Anal Bioanal Chem (2014) 406:7611–7621 DOI 10.1007/s00216-014-8150-8
RESEARCH PAPER
Identification and characterization of a phage display-derived peptide for orthopoxvirus detection Lilija Miller & Janine Michel & Guido Vogt & Jörg Döllinger & Daniel Stern & Janett Piesker & Andreas Nitsche
Received: 29 April 2014 / Revised: 17 July 2014 / Accepted: 1 September 2014 / Published online: 14 October 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Fast and reliable diagnostic assays are required for a resilient detection of clinical infections or biothreat-relevant pathogens. While PCR has proven to be the gold standard for nucleic acid detection, the identification of pathogen particles is still challenging and depends on the availability of wellcharacterized, chemically stable, and selective recognition molecules. Here, we report the screening of a phage display random peptide library for vaccinia virus-binding peptides. The identified peptide was extensively characterized using peptide-probe ELISA, surface plasmon resonance, nLC-MS/ MS, Western Blot, peptide-based immunofluorescence assay, and electron microscopy. Following identification, the phagefree, synthetic peptide, designated αVACVpep05, was shown to bind to vaccinia virus and other orthopoxviruses. We can demonstrate that the highly conserved orthopoxvirus surface protein D8 is the interaction partner of αVACVpep05, thus enabling the peptide to bind to other orthopoxviruses, including cowpox virus and monkeypox virus, viruses that cause clinically relevant zoonotic infections in humans. The process of phage display-mediated peptide identification has been optimized intensively, and we provide recommendations for the identification of peptides suitable for the detection of Lilija Miller and Janine Michel contributed equally to this work. Electronic supplementary material The online version of this article (doi:10.1007/s00216-014-8150-8) contains supplementary material, which is available to authorized users. L. Miller : J. Michel : G. Vogt : J. Döllinger : D. Stern : J. Piesker : A. Nitsche (*) Centre for Biological Threats and Special Pathogens, Highly Pathogenic Viruses, Robert Koch-Institut, Nordufer 20, 13353 Berlin, Germany e-mail: [email protected] L. Miller Junior Research Group 2, Novel Vaccination Strategies and Early Immune Responses, Paul-Ehrlich-Institut, Paul-Ehrlich-Straße 51-59, 63225 Langen, Germany
further pathogens. The peptide described here was critically characterized and seems to be a promising reagent for the development of diagnostic platforms for orthopoxviruses. We believe that our results will help to promote the development of alternative, nonantibody-based synthetic detection molecules for further pathogens. Keywords Phage display . Synthetic peptide . Pathogen detection . Vaccinia virus . Orthopoxviruses
Introduction The availability of rapid and reliable detection methods is an important factor in the detection of clinical infections, emerging pathogens, or biothreat agents. While numerous wellestablished detection methods are based on the amplification of nucleic acids, the detection of the pathogen particle itself is still challenging, and few approaches have been developed so far [1]. In the majority of particle detection approaches, specific binders are polyclonal or monoclonal antibodies. Despite numerous advantages, like specificity and affinity, antibodies also have limitations including tedious and expensive production, limited shelf-life, and the requirement of using animals [2]. Thus, there is an increasing interest in employing alternative recognition molecules to detect pathogens [3]. Ideally, these alternative molecules would display the benefits of antibodies by being more stable and easier to generate at the same time. This makes short proteins interesting candidates, or even peptides that can be selected by phage display [4]. Phage display is a molecular methodology by which heterologous peptides or short proteins are expressed on the surface of phage particles [5]. Applications include the construction and screening of combinatorial peptide libraries. For this purpose, synthetic oligonucleotides, fixed in length but
7612
with unspecified codons, are cloned as fusions to genes III or VIII of the filamentous phage M13 where they are multiply expressed as peptide–capsid fusion proteins. The resulting libraries can then be assayed for binding to target molecules in a process called “biopanning” [6]. A successful biopanning experiment results in the identification of consensus peptide sequences or amino acid motifs. The selected binding phage clones can be re-tested for their binding specificity to the target used for biopanning. For more detailed binding studies, a phage-free peptide can be synthesized and characterized more conveniently using various methods. In the past, phage display has been successfully utilized to identify pathogen-binding phage clones including viruses [7–9], bacterial spores [10], and bacterial cells [11, 12]. Despite some reports on subsequent application of the phages as well as phage-derived synthetic peptides for detection purposes, critical characterizations of the interaction between the phage-derived peptide and its target are still missing. Here, we performed biopanning against native vaccinia virus (VACV) particles to identify VACV particle-binding peptides. VACV is a complex virus belonging to the genus Orthopoxvirus (OPV) within the family Poxviridae. The genus OPV not only contains one of the most dangerous biothreat-relevant viruses, variola virus (VARV) [13], but its members cowpox virus (CPXV) and monkeypox virus (MPXV) are causing clinically relevant zoonotic infections today [14]. The biopanning against VACV resulted in the identification of a VACV-binding peptide, termed αVACVpep05. We further identified the VACV surface protein D8 as its viral target and determined the binding kinetics and affinity between the peptide αVACVpep05 and the D8 protein. Finally, we demonstrated that peptide αVACVpep05 can also detect other OPV members.
