A Genetic Screen to Identify Bacteriophage Lysins

Abstract

Lysins are phage-encoded, peptidoglycan (cell wall) hydrolases that accumulate in the bacterial cytoplasm during a lytic infection cycle. Late during infection, the lysins undergo holin-mediated translocation across the inner membrane into the peptidoglycan matrix where they cleave cell wall covalent bonds required for wall stability and allow bacterial lysis and progeny phage release. This potent hydrolytic activity is now the foundation of a powerful genetic-based screening process for the identification and analysis of phage lysin proteins. Here, we describe a method for identifying a lysin, PlyG, from a bacteriophage that specifically infects the Gram-positive organism Bacillus anthracis, however, the techniques described can be adapted to clone, express and analyze lysins from any phage infecting Gram-positive bacteria or possibly even Gram-negative bacteria.

1. Introduction

A lysin-based lytic mechanism (1) is widespread amongst phages infecting both Gram-positive and Gram-negative bacteria. Indeed, the lysins likely represent one of the most successful bacteriolytic agents used in nature (2). A major distinguishing feature of the lysin family of proteins concerns their modular design, which can vary dramatically depending on whether a lysin targets the thick, surface exposed cell wall of Gram-positive bacteria, or the very thin peptidoglycan of Gram-negative organisms that lies subjacent to and is protected by the external outer membrane. As such, lysins active against Gram-negative bacteria usually consist of a single 15 to 20 kDa catalytic domain, while those active against Gram-positive bacteria are 25 to 100 kDa and possess one or more distinct catalytic domains (multi-domain protein) fused to a cell wall-binding domain (CBD). The CBD’s of Gram-positive lysins are believed to recognize strain- or species-specific carbohydrate structures that decorate the surface of the Gram-positive cell wall (3), thus exerting a level of specificity over the binding of such lysins. A potent lytic activity, combined with an often extreme binding specificity, distinguish the lysins of Gram-positive microorganisms.

Unlike with Gram-negative peptidoglycan, the surface-exposed nature of Gram-positive peptidoglycan renders it vulnerable to hydrolysis by exogenously applied lysins. We have now demonstrated this “lysis from without” using an array of different lysins specifically active against a broad range of Gram-positive bacteria (4). Based on these findings, our laboratory and others are actively pursuing development of lysins as novel therapeutic agents (or alternatives to conventional antibiotics) active against Gram-positive pathogens. Additionally, we are seeking to exploit the binding capacity of these enzymes as the basis for sensitive diagnostic methods. The genetic and biochemical methods developed for these studies are presented.

2. Materials

2.1 Phage, plasmids, and bacteria

1. The γ phage, pBAD24 (5), and Bacillus cereus strain RSVF1 are part of The Rockefeller University Collection.

2. E. coli XL2-Blue ultracompetant cells (Stratagene; La Jolla, CA; http://www.stratagene.com/)

2.2 Equipment

1. Run One Electrophoresis Multicasting System (EmbiTec; San Diego, CA; http://www.embitec.com/)

2. Model C24 incubator shaker (New Brunswick Scientific; Edison, NJ.; http://www.nbsc.com/Main.asp)

3. SpectraMax Plus spectrophotometer (Molecular Devices; Sunnyvale, CA; http://www.moleculardevices.com/)

4. Model 5810R tabletop centrifuge (Eppendorf)

5. Orbital Shaker (Bellco Biotechnology Inc.; Vineland, NJ; http://www.bellcoglass.com/)

6. EchoTherm Model IC20 (Torrey Pines Scientific; San Marcos, CA; http://www.torreypinesscientific.com/) for incubations at 16°C and 65°C.

7. UV transluminator (Alpha Innotech Corp.; San Leandro, CA; http://www.alphainnotech.com/)

8. Dual chamber water bath (Precision)

2.3 Supplies

1. Lambda Maxi Kit (Qiagen Inc.; Valencia, CA; http://www1.qiagen.com/)

2. All DNA modifying enzymes and corresponding buffers were purchased from New England Biolabs (Ipswich, MA; http://www.neb.com/nebecomm/default.asp).

3. 1.5 ml Eppendorf tubes

4. Phenol, chloroform, NH₄OAc, ethanol, ethidium bromide, NaOH, Tris, L-arabinose, phosphate buffered saline (PBS), LB media, and ampicillin were all from Sigma-Aldrich (St. Louis, MO; http://www.sigmaaldrich.com/).

5. 1 kb DNA ladder (Invitrogen Corp.; Carlsbad, CA; http://www.invitrogen.com/)

6. TAE buffer

7. Agarose (Cambrex Corp.; http://www.cambrex.com/default.asp)

8. NucleoSpin DNA Extraction Columns and NucleoSpin Plasmid Kit (BD Biosciences; San Jose, CA; http://www.bdbiosciences.com/)

9. 150 × 15 mm polystyrene and 150 mm glass Petri dish (Fisher, Inc.)

