Expression and Functional Characterization of the First Bacteriophage-Encoded Chaperonin

Abstract

Chaperonins promote protein folding in vivo and are ubiquitously found in bacteria, archaea, and eukaryotes. The first viral chaperonin GroEL ortholog, gene product 146 (gp146), whose gene was earlier identified in the genome of bacteriophage EL, has been shown to be synthesized during phage propagation in Pseudomonas aeruginosa cells. The recombinant gp146 has been expressed in Escherichia coli and characterized by different physicochemical methods for the first time. Using serum against the recombinant protein, gp146's native substrate, the phage endolysin gp188, has been immunoprecipitated from the lysate of EL-infected bacteria and identified by mass spectrometry. In vitro experiments have shown that gp146 has a protective effect against endolysin thermal inactivation and aggregation, providing evidence of its chaperonin function. The phage chaperonin has been found to have the architecture and some properties similar to those of GroEL but not to require cochaperonin for its functional activity.

INTRODUCTION

Chaperonins are known to promote the correct folding of newly synthesized polypeptides and to prevent aggregation of proteins denatured under stress. Chaperonins are large cylindrical oligomers consisting of two rings arranged back to back. Each ring contains a central cavity, where unfolded proteins may be encapsulated and undergo productive folding in an ATP-dependent manner (19, 22, 23).

Based on protein sequence and structural features, chaperonins fall into two groups. Group I chaperonins (including heat shock proteins such as hsp60) are found in bacteria, endosymbiotic organelles (mitochondria and chloroplasts), and subsets of archaea (17, 22). Group II chaperonins are found in archaea and the eukaryotic cytosol (4, 14).

Chaperonin GroEL from Escherichia coli, which carries out its function with a cofactor, GroES, is the most studied of group I chaperonins (4, 14, 22, 45). GroEL consists of 14 identical 57-kDa subunits arranged in two heptameric rings (6, 41). GroES 10-kDa subunits also form a heptameric ring that acts as a detachable lid of the central cavity when GroEL forms a complex with GroES (37, 44). Group II chaperonins act independently of the detachable cofactor. The helical protrusion at the tip of the apical domain substitutes for the cofactor as a built-in lid of the central cavity (29). Group II chaperonins are almost always more complicated than group I chaperonins in the subunit composition of their complexes, which are hetero-oligomers rather than homo-oligomers (4, 14).

Some bacteriophages (λ, T4, RB49) are known to use host GroEL for folding their capsid proteins during morphogenesis. While phage λ uses both host GroEL and GroES, the GroES cochaperonin function in T4 and RB49 is performed instead by phage-encoded orthologs gp31 and CocO, respectively (2, 21, 42). As revealed by structural analysis and later by cryoelectron microscopy, the gp31 cap may create a somewhat larger folding cage with GroEL than with the GroES cap to accommodate the relatively large 56-kDa T4 major capsid protein (3, 12, 27).

The first chaperonin GroEL ortholog gene has been identified in the genome of Pseudomonas aeruginosa bacteriophage EL (211,215 bp, 201 open reading frames [ORFs]) (24), related to the giant phiKZ-like Myoviridae (11). The putative GroEL-like chaperonin protein (558 amino acids [aa]; GenBank accession number 5176674) encoded by gene 146 has only about 25% amino acid sequence identity with E. coli GroEL (547 aa) and P. aeruginosa GroEL (548 aa), while these bacterial chaperonins show 80% mutual sequence identity. Nevertheless, conservation of most of the residues involved in ATP/ADP and Mg²⁺ binding (5, 9, 44) suggests that gp146 belongs to this ubiquitous family of proteins. The two other GroEL orthologs encoded by phage genomes were predicted (30, 13), but none of them have been experimentally studied yet.

In the present work, chaperonin activity of a bacteriophage-encoded protein is demonstrated for the first time by functional and physicochemical characterization of the putative chaperonin, gp146, encoded by the genome of P. aeruginosa bacteriophage EL.

MATERIALS AND METHODS

Cloning, expression, purification, and antibody preparation of gp146.

The gene 146 was amplified from EL genomic DNA using the forward primer 5′-GTACGCATATGTCTCAAACGCTACTG-3′ and the reverse primer 5′-TCCTTTAGGATCCCCTTACCGCACCTT-3′, which contained recognition sequences for NdeI and BamHI (underlined), respectively. The amplicon was cloned into the pET-22b(+) vector (Novagene, United States) and expressed in E. coli BL21(DE3) as previously described (40). The transformants were cultured in 600 ml of 2× tryptone-yeast extract (TY) medium containing 200 μg/ml of ampicillin at 37°C. When the A₆₀₀ reached 0.7, the recombinant bacteria were induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and incubated for 3 h at 25°C. The cells were harvested by centrifugation at 2,500 × g for 10 min (Megafuge 2.0 R; Heraeus Instruments, Germany).

