Structure and conformational cycle of a bacteriophage-encoded chaperonin

Abstract

Chaperonins are ubiquitous molecular chaperones found in all domains of life. They form ring-shaped complexes that assist in the folding of substrate proteins in an ATP-dependent reaction cycle. Key to the folding cycle is the transient encapsulation of substrate proteins by the chaperonin. Here we present a structural and functional characterization of the chaperonin gp146 (ɸEL) from the phage EL of Pseudomonas aeruginosa. ɸEL, an evolutionarily distant homolog of bacterial GroEL, is active in ATP hydrolysis and prevents the aggregation of denatured protein in a nucleotide-dependent manner. However, ɸEL failed to refold the encapsulation-dependent model substrate rhodanese and did not interact with E. coli GroES, the lid-shaped co-chaperone of GroEL. ɸEL forms tetradecameric double-ring complexes, which dissociate into single rings in the presence of ATP. Crystal structures of ɸEL (at 3.54 and 4.03 Å) in presence of ATP•BeFx revealed two distinct single-ring conformational states, both with open access to the ring cavity. One state showed uniform ATP-bound subunit conformations (symmetric state), whereas the second combined distinct ATP- and ADP-bound subunit conformations (asymmetric state). Cryo-electron microscopy of apo-ɸEL revealed a double-ring structure composed of rings in the asymmetric state (3.45 Å resolution). We propose that the phage chaperonin undergoes nucleotide-dependent conformational switching between double- and single rings and functions in aggregation prevention without substrate protein encapsulation. Thus, ɸEL may represent an evolutionarily more ancient chaperonin prior to acquisition of the encapsulation mechanism.

Fig 1. Oligomeric state of ɸEL in solution.

(A-C) SEC-MALS analysis of ΦEL in absence of nucleotide (A), in presence of 2 mM ADP (B) or 2 mM ATP (C). The chromatographic absorbance traces at 280 nm wavelength are shown. The molecular mass determined for the protein peaks by static light scattering is indicated. The black and red traces were recorded in presence of 50 and 100 mM salt, respectively.

Fig 2. Concentration dependence of ATP hydrolysis by ɸEL.

(A, B) The curves show the concentration dependence of ATP hydrolysis by ɸEL in presence of 10 (A) and 100 mM KCl (B). Shown are the averages of three experiments. Error bars indicate standard deviations. The red lines represent the Hill curve fittings of the data.

Fig 3. Effect of GroES on GroEL and ɸEL.

(A, B) ATPase activity of GroEL and ɸEL in presence and absence of a two-fold excess of GroES or the model substrate DM-MBP. The buffer contained 10 mM (A) or 100 mM KCl (B) and the ATP concentration was 1 mM. The bar graph shows the averages from at least three experiments; error bars indicate standard deviations. (C) Proteinase K (PK) sensitivity of GroES alone, and of ɸEL and GroEL in presence and absence of GroES. The experiment was performed in buffer 20 mM Tris-HCl pH 7.5, 100 mM KCl, 5 mM MgCl₂ and 0.2 mM ADP [42]. The concentrations of GroEL and ɸEL were 1.5 μM; GroES was at 1.0 μM. After 3 min incubation with 0.2 g L^-1 PK at 25°C, the protease reaction was stopped by addition of phenylmethylsulfonyl fluoride (1 mM). The mixtures were analyzed by SDS-PAGE. *, proteolytic fragments of GroES; **, proteolytic fragments of ɸEL.

Fig 4. Molecular chaperone activity of ɸEL.

(A, B) Aggregation prevention activity of ɸEL and GroEL in presence and absence of ATP (A) or ADP (B). Rhodanese (Rho) aggregation was monitored by turbidity assay at 320 nm wavelength. The results of representative experiments are shown. (C) Rho refolding in presence of ATP. The chaperones GroEL, GroES and ɸEL were present when indicated. After quenching ATP hydrolysis by addition of CDTA at the indicated time points, Rho enzyme activity was determined. The averages from three experiments are shown, error bars indicate standard deviations.

Fig 5. Crystal structures of ɸEL.

(A) Ribbon representation of the ɸEL complex in crystal form I. Perpendicular views of the ɸEL complex are shown. The subunits are shown in rainbow colors. Bound ATP and Mg²⁺ is colored beige and shown in space-filling mode. Subunit chain identifiers and N-termini are indicated. (B) Structure of the ɸEL complex in crystal form II, using the same representation style. Bound ADP is colored silver and shown in space-filling mode. Domain movements in ɸEL subunits compared to crystal form I are indicated by curved arrows.

Fig 6. Conformations of ɸEL subunits observed in the crystal structures.

