Phage display is a powerful technique to capture proteins or protein moieties that interact with an immobilized molecule of interest. Once a decision of the type of phage cDNA library to create and screen has been made, the protocol described here permits efficient affinity selection leading to identification of interactors.
Using recombinant phage as a scaffold to present various protein portions encoded by a directionally cloned cDNA library to immobilized bait molecules is an efficient means to discover interactions. The technique has largely been used to discover protein-protein interactions but the bait molecule to be challenged need not be restricted to proteins. The protocol presented here has been optimized to allow a modest number of baits to be screened in replicates to maximize the identification of independent clones presenting the same protein. This permits greater confidence that interacting proteins identified are legitimate interactors of the bait molecule. Monitoring the phage titer after each affinity selection round provides information on how the affinity selection is progressing as well as on the efficacy of negative controls. One means of titering the phage, and how and what to prepare in advance to allow this process to progress as efficiently as possible, is presented. Attributes of amplicons retrieved following isolation of independent plaque are highlighted that can be used to ascertain how well the affinity selection has progressed. Trouble shooting techniques to minimize false positives or to bypass persistently recovered phage are explained. Means of reducing viral contamination flare up are discussed.
Why use phage display and affinity selection instead of the myriad other techniques available for discovering and investigating protein interactions with other molecules? Phage display can claim some unique advantages over other methods of detecting protein-ligand interactions1-3 including the following:
Very wide repertoire of bait molecules
The foremost reason is in the diversity of molecules capable of acting as bait in affinity selection 4. Phage display is a very powerful means of isolating protein fragments that interact with other proteins, nucleotides, carbohydrates, etc.5 Essentially, if a polymer/molecule can be attached to a recoverable support, it can be screened for affinity with phage displayed proteins. Additionally, the bait molecule can be accessed to determine if it retains biological activity when immobilized6, if conditions used to reduce/eliminate its activity are effective6 or, to introduce post-translational modifications to the bait prior to affinity selection.
Phage resistance to extraneous factors
A second reason for using phage display, is that it is possible that some baits may require environmental stress (heat, high osmolyte concentrations, specific cofactors, etc.) to capture their interacting proteins, or the phage displayed proteins may need to be somehow altered prior to affinity selection. The primary factor being studied is the interaction between bait and protein, not the condition that allows the interaction to occur or if it is lethal to the test organism. The T7 phage is particularly well suited to such studies because it can withstand harsh experimental conditions both intact and viable (e.g. the published thermal maximum for T7 viability is ~60 °C 7). As an example of altering phage displayed proteins, when examining the Arabidopsis thaliana seed proteome for protein substrates of the repair enzyme, Protein Isoaspartyl Methyl Transferease (PIMT), the virus used in these studies had been "aged" for one week, prior to each affinity selection round, to encourage introduction of isoaspartate (isoAsp) residues in susceptible proteins6 which is not possible in organisms that can recognize and repair/metabolize such abnormalities.
Metabolically inert
Furthermore, the phage are usually resistant to metabolic poisons and interfering small molecules that would, at the least, result in pleiotropic effects on metabolically active test organisms. Following the stringency washes, the poison is removed before a large volume of bacteria are introduced for infection so the poison is diluted to a range innocuous to the bacteria or subsequent phage replication. While investigating protein targets of the PIMT repair enzyme, S-Adenosyl Methionine (AdoMet) was used to activate the micro-titer-plate-well immobilized enzyme to permit target protein capture while relying on S-Adenosyl Homocysteine (AdoHcy) to inactivate the enzyme and provide a useful negative control secure in the knowledge that the virus would not be adversely influenced by either AdoMet or AdoHcy6. Additionally, members of certain Late Embryogenesis Abundant (LEA) protein families, investigated in this lab, are known to alter their shape in the presence of additives such as sucrose8, which can attain concentrations as high as 200 mM in soybean seeds at the point of physiological maturity9. The virus is not anticipated to be influenced by addition of 200 mM sucrose in each affinity selection round, possibly necessary for certain LEA-client protein interactions, which is not the case for autonomously viable test organisms10.
This lab has focused on discovering protein-protein interactions in mature, dehydrated or germinating seeds that underlie the mechanism of stored proteome protection during dehydration11 or repair of components of the stored proteome that are susceptible to isoAsp formation once the seed has imbibed6. Thus, the production and purification of the recombinant proteins required as bait in an active state, and ensuring that they remain so, before and after they are immobilized, although frequently difficult, is a cornerstone to our work. However, as each recombinant protein production scenario is different, optimizing conditions for recombinant protein production will not be addressed here. Users are urged to attempt, wherever possible, to determine if the immobilized protein is still functional (e.g. if it is an enzyme, do an enzyme assay in the microtiter plate wells). This will provide some confidence that the bait is biologically active and therefore, that any discovered interactions are somewhat more likely to have biological relevance.
