Here we present a method to select for novel variants of the E. coli biotin-protein ligase BirA that biotinylates a specific target peptide. The protocol describes the construction of a plasmid for the bacterial display of the target peptide, generation of a BirA library, selection and characterization of BirA variants.
Biotin is an attractive post-translational modification of proteins that provides a powerful tag for the isolation and detection of protein. Enzymatic biotinylation by the E. coli biotin-protein ligase BirA is highly specific and allows for the biotinylation of target proteins in their native environment; however, the current usage of BirA mediated biotinylation requires the presence of a synthetic acceptor peptide (AP) in the target protein. Therefore, its application is limited to proteins that have been engineered to contain the AP. The purpose of the present protocol is to use the bacterial display of a peptide derived from an unmodified target protein to select for BirA variants that biotinylates the peptide. The system is based on a single plasmid that allows for the co-expression of BirA variants along with a scaffold for the peptide display on the bacterial surface. The protocol describes a detailed procedure for the incorporation of the target peptide into the display scaffold, creation of the BirA library, selection of active BirA variants and initial characterization of the isolated BirA variants. The method provides a highly effective selection system for the isolation of novel BirA variants that can be used for the further directed evolution of biotin-protein ligases that biotinylate a native protein in complex solutions.
Biotinylation of a protein creates a powerful tag for its affinity isolation and detection. Enzymatic protein biotinylation is a highly specific post-translational modification catalyzed by biotin-protein ligases. The E. coli biotin-protein ligase BirA is extremely specific and covalently biotinylates only a restricted number of naturally occurring proteins at specific lysine residues1. The advantages of the BirA catalyzed biotinylation are currently harnessed by fusing the target protein with a small synthetic 15-amino-acid biotin acceptor peptide (AP) that is effectively biotinylated2 and allows for the highly specific and efficient in vivo and in vitro biotinylation by co-expression or addition of BirA3,4,5. Although the in vivo and in vitro BirA catalyzed biotin-protein ligation is an attractive labeling strategy, its application is limited to samples that contain AP-fused proteins. The purpose of this method is the development of new mutants of biotin-protein ligases that selectively biotinylate native unmodified proteins and, thereby expand the number of applications in which the enzymatic biotinylation strategy can be used.
Protein function can be evolved through iterative rounds of the gene mutation, selection, and amplification of gene variants with the desired function. A strong and efficient selection strategy is crucial for the directed evolution and biotin-protein ligase activity is readily selected due to the strong binding between biotin and streptavidin and its homologs6. Phage display technologies allow for the selection of phages that display biotinylated peptides7,8. Since amplification of isolated phages requires infection of a bacterial host, however, the phage selection with streptavidin creates a bottleneck in that the high-affinity binding of biotin to streptavidin is virtually irreversible under non-denaturing conditions. To ensure reversible binding of biotinylated phages, monomeric avidins with lower affinity were used which resulted in a modest ~10-fold enrichment7. We recently developed a bacterial display method for the isolation of novel BirA variants that eliminates the need for the elution from the affinity matrix and thereby removes a bottleneck from previous BirA selection systems9. Indeed, our bacterial display system allows for a >1,000,000-fold enrichment of active clones in a single selection step9, thus providing an effective selection system for the directed evolution of novel BirA variants.
Our bacterial display system consists of two components, BirA with a C-terminal 6xHis tag and a scaffold protein that allows for the surface display of a target peptide. We used the scaffold protein enhanced circularly permuted outer membrane protein X (eCPX) since the effective display of peptides can be observed at both the N- and C-termini10,11. The fusion of the target peptide sequence to the C-terminus of eCPX ensures biotinylation of bacteria expressing active BirA variants. The bacteria allow for the effective streptavidin selection as the biotinylated peptide now displays on the surface (Figure 1a).
The purpose of this method is to select for novel variants of BirA that biotinylates peptide sequences present in native proteins. The system is encoded by genes present on the plasmid pBAD-BirA-eCPX-AP, which contains an arabinose-inducible promoter controlling BirA (araBAD), and a T7 promoter controlling eCPX9 (Figure 1b). The present protocol describes the detailed procedure for 1) incorporation of a peptide derived from a target protein into the C-terminal of eCPX, 2) creation of a mutational library of BirA by error-prone PCR, 3) selection of streptavidin-binding bacteria by magnetic-activated cell sorting (MACS), 4) quantification of bacteria enrichment, and 5) initial characterization of isolated clones.