To purify gp146, the cell pellet was resuspended in 6 ml of 50 mM Tris-HCl (pH 7.5), sonicated for 2 to 3 min using a Virsonic 100 disintegrator (Virtis), and centrifuged at 12,000 × g (Eppendorf, Germany) for 10 min to remove the debris. Nucleic acids were precipitated by addition to the supernatant of 1/10 volume of 30% (wt/vol) streptomycin sulfate solution followed by centrifugation. The recombinant protein was precipitated from the supernatant by addition of saturated ammonium sulfate to a final concentration of 30% (wt/vol). The protein precipitate was pelleted by centrifugation as above, dissolved in 40 ml of 50 mM Tris-HCl (pH 7.5) containing 100 mM NaCl, and loaded on a Q-Sepharose (Bio-Rad, United States) column equilibrated with the same buffer. After the column was washed with 10 volumes of buffer, the protein was eluted by a linear gradient from 100 to 500 mM NaCl in 50 mM Tris-HCl (pH 7.5) and analyzed by SDS–10% PAGE (31). Fractions containing pure protein were concentrated using an Amicon Ultra-15 centrifugal filter device (molecular weight cutoff [MWCO], 100,000; Millipore, United States). The gp146 concentration was determined spectrophotometrically at 280 nm using a theoretical absorption coefficient of 35,870 M⁻¹ cm⁻¹.

To produce antibodies, female BALB/c mice (14 to 17 g) were immunized intraperitoneally with 0.2 ml of a suspension containing 50 to 100 μg of pure gp146 in 150 mM NaCl and an equal volume of complete Freund's adjuvant. Two booster injections were given with the same protein preparation and incomplete Freund's adjuvant at two- to three-week intervals. Serum was recovered 7 days after the third immunization. As determined by enzyme-linked immunosorbent assay (ELISA), the titer of antibody was 1:250,000.

Western blot analysis.

After separation by SDS-PAGE, proteins were transferred by electroelution from the gel onto a nitrocellulose membrane (Bio-Rad) in electroblotting buffer (0.1 M Tris, 0.1 M boric acid, 0.01 M EDTA) at 200 mA for 60 min. Antigen was detected with antiserum followed by rabbit anti-mouse horseradish peroxidase (HRP)-conjugated antibodies (1:6,000; Sigma, United States). Antibodies were diluted in phosphate-buffered saline (PBS) supplemented with 3% milk and 0.05% Tween 20, and incubations were for 1 h at room temperature with gentle rocking. Blots were developed with 0.02% 3′,3′-diaminobenzidine (Sigma) in hydrogen peroxide solution.

Preparation of lysate from the EL-infected cells.

The P. aeruginosa PAO1 cells were grown in 150 ml of 2× TY medium at 37°C to a density of 2 × 10⁸ cells/ml, infected by bacteriophage EL at a multiplicity of infection of 5, and incubated for 40 min at 37°C. The cells were harvested by centrifugation at 2,500 × g for 10 min. The pellet was resuspended in 50 mM Tris-acetate (pH 8) and 5 mM EDTA and sonicated. NaCl and phenylmethylsulfonyl fluoride (PMSF) were added to final concentrations of 140 mM and 1 mM, respectively. The lysate preparation was centrifuged at 12,000 × g for 10 min.

Immunoprecipitation of gp146 from the lysate.

Serum against gp146 was incubated with insoluble protein A (Sigma) in 50 mM Tris-acetate (pH 8) for 2 h at 4°C with gentle rocking. Protein A with attached antibodies was centrifuged at 4,000 × g for 5 min and subsequently washed in solution I (50 mM Tris-acetate [pH 8], 5 mM EDTA, 150 mM NaCl, 0.5% sodium deoxycholate, 1% Triton X-100), solution II (50 mM Tris-acetate [pH 8], 5 mM EDTA, 500 mM NaCl, 1% Triton X-100), solution III (50 mM Tris-acetate [pH 8], 5 mM EDTA, 82 mM KCl), and 50 mM Tris-acetate (pH 8) and 5 mM EDTA.

To precipitate phage chaperonin, protein A with attached antibodies against gp146 was incubated with the lysate of the EL-infected bacterial cells (45 min postinfection) at 4°C for 2.5 h. Nonbound proteins were removed by washing with solutions I to III. The pellet of protein A with bound immune complex was supplemented with SDS-sample loading buffer, incubated at 98°C for 5 min, and analyzed by SDS–10% PAGE. A sample prepared in parallel from the lysate of noninfected P. aeruginosa PAO1 cells was used as a control.

Coimmunoprecipitation of gp146-substrate complexes and substrate identification.