10. Replica transfer apparatus and velvet squares (Cat. No. 11DOTM001; MP Biochemicals; Solon, OH; http://www.mpbio.com/)

11. 10 ml round bottom culture tube (Sarstedt)

12. Brain Heart Infusion media (Difco)

13. 96-well microtiter plate (Costar)

14. 125 ml and 2 L Ehrlenmeyer flasks (VWR)

15. NucleoSpin DNA Extraction Columns (Clontech – a Takara Bio Company; Mountain View, CA; http://www.clontech.com/clontech/

3. Methods

The potent bacteriolytic activity of phage lysins provides the basis by which their coding sequences may be identified (6). Toward this end, we have developed an efficient means to detect lysin activity by screening induced plasmid expression libraries for agents that lyse live bacterial cells. We describe here one such activity screen for identification of a lysin, PlyG, from the purified bacteriophage, γ. The γ phage is a diagnostic tool used in clinical laboratories to identify the Gram-positive pathogen Bacillus anthracis, although it also infects some highly related B. cereus isolates like RSVF1 (7).

Following our description of lysin identification, we detail the processes by which we define and quantify lysin activity. Owing to the conserved, repeating structure of bacterial peptidoglycan, lysin catalytic domains are limited to one of three enzymatic activities: N-acetylglucosaminidase (lysozyme), N-acetylmuramoyl-L-alanine amidase, or endopeptidase. Unfortunately, because the catalytic domains from Gram positive phage lysins are not enzymatically active in vitro there is no assay involving simple chomogenic or flurogenic substrates that mimic target cell wall bonds and allow kinetic measurements or even simple activity quantitation. A higher order structure, perhaps one containing the CBD carbohydrate binding epitope in association with peptidoglycan, is apparently necessary for activity. Thus, the lysins require either purified cell walls, or whole cells as substrates for in vitro analyses. As result, we have adapted a turbidimetric assay (8) to titer lysin activity and define a standard unit. This method is based on observing zones of clearing on lawns of bacterial cells and produces results which are reproducible and allow direct comparison of activity between lysins active on different species.

It is important to note that the details of lysin identification, purification, and analysis will require slight, yet significant, alterations based on variables like the intended lysin source (i.e., is it encoded within a lytic phage, bacterial genome (prophage), or complex environmental sample) and activity target (i.e., Gram-positive vs. Gram-negative organism, fast or slow growing, etc.). Variations that may be incorporated into a lysin screen are described in the Notes. Otherwise, the methods described here were specifically developed to identify the PlyG lysin encoded within the purified genomic DNA of γ phage, although this method has also identified lysins active against Bacillus cereus, B. thuringiensis, Enterococcus faecalis, E. faecium, Listeria monocytogenes, Streptococcus pyogenes, S. pneumoniae, and S. agalactiae (7, 9–13).

3.1 Construction of a plasmid expression library encoding the partially digested γ phage genome

1. Prepare total genomic DNA from 250 ml of high-titer (~1×10⁹ PFU ml⁻¹) γ phage lysate using the Qiagen Lambda Maxi Kit (see ^{Note 1}). Resuspend purified DNA in distilled water at a final concentration of 500 ng μl⁻¹.

2. Prepare five 1.5-ml Eppendorf tubes, each containing 5 μg of phage DNA in 50 μl 1X New England Biolabs (NEB) Buffer 2. Add the NEB restriction endonuclease Tsp509I (see ^{Note 2}) to final enzyme unit amounts of 10, 5, 2.5, 1.0 and 0.5, respectively. Gently suspended the enzyme and incubate tubes for 5 min at 65°C.

3. Since Tsp509I cannot be heat inactivated, perform the following: Immediately add 50 μl phenol:chloroform (1:1 mixture) to each tube, vortex for 10 sec, and centrifuge for 5 min at 4°C in a table-top microcentrifuge (maximum speed). Recover the upper DNA phase, add 50 μl chloroform, and again vortex and centrifuge. To the resultant upper phase, add 50 μl 5M NH₄OAc, mix by inversion, add 250 μl 100% ethanol, and mix again. Centrifuge for 15 min at 4°C, wash each pellet with 70% ethanol, and resuspend in 15 μl dH₂0.

4. Prepare a 0.4 mm thick, 1% agarose gel. Add gel-loading buffer to each tube and load each (as well as a lane for 1 Kb DNA ladder) into gel with 1X TAE running buffer. The bromphenol blue dye front should be run to the bottom of the gel.

5. Stain gel for 15 min in 1X TAE containing 0.5 μg ml⁻¹ ethidium bromide and observe on a UV transilluminator, keeping exposure time to a bare minimum. With a clean razor blade, carefully excise gel segments containing partially digested DNA fragments in only the 500 to 2,000 bp range (see ^{Note 3}).