L. Miller et al.
bicarbonate buffer (Sigma), and lysed with IP buffer (Invitrogen) containing 100 mM NaCl and protease inhibitor (Thermo Scientific). Recombinant poxvirus proteins A27, A33, B5, F9, and L1 were from BEI Resources (Manassas, VA, USA). Recombinant proteins D8 (AA 2–260), F13 (full length), and H3 (AA 21–270) (all expressed in E. coli) were purchased from GenExpress (Berlin, Germany). Phage biopanning Affinity selections were performed using the Ph.D.-12™ Phage Display Peptide Library Kit as recommended by the manufacturer (New England Biolabs, NEB, complexity of the library on the order of 109 independent clones) with minor changes. MaxiSorp ELISA plate wells (Nunc) were coated with 5×106 pfu/well of VACV. One additional well was filled with blocking buffer for negative selection performed before every biopanning round. The Ph.D.-12 library was diluted to 2×1011 pfu/well (dilution buffer: PBS with 2 % [w/v] BSA and 0.5 % [v/v] Tween-20) and first added to BSA-coated wells for 2 h at room temperature (RT). The supernatant was transferred to target-coated wells and incubated for 2 h. Wells were washed 15 times after the first selection round and 20 times in all subsequent rounds with 300 μL/well of PBS with 0.5 % [v/v] Tween-20. Bound clones were eluted with 100 μL of 0.2 M glycine-HCl, pH 2.2, by incubating for 15 min at RT. The eluate was neutralized with 15 μL of 1 M Tris–HCl, pH 9.1, and then incubated in a water bath at 60 °C for 1 h to inactivate infectious poxvirus particles potentially present. Four selection rounds were performed altogether. After the fourth round, binding phages were eluted and used for a further negative selection against HEp-2 cell lysate. Sequence analysis of binding phage clones
Materials and methods Viral strains and other reagents VACV NYCBOH (VR-1536™) and porcine parvovirus (PPV; VR-742™) were purchased from ATCC®. The adenovirus (AdV) [15], CPXV strain GuWi [16], and MPXV (in press) were taken from RKI strain collection. Camelpox virus (CMLV CP-19), ectromelia virus (MP-Nü), and MPXV strain MSF#6-Mü were kindly provided by Hermann Meyer (Bundeswehr Institute of Microbiology, Munich). For preparation of HEp-2 cell lysate (used for negative selections), cells were lysed with RIPA buffer according to the manufacturer’s instructions (Thermo Scientific). For preparation of VACV lysate, VACV particles (propagated on HEp2 cells) were semi-purified over a sucrose cushion as described before [17], washed with 100 mM ammonium
For the identification of binding peptides, 10, 20, 20, and 68 phage clones of the first, second, third, and fourth eluate, respectively, were plugged and amplified. Single-stranded DNA was purified from phages and sequenced by the Sanger method using a “BigDye Terminator v3.1 Cycle Sequencing Kit” as recommended (Applied Biosystems). Nucleotide sequences were translated into amino acid sequences using “Library Insert Finder” software. Repeatedly identified sequences were defined as potential consensus peptide sequences, and the corresponding phage clones were characterized further. Phage ELISA Phage ELISA was performed as described (NEB) with minor changes. ELISA plates were coated with 105 pfu/well of VACV or PPV. Selected phage clones were added at
A phage display-derived peptide for orthopoxvirus detection
1010 pfu/well and incubated for 2 h at RT. Monoclonal phages were also added to uncoated, blocked wells to test for unspecific binders. Phage binding was detected by an anti-M13 antibody conjugated to horse radish peroxidase (HRP). A 3,3′,5,5′-tetramethylbenzidine (TMB, Sigma) solution was used as HRP substrate. Plates were assayed at 450/620 nm using an Infinite®200 PRO microplate reader (Tecan Group).
Peptide selection for synthesis Selected sequences were synthesized by PANATecs as phagefree peptides. Selection of suitable peptide sequences was based on DNA sequencing (most frequent) and phage ELISA results (specific target binding). To obtain similar properties to phage-attached peptides, synthetic peptides were ordered with the following features: (1) N-terminus acetylated to block positive charge, (2) C-terminus containing a linker sequence (Gly-Gly-Gly-Ser), followed by a lysine residue and a biotin molecule, and (3) peptide purity of >85 % (biochemistry grade). To rule out a lot-to-lot variability, the most specific peptide was reordered at PANATecs and at Biomatik, respectively.
Peptide-probe-based ELISA Synthetic peptides were tested using ELISA. To this end, plates were coated with 5×105 pfu/well of VACV, PPV, or AdV as specificity controls. Plates were washed twice with 300 μL/well of washing buffer (PBS with 0.05 % [v/v] Tween20), then completely filled with blocking buffer (PBS with 3 % [w/v] BSA) and incubated for 2 h at RT. Synthetic peptides were twofold serially diluted, added at 100 μL/well, and incubated for 2 h at RT. Subsequently, plates were washed six times, and 100 μL/well of streptavidin peroxidase (SAPOD) (Dianova) were added for 1 h at RT. Wells were washed six times, a TMB substrate solution was added, and the color reaction was detected. For selectivity analysis of the synthetic peptides towards different OPV species, experiments were performed as described above with minor modifications. Wells coated with parapoxvirus (PPXV), HEp-2 cell lysate, and BSA were utilized as controls. Peptides were added at 400 nM and incubated for 2 h at RT. To assess functionality of peptides as capture molecules for OPV, the biotinylated peptides (100 μL of 400 nM solution) were immobilized on neutravidin-coated plates (Pierce) for 2 h at RT. Subsequently, VACV, CMLV, CPXV, ECTV, PPXV, and MPXV were added to individual wells and incubated for 1 h at RT. After washing, poxvirus particles were detected by a polyclonal rabbit anti-VACV antibody (Acris Antibodies) and an HRP-labeled anti-rabbit serum (Dianova).
