|Home | About | Journals | Submit | Contact Us | Français|
The use of phage display peptide libraries allows rapid isolation of peptide ligands for any target selector molecule. However, due to differences in peptide expression and the heterogeneity of the phage preparations, there is no easy way to compare the binding properties of the selected clones, which operates as a major “bottleneck” of the technology. Here, we present the development of a new type of library that allows rapid comparison of the relative affinity of the selected peptides in a high-throughput screening format. As a model system, a phage display peptide library constructed on a phagemid vector that contains the bacterial alkaline phosphatase gene (BAP) was selected with an antiherbicide antibody. Due to the intrinsic switching capacity of the library, the selected peptides were transferred “en masse” from the phage coat protein to BAP. This was coupled to an optimized affinity ELISA where normalized amounts of the peptide–BAP fusion allow direct comparison of the binding properties of hundreds of peptide ligands. The system was validated by plasmon surface resonance experiments using synthetic peptides, showing that the method discriminates among the affinities of the peptides within 3 orders of magnitude. In addition, the peptide–BAP protein can find direct application as a tracer reagent.
The development of phage-displayed peptide libraries (1) allows millions of peptides to be screened, simultaneously, for binding to a target of interest. In these libraries, any one bacteriophage displays a distinctive peptide sequence typically fused to either the pIII or pVIII viral coat protein. The peptide sequence can be encoded in the phage DNA, or most commonly in a phagemid system, consisting of a vector that codes for the peptide–viral coat protein fusion, and utilizes a “helper phage” to allow recombinant phage production. Since even a single peptide–target binding event can be detected and amplified, the technique has the potential to discover peptides that specifically bind to almost any target. For this reason, the technology has been applied in many fields (2–5).
Typically, the outcome of the selection procedure (panning) is analyzed by the so-called phage ELISA, but this test does not provide any information other than a negative/positive binding result. Further evaluation of positive clones requires DNA sequencing of the encoded peptide and the preparation of the encoded sequences as synthetic peptides, which must then be individually analyzed by competitive ELISA or surface plasmon resonance experiments in order to identify the peptide sequences that bind with the highest affinities. For most purposes, not only the specificity, but very importantly, the affinity of the selected peptides is a major issue. This approach is a time-consuming and costly step that constitutes a major bottleneck for the technique, because it restricts the number of clones that can be assayed, and thereby contributes to the loss of low-abundancy but valuable clones.
In order to overcome these limitations, several works have been published in which model peptides selected from a phage library are transferred by different strategies to a partner molecule (i.e., maltose binding protein, bacterial alkaline phosphatase, etc.) to analyze the reactivity of the peptide out of the context of the phage particle (6–9). Due to its enzymatic activity and good expression in bacteria, BAP has been one of the most studied scaffolds. In this regard, Han et al. (9) devised a convenient method for the transference of the selected peptides to a highly active form of BAP using a ligation-independent cloning protocol, but the information provided by their procedure was only in terms of positive or negative binding. Other authors worked to obtain additional information on the peptide reactivity, particularly with regard to their relative affinity for the selector molecule (6, 8), but their approaches required one by one processing of the peptide–BAP candidates. Indeed, these methods require the purification of the peptide–BAP protein for each individual clone, followed by adjustment of the BAP activity of each clone in order to assay comparative amounts of individual peptides, which greatly restricts the number of clones that can be examined.