1. Insertion of Peptide Coding Sequencing Sequence in pBAD BirA-eCPX-AP
NOTE: To select for BirA variants that biotinylate a native target protein, start by identifying a 15-amino acid peptide sequence in the proteins primary sequence that contains at least one lysine (K) residue.
2. Generation of a BirA Library
NOTE: The initial BirA mutational library (Figure 1c, step 1) is created by error-prone PCR. Other methods to generate the BirA mutational library are likely to work as well.
3. Selection of Bacteria Expressing Biotinylated Peptide
NOTE: This part of the protocol covers step 2-5 of Figure 1c. It is highly recommended that the selection approach is setup using pBAD-BirA-eCPX-AP and pBAD-BirA-eCPX-AP(K10A) as positive and negative controls.
4. Quantification of Enrichment
NOTE: Quantification of the live bacteria in the "input" and "output" samples are performed after each selection round by plating of serial dilutions of the samples and subsequent counting of colony forming units (CFUs).
5. Characterization of Selected BirA Variant
NOTE: The characterization can be performed after selecting BirA variants from the first BirA library; however, the BirA variants generally have low activity towards the peptide. Therefore, an additional round of mutation and selection can also be performed before the characterization. Usually, 10 clones from the final selection round are isolated for further characterization.
Western blot of pBAD-BirA-eCPX-AP expressing bacteria produces a ~22 kDa streptavidin-reacting band consistent with the molecular weight of eCPX (Figure 2a). Unlike BirA-6xHis, biotinylated eCPX-AP was present in both uninduced and induced cultures (Figure 2a) due to a small degree of T7 promoter activity even in uninduced cultures and subsequent biotinylation of the AP by endogenous BirA. In BirA-eCPX-AP(K10A) expressing cultures, no biotinylated eCPX band was detected (Figure 2a). The strong surface biotinylation in the eCPX-AP expressing bacteria causes aggregation upon addition of streptavidin magnetic beads and the formation of a pellet at the bottom of a tube (Figure 2b). In the eCPX-AP(K10A) expression bacteria, streptavidin-bead aggregation and precipitation was not observed (Figure 2b). Analysis of the precipitate from the streptavidin pulldown, displays a clear 22-kDa streptavidin-reacting and anti-6xHis band in the samples from eCPX-AP cultures, but not eCPX-AP(K10A) cultures (Figure 2c). Similarly, the count of bacteria bound to the streptavidin-beads was significantly higher in the eCPX-AP than the eCPX-AP(K10A) cultures (Figure 2d).
To select for BirA variants that biotinylate a target peptide, its DNA sequence was incorporated into the C-terminal of eCPX by PCR using the primers designed in step 1.10 and 1.11. An example of the primers designed for the incorporation of a peptide sequence derived from the α-subunit of the epithelial Na+ channel (ENaC) is shown in Figure 3a. After PCR, a 5-µL aliquot was analyzed by agarose gel electrophoresis and a clear and strong band at ~5900 bp was observed (Figure 3b).
After the generation of the BirA mutation library, the selection of active BirA variants was initiated. A low degree of streptavidin-bound vs. input bacteria was expected after the first selection round. However, after 2nd and 3rd selection rounds a clear enrichment was observed in the degree of streptavidin-bound bacteria Figure 4a). If a clear enrichment is not detected (Figure 4b), it is indicative of the failure of the BirA variants to biotinylate the peptide and another peptide sequence should, therefore, be tested.
After the final selection round, 10 clones were characterized by western blotting for their ability to biotinylate the displayed peptide (Figure 4c). In the positive control (i.e., AP), ~22 kDa band corresponding to the biotinylated eCPX-AP was observed in a western blot probed with streptavidin-HRP (Figure 4c). In the tested clones, a band at similar size was indicative of biotinylation of the displayed peptide fused to eCPX (Figure 4c). The intensity of the ~22 kDa bands was lower than the intensity of the eCPX-AP band in the positive control, indicating a lower activity of the isolated BirA variants. The isolated clones can, therefore, be used as a template for another round of mutations and selection, yielding highly active clones. Additional bands indicate the isolated clones were not specific towards the displayed peptide and that additional targets were also biotinylated (Figure 4c).
Reagent | volume (µL) |
5x Reaction Buffer | 4 |
10 mM dNTP | 0.4 |
10 µM Forward Primer | 1 |
10 µM Reverse Primer | 1 |
pBAD-BirA-eCPX-AP | Variable (~25 ng) |
High-Fidelity DNA polymerase | 0.20 |
Nuclease-Free Water | to 20 |
Table 1: PCR reagents. Units and volumes may vary between manufacturers.