7613
Pull down assay, SDS-PAGE, and nLC-MS/MS analysis A pull down assay with subsequent nano-scale liquid chromatographic tandem mass spectrometry (nLC-MS/ MS) analysis using peptide-coated magnetic beads and VACV was employed to identify the interaction partner of the VACV-binding phage-free synthetic peptide designated as αVACVpep05. For pull down assay, magnetic SA-coupled Dynabeads® (Invitrogen Dynal) were incubated with 600 nM peptide for 30 min at RT. Beads were then washed five times, incubated with 100 μL of VACV lysate [17] for 1 h at RT, and again washed six times. These washing fractions 1–6 (500 μL) were collected and their volume reduced with a Speedvac to the same volume as the final eluate (50 μL). For elution of bound protein, beads were first incubated with 50 μL of RIPA buffer for 15 min on ice, followed by a 5-min incubation step at 95 °C. For SDS-PAGE, virus lysate, collected washing fractions 1, 4, 5, and 6, and the eluate were each mixed with 6× Laemmli buffer (Bio-Rad) and incubated for 5 min at 95 °C. Subsequently, 50 μL of each sample were run on an 8–16 % SDS gel (Thermo Scientific). The washed gel was stained overnight with the Flamingo™ fluorescent gel stain (Bio-Rad) and the images acquired using the Image Lab software (v4.0.1). For mass spectrometric analysis, detected protein bands were excised from gel and prepared for analysis as described elsewhere [18]. Peptides were separated with Easy-nanoLC (Thermo Scientific) on a 10 cm, ID 75 μm, 3 μm, C18 column using a linear 30 min gradient of 2–40 % acetonitrile in 0.1 % formic acid at 300 nL/ min. The LC was coupled online to an LTQ Orbitrap Discovery™ mass spectrometer (Thermo Scientific), which was operated in a data-dependent manner in the m/z range of 300–1,700 by selecting the Top7 2+ and 3+ charged ions for CID type fragmentation. Proteins were identified using SEQUEST algorithm via Proteome Discoverer 1.3 (Thermo Scientific). Therefore, a custommade database was generated from all VACV entries of the UniProt Knowledge database. The quality thresholds for peptide identification were XCorr values of ≥2 (2+ ions), ≥2.5 (3+ ions) and a maximum deltaCn of 0.05. The false discovery rate was validated on q values and set to be below 1 %. Surface plasmon resonance measurements For surface plasmon resonance (SPR) analysis, an SAcoated sensor chip was utilized in a Biacore X 100 SPR unit (GE Healthcare). To prepare the SA chip for binding, 1 M NaCl in 50 mM NaOH was injected three times. Biotinylated peptides diluted in running buffer (HBS-EP+
7614
L. Miller et al.
Table 1 Consensus peptide sequences of potential VACV-binding phage clones Phage Amino acid sequencea clone no.
Frequencyb
1 2 3 4 5 6 7 8 9 10 11 12 13
KVWLPPRHEHQY K V F Y PA A A N P N Q QALLEGNAKGGN NRPDSAQFWLHH TA D K L LY G L F K S DEWDALLMRIRT NPTPYPMLPLRG K P T Y S W D PA Q L K GPTFSWDHLRGQ KIFQLPQISPPM GDLASWIITSFK NMELHPHSLPRP ANTTKHSVLAAI
2nd eluate 2 1 – – – – – – – – – – –
14 15 16 17
EALNDWVNDSEY A P TAY N K N D WA L DPWWRGNEARAA TPVWSWEPPLQE
– – – –
3rd eluate 1 1 1 3 1 1 – – – – – – –
4th eluate 3 – 2 – 6 3 8 2 2 4 8 2 2
– – – –
2 2 2 4
a
Sequences are provided using the one-letter amino acid code from Nterminus to C-terminus
b Frequency of how often the peptide sequence was found in the respective phage eluate
, 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.05 % Tween 20) were immobilized at surface densities of 269 RU αVACVpep01 (non-binding peptide) to flow cell 1 and 67 RU αVACVpep05 to flow cell 2.
Fig. 1 Estimation of binding specificities of selected phage clones containing consensus peptide sequences by phage ELISA. To test the binding specificity of the 17 frequently found phage clones, MaxiSorp ELISA plates were coated with 105 pfu/well of VACV or PPV particles. Wells filled with blocking buffer (BSA) were utilized to test for unspecific binders. Selected phage clones (1–17) were diluted and added at 1010 pfu/well. Additionally, the wild-type M13KO7 phage (WT) and the
To find the binding partner of αVACVpep05, 50 μg/ mL of various recombinant VACV surface proteins were injected over the sensor chip for 120 s at 5 μL/min. To remove bound proteins from immobilized peptides between measurement cycles, flow cells were regenerated by injection with 10 mM glycine-HCl pH 1.7 (GE Healthcare) for 30 s at 10 μL/min. After identification of D8 (MW 30 kDa) as binding partner, affinity to αVACVpep05 was determined by application of twofold serially diluted (3.25 nM– 1.67 μM; duplicate injection of 1.67 μM) recombinant D8 protein to both flow cells of the SA chip with a flow rate of 30 μL/min. Association measurement was performed for 120 s, and dissociation was assessed for 60 s. All measurements were performed at 25 °C. For analysis, double-referenced [19] response signals were fit with BIAEvaluation Software 1.0 to either a 1:1 Langmuir binding model to determine binding kinetics or a Steady-State-Affinity model to determine binding affinity.