To date, no practical procedure for high-throughput assessment of the relative affinities of the peptides selected from phage display libraries has been reported. Here, we present a novel method devised to solve the critical steps that limit this analysis. For this, we constructed a new type of phage display library with the intrinsic capacity of being easily “switched” into peptide–mBAP expression libraries. In this way, after initial selection of the phage library (usually three to four rounds of panning) the pool of selected phagemids is digested to remove the gene encoding the viral coat protein pIII, which allows fusion of the encoded peptides to the N-terminus of a highly active mutant bacterial alkaline phosphatase (mBAP). This switching process is coupled to a fast normalized selection procedure that does not require purification of the recombinant peptide–BAP fusions, and allows simultaneous direct investigation of hundreds of peptide candidates in terms of relative affinities and binding characteristics to the selector molecule. By this means, even if they are underrepresented, the most valuable peptide ligands can be rapidly identified. The method was developed and validated using a monoclonal antibody against the herbicide molinate as a model-selector molecule. In addition to providing information about the relative binding strengths of the peptides, the method yields ligand–enzyme recombinant proteins that can be used as tracer bioconjugates in different applications.
Vector pAFF2 (10) and E. coli cell strains ARI292 and ARI236 were kindly provided by Dr Peter Schatz; Affymax Research Institute, Palo Alto, CA. Vector pUY10 was a gift from Dr. Magela Laviña, Facultad de Ciencias, Montevideo, Uruguay. E. coli strain LMG194 (11), a strain derived from E. coli KS272 (12), a negative mutant for alkaline phosphatase, was a gift from Dr. Carina Gaggero, IIBCE, Montevideo, Uruguay.
Plasmid pUY10 encoding wild type BAP was used as template for PCR amplification using primers (5′-TCAGTCGACACTAGTGCTGCTCAGGGCGATAT TACT-3′) and (5′-CAGAAGCTTTTATTAGTGATGGTGATGGTGATGTTTCAGCC CCAGAGCGGCTTTCATGGTGT- 3′). SalI and HindIII restriction sites (italics) were introduced for cloning, as well as a SpeI site (underlined) at the 5′ end of the BAP gene to allow later elimination of pIII phage coat protein gene. After PCR amplification using pfu DNA polymerase (Stratagene), the BAP product was digested with SalI and HindIII (New England Biolabs), and cloned into the arabinose regulatable phagemid vector pAFF2. Quickchange Site-Directed Mutagenesis Kit (Stratagene) was used both for removing a BstXI restriction site within BAP, as well as for introduction of a Asp101→Ser mutation in the BAP gene cloned in pAFF2. This amino acid change has been reported to possess enhanced BAP catalytic activity (mBAP) (13). The obtained phagemid was named pAFF/mBAP (Figure 2).
pAFF/mBAP vector was digested with BstXI restriction enzyme (New England Biolabs). The digested vector was mixed with two “half sites” adaptor oligonucleotides, ON-28 and ON-29, and the long DNA oligo coding for the cyclic peptide CAREGYPARYC, which binds to EB7 mAb (14) followed by a SpeI site (Figure 2). The three oligonucleotides and the vector were annealed and later ligated with T4 DNA ligase (Invitrogen) using standard DNA procedures. ARI 236 competent cells were transformed with the ligation mix. Transformed cells were plated on LB agar ampicillin plates. DNA sequencing of clones confirmed the correct cloning of the model peptide fused to pIII phage coat protein. One of these clones was grown and a miniprep DNA extraction (Qiagen) was performed. Phagemid DNA was then digested with SpeI (New England Biolabs) restriction enzyme, removing pIII gene. Digestion was run on an agarose gel, and the band corresponding to the phagemid without the pIII gene was purified using QIAGEN Quiaquick gel extraction kit and ligated as described above. ARI 236 competent cells were transformed with the ligation mix. A direct screening for successful fusion of peptides–mBAP was performed by plating the transformed cells in LB plates with ampicillin (100 μg/mL), 0.2% arabinose, and BCIP (5-bromo, 4-chloro, 3-indolylphosphate (Pierce) at 40 μg/mL). After overnight incubation of the plates at 37 °C, bacterial colonies expressing peptide–BAP fusions could be clearly distinguished from colonies that were BAP by simply checking for blue colonies.