Reagent | volume (µL) |
2x Enzyme mix | 25 |
pBAD-BirA-eCPX-AP with target peptide sequence* | Variable (~50 ng) |
Mutant megaprimer | 250 ng |
Buffer | 3 |
Nuclease-Free Water | to 50 |
Table 2: Error-prone PCR reagents. Units and volumes may vary between manufacturers. * prepared in section 1 of the protocol.
Figure 1: The bacterial display system for BirA selection. (a) The system was based on the co-expression of 2 components: BirA and eCPX fused with the acceptor peptide (AP). eCPX is transported to the surface and, if the BirA variant biotinylates the AP, the biotin (red B) attached to the eCPX-AP is displayed on the surface. (b) The system was expressed from the plasmid pBAD-BirA-eCPX-AP, where BirA expression is controlled by an arabinose-inducible promoter and eCPX-AP expression is driven by the T7 promoter. (c) After generation of a randomly mutated library of BirA variants (step 1), BirA and eCPX-AP expression was induced (step 2). Bacteria were incubated with affinity reagent (step 3), unbound bacteria were discarded (step 4) and selected bacteria were amplified (step 5). This figure has been modified from Granhøj et al.9 Please click here to view a larger version of this figure.
Figure 2: Representative results from model selection with pBAD-BirA-eCPX-AP and pBAD-BirA-eCPX-AP(K10A). (a) By western blotting, eCPX-AP was observed to be biotinylated in both uninduced and induced bacteria, while no biotinylation of eCPX-AP(K10A) was detected even after the induction of BirA. BirA expression was detected by anti-6xHis antibody. * Indicates an unspecific streptavidin-reacting protein when BirA was induced. (b) Bacterial cultures with induced expression of BirA and eCPX-AP aggregate rapidly after addition of magnetic streptavidin-beads (arrow), while no aggregation was observed in AP(K10A) bacteria. (c) BirA was present in bacteria expressing eCPX-AP and eCPX-AP(K10A) before thestreptavidin-pulldown (input), but only BirA in eCPX-AP expressing bacteria was pulled down by streptavidin. In agreement, (d) viable bacteria were precipitated effective in eCPX-AP, but not eCPX-AP(K10A), expressing bacteria. This figure has been modified from Granhøj et al.9 Please click here to view a larger version of this figure.
Figure 3: Primer design and incorporation of target peptide coding sequence into pBAD-BirA-eCPX-AP by PCR. (a) An example of the primers design used for the incorporation of a peptide sequence derived from αENaC into the C-terminal of eCPX. The biotin accepting lysine is shown in red. The target peptide sequence was reverse translated to DNA, and forward and reverse primers were designed by ensuring a ~15 base overlap between the primers. (b) Representative agarose gel electrophoresis of PCR with pBAD-BirA-eCPX-AP as template and primers specific for α, β, and γ-ENaC derived peptide sequences, respectively. A clear and strong DNA product at ~5900 bp was indicative of a successful PCR. "M" indicates marker lane. Please click here to view a larger version of this figure.
Figure 4: Representative results from selection and characterization of bacteria displaying peptides. Bacteria displaying a peptide derived from (a) TagRFP showed a clear enrichment after 3 selection rounds, while a peptide derived from (b) EGFP showed no enrichment of streptavidin-bound bacteria even after 4 selection rounds. (c) 10 clones of bacteria displaying a peptide from γENaC through 5 selection rounds were tested for their ability to biotinylate the γENaC-peptide. All 10 clones showed a streptavidin-reacting band consistent with the size of eCPX-AP, indicating that the isolated clones contain BirA variants that biotinylate the displayed peptide. Additional streptavidin-reacting bands were also observed, indicating that other proteins, besides the displayed peptide, were also biotinylated. * Indicates an endogenous E. coli protein biotinylated by BirA. This figure has been modified from Granhøj et al.9 Please click here to view a larger version of this figure.
As for all selection methods, the stringency of the washing steps is of utmost importance. Since bacteria do not need to be eluted from the beads before the amplification of the selected clones, the high affinity binding between biotin and streptavidin can be used instead of using lower affinity avidins, as previously done with the phage display system, for the selection of BirA variants7,8. This ensures that rare clones are selected and that non-biotinylated bacteria are discarded. Another advantage of using bacterial display, as compared to phage display, is that bacterial display is quantitative11 and, therefore, allows for the selection of the bacteria based on the enzymatic activity.