Western Blot For Western Blot analysis, 500 ng of recombinant proteins A27, A33, B5, D8 (native and heat-treated), H3, F9, F13, and L1 and 3 μg of a VACV lysate (native and heat-treated) were separated on a 4–20 % premade SDS gel (Thermo Scientific) and transferred to a nitrocellulose membrane (Thermo Scientific). The membrane was then incubated with 500 ng/mL of the biotinylated αVACVpep05. After washing, proteins were detected by incubation with SA-POD (1:50,000) and subsequent substrate reaction.
original Ph.D.-12 library (Lib) were tested as controls. The phage binding was detected with a monoclonal anti-M13 antibody. All samples were measured in duplicate once; error bars are provided as standard deviations. Arrows point at the five phage clones that were selected for peptide synthesis, according to the signal intensity and the difference in signal between VACV and the control virus PPV
A phage display-derived peptide for orthopoxvirus detection
7615
Table 2 Criteria for the selection of phage clones with peptides to be synthesized Phage clone no.
ODVACV-ODPPVa
ODVACV*Clone-ODVACV*Libb
1 2 3 4
0.50 0.01 0.39 0.24
0.00 −0.52 −0.02 −0.01
5 6 7 8 9 10 11 12 13 14 15 16 17
0.79 0.15 0.10 0.48 0.24 0.08 0.28 0.09 0.38 0.51 0.08 0.18 0.39
1.06 0.46 −0.40 0.06 −0.23 0.19 0.94 −0.30 0.26 0.53 0.03 −0.12 0.99
colloidal gold particles and negatively stained as previously described [20].
Peptide-based fluorescence assay For visualization of the specific interaction of peptide αVACVpep05 with OPV-infected cells, CPXV-infected HEp-2 cells fixed onto glass slides were incubated with peptide αVACVpep05 (50 μg/mL) preincubated with streptavidin-fluorescein isothiocyanate (SA-FITC, 1:80, Invitrogen) for 1 h at 37 °C. Cells were additionally incubated with a polyclonal anti-D8 antibody coupled to DyLight 649 (RKI) to assess colocalization. The slides were analyzed using a confocal laser scanning microscope (Zeiss).
Results
a
Differences in optical density (OD) values between VACV signal (ODVACV) and PPV signal (ODPPV) as determined with the respective phage clone using phage ELISA b Differences in the OD values between the VACV signal of the individual phage clone (ODVACV*Clone) and the VACV signal of the native phage display library (ODVACV*Lib) as determined by phage ELISA
Electron microscopy For visualization of specific binding of the αVACVpep05 to OPV particles by electron microscopy, semi-purified [17] ectromelia virus (ECTV) particles were fixed with 2 % paraformaldehyde in 0.05 M HEPES buffer and adsorbed onto a sample support (400 mesh copper grid, coated with a plastic film). Then, samples were incubated with biotinylated peptides, followed by an anti-biotin antibody coupled to 10 nm
Biopanning against native VACV particles resulted in the identification of multiple consensus peptide sequences To isolate binders to VACV, a commercially available filamentous phage display library (∼109 independent clones) was used. Four rounds of biopanning were carried out against infectious VACV particles immobilized on wells of a 96well plate. In total, 118 plaques were plugged and the ssDNA prepared from the phages and subsequently sequenced. Among the peptide sequences identified, 17 sequences were present two to eight times and were thus defined as consensus peptide sequences (Table 1). The presence of such consensus sequences after multiple selection rounds indicates an enrichment of specifically binding phage clones. Nevertheless, to decide which of the identified phage clones encodes the most suitable peptide for virus detection, the binding specificities of the selected phage clones were estimated.
Table 3 Potential VACV-binding peptides selected for synthesis Phage clone no.
Peptide designation
Amino acid sequencea
Frequencyb
1 5 8 13 14
αVACVpep01 αVACVpep05 αVACVpep08 αVACVpep13 αVACVpep14
Ac-KVWLPPRHEHQYGGGSK-NH2 (Biotin)c Ac-TADKLLYGLFKSGGGSK-NH2 (Biotin) Ac-KPTYSWDPAQLKGGGSK-NH2 (Biotin) Ac-ANTTKHSVLAAIGGGSK-NH2 (Biotin) Ac-EALNDWVNDSEYGGGSK-NH2 (Biotin)
6 7 2 2 2
a
Sequences are provided using the one-letter amino acid code from N-terminus to C-terminus
b
Frequency of a peptide sequence found in all phage eluates
c
The abbreviations Ac and NH2 refer to the presence of an N-terminal acetyl group and a C-terminal amide; all peptides were synthesized with a biotin at the C-terminus
7616
Estimation of binding specificities of phage clones enriched during biopanning using a phage ELISA
L. Miller et al.
clones which fulfill the two criteria (i) ODVACV −ODPPV ≥0.3 and (ii) ODVACV*Clone −ODVACV*Lib ≥0 (Table 2) and to subsequently test the binding characteristics of the corresponding synthetic peptides. According to this, five peptide sequences derived from phage clones number 1, 5, 8, 13, and 14 were selected for synthesis (Table 3).