ARI 236 cells expressing the model fusion CAREGY-PARYC-mBAP were cultured for 6 h at 37 °C after inducing with 0.2% arabinose. Cells were then pelleted by centrifugation (in aliquots of 1 mL) and the pellets were frozen at −80 °C. The cell pellet was lysed with 300 μL of B-PER reagent (Pierce) plus 1/25 volume of Complete EDTA-FREE Protease Inhibitor Cocktail (Roche Diagnostics, GmbH). Cell lysates were centrifuged and the supernatants were transferred to low-binding, nontreated, polystyrene plates (100 μL lysate + 100 uL AP wash buffer (50 mM Tris–Hcl pH = 8; 150 mM NaCl; 0.05% Tween-20)) and serial dilutions (twofold) were then performed in AP wash buffer. Dilutions were transferred to a 96-well microtiter plate (Nunc-Immuno Plate Maxi-Sorp) precoated with anti-BAP monoclonal antibody (mAb) (Sigma) at 1.25 μg/mL of IgG1 in 1 × PBS (phosphate-buffered saline), polyclonal anti-BAP antibody (Polysciences, Inc.) at 2.5 μg/mL of total IgG in PBS 1×, both of them preblocked with BSA 1%. These serial dilutions were also transferred to a microtiter plate coated with Ni-NTA (Qiagen) and blocked with 1% bovine serum albumin (Sigma) in PBS buffer. Wells were washed 5 times with AP wash buffer and mBAP activity was detected by adding PNPP (p-nitrophenyl phosphate at 1 mg/mL) substrate (Pierce) in 50 mM Tris pH = 10, 200 mM NaCl, 1 mM MgCl2. Plates were read at 405 nm in a microtiter plate reader (Multiskan MS, Labsystems).
A peptide phage library was constructed on pAFF/mBAP vector and harvested as described by Cwirla et al. (15). Library diversity was estimated to be around 2 × 108 independent clones. This library was panned against an anti-molinate (herbicide) (16) mAb 14D7. Microtiter plates (Nunc-Immuno Plate Maxi-Sorp) were coated with mAb 14D7 at 5 μg/mL in PBS by incubating overnight at 4 °C (100 μL per well). Blocking was performed by completely filling the wells with 1% BSA in PBS and incubating at 37 °C for 1 h. After blocking, each well was washed 5 times with PBS–0.5% Tween-20 (PBST) and peptide library in PBS 1 × –BSA1% was added to 6 precoated wells (approximately a total of 1011 phage particles). The plate was incubated for 2 h at 4 °C; wells were washed 10 times with cold PBST to remove unbound phages. Bound phages were eluted by adding 100 μL per well of elution buffer (0.1 N glycine, pH = 2.2 adjusted with HCl) and incubating at room temperature for 10 min. Neutralization of pH was done by adding 35 μL of 2 M Tris (pH unadjusted). The eluted phage (300 μL) were added to 10 mL of log-phase E. coli ARI 292 cells and amplified in LB media containing 0.25% K2HPO4, 0.1% MgSO4, 0.1% glucose, and 100 μg/mL ampicillin to an OD600 = 0.4. Helper phage M13KO7 at a multiplicity of infection 10:1 was added. After a period of 30 min at 37 °C without shaking, arabinose and kanamycin were added at a final concentration of 0.02% and 40 μg/mL, respectively, and the cultures incubated overnight at 37 °C with vigorous shaking. This panning protocol was then repeated twice, and after three rounds of panning, individual amplified phage clones were tested for their ability to bind 14D7 coated wells. Phage ELISAs were performed as reported before (5).
ARI 292 cells were infected with approximately 5 × 107 phage particles from the third output of panning and grown in LB–ampicillin as described. Phagemid DNA was obtained using the Qiagen miniprep kit (Qiagen). The encoded peptides were fused to mBAP by switching with SpeI as described above. Competent LMG194 was transformed with the ligation mix and plated as described above. Blue colonies were individually inoculated in 0.5 mL of LB–ampicillin with 0.1% glucose in wells of a 2 mL 96-well block (Qiagen). The plate was sealed with airpore tapesheets (Qiagen) and grown with shaking at 37 °C. The next day, 80 μL of glycerol were added, and the block was frozen at −80 °C (master plate).
