of the membrane insertion of gp9 in detail. Previously, it had been shown by FTIR spectroscopy that the membrane-inserted protein has a high α-helical conformation and adopts a transmembrane conformation . In a short pulse, the synthesised gp9 was radioactively labelled and analysed for membrane insertion by protease added to the outside of the membrane (Figure 5). Indeed, the protease removed the antigenic tag at the N-terminus of gp9, whereas the cytoplasmic GroEL protein was protected from proteolysis. When the same experiment was performed in cells that were depleted for YidC, gp9 was not digested suggesting that it was not inserted into the membrane under these conditions. We conclude, that gp9 uses the YidC-only pathway for insertion similar to gp8 [4, 5]. In contrast to our in vivo experiments, earlier in vitro data with artificial liposomes consisting of DOPC and DOPG had suggested that the gp9 protein inserts sponanteously into the membrane .
Very recently, similar gp9 variants to our gp9 fusion proteins were described that allowed a display on the phage . In contrast to our work, a phagemid system was used and the N-terminus of gp9 fusion protein had a pelB signal sequence attached. This likely changes the route of membrane insertion to the Sec-translocase and allows the translocation of large N-terminal domains across the cytoplasmic membrane. Compared to the phagemid system used in previous reports [10, 13–15], we present a new method of gp9 phage display which allows a polyvalent phage display without the need of an N-terminal signal sequence and helper phage infection. In our system the only gp9 copy available is the modified gp9 protein on a plasmid when amber 9 phage was used. Therefore, all gp9 proteins on the phage particle possess the modified N-terminus. Further, our system allows to clearly determine the extend of interference of the modified protein with the propagation cycle of the phage. When no interference is observed and the gp9 fusion protein can substitute for the wild-type, amber 9 phage can be used to generate modified phage particles without a modification of the phage genome. This application might be useful for systems that are sensitive to genetically modified organisms according to (GMO)-rules.
Bacteriophage M13 is suitable for phage display not only with a modified gp3 but also with a modified gp9 which is a minor coat protein at the phage tip. The modified gp9 protein can be supplied in trans from a plasmid and fully complements an amber 9 phage mutant. The modified phage tip is very well accessible to specific antibodies.
M13 phage was from our lab collection . M13am9 with an amber mutation in the second codon of gIX was constructed by site-directed mutagenesis . For the construction of gp9-T7, gp9-DT7, gp9-HA and gp9-DHA RF-DNA of M13mp19 served as template for PCR amplification. The PCR amplified gIX was subcloned into pMS119  and an unique MunI restriction site was introduced by QuikChangeTM in vitro mutagenesis between the codons 2 and 3. Into this site RF-DNA of M13mp19 served as template for the amplification of gIX by PCR. The gIX fragment was subcloned into pMS119, DNA fragments encoding the T7 and HA tag sequences were introduced by ligation, resulting in pMS-g9-T7 and pMS-g9-HA. Also, longer epitopes were introduced to construct pMS-g9-DT7 and pMS-g9-DHA, respectively. For protein expression and complementation experiments E. coli K38 (HfrC T2R relA1 pit-10 spoT1 tonA22 ompF627 phoA4 λ-)  was transformed as a non-suppressor strain. E. coli K37 (HfrC supD32 relA1 pit-10 spoT1 tonA22 ompF627 phoA4 T2R λ-) [19, 20] was used as a suppressor strain and E. coli JS7131 (MC1060 ΔyidC attB::R6Kori ParaBADyidC + Specr) as a depletion strain of the membrane insertase YidC .
On agar plates 4 mL melted LB top agar (47°C) containing 1 mM IPTG was mixed with 500 μL of a fresh E. coli K38 overnight culture bearing either pMS-g9/7 pMS-g9-T7, pMS-g9-DT7, pMS-g9-HA or pMS-g9-DHA. After solidification of the top agar, 10 μL of a phage suspension was applied on top of the agar from serial dilutions of a phage stock. Plaque formation was observed after incubation at 37°C overnight.
2 mL cultures of E. coli K38 bearing plasmids encoding a respective gp9 variant were grown at 37°C to the early exponential phase in M9 minimal medium. Protein expression was induced by adding 1 mM IPTG and 10 min later the newly synthesised proteins were pulse-labelled for 10 min with 20 μCi 35S-methionine. To remove the non-incorporated 35S-methionine the total bacterial proteins were precipitated with 12% TCA on ice overnight, washed with cold acetone and resuspended in 10 mM Tris/HCl 2% SDS, pH 8.0. The samples were immunoprecipitated with antiserum to the T7 tag (Novagen) or to the HA tag (Roche), respectively, and analysed by SDS tricine PAGE and phosphorimaging.