A graphic depiction of the procedure described below (Figure 1) highlights the two primary components for affinity selection using a phage display library: A) a phage display cDNA library likely to encode proteins with affinity for the bait and; B) a purified recombinant bait protein. Production of bait (recombinant protein) has been extensively examined and literature outlining best practices for securing soluble, active recombinant protein from E. coli12-13, eukaryotic yeast14, insect15-16, plant17-18, or mammalian19-20 cells abound.
In the following protocol, a hexahistidyl tagged recombinant protein has been used as bait. This allows verification that the bait proteins remain in the wells after overnight incubation and washing steps.
1. ELISA for Recombinant Protein Retention in the Microtiter Plate Wells
Note: If the recombinant protein of interest does not attach itself to the wells, it is possible to alter the composition/pH of the buffer considerably (Carbonate buffer pH 10.0) or add chaotropes (urea) to attempt to assist protein attachment to the microtiter plate wells. However, re-establishing that the recombinant protein: 1) remains bound to the microtiter plate wells under the conditions for the affinity selection and; 2) retains its biological activity following removal of the high pH/chaotrope is recommended.
2. Growing Bacterial Host (BLT5403) for Titering
Note: The host bacterial cells BLT5403, expressing a plasmid-borne source of the T7 native capsid protein without which the T7 mid-, and low-copy vectors cannot replicate successfully, are extremely susceptible to the virus. It is imperative to avoid contamination. Contamination will eventually occur at which point all surfaces in contact with the virus/infected bacteria must be wiped with 70% (v/v) ethanol or scrubbed using detergent (Figure 2I). If this is ineffective, a thorough decontamination of all surfaces that can withstand Envirocide (Figure 2J) is required. Start BLT5403 cells from freezer stock each new round of affinity selection to help mitigate contamination of the stock in 4 °C, and ensure vigorous cells are being used in the protocol. Filter barrier pipette tips are essential.
3. Affinity Selection
DAY ONE
For example: Start labeling 1.5 ml Eppendorf tubes for three 10-2 dilutions of phage from the BSA replication one well as 1, 2, and 3 (three triplicate readings of phage titer for replication one), and then mark borosilicate test tubes 1, 2, and 3 in which the infected bacteria will be mixed with uninfected bacteria and top agarose before pouring on LB100 AMP agar plates (Figure 3C).
Note: Round three and four may require approximate dilutions of as much as 10-10 volume of phage to volume of to LBAMP100.
DAY TWO
Note: Start the culture soon enough that, by the time the phage infected-BLT5403 from the completion of the affinity selection are ready to add, the cells in the 500 ml are between 0.6 and 1.0 OD600. If they grow much beyond 1.0, lysis may not occur due to resource restrictions as cells approach stationary phase. If the cells do not attain an OD600 of 0.6 at least prior to the introduction of phage, the multiplicity of infection can be too high, rapidly killing the cells, which can influence the representation of the lysate. If the cells are not yet approaching an OD600 of 0.6 by the completion of the affinity selection, inoculate the flask with more BLT5403 from the second (or even from both second and third) test tubes shaking at 37 °C (Figure 2F). If an OD600 of 1.0 is approaching too fast, turn off the heater and the shaker and open the lid to cool the culture and starve the bacteria for oxygen. Recommence heating/shaking once the phage have been added.
From here use filter barrier tips to avoid phage cross-contamination.
4. Day Three
5. Titering
6. Determining and Calculating Titer
7. Plaque Isolation, PCR, and Sequencing
Being able to impair the capacity of the bait to interact with phage-displayed proteins (metabolically poisoning the bait) provides a potent negative control for this technique. It is also advisable to determine if the bait, when bound to the microtiter plate well, retains its function. Both of these checks will increase the confidence that interacting, phage-displayed proteins recovered by the nonpoisoned bait are legitimate.
Sampling three triplicates from each well adds considerably to the time and work involved in the affinity selection but provides a more accurate estimate of the titer within each well than sampling only once. While most of the titers for the triplicates are in good agreement, note that the triplicates for replication 2 of the BSA in round four vary widely (Table 1). By sampling several times, the final estimated tallies are not as susceptible to pipetting errors and a more accurate estimate of the actual titer and the variation around this estimate, well-to-well, are obtained (Table 1, Figure 5C).