(A) Ribbon representation of ɸEL subunit conformation I as observed in chain A of crystal form I. Perpendicular views of ɸEL are shown. Apical, intermediate and equatorial domain are shown in gold, red and blue, respectively. Bound ATP is shown in stick representation. The intramolecular contact between apical and equatorial domain is highlighted by a box. The αK-αL hairpin is indicated. N- and C-termini are indicated. (B) Zoom-in on the bound ATP and Mg²⁺ in chain A of crystal form I (conformation I). The representation is equivalent to panel A. Unbiased Fo-Fc omit density at 2.85σ for the nucleotide is shown in green as meshwork. Secondary structure elements involved in contacts to the nucleotide are indicated. (C) Ribbon representation of ɸEL subunit conformation II as observed in chain D of crystal form II. The rotation of the intermediate domain (red) compared to conformation I is indicated by a curved arrow. The intramolecular contact between apical and equatorial domain is highlighted by a box. Bound ADP is shown in stick representation. (D) Ribbon representation of ɸEL subunit conformation III as observed in chain A of crystal form II. The rotation of the apical domain (gold) compared to conformations I and II is indicated by a curved arrow. Bound ATP is shown in stick representation. (E) Contact between subunits in conformations II and III, as observed between chains E and D in crystal form II. The characteristic intermediate-apical domain intermolecular contact is highlighted by a box. (F) Contact between subunits in conformations III and II, as observed between chains B and A in crystal form II. The contacts are almost exclusively formed between the equatorial domains (boxed area). (G) Zoom-in on the bound ADP in chain D of crystal form II (conformation II). Unbiased Fo-Fc omit density at 1.7σ for the nucleotide is shown in green as meshwork. Movement of the intermediate domain and remodeling of the αA-β1 loop, respectively, are indicated by arrows. (H) Zoom-in on the bound ATP and Mg²⁺ in chain A of crystal form II (conformation III). Unbiased Fo-Fc omit density at 2.85σ for the nucleotide is shown in green as meshwork.

Fig 7. Cryo-EM structure of apo-ɸEL.

(A) Cryo-EM density map of apo-ɸEL at 3.45 Å resolution. Perpendicular views of an isocontour surface at 4.8σ is shown. The top view is along the two-fold symmetry axis. The box highlights the inter-subunit interface at the equator of the complex. (B) Atomic model of apo-ɸEL. Backbone traces of the subunits are shown. Symmetry-equivalent subunits are shown in the same color. (C, D) Zoom on the symmetrical inter-subunit interface at the equator of the complex. Panel C shows the cryo-EM density. Panel D the model in ribbon representation. Contact sidechains are indicated and shown in stick representation (Glu466 is hidden behind the αO helix ribbon in this projection). Secondary structure elements participating in the interactions are indicated.

Fig 8. CryoEM structures of ɸEL in complex with ADP or ATPγS.

(A) Superposition of the cryo-EM density map and the structural model of the ɸEL•ADP complex. The pseudo-atomic model is shown as backbone trace, with the subunits colored individually. ADP is shown in stick representation. The cryo-EM density map is shown as an isocontour surface at 4.5 σ. Two perpendicular views are shown. In the bottom view, only the equatorial domain section is shown to demonstrate the quality of the fit. (B) Superposition of the ɸEL•ADP complex with apo-ɸEL, showing the unaltered ring-ring interface. Backbone traces are shown. The ɸEL•ADP complex is shown in rainbow colors, and apo-ɸEL in grey. (C) Superposition of the cryo-EM density map and the structural model of the ɸEL•ATPγS complex. The representation is the same as in panel A. (D) Superposition of the ɸEL•ATPγS complex with apo-ɸEL, showing the changes at the ring-ring interface. Backbone traces are shown. The ɸEL•ATPγS complex is shown in rainbow colors, and apo-ɸEL in grey.

Fig 9. Comparison of the structures of ɸEL and GroEL.

(A) Subunit structure of ɸEL (conformation I). The representation is the same as in Fig 6. (B, C) Structures of GroEL subunits from the cis (B) and trans rings (C) of the symmetric GroEL:GroES₂ complex (PDB 1PCQ) [51]. The same representation style as in panel A. (D, E) Structures of the apical domains in ɸEL (D) and GroEL (E). Ribbon representations are shown. Remodeled regions in ɸEL discussed in the text are highlighted in boxes. Chain termini and α-helices are indicated. Arrows point to the proposed substrate binding site in group I chaperonins. For GroEL, the PDB dataset 1XCK (apo, open state) [52] was used. (F, G) Structures of the intermediate domains in ΦEL (F) and GroEL (G). Chain termini and selected secondary structure elements are indicated. (H, I) Structures of the equatorial domains in ɸEL (H) and GroEL (I). Ribbon representations are shown. A view from the ring-ring interface is shown. Large insertions in the respective structure involved in inter-ring contacts are highlighted in boxes.

Fig 10. Hypothetical model for the ATPase and substrate interaction cycle of ɸEL.

ɸEL is shown schematically as a top view with the apical domains in yellow. The double-ring apo-state with asymmetric apical domain orientation has only low affinity for SP. It is converted by ATP binding (1) to the single-ring characterized by symmetric apical domain topology and high binding affinity for SP (2). ATP hydrolysis in three alternating subunits (3) may result in partial SP release, with ATP hydrolysis in the remaining four subunits (4) completing SP release and folding.