6. Recover DNA from agarose slices using 4–8 NucleoSpin DNA Extraction Columns and elute each in a final volume of 50 μl distilled water (dH₂0). Pool samples and run 10 μl on a 1% agarose gel for quantitation of DNA recovery - a DNA smear should be seen in the 500–2000 bp range.

7. Set up three 15 μl ligations in 0.2 ml tubes containing 1X ligation buffer, 10 units of T4 DNA ligase and ~10 ng of plasmid pBAD24 (previously linearized with EcoRI and dephosphorylated with antarctic phosphatase) and either 50, 100, or 200 ng, respectively, of Tsp509I-digested phage DNA. Set up a fourth, ligation, in which all phage DNA is omitted, is established as a self-ligation control.

8. After overnight incubation at 16°C, transform 2 μl of each ligation into 100 μl XL2-Blue ultracompetant E. coli following the manufacturer’s protocol (Stratagene, Inc.).

9. Plate cells on LB agar supplemented with 100 μg ml⁻¹ ampicillin, at a density yielding ~300 distinct colonies per 150 × 15 mm polystyrene Petri dish after overnight incubation at 37°C. Ultimately, 10 plates, each containing ~300 colonies, should be obtained (see ^{Note 4}).

10. The quality of the γ phage expression library, with respect to insert size and diversity, must be assessed. First, isolate total DNA from 25 random transformants by suspending each colony in 25 μl 0.5 M NaOH, and adding 25 μl 1 M Tris pH 8.0, 450 μl dH₂0, and mixing vigorously. Next, use 1.0 μl DNA samples as template in PCR reactions with primers that flank the pBAD24 MCS (see ^{Note 5}). Run 7.5 μl of each reaction on a 1% agarose gel to assess the range of insert sizes. We observed that 24 of 25 transformants had γ phage inserts (from 0.3 to 2.5 kb in size) with no bias toward any particular size as expected for a truly random library (see ^{Note 6}).

3.2 Screening the γ phage library for lysin activity

1. The γ phage expression library consists of ~3000 E. coli transformants on ten 150 × 15 mm LB plates containing ampicillin.

2. Using 150 mm velvet squares and a replica transfer apparatus, replica plate library onto ten 150 mm glass Petri dishes, each consisting of 90 ml LB agar with 100 μg μl⁻¹ ampicillin and 0.2% L-arabinose (see ^{Note 7}).

3. Mark the master and replica plates with an alcohol-resistant marker at such a position to allow accurate alignment of plate pairs when recovering positive clones.

4. Incubate plates overnight at 37°C for the induced library to grow in.

5. On the day prior to the activity screen, set up a 5 ml LB liquid culture with B. cereus strain RSVF1 (see ^{Note 8}) and incubate overnight at 30°C shaking at 150 rpm.

6. On the day of the activity screen, have available: 55°C LB soft agar (0.75% agar) and a set of ten 10 ml round-bottom culture tubes each containing 100 μl of B. cereus RSVF1 overnight culture and kept at 37°C.

7. Permeabilize induced E. coli expression clones with chloroform vapors as follows: in a chemical fume hood, add 10 ml chloroform to the inverted lid of each glass plate, then invert E. coli clones over the chloroform and incubate for 10 min, then place the clones face up for 10 min to allow chloroform evaporation (see ^{Note 9}).

8. Add 7.5 ml LB soft agar to an RSVF1-containing tube and pour contents over the clones on the glass plate, rapidly rotating to allow overlay of the entire surface (see ^{Note 10}). Repeat for each glass plate.

9. Incubate plates at room temperature for 4 hours (the RSVF1 lawn may become visible as a very faint haze at this point). Place at 4°C overnight and return to room temperature the following day to allow the RSVF1 to grow in.

10. Lysin-encoding clones are distinguished by the appearance of distinct RSVF1 clearing zones in the overlay surrounding such clones (see Figure 1). If no zones become apparent after incubation at room temperature, incubate overnight at 4°C again, then return to room temperature and continue search for clearing zones (see ^{Note 11} for alternative methods to screen for lysins). We detected 52 positive clones in our screen (1.7% of the library). In similar screens for other Siphoviridae lysins, we observed rates between 0.1 and 2%.

11. Recover master E. coli colonies corresponding to positive clones by aligning glass plate-master polystyrene plate pairs. Streak to single colonies on LB agar plates containing 100 μg μl⁻¹ ampicillin and incubate overnight at 37°C.

12. Restreak each clone on LB agar with ampicillin and inoculate 5 ml LB liquid with ampicillin and incubate overnight at 37°C shaking at 200 rpm.

13. Use the LB agar plate cultures to generate frozen stocks (see ^{Note 12}).

14. Use the LB liquid cultures to prepare plasmid using the NucleoSpin Plasmid Kit. Sequence resulting plasmids with both BAD1 and BAD3 primers (see ^{Note 5}).

