GroEL/S substrate specificity based on substrate unfolding propensity

Abstract

Phage P22 wild-type (WT) coat protein does not require GroEL/S to fold but temperature-sensitive-folding (tsf) coat proteins need the chaperone complex for correct folding. WT coat protein and all variants absolutely require P22 scaffolding protein, an assembly chaperone, to assemble into precursor structures termed procapsids. Previously, we showed that a global suppressor (su) substitution, T1661, which rescues several tsf coat protein variants, functioned by inducing GroEL/S. This led to an increased formation of tsf:T1661 coat protein:GroEL complexes compared with the tsf parents. The increased concentration of complexes resulted in more assembly-competent coat proteins because of a shift in the chaperone-driven kinetic partitioning between aggregation-prone intermediates toward correct folding and assembly. We have now investigated the folding and assembly of coat protein variants that carry a different global su substitution, F170L. By monitoring levels of phage production in the presence of a dysfunctional GroEL we found that tsf:F170L proteins demonstrate a less stringent requirement for GroEL. Tsf:F170L proteins also did not cause induction of the chaperones. Circular dichroism and tryptophan fluorescence indicate that the native state of the tsf: F170L coat proteins is restored to WT-like values. In addition, native acrylamide gel electrophoresis shows a stabilized native state for tsf:F170L coat proteins. The F170L su substitution also increases procapsid production compared with their tsf parents. We propose that the F170L su substitution has a decreased requirement for the chaperones GroEL and GroES as a result of restoring the tsf coat proteins to a WT-like state. Our data also suggest that GroEL/S can be induced by increasing the population of unfolding intermediates.

Fig 1.

The F170L suppressor substitution has a less stringent GroEL requirement than T166I. Phage that had tsf:su substitutions or were WT in coat protein were grown on DW720 cells (WT GroEL), or on DW716 cells (dysfunctional GroEL44). The relative titer is the titer of phage produced at each experimental temperature and condition divided by the titer of the phage at the permissive temperature and condition (28°C, DW720 cells). Open squares are phage production with tsf:T166I coat proteins grown on DW716 cells (GroEL44); closed squares are phage production with tsf:T166I coat proteins on DW720 cells (WT GroEL); open circles are phage production with tsf:F170L coat proteins grown on DW716 cells; closed circles are phage production with tsf:F170L coat proteins grown on DW720 cells

Fig 2.

The F170L global suppressor substitution does not induce GroEL expression in vivo. Pulse-chase experiments were done with cells infected with phage that carried tsf, tsf:su, or WT coat protein. Lysates were applied to 5–20% linear sucrose gradients and fractionated after centrifugation. The fractions were run on 10% SDS-acrylamide gels. Protein bands from the autoradiographs were quantified using densitometry. The total amount of labeled soluble coat protein, as well as the total amount of labeled GroEL for each variant was calculated. (A) The ratio of labeled GroEL per labeled soluble coat protein for each mutant was determined and normalized to the WT ratio. (B) The amount of soluble coat proteins bound to GroEL. In both panels, open bars represent WT data, gray bars represent tsf data, and black bars represent tsf:F170L data. The striped bars represent tsf:T166I data taken from Parent et al (2004). Although only a representative experiment is shown here, errors for these experiments are typically ±5% (Parent et al 2004)

Fig 3.

Aggregation is decreased for some, but not all, tsf coat proteins when the F170L substitution is also present. The propensity for aggregation while folding was determined by refolding urea-denatured coat proteins by rapid dilution with buffer at various temperatures. Aliquots were taken with time after dilution and run on native polyacrylamide gels silver stained to detect the presence of native monomers and aggregates. Native gels of a time course ranging from 0.5–15 minutes at 15°C and 33° C are shown. (A) A representative sample of aggregation of tsf coat protein during folding; D174N is shown because F353L and S223F were previously published (Aramli and Teschke 2001). (B) A representative sample of tsf:F170L coat protein aggregation

Fig 4.

The native state is stabilized for tsf:F170L coat proteins. Temperature shift experiments were performed as described in the Materials and Methods. In brief, natively folded coat protein held on ice was transferred to 33°C for 0–30 minutes. Samples were run on native polyacrylamide gels and silver stained to detect the presence of native monomers and aggregated forms. (A) Densitometry data taken from the native gels. The native fraction remaining is the intensity of the native band at the experimental time divided by the intensity at time 0. The lines drawn do not reflect a fit of the data to any model; they are drawn to aid the eye. Closed squares are data for WT coat protein; open circles are tsf coat protein; closed circles are tsf:F170L coat proteins. (B) The average fraction of native protein remaining after 30 minutes at 33°C; error bars are the standard deviation taken from 5 data sets

Fig 5.

Tsf:F170L coat proteins have WT-like secondary structures compared with their tsf parents. Circular dichroism spectra of WT, tsf, and tsf:F170L coat proteins were done as described in the Materials and Methods and are shown. Closed diamonds are WT coat protein; closed squares are tsf coat proteins; open squares are tsf: F170L coat proteins

Fig 6.

The F170L substitution results in coat proteins that are more assembly competent. In vitro assembly reactions were performed as described in Materials and Methods. (A) The light scattering data for WT and variant coat proteins at 20°C. Solid lines represent tsf:F170L coat protein assembly, and dashed lines represent tsf coat protein assembly. (B) The percentage of procapsids formed during these experiments was determined by the fraction of coat protein sedimenting in the procapsid position of sucrose gradient sedimentation

Fig 7.

Our current model for P22 coat protein folding and assembly. See Discussion for a description of this model. The dashed boxes show the predominant forms for each folding reaction, and arrows represent the flux of the folding reaction determined by equilibrium and kinetic experiments (Anderson and Teschke 2003; Doyle et al 2003, 2004). The GroEL dependence is based on the ts temperature when GroE44 is present ( Gordon et al 1994; Nakonechny and Teschke 1998; Aramli and Teschke 1999). A minus sign (−) means no GroEL dependence, and the number of plus signs (+) indicates the degree of dependence