Coimmunoprecipitation of chaperonin-substrate complexes was performed as described above, except that specifically bound proteins were eluted from the protein A with 50 mM Tris-acetate (pH 8) in the presence of 5 mM MgCl₂ and 3 mM ATP at 37°C and then separated by SDS–10% PAGE. Gel lanes were cut into slices and subjected to in-gel trypsin digestion. Peptides were extracted and subsequently identified by liquid chromatography-electrospray ionization-tandem mass spectrometry (MS) analysis on an LCQ Classic (Thermo Electron Corporation, San Jose, CA) as described previously (16, 32).

Edman degradation.

The individual protein band of immunoprecipitated gp146, after separation by SDS-PAGE followed by transfer onto an Immobilon membrane (Millipore), was subjected to N sequencing. The N-terminal amino acid sequence was determined with a Procise cLC 491 protein sequencing system (PE Applied Biosystems, United States). Phenylthiohydantoin derivatives of the amino acids were identified by a 120A PTH analyzer (PE Applied Biosystems).

Cloning, expression and purification of gp188.

The gene 188 was amplified from EL genomic DNA using the forward primer 5′-TAAATACATATGAACTTCCGGACGAAG-3′ and the reverse primer 5′-GGAAAACTCGAGGAATCAATACGAAATAACGTG-3′ (NdeI and XhoI recognition sites are underlined). The PCR product was cloned into the pET-28b(+) vector (Novagene) and expressed in E. coli cells at 25°C. The recombinant protein containing a 6×His affinity tag at the N terminus (gp188*) was purified on His-Select nickel affinity gel (Sigma), dialyzed against 50 mM Tris-HCl (pH 7.5), and concentrated using an Amicon Ultra-15 centrifugal filter device (MWCO, 30,000). Protein concentration was determined spectrophotometrically at 280 nm using a theoretical absorption coefficient of 58,455 M⁻¹ cm⁻¹.

Endolysin activity assay.

The muralytic activity of endolysin (gp188*) was determined as previously described (7). The P. aeruginosa PAO1 cells with the outer membrane permeabilized by chloroform treatment were used as a substrate. A 30-μl sample of protein solution was mixed with 270 μl cells, and optical density was measured at 490 nm for 30 min using a Wallac 1420 microbiology reader (PerkinElmer, United States). The activity was determined from the decrease in absorbance as a function of time using a standardized calculation method. Negative controls (buffer without enzyme) were subtracted from sample measurements.

Analytical ultracentrifugation.

Sedimentation-velocity experiments were carried out at 22°C using a model E analytical ultracentrifuge (Beckman Instruments, United States) equipped with photoelectric scanning absorption optical system, with spinning rates of 20,000 rpm for gp146 and 48,000 rpm for gp188* and scanning at 280 nm. Protein concentration was 1 mg/ml. Data analysis was carried out using the SEDFIT program (39).

Hydrodynamic diameter determination.

The hydrodynamic diameters of proteins were determined using dynamic light scattering (DLS). All experiments were carried out at 25°C on a ZetaSizer NanoZS instrument (Malvern) with a laser wavelength of 633 nm. Data were analyzed using Dispersion Technology software, version 5.10.

Estimated protein density was calculated on the assumption of spherical particles from the equation ρ = M_w/4πN_ar_H³, where M_w is molecular weight, N_a is Avogadro's number, and r_H is the hydrodynamic radius.

Electron microscopy.

Purified protein was loaded on a Pioloform-coated copper grid. The specimen was then contrasted with 1% uranyl acetate and observed in a Jeol (Japan) JEM 1400 electron microscope.

DSC.

The thermal denaturation of proteins was investigated by differential scanning calorimetry (DSC) using a DASM-4 microcalorimeter (Biopribor, Russia) with 0.47-ml capillary platinum cells. All of the measurements were carried out at a heating rate of 1°C min⁻¹ in the temperature range from 5 to 90°C and at a constant pressure of 2 atm. Curves were corrected for time response as earlier described (35). The second heating was used as the instrument baseline because of irreversible denaturation found for all samples. The chemical baseline was calculated and subtracted using Origin 1.16 software (MicroCal, Inc., United States). The standard reaction buffer used in the experiments was 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, and 100 mM KCl. The concentration of gp146 was 0.58 μM. Protein concentration and molar excess heat capacity were calculated on a tetradecameric complex with a molecular mass of 863.8 kDa. To investigate the influence of nucleotide binding on gp146 structure, ATP or ADP (Sigma) was added to the protein solution to a final concentration of 1 mM.

Fluorescence measurements.