Seventeen monoclonal phage clones were assayed by phage ELISA to estimate their binding to the target. Of note, the estimation of the binding specificities of selected phage clones has been frequently recommended for exclusion of biopanning process artifacts. However, no defined criteria for the choice of the “right” phage clone based on the obtained phage ELISA results have been defined so far. Exactly those criteria could help to select a phage clone bearing a peptide with the best binding characteristics. Most frequently, binding characteristics of the selected phage clone are tested against target- and BSA-coated wells in a phage ELISA. More stringent criteria, such as testing the binding against an alternative target, have to our knowledge, not been applied so far. To be able to subsequently define those criteria, we decided to additionally (i) include wells coated with PPVas an unrelated virus, and (ii) to compare the binding characteristics of enriched phage clones with those of the native phage display library. Nearly all individual phage clones bound better to VACV than to the control virus (PPV), with ELISA signals ranging from 0.4 to 2.0 optical density (OD) for VACV-coated wells, and 0.2 to 1.6 OD for PPV-coated wells (Fig. 1). The addition of phage clones to uncoated, BSA-blocked wells resulted in low signals of
RESEARCH PAPER
Identification and characterization of a phage display-derived peptide for orthopoxvirus detection Lilija Miller & Janine Michel & Guido Vogt & Jörg Döllinger & Daniel Stern & Janett Piesker & Andreas Nitsche
Received: 29 April 2014 / Revised: 17 July 2014 / Accepted: 1 September 2014 / Published online: 14 October 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Fast and reliable diagnostic assays are required for a resilient detection of clinical infections or biothreat-relevant pathogens. While PCR has proven to be the gold standard for nucleic acid detection, the identification of pathogen particles is still challenging and depends on the availability of wellcharacterized, chemically stable, and selective recognition molecules. Here, we report the screening of a phage display random peptide library for vaccinia virus-binding peptides. The identified peptide was extensively characterized using peptide-probe ELISA, surface plasmon resonance, nLC-MS/ MS, Western Blot, peptide-based immunofluorescence assay, and electron microscopy. Following identification, the phagefree, synthetic peptide, designated αVACVpep05, was shown to bind to vaccinia virus and other orthopoxviruses. We can demonstrate that the highly conserved orthopoxvirus surface protein D8 is the interaction partner of αVACVpep05, thus enabling the peptide to bind to other orthopoxviruses, including cowpox virus and monkeypox virus, viruses that cause clinically relevant zoonotic infections in humans. The process of phage display-mediated peptide identification has been optimized intensively, and we provide recommendations for the identification of peptides suitable for the detection of Lilija Miller and Janine Michel contributed equally to this work. Electronic supplementary material The online version of this article (doi:10.1007/s00216-014-8150-8) contains supplementary material, which is available to authorized users. L. Miller : J. Michel : G. Vogt : J. Döllinger : D. Stern : J. Piesker : A. Nitsche (*) Centre for Biological Threats and Special Pathogens, Highly Pathogenic Viruses, Robert Koch-Institut, Nordufer 20, 13353 Berlin, Germany e-mail: [email protected] L. Miller Junior Research Group 2, Novel Vaccination Strategies and Early Immune Responses, Paul-Ehrlich-Institut, Paul-Ehrlich-Straße 51-59, 63225 Langen, Germany
further pathogens. The peptide described here was critically characterized and seems to be a promising reagent for the development of diagnostic platforms for orthopoxviruses. We believe that our results will help to promote the development of alternative, nonantibody-based synthetic detection molecules for further pathogens. Keywords Phage display . Synthetic peptide . Pathogen detection . Vaccinia virus . Orthopoxviruses
Introduction The availability of rapid and reliable detection methods is an important factor in the detection of clinical infections, emerging pathogens, or biothreat agents. While numerous wellestablished detection methods are based on the amplification of nucleic acids, the detection of the pathogen particle itself is still challenging, and few approaches have been developed so far [1]. In the majority of particle detection approaches, specific binders are polyclonal or monoclonal antibodies. Despite numerous advantages, like specificity and affinity, antibodies also have limitations including tedious and expensive production, limited shelf-life, and the requirement of using animals [2]. Thus, there is an increasing interest in employing alternative recognition molecules to detect pathogens [3]. Ideally, these alternative molecules would display the benefits of antibodies by being more stable and easier to generate at the same time. This makes short proteins interesting candidates, or even peptides that can be selected by phage display [4]. Phage display is a molecular methodology by which heterologous peptides or short proteins are expressed on the surface of phage particles [5]. Applications include the construction and screening of combinatorial peptide libraries. For this purpose, synthetic oligonucleotides, fixed in length but
7612
with unspecified codons, are cloned as fusions to genes III or VIII of the filamentous phage M13 where they are multiply expressed as peptide–capsid fusion proteins. The resulting libraries can then be assayed for binding to target molecules in a process called “biopanning” [6]. A successful biopanning experiment results in the identification of consensus peptide sequences or amino acid motifs. The selected binding phage clones can be re-tested for their binding specificity to the target used for biopanning. For more detailed binding studies, a phage-free peptide can be synthesized and characterized more conveniently using various methods. In the past, phage display has been successfully utilized to identify pathogen-binding phage clones including viruses [7–9], bacterial spores [10], and bacterial cells [11, 12]. Despite some reports on subsequent application of the phages as well as phage-derived synthetic peptides for detection purposes, critical characterizations of the interaction between the phage-derived peptide and its target are still missing. Here, we performed biopanning against native vaccinia virus (VACV) particles to identify VACV particle-binding peptides. VACV is a complex virus belonging to the genus Orthopoxvirus (OPV) within the family Poxviridae. The genus OPV not only contains one of the most dangerous biothreat-relevant viruses, variola virus (VARV) [13], but its members cowpox virus (CPXV) and monkeypox virus (MPXV) are causing clinically relevant zoonotic infections today [14]. The biopanning against VACV resulted in the identification of a VACV-binding peptide, termed αVACVpep05. We further identified the VACV surface protein D8 as its viral target and determined the binding kinetics and affinity between the peptide αVACVpep05 and the D8 protein. Finally, we demonstrated that peptide αVACVpep05 can also detect other OPV members.