Replicates of this plate were established by touching the surface of the master plate frozen cultures with a multichannel micropipette using sterile tips. The tips were then used to inoculate 1 mL cultures of LB–ampicillin, 0.1% glucose, that were grown at 37 °C in a 96-well block. The next day, 5 μL was used to inoculate wells of a fresh block containing 1 mL of LB–ampicillin and 0.2% arabinose. The block plate was incubated for 6 h at 37 °C with vigorous shaking and was then centrifuged for 30 min at 1200 × g at 4 °C. The supernatant was discarded, and the pelleted cells were frozen at −80 °C. Each cell pellet was lysed with 300 μl of B-PER reagent (Pierce) plus 1/25 volume of Complete EDTA-FREE Protease Inhibitor Cocktail (Roche Diagnostics, GmbH) by gently pipetting up and down and using a multichannel micropipette. Cell lysates were centrifuged as previously, and 100 μL of each clone supernatant were transferred with a multichannel micropipette to each of two ELISA plates precoated with polyclonal anti-BAP antibody, and preblocked with 1% BSA. After 1 h incubation, one of the plates was washed 10 times with AP wash buffer, followed by addition of 100 μL of PNPP in substrate buffer (50 mM Tris, 200 mM NaCl, 1 mM MgCl2, pH 10), and the reaction was incubated at 37 °C. The plate was read at 405 nm in a microtiter plate reader (Multiskan MS, Labsystems). The second plate was washed with PBS, 0.05% Tween 20 (PBST), and probed with 100 μL of biotinylated mAb 14D7 (0.6 μg/mL) in PBST for 1 h at room temperature. After washing, streptavidin–peroxidase (Roche Diagnostics) conjugate was added (1:1000 dilution) and incubation was continued for 30 min at room temperature. The plate was washed with PBST, and 100 μL of the peroxidase substrate (0.4 mL of a 6 mg/mL DMSO solution of 3,3′,5,5′-tetramethylbenzidine (TMB) and 0.1 mL of 1% H202 in water in a total of 25 mL of 0.1 M citrate acetate buffer, pH 5.5) was dispensed into each well. The enzymatic reaction was stopped by the addition of 50 μL of 2 M H2SO4, and the absorbance was read at 450 nm. Three replicates were prepared in this way to perform three independent experiments for plates A and B. Values obtained for clone number 1 were used as an internal control for normalization of the ELISA values.
Interaction analyses were performed using a Biacore 3000 optical biosensor equipped with research-grade CM5 sensor chips (Biacore AB, Uppsala, Sweden). Data were collected at the highest collection rate, with the biosensor instrument thermostatted to 25 °C. HBS-P (10 mM Hepes, 0.15 M NaCl, 0.005% P20 surfactant, pH 7.4) running buffer, amine coupling reagents (N-ethyl-(3- dimethylaminopropyl) carbodiimide (EDC), N-hydroxysuccinimide (NHS), and ethanolamine HCl) were purchased from Biacore AB. mAb14D7 was immobilized on a CM5 sensor chip using standard amine-coupling chemistry. The carboxymethyl dextran surface was activated with a 7 min injection of a 1:1 ratio of 0.4 M EDC/0.1 M NHS. The antibody was coupled to the surface with a 7 min injection of 14D7 mAb diluted in 10 mM sodium acetate (pH = 5.0). Remaining activated groups were blocked with a 7 min injection of 1 M ethanolamine (pH = 8.5). The immobilization procedure was performed at 25 °C and a 14D7 surface density of approximately 12 000 RU was obtained. Peptides were assayed on the sensor chip in the 20 nM to 100 μM range.