Fluorescence spectra of protein solutions were measured in a 2-ml quartz cell at various temperatures with a FluoroMax-3 spectrofluorometer (Horiba Jobin Yvon, France) using an excitation wavelength of 280 nm. The samples were loaded in the cell and heated to 75°C at the average rate of 2°C/min. The emission spectra were recorded at 300 to 400 nm at the corresponding temperatures. Experimental curves representing the temperature dependence of wavelength of the emission maximum (λ_max) were fitted using a Boltzmann sigmoid curve.

CD spectroscopy.

Circular dichroism (CD) measurements were performed using a Chiroscan CD spectrometer (Applied Photophysics, United Kingdom) in a 0.5-mm cell. Spectra were recorded in the range of 200 to 260 nm and were baseline corrected by subtracting the buffer spectrum. Each point was measured for 1 s. The observed value was converted to mean residue ellipticity.

Isothermal titration calorimetry (ITC).

Nucleotide binding was investigated on a VP-ITC instrument (MicroCal Ltd., United States) with a 1.4-ml cell. ATP and ADP concentrations were measured spectrophotometrically using a UV-1601 instrument (Shimadzu Scientific Instruments Inc., Japan) at 259 nm with an extinction coefficient of 15,400 M⁻¹ cm⁻¹. All experiments were carried out at 10°C in 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, and 100 mM KCl. Titration experiments were performed by successive 10-μl injections of 40 μM ATP or 60 μM ADP solution into gp146 (0.58 μM for the tetradecamer), and the interval between injections was 5 min. All samples were degassed before the experiment. Binding isotherms were corrected by subtracting the ligand dilution isotherms, determined by titrating ATP and ADP solutions into buffer. Data analysis was carried out using MicroCal Origin 7.0 software with the “one set of sites” model. Resulting coefficients of determination were calculated from the equation, R² = 1 − $\frac{{\displaystyle {\sum }_i{\left({\text{Exp}}_i\right)}^2}}{{\displaystyle {\sum }_i{\left({\text{Exp}}_i-\text{mean}\left(\text{Exp}\right)\right)}^2}}$, where Exp and Fit are experimental and fitted ordinate arrays, respectively, and exceeded 0.99.

gp188* thermal inactivation.

The gp188* samples were incubated at 0.1 μM in 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, and 100 mM KCl in the absence or presence of 0.2 μM gp146 (for the tetradecamer) and 3 mM ATP at 37 and 50°C. At various times, aliquots were withdrawn and chilled on ice and the muralytic activity was determined at 23°C as described above.

gp188* aggregation assay.

The effect of gp146 on gp188* aggregation was investigated in 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, and 100 mM KCl at 45°C using DLS. The gp188* samples at a constant concentration of 3 μM were incubated with different concentrations of gp146 in the absence or presence of 1.2 mM ATP for a few hours, and the hydrodynamic diameters of particles were continuously measured. The molar ratio of gp188* to gp146 was calculated for the monomer and tetradecamers, respectively.

RESULTS

Expression and purification of gp146.

A plasmid construct for expression of gene 146 in E. coli cells was designed. The bacterial expression of gp146 (61.7 kDa) was carried out at a lowered temperature (25°C) to increase the yield of soluble protein. Low-temperature induction is known to result in a significant improvement in solubility by slowing down the rate of protein expression and allowing time for folding (43). The recombinant protein was precipitated from the cell lysate by adding ammonium sulfate, followed by purification on a Q-Sepharose column (Fig. 1), and used for antibody preparation. It has been shown by Western blot analysis that the polyclonal antibodies obtained specifically bind to the recombinant gp146 but recognize neither bacterial nor EL structural proteins (not shown).

Fig 1.

SDS–10% PAGE of expressed gp146 (lane 1) and gp146 purified by precipitation with 30% ammonium sulfate (lane 2) and subsequent chromatography on Q-Sepharose (lane 3). Lane M, protein marker.

Analysis of gp146 expression in vivo.

According to the structural proteome analysis of bacteriophage EL, gp146 is not part of the mature phage particle (33). However, it has been found during phage EL propagation. To investigate gp146 expression in vivo, the proteins from EL-infected bacterial cells at different infection stages were resolved by SDS-PAGE followed by Western blot analysis. An apparent band (about 60 kDa) likely corresponding to gp146 was detected from 15 min postinfection onwards with serum against the recombinant protein (not shown). In vivo-expressed native gp146 was immunoprecipitated from the clarified lysate of the EL-infected bacterial cells (45 min postinfection) using antibodies against the recombinant protein attached to insoluble protein A. A similar procedure was performed simultaneously with an equivalent amount of lysate from noninfected cells. The immunoprecipitated complexes were supplemented by SDS-sample loading buffer and heated. Upon this treatment, the associated proteins were released from protein A into solution and separated by SDS-PAGE (Fig. 2). An additional band with an electrophoretic mobility coinciding with that of the recombinant gp146 was found in the protein pattern from the EL-infected cells (Fig. 2A, lane 3) compared to a control. No additional bands, except for the bands typical of the heavy and light chains of antibodies (Fig. 2A, lane 1), were observed in the precipitate from noninfected cells (Fig. 2A, lane 2). This band was clearly revealed by immunoblotting with serum against the recombinant gp146 (Fig. 2B). The identity of the immunoprecipitated protein to gp146 was confirmed by N-terminal amino acid sequencing. Six amino acid residues were determined, and all of them were completely identical to the gp146 primary structure. Thus, gp146 is really synthesized in vivo during bacteriophage EL propagation in P. aeruginosa cells. Further investigations of the recombinant protein were performed. To test whether gp146 can function as a chaperonin, it was necessary to find its native substrate.