L. Miller et al.
bicarbonate buffer (Sigma), and lysed with IP buffer (Invitrogen) containing 100 mM NaCl and protease inhibitor (Thermo Scientific). Recombinant poxvirus proteins A27, A33, B5, F9, and L1 were from BEI Resources (Manassas, VA, USA). Recombinant proteins D8 (AA 2–260), F13 (full length), and H3 (AA 21–270) (all expressed in E. coli) were purchased from GenExpress (Berlin, Germany). Phage biopanning Affinity selections were performed using the Ph.D.-12™ Phage Display Peptide Library Kit as recommended by the manufacturer (New England Biolabs, NEB, complexity of the library on the order of 109 independent clones) with minor changes. MaxiSorp ELISA plate wells (Nunc) were coated with 5×106 pfu/well of VACV. One additional well was filled with blocking buffer for negative selection performed before every biopanning round. The Ph.D.-12 library was diluted to 2×1011 pfu/well (dilution buffer: PBS with 2 % [w/v] BSA and 0.5 % [v/v] Tween-20) and first added to BSA-coated wells for 2 h at room temperature (RT). The supernatant was transferred to target-coated wells and incubated for 2 h. Wells were washed 15 times after the first selection round and 20 times in all subsequent rounds with 300 μL/well of PBS with 0.5 % [v/v] Tween-20. Bound clones were eluted with 100 μL of 0.2 M glycine-HCl, pH 2.2, by incubating for 15 min at RT. The eluate was neutralized with 15 μL of 1 M Tris–HCl, pH 9.1, and then incubated in a water bath at 60 °C for 1 h to inactivate infectious poxvirus particles potentially present. Four selection rounds were performed altogether. After the fourth round, binding phages were eluted and used for a further negative selection against HEp-2 cell lysate. Sequence analysis of binding phage clones
Materials and methods Viral strains and other reagents VACV NYCBOH (VR-1536™) and porcine parvovirus (PPV; VR-742™) were purchased from ATCC®. The adenovirus (AdV) [15], CPXV strain GuWi [16], and MPXV (in press) were taken from RKI strain collection. Camelpox virus (CMLV CP-19), ectromelia virus (MP-Nü), and MPXV strain MSF#6-Mü were kindly provided by Hermann Meyer (Bundeswehr Institute of Microbiology, Munich). For preparation of HEp-2 cell lysate (used for negative selections), cells were lysed with RIPA buffer according to the manufacturer’s instructions (Thermo Scientific). For preparation of VACV lysate, VACV particles (propagated on HEp2 cells) were semi-purified over a sucrose cushion as described before [17], washed with 100 mM ammonium
For the identification of binding peptides, 10, 20, 20, and 68 phage clones of the first, second, third, and fourth eluate, respectively, were plugged and amplified. Single-stranded DNA was purified from phages and sequenced by the Sanger method using a “BigDye Terminator v3.1 Cycle Sequencing Kit” as recommended (Applied Biosystems). Nucleotide sequences were translated into amino acid sequences using “Library Insert Finder” software. Repeatedly identified sequences were defined as potential consensus peptide sequences, and the corresponding phage clones were characterized further. Phage ELISA Phage ELISA was performed as described (NEB) with minor changes. ELISA plates were coated with 105 pfu/well of VACV or PPV. Selected phage clones were added at
A phage display-derived peptide for orthopoxvirus detection
1010 pfu/well and incubated for 2 h at RT. Monoclonal phages were also added to uncoated, blocked wells to test for unspecific binders. Phage binding was detected by an anti-M13 antibody conjugated to horse radish peroxidase (HRP). A 3,3′,5,5′-tetramethylbenzidine (TMB, Sigma) solution was used as HRP substrate. Plates were assayed at 450/620 nm using an Infinite®200 PRO microplate reader (Tecan Group).
Peptide selection for synthesis Selected sequences were synthesized by PANATecs as phagefree peptides. Selection of suitable peptide sequences was based on DNA sequencing (most frequent) and phage ELISA results (specific target binding). To obtain similar properties to phage-attached peptides, synthetic peptides were ordered with the following features: (1) N-terminus acetylated to block positive charge, (2) C-terminus containing a linker sequence (Gly-Gly-Gly-Ser), followed by a lysine residue and a biotin molecule, and (3) peptide purity of >85 % (biochemistry grade). To rule out a lot-to-lot variability, the most specific peptide was reordered at PANATecs and at Biomatik, respectively.
Peptide-probe-based ELISA Synthetic peptides were tested using ELISA. To this end, plates were coated with 5×105 pfu/well of VACV, PPV, or AdV as specificity controls. Plates were washed twice with 300 μL/well of washing buffer (PBS with 0.05 % [v/v] Tween20), then completely filled with blocking buffer (PBS with 3 % [w/v] BSA) and incubated for 2 h at RT. Synthetic peptides were twofold serially diluted, added at 100 μL/well, and incubated for 2 h at RT. Subsequently, plates were washed six times, and 100 μL/well of streptavidin peroxidase (SAPOD) (Dianova) were added for 1 h at RT. Wells were washed six times, a TMB substrate solution was added, and the color reaction was detected. For selectivity analysis of the synthetic peptides towards different OPV species, experiments were performed as described above with minor modifications. Wells coated with parapoxvirus (PPXV), HEp-2 cell lysate, and BSA were utilized as controls. Peptides were added at 400 nM and incubated for 2 h at RT. To assess functionality of peptides as capture molecules for OPV, the biotinylated peptides (100 μL of 400 nM solution) were immobilized on neutravidin-coated plates (Pierce) for 2 h at RT. Subsequently, VACV, CMLV, CPXV, ECTV, PPXV, and MPXV were added to individual wells and incubated for 1 h at RT. After washing, poxvirus particles were detected by a polyclonal rabbit anti-VACV antibody (Acris Antibodies) and an HRP-labeled anti-rabbit serum (Dianova).