Increasing phage titer, preferentially for the nonpoisoned bait, as the number of affinity selection rounds increase, also provides confidence that the technique is working (Figure 5C). The retention of an increasing percentage of phage containing insert in nonpoisoned bait wells as the number of affinity selections increases is also auspicious (Figures 6A and 6B).
After advanced rounds of affinity selection, an increase in the number of independently recovered phage that have insert (discernible as PCR amplicons greater in size than ~100 bp; Figure 6B), relative to the first round (Figure 6A), and the settling of these amplicons on a few identically sized bands (note lanes 5-9, 11-18 in Figure 6B), is a good indication that particular clones are being selected in the affinity selection (Figures 6A and 6B). After sequencing a number of plaque obtained in the final affinity selection round, retrieval of a similar region (different clones) of the same protein is independent verification that the portion of the displayed protein has a bona fide affinity for the bait. These independently recovered phage also provide information on what part of the entire protein is capable of interacting with the bait. If the phage displayed cDNA library has been constructed using random priming, a great diversity of partial- and full-length-protein coverage can be anticipated (within the limits of the phage to tolerate the cDNA insertion size), leading to multiple, independent clones covering the same portion of the protein. Even the contiguous out-of-frame hits should be scrutinized to determine if they encode a similar motif to which the bait binds. (This assumes that the libraries were not constructed using ORF-technology3).
Table 1. Calculations for the average titer/ml infected bacteria from the wells of the fourth round of affinity selection for the three treatments. The numbers for each triplicate of the 10-3 dilution of the plaque taken from the first replication of the bait well are taken from Figure 5B. Click here to view larger table.
Figure 1. Overall graphic depiction of the phage display process. Two primary requirements include; A) access to a phage displayed library, preferably constructed using randomly primed, poly-A selected RNA, for generating the cDNAs; and B) a bait that is, in so far as possible, verifiable as biologically active, even when immobilized on a solid support and, if possible, rendered inactive by addition of an inhibitor (e.g. competitively inhibited from binding bait protein). RT: reverse transcription.
Figure 2. Some of the equipment and material necessary for performing phage display. Sterile technique requires access to both: A) an autoclave and; B) a laminar flow hood. Phage libraries and BLT5403 bacterial stock require -80 °C temperatures for prolonged storage C) Growing phage and bacteria requires, D-F) 37 °C incubators, one of which, E-F) is capable of growing liquid cultures (orbital). Optimizing the experiment through color coding, F) minimizes mistakes. BLT5403 cells for titering can be stored for up to one week, G) at 4 °C. Optimizing the placement of the, H) spectrophotometer for following cell densities next to the orbiting shaker can save time. Frequently treating surfaces and pipettes with, I) powerful detergent and J) Envirocide, can help to minimize the risk of viral contamination. Click here to view larger image.
Figure 3. One possible set-up useful for permitting smooth affinity selection. A) The BLT5403 are grown a day or two prior to titering and are plated, without viral introduction, to determine that they are not contaminated with virus. Titering on the day of the affinity selection can be facilitated by preparing beforehand by: B) prelabeling the Eppendorf and; C) borosilicate tubes to be used. B) Aliquoting the dilution solutions of LB100AMP and storing them at 4 °C overnight. D) Place the prenumbered, LB100AMP agar-containing, Petri dishes at 37 °C to warm approximately 1 hr before titering. E) Melting the sterile top agarose (loosen the cap) and placing it in a preheated 50 °C water bath to cool to this temperature an hour before titering commences. Use a lead donut to avoid the bottle capsizing as top agarose is removed. Click here to view larger image.
Figure 4. Titering the phage recovered from a round of affinity selection. Plating commences as soon as possible after infection to prevent artificial inflation of the titer. A) The first set of virus-infected BLT5403 dilutions and the borosilicate tubes assembled in racks. Using filter barrier tips: B-C) 250 μl BLT5403 cells from the stock from 4 °C are transferred to the borosilicate tubes. Using filter barrier tips: D-E) The BLT5403 in the first borosilicate tube is infected with the first virus dilution. F) The borosilicate pipette is heated over an open flame to prevent clogging and to keep the top agarose warm. Do not heat the pipette excessively or the bacteria will be scalded and die. G) 3 ml molten top agarose are quickly retrieved and poured over the bacteria in the borosilicate tube. H) The tube is flicked briefly to mix the contents and; I) The contents poured onto the prewarmed LB100AMP agar plates, using the tube to move bubbles to the sides of the plate and to ensure the top agarose covers the whole plate. Click here to view larger image.