Fig 2.

Immunoprecipitation of native gp146 expressed in vivo from the lysate of EL-infected P. aeruginosa cells (45 min postinfection) using serum attached to protein A. Proteins were resolved by SDS–10% PAGE and visualized in the gel by staining with Coomassie brilliant blue (A) and serum against gp146 after being transferred from the gel onto a nitrocellulose membrane (B). Shown are protein patterns of serum (lanes 1) and immunoprecipitated complexes from the lysate of noninfected (lanes 2) and EL-infected (lanes 3) bacterial cells. Lane M, protein marker; lanes c, recombinant gp146.

Isolation and identification of substrate for gp146.

Antibodies against gp146 attached to protein A were further used for coimmunoprecipitation of the putative chaperonin along with the substrate proteins associated with it from the lysate of EL-infected bacteria. To rule out nonspecific protein binding to antigen, the immune complexes were thoroughly washed with buffers containing detergents and NaCl (see Materials and Methods). To ensure ATP-dependent release of substrate proteins from gp146, the precipitated complexes were incubated with ATP- and Mg²⁺-containing buffer at an elevated temperature. Eluted proteins were separated by SDS-PAGE. The gel lane was cut into slices and subjected to in-gel trypsin digestion. Peptides were extracted and separated by liquid chromatography followed by MS analysis. One of the phage EL proteins, gp188 (292 aa, 32.5 kDa), was identified. A recombinant protein with a 6×His affinity tag at the N terminus (gp188*, 34.3 kDa) was expressed in E. coli cells in a soluble, enzymatically active form and purified by metal chelate chromatography for further study. Using serum against the recombinant gp188*, gp188 expression was observed during phage EL multiplication in bacterial cells, but 45 min later than that of gp146 (not shown).

Architecture of gp146 and gp188*.

The structural characteristics and homogeneity of the purified recombinant gp146 and gp188* were investigated using analytical ultracentrifugation and DLS. One-component raw sedimentation profiles were observed for both proteins (Fig. 3A). The sedimentation coefficients equaled 23.5 ± 0.3 S and 3.3 ± 0.3 S for gp146 and gp188*, respectively (Fig. 3B and C). The fitting of the frictional coefficient for gp146 yielded a value of 1.39, suggesting a nonspherical form of the protein. Based on the calculated values of the sedimentation and frictional coefficients, the protein's molecular mass was estimated. Its value (872 kDa) most likely corresponds to a tetradecamer, considering that the calculated molecular mass of the gp146 monomer is 61.7 kDa. The estimated molecular mass of gp188* (35 kDa) is close to that of a protein monomer. According to DLS data, the hydrodynamic diameters of gp146 and gp188* are about 13.2 and 5.6 nm, respectively (not shown). Estimation of the gp146 density was 1.191 g/ml, which is lower than the mean density of a typical protein with the same molecular weight (18), suggesting the presence of an inner cavity in the complex. An electron microscopy assay confirmed that the phage chaperonin is composed of double-stacked heptameric rings with a central cavity (Fig. 4). Thus, the architecture of the recombinant gp146 is close to that of GroEL from E. coli (6).

Fig 3.

Sedimentation data of recombinant gp146 (1.2 μM) and gp188* (14.6 μM) at 22°C. (A) Raw data for gp146; every third curve is shown. (B and C) Fitting of sedimentation coefficients for gp146 and gp188*, respectively.

Fig 4.

Electron microscopic images of the purified recombinant gp146 negatively stained with 1% uranyl acetate. The inset demonstrates the 7-fold symmetry of the chaperonin complex.

Thermal denaturation of gp146 and gp188*.

Thermal denaturation of the recombinant proteins was analyzed by DSC. The DSC profile of the putative chaperonin showed a cooperative endothermic transition with a maximum at 62.1°C and a calorimetric enthalpy change of about 12,600 kJ/mol for the complex (Fig. 5A; Table 1). The calorimetric transition was found to be irreversible (after heating to 90°C and a subsequent cooling to 5°C, no transition was observed during the second scan) and concentration independent (2 to 16 μM for a protein subunit), which indicates an absence of dissociation stages in melting.