7613
Pull down assay, SDS-PAGE, and nLC-MS/MS analysis A pull down assay with subsequent nano-scale liquid chromatographic tandem mass spectrometry (nLC-MS/ MS) analysis using peptide-coated magnetic beads and VACV was employed to identify the interaction partner of the VACV-binding phage-free synthetic peptide designated as αVACVpep05. For pull down assay, magnetic SA-coupled Dynabeads® (Invitrogen Dynal) were incubated with 600 nM peptide for 30 min at RT. Beads were then washed five times, incubated with 100 μL of VACV lysate [17] for 1 h at RT, and again washed six times. These washing fractions 1–6 (500 μL) were collected and their volume reduced with a Speedvac to the same volume as the final eluate (50 μL). For elution of bound protein, beads were first incubated with 50 μL of RIPA buffer for 15 min on ice, followed by a 5-min incubation step at 95 °C. For SDS-PAGE, virus lysate, collected washing fractions 1, 4, 5, and 6, and the eluate were each mixed with 6× Laemmli buffer (Bio-Rad) and incubated for 5 min at 95 °C. Subsequently, 50 μL of each sample were run on an 8–16 % SDS gel (Thermo Scientific). The washed gel was stained overnight with the Flamingo™ fluorescent gel stain (Bio-Rad) and the images acquired using the Image Lab software (v4.0.1). For mass spectrometric analysis, detected protein bands were excised from gel and prepared for analysis as described elsewhere [18]. Peptides were separated with Easy-nanoLC (Thermo Scientific) on a 10 cm, ID 75 μm, 3 μm, C18 column using a linear 30 min gradient of 2–40 % acetonitrile in 0.1 % formic acid at 300 nL/ min. The LC was coupled online to an LTQ Orbitrap Discovery™ mass spectrometer (Thermo Scientific), which was operated in a data-dependent manner in the m/z range of 300–1,700 by selecting the Top7 2+ and 3+ charged ions for CID type fragmentation. Proteins were identified using SEQUEST algorithm via Proteome Discoverer 1.3 (Thermo Scientific). Therefore, a custommade database was generated from all VACV entries of the UniProt Knowledge database. The quality thresholds for peptide identification were XCorr values of ≥2 (2+ ions), ≥2.5 (3+ ions) and a maximum deltaCn of 0.05. The false discovery rate was validated on q values and set to be below 1 %. Surface plasmon resonance measurements For surface plasmon resonance (SPR) analysis, an SAcoated sensor chip was utilized in a Biacore X 100 SPR unit (GE Healthcare). To prepare the SA chip for binding, 1 M NaCl in 50 mM NaOH was injected three times. Biotinylated peptides diluted in running buffer (HBS-EP+
7614
L. Miller et al.
Table 1 Consensus peptide sequences of potential VACV-binding phage clones Phage Amino acid sequencea clone no.
Frequencyb
1 2 3 4 5 6 7 8 9 10 11 12 13
KVWLPPRHEHQY K V F Y PA A A N P N Q QALLEGNAKGGN NRPDSAQFWLHH TA D K L LY G L F K S DEWDALLMRIRT NPTPYPMLPLRG K P T Y S W D PA Q L K GPTFSWDHLRGQ KIFQLPQISPPM GDLASWIITSFK NMELHPHSLPRP ANTTKHSVLAAI
2nd eluate 2 1 – – – – – – – – – – –
14 15 16 17
EALNDWVNDSEY A P TAY N K N D WA L DPWWRGNEARAA TPVWSWEPPLQE
– – – –
3rd eluate 1 1 1 3 1 1 – – – – – – –
4th eluate 3 – 2 – 6 3 8 2 2 4 8 2 2
– – – –
2 2 2 4
a
Sequences are provided using the one-letter amino acid code from Nterminus to C-terminus
b Frequency of how often the peptide sequence was found in the respective phage eluate
, 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.05 % Tween 20) were immobilized at surface densities of 269 RU αVACVpep01 (non-binding peptide) to flow cell 1 and 67 RU αVACVpep05 to flow cell 2.
Fig. 1 Estimation of binding specificities of selected phage clones containing consensus peptide sequences by phage ELISA. To test the binding specificity of the 17 frequently found phage clones, MaxiSorp ELISA plates were coated with 105 pfu/well of VACV or PPV particles. Wells filled with blocking buffer (BSA) were utilized to test for unspecific binders. Selected phage clones (1–17) were diluted and added at 1010 pfu/well. Additionally, the wild-type M13KO7 phage (WT) and the
To find the binding partner of αVACVpep05, 50 μg/ mL of various recombinant VACV surface proteins were injected over the sensor chip for 120 s at 5 μL/min. To remove bound proteins from immobilized peptides between measurement cycles, flow cells were regenerated by injection with 10 mM glycine-HCl pH 1.7 (GE Healthcare) for 30 s at 10 μL/min. After identification of D8 (MW 30 kDa) as binding partner, affinity to αVACVpep05 was determined by application of twofold serially diluted (3.25 nM– 1.67 μM; duplicate injection of 1.67 μM) recombinant D8 protein to both flow cells of the SA chip with a flow rate of 30 μL/min. Association measurement was performed for 120 s, and dissociation was assessed for 60 s. All measurements were performed at 25 °C. For analysis, double-referenced [19] response signals were fit with BIAEvaluation Software 1.0 to either a 1:1 Langmuir binding model to determine binding kinetics or a Steady-State-Affinity model to determine binding affinity.
Western Blot For Western Blot analysis, 500 ng of recombinant proteins A27, A33, B5, D8 (native and heat-treated), H3, F9, F13, and L1 and 3 μg of a VACV lysate (native and heat-treated) were separated on a 4–20 % premade SDS gel (Thermo Scientific) and transferred to a nitrocellulose membrane (Thermo Scientific). The membrane was then incubated with 500 ng/mL of the biotinylated αVACVpep05. After washing, proteins were detected by incubation with SA-POD (1:50,000) and subsequent substrate reaction.
original Ph.D.-12 library (Lib) were tested as controls. The phage binding was detected with a monoclonal anti-M13 antibody. All samples were measured in duplicate once; error bars are provided as standard deviations. Arrows point at the five phage clones that were selected for peptide synthesis, according to the signal intensity and the difference in signal between VACV and the control virus PPV
A phage display-derived peptide for orthopoxvirus detection
7615
Table 2 Criteria for the selection of phage clones with peptides to be synthesized Phage clone no.