Figure 5. Typical titer results. A) The lower dilutions (10-2) almost invariably result in confluent lysis for all treatments in the first affinity selection round (the virus in the figure have been grown overly long resulting in excessive plaque size to facilitate photography). B) The appropriate dilutions of plaque are counted using a sharpie to keep track of tallied plaque. C) The titer increases drastically for bait wells while remaining more-or-less stable for BSA or poisoned bait. In later rounds, titer from bait wells plateaus and further affinity selection is unproductive. D) A well isolated plaque has been chosen for sequencing and collected by wetting the Pasture pipette with solution from the Eppendorf tube, being sure to eliminate the excess liquid least it run over multiple plaque and contaminate the core. E) The pipette has been pushed through the plaque with the pipette bulb already compressed. F) The bulb has been released while the pipette is still in the top agarose, sucking the plaque up into the pipette (arrow). G) The cored plaque is about to be delivered into the liquid in the appropriate Eppendorf tube. Phage dilutions (volume of infected bacteria per volume of LB100AMP) are depicted on the Petri dishes (e.g. 10-3 = 1/1,000 dilution). The three triplicate samples from replication well 1 are abbreviated as Trip. 1, Trip. 2, and Trip. 3, all for replication 1 (Rep. 1= well 1). The numbers at the bottom of the plates in B) are the actual plaque numbers tallied on the plates. These are for the 10-3 dilution of the bacteria from the fourth round of affinity selection (R4 in figure) over the bait. Click here to view larger figure.
Figure 6. Typical PCR results for affinity selection over an appropriate bait. A) The percentage of empty vector plaque (arrows) is usually quite high initially but, in bait wells; B) decreases in subsequent rounds. The plaque containing insert has tended to settle on those containing identically sized insert in this later affinity selection round. MWM: Molecular Weight Marker.
By running the experiment in three replicated wells, independently acquired phage of the same protein binding to the bait can be distinguished even if they are the same clone (i.e. no difference in the nucleotide sequence of the CDS region that has been acquired but these have been retrieved from independent wells). Otherwise, the only way to distinguish among phage encoding the same protein binding to the bait is if they are independently reverse transcribed regions that differ in some part of the nucleotide sequence.
The use of one Fernbach flask in which to grow all of the bacteria for use in amplifying the affinity selected phage permits a much less hectic monitoring of bacterial growth leading up to infection following affinity selection than trying to monitor 9 Erlenmeyer flasks. Decanting these bacteria, just prior to infection, into prewarmed Erlenmeyer flasks also eliminates sublibrary titer discrepancies due to alterations in the multiplicity of infection due to poor bacterial growth in specific flasks. Thus, alterations in phage titer from round-to-round should depend on the number of phage retained in the wells and the variation in titer among replicated wells should be minimal.
One exceptional advantage using phage display is the great power the technique has in identifying phage-displayed-protein that have an affinity for a particular bait. Due to the efficiency with which the phage replicate in the BLT5403 host, even a single phage representing a rare CDS that is retained in a well in the first affinity selection round, should the coat-fused protein have a high affinity for the bait, will be represented in the second round by much greater numbers. By the fourth round, this phage should constitute the majority of the phage present in the lysed culture.
This capacity of the phage to replicate to high titer is also a limitation of the technique. If a particular bait has an interacting protein partner for which it has high affinity in the library, it is very difficult to capture proteins that may legitimately bind the bait but at a lower affinity as these will tend to be outcompeted by the high affinity binding protein. The use of next generation sequencing techniques to identify such rare phage is one possible means of circumventing this limitation1. Furthermore, the susceptibility of the BLT5403 to the T7 phage results in bouts of "contamination" throughout the lab that require perseverance, rather harsh decontaminants (e.g. Envirocide) and excellent sterile technique to overcome. Contamination can result in a considerable loss of time in the progress of affinity selection as one tries to identify and eliminate the source.
One means that has limited success in partitioning phage into tight and loose "binders" is to try and recover the loose binders first by introducing into the wells a solution (1% SDS or similar chaotrope) designed to remove phage that are less tightly bound to the bait. This solution can then be removed first, prior to the introduction of the bacteria into the wells which are anticipated to be infected by those phage remaining bound to the bait. This technique can also reduce background, spuriously binding phage, considerably. However, if this technique is pursued, the number of sublibraries, replicates and duplicates affinity selected in subsequent rounds is doubled, which adds considerably to the time and expense.