Fig 5.

DSC profiles of recombinant gp146 (A) and gp188* (B) in 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, and 100 mM KCl showing the dependence of the heat capacity on temperature for gp146 (0.58 μM) and gp188* (7.3 μM) in the absence of nucleotides (thick solid lines) and in the presence of ATP (A, dashed line) and ADP (A, thin solid line).

Table 1.

Thermodynamic parameters of gp146 thermal denaturation: effect of nucleotide binding

Sample	ΔHcal (kJ/mol)	Tm (°C)
gp188*	608 ± 30	54.8 ± 0.1
gp146	12,619 ± 939	62.1 ± 0.5
gp146/ATP	12,737 ± 517	60.6 ± 0.3
gp146/ADP	14,036 ± 380	62.2 ± 0.3

Since chaperonins assist the folding of polypeptide chains in an ATP-dependent manner, we examined the effect of ATP and ADP binding on thermal denaturation of gp146 (Fig. 5A; Table 1). Addition of ADP increased cooperativeness and enthalpy of the gp146 denaturation, which suggests chaperonin stabilization. In contrast, ATP binding results in complex destabilization, which is indicated by a slight decrease of the temperature of the heat capacity maximum (T_m). This result closely correlates with the effect of low ATP concentrations on GroEL thermal denaturation (20).

Thermal denaturation of gp188* is also irreversible and takes place at temperatures lower than those for gp146 (Fig. 5B; Table 1).

Fluorescence and CD spectroscopies.

Fluorescence and CD spectroscopies were used to detect changes in the tertiary and secondary structures of proteins, respectively, at different temperatures. Red shifts in the wavelength of the emission maximum (λ_max) without appreciable changes in fluorescence intensity were observed for both proteins when the temperature was increased from 25 to 75°C. Figure 6A shows the temperature dependence of λ_max of the emission spectra obtained by exciting gp146 and gp188* solutions at 280 nm. The far-UV CD spectra of proteins have two minima near 208 and 220 nm (Fig. 6B, inset), which is characteristic of proteins with a significant amount of α-helical content. When temperature was increased, the ellipticity at 208 nm was decreased dramatically with no change in the peak position for either protein (Fig. 6B). All experimental curves indicating changes in the tertiary and secondary structures of proteins upon their heating have an abrupt sigmoid form, which means a cooperative protein transition from its native structure to a disordered one. However, the destruction of the secondary structures of gp146 and gp188* was observed at temperatures higher than that at which the tertiary structure was destroyed (Fig. 6). When gp188* was heated to 50°C, half of its tertiary structure was lost, whereas its secondary structure did not change. Both methods have shown that gp188* undergoes structural changes at lower temperatures than gp146. The recombinant gp146 remains intact at temperatures up to 52°C. These results coincide with those obtained by DSC.

Fig 6.

Temperature-induced changes in the tertiary (A) and secondary (B) structures of gp188* (thick line, filled circles) and gp146 (thin line, open circles) in 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, and 100 mM KCl observed by fluorescence and CD spectroscopies. (A) Temperature dependence of wavelength of the emission maximum (λ_max) of the fluorescent spectra of gp146 (0.12 μM) and gp188* (1.4 μM). (Inset) Fluorescence emission spectra of gp146 (thin line) and gp188* (thick line) at 25°C by excitation at 280 nm. (B) Temperature dependence of the mean residue ellipticity at 208 nm of the far-UV CD spectra of gp146 (0.12 μM) and gp188* (2.9 μM). (Inset) Far-UV CD spectra of gp188* at 25, 55, 58, and 73°C and gp146 at 25, 60, 63, and 75°C (the ellipticity decreases with the rising temperature).

Isothermal titration calorimetry of phage chaperonin.

To study ATP and ADP binding to the recombinant gp146 more accurately, nucleotide titration was carried out using ITC. Typical raw traces of the calorimetric titration of ADP and ATP to pure gp146 are shown in Fig. 7A and C, respectively. The downward peak represents the heat release at each injection of the nucleotide solution. The binding isotherm of the nucleotide was obtained by integrating each peak and subtracting the dilution heat. Such isotherms for ADP and ATP are shown in Fig. 7B and D, respectively. The recombinant gp146 has been shown to bind both ATP and ADP. The observed binding isotherms were fitted to the “one set of sites” model. In this model, the putative chaperonin has several identical nucleotide binding sites, which are independent of each other and have a uniform binding constant, K_a, and enthalpy change, ΔH. The parameter values fitting this model best are listed in Table 2. As displayed in Fig. 7B and D, the theoretical curves (solid lines) show a good match with the experimental data (squares) for ADP and ATP, respectively. Our results indicate that nucleotide binding to gp146 could be well interpreted by the noncooperative model, while ATP is known to induce cooperative conformational changes in GroEL (28). The ADP binding constant and enthalpy are a little higher than those for ATP. Observed stoichiometry of nucleotide binding to the recombinant gp146 is close to 7 for both ADP and ATP (Fig. 7B and D; Table 2). In addition, the data for the competitive titration indicated that the binding sites are universal and able to bind both ATP and ADP (not shown). Considering the binding stoichiometry, two heptameric rings may differ in their abilities for nucleotide binding. Another explanation points to sample heterogeneity, suggesting that various structural conformations of the chaperonin are available for nucleotide binding.