Selected clones expressing different fusions were grown expressing the recombinant protein. Recombinant proteins were purified by standard Ni-NTA agarose (Qiagen) following manufacturers’ instructions. An aliquot of the purified proteins was run on a 12% SDS-PAGE and stained with Coomasie blue.
Nunc Maxisorp ELISA plates were coated with 0.6 ug/mL of 14D7 mAb and blocked as described above. After that, a fixed amount of peptide–mBAP previously assessed by checkerboard titration was mixed with different amounts of molinate and transferred to the 14D7 mAb coated wells, to allow competition for binding to the antibody. After 1 h incubation, plates were washed with AP Wash buffer and PNPP substrate was added. Plates were read as described above.
The basic principles underlying the use of phage/mBAP switch libraries for the selection and evaluation of peptides are schematized in Figure 1. As shown in the figure, the system relies on two critical steps: first, it must allow massive transfer of the individual peptides selected on the phages to create a large peptide–mBAP library. Second, the peptide–mBAP expression library must be easily examined for the desired binding properties of the peptides. The developments required to accomplish these two aims are described in the following sections.
Vector pAFF/mBAP (Figure 2), which contains a truncated version of the mutated E. coli BAP, was constructed as described in the Methods section. A shortened version of BAP (lacking residues 1–10) was used because it has been reported to be more resistant to proteolysis (17) with no loss of catalytic activity (18). A 6 × His tag was appended to its C-terminus to facilitate the purification of the recombinant protein. The vector contains two noncomplementary BstXI sites that are used to insert the random oligonucleotide sequences encoding the peptides. Oligonucleotide cloning introduces an additional SpeI site that allows later switching of peptide fusions from pIII to BAP.
The system was initially explored using the model cyclic peptide CAREGYPARYC, which binds to mAb EB7 (14). The DNA sequence encoding for the peptide was cloned in frame with pIII in the pAFF-mBAP vector. Later, this construct was digested with SpeI (Figure 2), then ligated and used for transformation of E. coli ARI236. Peptides correctly fused to mBAP could be identified as blue colonies on LB ampicillin agar plates after addition of arabinose and the chromogenic substrate for alkaline phosphatase, BCIP (Figure 1d). Although the cells used for transformation (ARI236) have the wild-type alkaline phosphatase gene in their chromosome, streaking of this strain in control plates produced white colonies suggesting that endogenous expression of alkaline phosphatase is negligible. Fusion of peptides to mBAP through SpeI switching proved to be highly efficient, as nearly 99% of the colonies were positive by the BCIP assay.
The expression of the peptide–mBAP fusion was studied using cell cultures of ARI 236 expressing the CAREGYPARYC–mBAP protein. After 6 h induction with arabinose, the expression of the recombinant peptide–BAP protein was analyzed by Western blot. The immunoreactivity of the peptide was easily detectable with mAb EB7 (14) and comparable to that of BAP with the anti-BAP antibody (Supporting Information, Figure A). Since BAP spontaneously forms dimers by self-association, two copies of the peptide are expected to be displayed when the enzyme is used as scaffold. Further analysis of the CAREGYPARYC–mBAP expression was tested by growing the clone in 1 mL cultures in 96-well culture blocks (supplementary methods). The cell pellet was lysed with B-PER reagent, and the supernatant was incubated into microtiter wells precoated with different antibodies or Ni-NTA. Saturation was directly examined by measuring BAP activity (Supporting Information, Figure B). The best capture efficiency was obtained with polyclonal anti-BAP antibodies. The supernatant could be diluted over 100 times without significant change in signal, showing that the amount of recombinant CAREGYPARYC–mBAP obtained from 1 mL culture was in large excess of the amount required to saturate the wells. This result proved to be valid for other peptide sequences as shown in the following section. The endogenous alkaline phosphatase activity, although detectable, was negligible and about 500-fold less than that of the recombinant protein. Nevertheless, to rule out any possible interference of this activity, we adopted the strain LMG194 for all of the following expression experiments. This strain is mutant for alkaline phosphatase, and also for the catabolic arabinose genes, and hence induction of gene expression can be done without arabinose being consumed by E. coli.