15. Assemble DNA sequences and examine in silico to determine whether a lysin has been identified (see ^{Note 13}).

Fig. 1.

Lytic activity screen for the identification of γ phage lysin, PlyG. The induced and permeabilized E. coli expression library is overlaid with agar containing B. cereus strain RSVF1. Owing to the presence of ampicillin in the LB agar, the ampicillin-sensitive RSVF1 grows only near E. coli colonies expressing plasmid-encoded β-lactamase. A clearing zone in the RSVF1 overlay, surrounding one of the library clones (indicated by an arrow), identifies the presence of a cloned lysin.

3.3 Expression of the PlyG lysin

1. To create PlyG expression strain, clone the entire 702 bp plyG ORF (include no flanking γ phage sequence) into plasmid expression vector pBAD24 and transform E. coli strain XL1-Blue (see ^{Note 14}).

2. Inoculate XL1-Blue/pBAD24::plyG into 15 ml LB liquid culture with 100 μg ml⁻¹ ampicillin (in a 125 ml Ehrlenmeyer flask) and shake at 30°C overnight at 150 rpm.

3. Dilute overnight culture 1:100 into 1 liter of LB liquid with ampicillin (in a 2 liter Ehrlenmeyer flask) and shake at 225 rpm for 3 h at 37°C.

4. Add the L-arabinose inducer to a final concentration of 0.2% and continue growth at 30°C overnight (~12 h) (see ^{Note 15}).

5. Wash culture by spinning down cells at 4000 rpm in an Eppendorf tabletop centrifuge and then adding a 1X volume of 1X PBS to the pellet, remove the supernatant and finally resuspend in 50 ml 1X PBS.

6. For bacterial lysis, add chloroform to a final concentration of 20%, vortex for 1 min, gently agitate in an orbital shaker for 1 hour at 4°C, and vortex again for 1 min (see ^{Note 16}).

7. Pellet bacterial debris for 15 min at 4°C and 4000 rpm in an Eppendorf tabletop centrifuge and carefully remove supernatant.

8. The supernatant contains crude PlyG, which can be assayed for activity according to section 3.4 below. (see ^{Note 17}).

3.4 Quantifying lysin activity

1. Grow B. cereus strain RSVF1 overnight in 10 ml Brain Heart Infusion (BHI) broth at 30°C shaking at 150 rpm.

2. The next morning, start a 50 ml culture of RSVF1 from a 1:100 dilution of the overnight in fresh BHI media and incubate as above.

3. At exactly 3 hours of incubation for this culture (see ^{Note 18}), harvest the cells by centrifugation in 50 ml Falcon tubes for 15 min at 4°C and 4000 rpm in an Eppendorf tabletop centrifuge.

4. Resuspend the pellet and wash 1X in 10 mls of phosphate buffered saline (PBS) (see ^{Note 19}).

5. Resuspend the pellet and adjust the final OD₆₀₀ to 1.0 by the addition of PBS.

6. Prepare PlyG for assay by diluting PlyG 1:10 in PBS and making additional 2 fold dilutions in PBS (i.e. final dilutions will be 1:10, 1:20, 1:40, 1:80, etc.).

7. Pre-warm a 96-well microtiter plate and Molecular Devices spectrophotometer to 37°C (see ^{Note 20}).

8. Add 100 μl of the PlyG dilutions in duplicate to the 96-well plate. Include a “no PlyG” control with 100 μl PBS only.

9. At time zero, add 100 μl aliquots of the B. cereus strain RSVF1 to each well using a 12-channel micropipette.

10. Agitate for 5 sec and take an OD₆₀₀ endpoint reading. Since the starting OD₆₀₀ was adjusted to 1.0, and these cells were mixed 1:1 with PlyG/PBS, all wells should now have a reading of ~0.5.

11. Allow the door of the spectrophotometer to close, incubate at 37°C with continual agitation, and take another OD₆₀₀ endpoint reading at time=15 min.

12. Measurement of lysin activity is based upon turbidimetric determination of cell lysis. 15 min endpoint readings should form a range from 0.5 (for controls or very dilute PlyG wells) to 0.1 for concentrated PlyG wells. Thus, we define the standard lysin “unit” as the highest dilution that produces a half-drop in OD in the 15 min assay. For example, if an 80-fold dilution of PlyG produced a drop in OD₆₀₀ of B. cereus strain RSVF1 from 0.5 to 0.25 in 15 min, then we say that PlyG has a titer (or activity) of 80 U/ml for this strain (see ^{Note 21}).

13. It is our experience that most lysins, when purified to homogeneity, will have a specific activity of ~ 1 U per μg protein, although some very active lysins may have a specific activity 100 or 1000 times higher.