ODVACV-ODPPVa
ODVACV*Clone-ODVACV*Libb
1 2 3 4
0.50 0.01 0.39 0.24
0.00 −0.52 −0.02 −0.01
5 6 7 8 9 10 11 12 13 14 15 16 17
0.79 0.15 0.10 0.48 0.24 0.08 0.28 0.09 0.38 0.51 0.08 0.18 0.39
1.06 0.46 −0.40 0.06 −0.23 0.19 0.94 −0.30 0.26 0.53 0.03 −0.12 0.99
colloidal gold particles and negatively stained as previously described [20].
Peptide-based fluorescence assay For visualization of the specific interaction of peptide αVACVpep05 with OPV-infected cells, CPXV-infected HEp-2 cells fixed onto glass slides were incubated with peptide αVACVpep05 (50 μg/mL) preincubated with streptavidin-fluorescein isothiocyanate (SA-FITC, 1:80, Invitrogen) for 1 h at 37 °C. Cells were additionally incubated with a polyclonal anti-D8 antibody coupled to DyLight 649 (RKI) to assess colocalization. The slides were analyzed using a confocal laser scanning microscope (Zeiss).
Results
a
Differences in optical density (OD) values between VACV signal (ODVACV) and PPV signal (ODPPV) as determined with the respective phage clone using phage ELISA b Differences in the OD values between the VACV signal of the individual phage clone (ODVACV*Clone) and the VACV signal of the native phage display library (ODVACV*Lib) as determined by phage ELISA
Electron microscopy For visualization of specific binding of the αVACVpep05 to OPV particles by electron microscopy, semi-purified [17] ectromelia virus (ECTV) particles were fixed with 2 % paraformaldehyde in 0.05 M HEPES buffer and adsorbed onto a sample support (400 mesh copper grid, coated with a plastic film). Then, samples were incubated with biotinylated peptides, followed by an anti-biotin antibody coupled to 10 nm
Biopanning against native VACV particles resulted in the identification of multiple consensus peptide sequences To isolate binders to VACV, a commercially available filamentous phage display library (∼109 independent clones) was used. Four rounds of biopanning were carried out against infectious VACV particles immobilized on wells of a 96well plate. In total, 118 plaques were plugged and the ssDNA prepared from the phages and subsequently sequenced. Among the peptide sequences identified, 17 sequences were present two to eight times and were thus defined as consensus peptide sequences (Table 1). The presence of such consensus sequences after multiple selection rounds indicates an enrichment of specifically binding phage clones. Nevertheless, to decide which of the identified phage clones encodes the most suitable peptide for virus detection, the binding specificities of the selected phage clones were estimated.
Table 3 Potential VACV-binding peptides selected for synthesis Phage clone no.
Peptide designation
Amino acid sequencea
Frequencyb
1 5 8 13 14
αVACVpep01 αVACVpep05 αVACVpep08 αVACVpep13 αVACVpep14
Ac-KVWLPPRHEHQYGGGSK-NH2 (Biotin)c Ac-TADKLLYGLFKSGGGSK-NH2 (Biotin) Ac-KPTYSWDPAQLKGGGSK-NH2 (Biotin) Ac-ANTTKHSVLAAIGGGSK-NH2 (Biotin) Ac-EALNDWVNDSEYGGGSK-NH2 (Biotin)
6 7 2 2 2
a
Sequences are provided using the one-letter amino acid code from N-terminus to C-terminus
b
Frequency of a peptide sequence found in all phage eluates
c
The abbreviations Ac and NH2 refer to the presence of an N-terminal acetyl group and a C-terminal amide; all peptides were synthesized with a biotin at the C-terminus
7616
Estimation of binding specificities of phage clones enriched during biopanning using a phage ELISA
L. Miller et al.
clones which fulfill the two criteria (i) ODVACV −ODPPV ≥0.3 and (ii) ODVACV*Clone −ODVACV*Lib ≥0 (Table 2) and to subsequently test the binding characteristics of the corresponding synthetic peptides. According to this, five peptide sequences derived from phage clones number 1, 5, 8, 13, and 14 were selected for synthesis (Table 3).
Seventeen monoclonal phage clones were assayed by phage ELISA to estimate their binding to the target. Of note, the estimation of the binding specificities of selected phage clones has been frequently recommended for exclusion of biopanning process artifacts. However, no defined criteria for the choice of the “right” phage clone based on the obtained phage ELISA results have been defined so far. Exactly those criteria could help to select a phage clone bearing a peptide with the best binding characteristics. Most frequently, binding characteristics of the selected phage clone are tested against target- and BSA-coated wells in a phage ELISA. More stringent criteria, such as testing the binding against an alternative target, have to our knowledge, not been applied so far. To be able to subsequently define those criteria, we decided to additionally (i) include wells coated with PPVas an unrelated virus, and (ii) to compare the binding characteristics of enriched phage clones with those of the native phage display library. Nearly all individual phage clones bound better to VACV than to the control virus (PPV), with ELISA signals ranging from 0.4 to 2.0 optical density (OD) for VACV-coated wells, and 0.2 to 1.6 OD for PPV-coated wells (Fig. 1). The addition of phage clones to uncoated, BSA-blocked wells resulted in low signals of