A library coding for 8-mer cyclic peptides with an estimated diversity of 2 × 108 independent clones was constructed on phagemid pAFF-mBAP. As model selector molecule we used mAb 14D7 (Figure 1a), which is specific for the herbicide molinate, and is the basis for a competitive assay for this herbicide (16). Since competitive elution with the analyte would not warrant elution of the highest-affinity clones, we used acid conditions in order to ensure that all low- and high-affinity clones were selected. After three rounds of panning, individual phage clones were tested for binding to the antibody by ELISA on plates coated with either mAb 14D7 or BSA (negative control). Ten out of ten clones examined were specific for the antibody, and six different sequences were identified (Table 1). Thus, at this point of the selection process most clones were specific and diverse, which was very convenient for the purpose of developing a method for ranking the affinities of a high number of clones. DNA from this pool of phage was “switched” with SpeI as described in the methods section (Figure 1b,c) and electroporated into LMG194. Blue colonies were picked from an arabinose-BCIP plate (Figure 1d) and were used to inoculate the “master plate” described in the Methods section.
As mentioned before, the intrinsic affinity of the peptides for the selector molecule can not be estimated by phage ELISA (6). To overcome this limitation, we devised a method that rapidly normalizes the amount of peptide ligand to be tested by capturing saturating amounts of peptide–mBAP fusion on each of the microtiter wells to be analyzed. The method was applied to screen the third output of the panning experiment. Individual clones expressing peptide–mBAP fusions were picked from the “master plate”, grown, and later lysed in 96-well square culture plates as described in the Methods section. Cell lysates were transferred to two identical plates coated with polyclonal anti-BAP antibody. One of the plates was used to verify that the same molar amount of each peptide–mBAP fusion (same BAP activity under conditions of initial velocity) was retained in each well, Figure 3. This was valid for all but two of the clones examined, clones 74 and 84. Once it was verified that the same amount of the peptides were retained in each well, mAb 14D7 (0.6 μg/mL) was used to examine the reactivity of the peptides. Under these conditions, the signal of each well is expected to be proportional to the affinity of the peptide for the antibody (Figure 1f). On the basis of the ELISA signal, the clones were grouped in five categories (Figure 4). Excellent reproducibility was attained for individual clones analyzed in three independent experiments. Sequence analysis of five clones from each category revealed peptide sequences characteristic of each group, supporting the assumption that the method sorts the peptides according to their reactivity with the antibody. Notice that, in none of the cases, was a peptide sequence belonging to a particular affinity group found in another group. To confirm this, four representative peptides belonging to each of the different affinity groups were synthesized (including disulfide formation). Binding of the soluble peptides to mAb 14D7 was studied using surface plasmon resonance. The results confirmed that the method properly sorts the peptides in a range of up to 3 orders of magnitude as shown in Table 2. Peptide CNNPW-LAKYC, which was the most abundant after selection, exhibited immeasurable interaction on the Biacore chip, showing that the avidity effect due to the two peptide copies expressed in the peptide–mBAP protein may be the cause of the moderate binding observed in the affinity ELISA. In spite of its low affinity for the antibody, this peptide was most probably preferentially selected due to a favorable expression in E. coli, which resulted in an over-representation of this phage clone in the amplified phage stock. This may also promote a multivalent display of the peptide in the phage particle reinforcing the avidity of the interaction with the antibody. Supporting that, experimental evidence of the preferential expression of this peptide was obtained during the preparation of the recombinant peptide–mBAP fusion proteins (not shown).
Finally, notice that, in addition to favor the selection of the desired peptides, the method produces peptide–mBAP fusions that have the potential of being used directly as probe or tracer reagents in different applications. Supporting that, the peptide–mBAP fusions of three of the peptides selected here could be used to set up sensitive competitive assays for molinate as shown in Figure 5. Similar IC50 values were obtained with the three peptide–mBAP tracers, and these values were similar to that obtained for the competitive assay set up using the chemically synthesized hapten (16) (69 ng/mL).