Following the course of phage titer increase over several rounds of affinity selection can run through a large volume of LB and a large number of LB100 AMP plates, filter barrier tips, Eppendorf tubes, etc. This is expensive as is the sequencing required to examine even a modest number of phage. Titering a large number of dilutions of several replicates of treatments in triplicate requires considerable time so the importance of setting up as much of the material as possible the day prior to the affinity selection is paramount.
One exciting possibility that is currently being explored is to use the location of the independent phage proteins, binding to a bait, to model the physical attributes and the three dimensional structure of the region of the protein interacting with the bait. For example, examining the attributes of protein targets that bound to a set of homologous Late Embryogenesis Abundant (LEA) proteins from Arabidopsis (Arabidopsis thaliana) and Soybean (Glycine max), one important conclusion was that the LEA proteins bound to the region of the proteins that were the most (or among the most) hydrophilic of the binding proteins11. Retrospectively, given that the LEA proteins are hypothesized to protect their client molecules from denaturation during dehydration, it can be surmised that the region of the client molecules most susceptible to dysfunction without LEA protection might be those most capable of binding water (hydrophilic).
It is advisable to minimize the time between the introduction of the bacteria into the wells and their retrieval so the titer is not inflated. Ensure that the dilutions are sufficient by plating some BLT5403 without viral infection to verify that the bacterial stock is not contaminated and that the greatest dilutions are sufficient to avoid confluent lysis.
The authors have nothing to disclose.
This project was partially funded by an NSF IOS (0849230); Hatch, McIntire-Stennis (AD421 CRIS), USDA Seed Grant (2011-04375), and Sir Frederick McMaster Research Fellowship to ABD.
Name of Reagent/Material | Company | Catalog Number | Comments |
Tryptone | Becton Dickinson Co. | 211705 | |
Yeast Extract | Becton Dickinson | 212750 | |
Agar | Becton Dickinson | 214010 | |
Sodium Chloride | Fisher Scientific | BP358-10 | |
Agarose | Research Products International | A20090 | |
Blocking Reagent | EMD Chemicals Inc. | 69064 | |
Tween 20 | Fisher Scientific | BP337-100 | |
Sodium Hydroxide | Fisher Scientific | BP359-212 | |
Sodium Dodecyl Sulfate | Fisher Scientific | BP166-500 | |
Tris Base | Fisher Scientific | BP152-5 | |
Chloroform | Fisher Scientific | BP1145-1 | |
Escherichia coli BLT5403 | EMD Chemicals Inc. | BLT5403 | Genotype: F- ompT hsdSB (rB – mB -) gal dcm pAR5403 (AmpR) |
ampicillin, Sodium Salt | Research Products International | A40040 | |
ethidium bromide | Fisher Scientific | BP102-1 | |
Bovine Serum Albumin (BSA) | Sigma-Aldrich Corp. | A-9647 | |
penta-HIS primary antibody | Qiagen Inc. | 34660 | |
Goat anti mouse alkaline phosphatise conjugate | Sigma-Aldrich Corp. | A-5153 | |
para-nitrophenylphosphate | Sigma-Aldrich Corp. | N-7653 | |
T7-UP primer | EMD Chemicals Inc. | 5′-GGAGCTGTCGTATTCCAGTC-3′ | |
T7-DOWN primer | EMD Chemicals Inc. | 5′-AACCCCTCAAGACCCGTTTA-3′ | |
Qiaquick PCR Purification kit | Qiagen Inc. | 28104 | |
Qiaquick Gel Extraction kit | Qiagen Inc. | 28706 | |
Big Dye Terminator v3.1 Cycle Sequencing Kit | Invitrogen Life Technologies | 4336917 | |
Heating Block | Pierce Chemical Co. (Thermo Fisher Scientific Co.) | 18780 | |
96 Well Cell Culture Plates | Corning | Costar 3590 | |
Isotemp Incubator Oven | Fisher Scientific | 516D | |
Pasteur Pipets, 5 ¾” | Fisher Scientific | 13-678-20A | |
ChromaView Transilluminator | UVP Inc. | TS-15 | |
UV Shield | Oberon | 071AF | |
Gloves, Nitrile | Fisher Scientific | 19-170-010C | |
Sterile 1.5 ml tubes | USA Scientific Inc | 1615-5500 | |
10 ml borosilicate Pipets | Fisher Scientific | 13-678-25E | |
Borosilicate culture tubes | Fisher Scientific | 14-961-27 | |
Uniskan I, ELISA plate reader | Labsystem and Flow Laboratories, Helsinki, Finland | Type 362 |