Fig 7.

ITC profiles of the binding of ADP (A and B) and ATP (C and D) to recombinant gp146. (A and C) Raw data for successive injections of ADP (ATP) into gp146; (B and D) integrations of each injection (squares) and the best fit by the “one set of sites” model (solid lines) versus molar ratio of ADP (ATP) to gp146.

Table 2.

Best-fit parameters obtained by the “one set of sites” model

Nucleotide	Stoichiometry	Ka (μM−1)	ΔH (kJ/mol)	ΔS (J · mol−1 · °C−1)
ATP	7.33 ± 0.03	5.48 ± 0.24	−36.17 ± 0.23	1.7
ADP	8.76 ± 0.03	7.76 ± 0.36	−42.50 ± 0.22	−17.6

Effect of gp146 on gp188* thermal inactivation in vitro.

Chaperonins are thought to play a cellular role not only in assisting the folding of nascent proteins but also in protecting mature proteins undergoing stress. It is known that chaperonin GroEL can stabilize enzymes against thermal inactivation (36, 38). To find out whether gp146 can carry out a similar role, we investigated the effect of gp146 on heat inactivation of its substrate, gp188*, in vitro. According to DSC profiles (Fig. 5) and fluorescence and CD data (Fig. 6), gp188* denaturation starts at approximately 47°C, while gp146 seems to remain native at temperatures up to 52°C, even in the presence of ATP. Therefore, gp188* inactivation was examined under physiological conditions at 50°C. Aliquots were removed at appropriate intervals and assayed for residual muralytic activity. Inactivation of the free gp188* followed first-order kinetics with a half-time of about 16 min (not shown). As shown in Fig. 8A, the loss of the gp188* enzymatic activity during the incubation was retarded by gp146 in the presence of ATP. The enzymatic activity still remaining 60 min later was more than twofold higher in the presence of gp146 than in its absence. It was found that the free gp188* was also inactivated at 37°C, while in the presence of gp146 it remained intact for a long time (Fig. 8B). It should be noted that gp188* thermal inactivation is also slowed down by gp146 in the absence of ATP, but to a lesser extent than in its presence (Fig. 8B).

Fig 8.

Activity of gp188* (0.1 μM for monomer) during thermal inactivation in 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, and 100 mM KCl at 50°C (A) and 37°C (B). Shown are free gp188* (white bars) and gp188* in the presence of gp146 (0.2 μM for the tetradecamer) with 3 mM ATP (gray bars) and without ATP (hatched bars). The graph shows the results of three independent thermal inactivation assays, with error bars indicating the standard deviations.

Effect of gp146 on gp188* aggregation in vitro.

Heat inactivation of gp188* was found to be accompanied by irreversible aggregation. The effect of gp146 on the thermal aggregation of gp188* under physiological conditions was studied using DLS, which allows sizing of the protein aggregates. To observe the formation of aggregates, the gp188* concentration was increased to 3 μM in comparison with the concentration used in the thermal inactivation assay. Figure 9 shows DLS curves of the gp188* aggregation at 45°C. The free gp188* aggregation to large-sized particles takes less than 30 min, irrespective of the presence or absence of ATP. It was found that gp146 had the capacity to suppress denatured enzyme aggregation both in the presence and absence of ATP. The effect of gp146 on the aggregation of gp188* in the presence of ATP is shown in Fig. 9A. The degree of suppression of the aggregation correlates with the chaperonin content in the solution. This is indicated by the shifting of the curves from the reference to the right when the molar ratio of gp146 to substrate increases. At an equimolar ratio, gp146 was found to completely suppress gp188* aggregation. Indeed, the related curve has no difference to that for free gp146 (Fig. 9A). The replacement of gp146 by bovine serum albumin (BSA) did not result in a protective effect (not shown).

Fig 9.

Effect of gp146 on the aggregation of gp188* (3 μM) at 45°C in the presence of 1.2 mM ATP (A) and in the absence of ATP (B). Thin lines mean DLS trends of free proteins. Thick lines correspond to their mixture at various molar ratios (the number near each of the curves indicates a molar ratio of monomeric enzyme to tetradecameric gp146). The dashed line (B) corresponds to the gp188*-gp146 mixture at a molar ratio of 1:0.2; the arrow shows the time when ATP was added.