In the protocol, we used MACS to select for bacteria creating a binary selection system based on the presence or absence of biotin on the surface. However, by using quantitative fluorescence activated cell sorting, instead, it should be possible to select for bacteria that express the most active variants of BirA. This will be important in the future development of the novel BirA variants as it will allow an effective selection for the most active BirA variants.
We have, so far, used the bacterial display of 14 different peptides and, of those, 13 produced a clear enrichment9, indicating that our selection system provides a robust method to select for the novel BirA variants. In the current setup, we have only tested the selection of BirA variants that are active towards 15-amino acid peptides and, thereby we preferentially selected for the BirA variants that are active towards the primary sequence of the target protein. The targeted lysine can, however, be buried inside the 3D structure of a protein or not be otherwise accessible for BirA, yielding BirA variants that are not active against their target protein. A potential solution would be to display the larger protein fragment on eCPX. The eCPX scaffold is versatile with respect to the peptide display11; however, it is not known whether larger proteins can be displayed.
We used the selection system to isolate a BirA variant that biotinylates native TagRFP9. The tested BirA variant specifically biotinylated TagRFP on the targeted lysines, but the activity of the isolated variant was low9. Therefore, further rounds of directed evolution should be performed to improve its activity. The target peptide is in the C-terminus of TagRFP, where the structural similarity between the displayed peptide and the protein region is more likely. Bioinformatic analysis of all human and mouse proteins shows that ~75% of the proteins contain one or more lysine within their first and/or last 30 amino acids9. Thus, the bacterial display system of peptides can potentially be used to isolate active BirA variants towards a large fraction of native proteins.
The authors have nothing to disclose.
The authors thank Mohamed Abdullahi Ahmed for the expert technician assistance. This work was supported by grants from the Lundbeck Foundation, the Novo Nordisk Foundation, the Danish Kidney Association, the Aase og Ejnar Danielsen Foundation, the A.P. Møller Foundation for the Advancement of Medical Science, and Knud and Edith Eriksen Memorial Foundation.
10% precast polyacrylamide gel | Bio-Rad | 4561033 | |
Ampicilin | Sigma-Aldrich | A1593 | |
ApE – A plasmid editor v2.0 | NA | NA | downloaded from http://jorgensen.biology.utah.edu/wayned/ape/ |
Arabinose | Sigma-Aldrich | A3256 | |
Biotin | Sigma-Aldrich | B4501 | |
DMSO | Sigma-Aldrich | D2650 | |
DPBS (10X), no calcium, no magnesium | ThermoFischer Scientific | 14200083 | |
DpnI restriction enzyme | New England BioLabs | R0176 | |
Dynabeads MyOne Streptavidin C1 | ThermoFischer Scientific | 65001 | |
GenElute Plasmid Miniprep Kit | Sigma-Aldrich | PLN350 | |
GeneMorph II EZClone Domain Mutagensis kit | Agilent Technologies | 200552 | |
Glucose | Sigma-Aldrich | G8270 | |
Glycerol | Sigma-Aldrich | G5516 | |
Immobilon-P PVDF Membrane | Millipore | IPVH15150 | |
IPTG | Sigma-Aldrich | I6758 | |
LS Columns | Miltenyi Biotec | 130-042-401 | |
NaCl | Sigma-Aldrich | S7653 | |
NEB 5-alpha Competent E. coli | New England BioLabs | C2987 | |
NuPAGE LDS Sample Buffer (4X) | ThermoFischer Scientific | NP0007 | |
NuPAGE Sample Reducing Agent (10X) | ThermoFischer Scientific | NP0009 | |
pBAD-BirA-eCPX-AP | Addgene | 121907 | Used a template and positive control |
pBAD-BirA-eCPX-AP(K10A) | Addgene | 121908 | negative control |
Q5 High-Fidelity DNA Polymerase | New England BioLabs | M0491 | For insertion of peptide sequence in pBAD-BirA-eCPX-AP, any high fidelity polymerase will do |
QuadroMACS Separator | Miltenyi Biotec | 130-090-976 | |
Skim Milk Powder | Sigma-Aldrich | 70166 | |
Streptavidin-HRP | Agilent Technologies | P0397 | |
T7 Express lysY/Iq Competent E. coli | New England BioLabs | C3013 | |
Tryptone | Millipore | T9410 | |
Tween-20 | Sigma-Aldrich | P9416 | |
Western Lightning Plus-ECL | PerkinElmer | NEL103001EA | |
Yeast extract | Sigma-Aldrich | Y1625 |