To test the membrane insertion of gp9, E. coli K38 bearing pMS-g9-T7 was grown to the early exponential phase in M9 minimal medium. Cells were induced for 10 min with 1 mM IPTG and labelled with 35S-methionine for 10 min. To generate spheroplasts, the cells were centrifuged at 12 000 g for 3 min and resuspended in 500 μL of ice-cold spheroplast buffer (40% w/v sucrose, 33 mM Tris/HCl, pH 8.0). Lysozyme (5 μg/mL, final concentration) and 1 mM EDTA were added for 15 min. Aliquots of the spheroplast suspension were incubated on ice for 1 h either in the presence or absence of 0.5 mg/mL proteinase K. The samples were precipitated with 12% TCA, washed with cold acetone and resuspended in 10 mM Tris/HCl, 2% SDS, pH 8.0 and immunoprecipitated with antibodies against T7, OmpA (a periplasmic control), or GroEL (a cytoplasmic control). Samples were analysed by SDS tricine PAGE and phosphorimaging.
To test the requirement of YidC for the membrane insertion of gp9-T7, the YidC depletion strain E. coli JS7131 bearing pMS-g9-T7 was grown to the early exponential phase in LB with 0.2% arabinose. After back-dilution, the cells were grown in M9 minimal medium with either 0.2% arabinose (YidC+) or 0.2% glucose (YidC-) for 2 h. To induce expression of gp9-T7, 1 mM IPTG was added and after 10 min the cells were pulse-labelled with 35S-methionine for 10 min and then converted to spheroplasts by lysozyme treatment as described above. Samples were immunoprecipitated with antibodies to T7, OmpA (a periplasmic control), or GroEL (a cytoplasmic control). For testing the YidC depletion, samples of the cultures were drawn and precipitated with TCA (12%, final concentration), washed with cold acetone, resuspended in 10 mM Tris/HCl, 2% SDS, pH 8.0 and analysed by SDS/PAGE and Western blot using YidC antiserum.
50 mL cultures of E. coli K38 cells harbouring either pMSg9-T7, pMSg9-DT7, pMSg9-HA or pMSg9-DHA were grown at 37°C in LB-medium to a density of 2 × 108 cells/mL. The expression of the gp9 variant proteins was induced by adding 1 mM IPTG and the cells were infected with M13am9 at m.o.i 10. Adsorption of the phage was allowed for 5 min at room temperature without shaking. Subsequently, the infected cells were shaken overnight at 37°C. The phage was harvested from the supernatant after removing the cells by centrifugation. Then, the phage titer was determined by serial dilutions on E. coli K37. Every dilution was plated three times on LB agar plates to control variations in plating and pipetting. The agar plates were incubated at 37°C overnight and the plaques were counted and averaged for each dilution step.
For detection of the plasmid-encoded variants on the phage via dot-blot, serial dilutions of the above described phage stocks were prepared resulting in equal amounts of phage particles/400 μL for every variant. 400 μL of each suspension was adsorbed on a nitrocellulose membrane (Hybond ECL Nitrocellulose, Amersham) via dot-blot equipment (MiniFold®, Schleicher & Schuell) and treated overnight with blocking solution (1x Tris-buffered saline (TBS) pH 8, 5% non-fat dry milk w/v). The blot was washed three times with 1x TBS and incubated with antiserum to M13 gp8, to T7 or to HA tag, respectively. The presence of gp9 variants was analysed with a secondary peroxidase-coupled antibody by chemoluminescence.
For testing the exposure of an antigenic epitope 50 μL of each phage stock solution (about 1011 phage/mL) of M13gp9-DT7 and M13gp9-DHA was incubated with 1 × TBS containing 0.1% BSA for 30 min to avoid unspecific binding of the primary antibody to the sample. Each sample was then incubated with the respective serum (diluted 1:20 in 1x TBS) for 1 h. Then, protein A coupled immunogold particles (Protein A - 20 nm colloidal gold, Sigma-Aldrich) was added 1:20 in 1x TBS for 1 h. After immunogold labelling, 10 μL of the phage stock solution was adsorbed on carbon-coated copper grids (Athene 200, Plano, Wetzlar/Germany) that had been glow discharged shortly before use . The suspensions were allowed to adsorb for 5 min, unbound material was removed by touching the grid to filter paper. The grid was then washed by touching the surface of a drop of distilled water for 2 sec. The excess water was removed by touching the grid to filter paper. A drop (5 μL) of 5% phosphotungstic acid (pH 7) was then applied to the grid and after 30 sec the excess stain was removed by touching the grid to a drop (50 μL) of ddH20 for 2 sec. The excess liquid was drawn off with filter paper. The grid was dried at room temperature and examined by electron microscopy.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.