A similar influence of gp146 on the gp188* aggregation was observed under the same conditions in the absence of ATP (Fig. 9B). However, in this case the effect seems to be caused by the formation of highly stable long-lived binary complexes between gp146 and the thermally unfolded substrate. Bound to gp146, gp188* cannot undergo unspecific aggregation. Control experiments with BSA at molar excess do not show a comparable suppression of gp188* aggregation under these conditions (not shown), indicating a specific effect of gp146 similar to that earlier observed for chaperonin GroEL (10, 25). The complex of thermally inactivated substrate and gp146 can be easily dissociated by adding ATP. ATP-dependent dissociation results in the release of gp188*, which denatures and aggregates rapidly, as indicated by the increase in large protein particles (Fig. 9B).

DISCUSSION

In this study, we have characterized one of the GroEL orthologs predicted from phage genomes (13, 24, 30) and demonstrated its chaperonin activity for the first time. The phage chaperonin was shown to be synthesized during bacteriophage EL propagation in P. aeruginosa cells. The native protein was isolated by immunoprecipitation from EL-infected bacterial cells using serum against recombinant gp146 and identified by N-terminal amino acid sequencing. Physicochemical characterization of the phage chaperonin by a number of different techniques was carried out on the recombinant gp146 produced by E. coli cells. It was found that the recombinant gp146 self-assembled into a large oligomeric complex with the architecture typical of chaperonins. Similar to GroEL from E. coli, the phage chaperonin is a homotetradecamer composed of two stacked seven-member rings, each with a central cavity.

A native phage substrate of gp146, gp188, was isolated and identified. Characterization of recombinant proteins by different methods revealed that gp188* was less thermostable than gp146. The recombinant gp188* underwent inactivation and aggregation in vitro under physiological conditions at elevated temperatures below 50°C, while gp146 still remained native under these conditions. It was demonstrated that gp146 can protect the enzymatic activity of gp188* and suppress the irreversible aggregation of the thermally unfolded substrate molecules in vitro. These results provided evidence of the gp146 chaperonin function.

Whether the phage chaperonin plays a role in the EL life cycle is still unclear. We suppose that gp146 somehow affects lysis, because its substrate, gp188, appears to be an endolysin (8). Endolysins are known to be phage-encoded enzymes produced during the late phase of gene expression in the lytic cycle to degrade peptidoglycan, the main constituent of the bacterial cell wall, thereby enabling progeny virions to be liberated (34, 46). We confirmed that gp188 synthesis actually started during the late stage of the virus reproduction cycle, much later than gp146 synthesis. It is known that under physiological conditions many proteins undergo continuous denaturation in the cell, resulting in the loss of their biological activity (15). Considering this fact, we suppose that, during EL propagation in bacterial cells, the phage chaperonin could perform the same protective role in relation to its substrate as it does in vitro and thereby increase the half-life of endolysin, which otherwise might be inactivated and might aggregate. Further application of genetic approaches is required to prove the correctness of this hypothesis.

The phage chaperonin has shown to function in both ATP-dependent and ATP-independent manners. The first one seems to be similar to the chaperonin functional ATPase cycle, which results in the release of a properly folded and catalytically active enzyme. In the absence of ATP, gp146 presumably binds unfolded substrate molecules and forms stable binary complexes. Therefore, in the second manner, gp146 only passively prevents the trapped substrate from aggregation without renaturation. The passive model, whereby the chaperonin assists refolding simply by preventing aggregation of the misfolded substrates, was considered earlier for GroEL activity (26).

Thus, some structural and functional characteristics of the phage chaperonin are probably similar to those of bacterial GroEL. However, unlike GroEL, the phage chaperonin does not require a cochaperonin for its activity. Indeed, we have shown by in vitro experiments using purified recombinant proteins (gp146 and gp188*) that the phage chaperonin is able to function without any additional protein cofactors. It should be noted that no phage or host cochaperonin was found to coimmunoprecipitate together with the substrate protein from the EL-infected bacteria. Therefore, it is possible to assume that, like group II chaperonins, the phage chaperonin appears to have a helical protrusion, which can play a role equivalent to GroES, sealing off the central cavity from the outside. According to our preliminary observations of the gp146 shape in solution by small-angle X-ray scattering (SAXS), the chaperonin complex can adopt either an open or closed conformation that appears to be regulated by the position of the helical protrusion (1). Further structural and biochemical investigations are required to reveal the detailed mechanism of its functioning, which can be distinct from the mechanisms of the other known chaperonins. Further analysis of its functional and structural characteristics may provide important insights into the nature of the phage chaperonin, which is probably a representative of a new group of chaperonins.