In the past decades, phage display has been consolidated as a reliable and rapid technology for the identification of bioactive peptides and antibody fragments selected from high-diversity libraries. For most purposes, not only the specificity, but very importantly, the affinity of the selected peptides is a major issue. This is particularly true for bioactive peptides that must show agonist or antagonist activity at the low concentration reached at the place of action. For this reason, the affinity of the peptides obtained after the initial selection from fully random peptide libraries is often improved through the construction of a secondary phage display library (mutagenesis library). In the mutagenesis libraries, the consensus residues of the initial phage pool are over-represented to favor the optimization of their flanking residues along the selection process (10, 19). In both cases (random and mutagenesis libraries), the outcome of the panning experiments should be easily examined to identify the higher-affinity clones. However, the identification of high-affinity binders among the pool of selected phage-borne peptides is not an easy task, and there is currently no systematic, yet practical, method to do this.
To deal with this problem, several alternative methods have been developed to facilitate the evaluation of the specificity and relative affinity of the isolated peptides; some of these methods use BAP as a partner fusion protein to study the peptide out of the context of the phage (69), but they are not compatible with high-throughput analysis. Here, we present a simpler method for the generation of large collections of clones expressing peptide–BAP fusions, coupled with a new method for rapid and simultaneous evaluation of the relative affinity of a large numbers of clones. In this method (affinity ELISA), the BAP activity works as an internal control to verify that equal amounts of peptides are assayed in each well of the affinity ELISA. Notice that the use of a much simpler affinity ELISA based on the capture of the phage particle bearing the peptide with an antiphage antibody would not produce reliable results due to the differential expression of individual peptides, which, as discussed above, would affect the average copy number of the displayed peptides.
The method showed excellent reproducibility and allowed us to sort the peptides within a wide range of affinities (up to 3 orders of magnitude) as shown by surface plasmon resonance analysis. Although we cannot rule out that, depending on their sequences, the peptides may show occasional interactions with the fusion partner that affect their binding to the target, in the four cases examined here there was a complete agreement between the affinity ELISA and the surface plasmon resonance analysis. By modification of the affinity ELISA conditions, it may be possible to shift this range to the target affinity (i.e., using stringent washing, lower amount of selector molecule, lowering the incubation time with the substrate, etc.). Interestingly, the highest-affinity clone (CKGLHMWFNC) which happened to be a low-abundance clone (about one in ten in the third output) could be easily spotted. This shows the potential of the method as a tool to identify valuable clones that are underrepresented in the selected phage pool. These “rare” nonconsensus fitting and low-frequency clones would be probably lost by the standard procedure of DNA sequencing and chemical peptide synthesis based on consensus or single sequences.
An additional benefit of the method is that the selected peptides can be directly produced as recombinant peptide–protein fusions. In our model, the recombinant peptide is linked to a high-turnover mutated form of BAP, in a fixed stoichiometric relationship, constituting a highly defined bioconjugate reagent that can be promptly used, for instance, as an immunoassay reagent. The applications of the concept of switch phage libraries can be extended to fusion partners others than BAP. In principle, any gene that is properly expressed in the periplasmic compartment of E. coli could be used to substitute the BAP gene in the pAFF/mBAP-PhoA vector, allowing the creation of sublibraries carrying the desired gene. This may include a different tracer activity such as firefly luciferase or green fluorescent protein to monitor intracellular binding of the peptides, or intein genes, that may help to produce the free recombinant peptide.
This work was supported in part by National Institute of Health, Fogarty International Center Grant (TW05718); Dirección de Innovación Ciencia y Tecnología, Uruguay (PDT54-41); the National Institute of Environmental Health Sciences (Superfund Basic Research program P42 ES04699, Center P30 ES05707).