Binding specificity of Escherichia coli trigger factor

Abstract

The ribosome-associated chaperone trigger factor (TF) assists the folding of newly synthesized cytosolic proteins in Escherichia coli. Here, we determined the substrate specificity of TF by examining its binding to 2842 membrane-coupled 13meric peptides. The binding motif of TF was identified as a stretch of eight amino acids, enriched in basic and aromatic residues and with a positive net charge. Fluorescence spectroscopy verified that TF exhibited a comparable substrate specificity for peptides in solution. The affinity to peptides in solution was low, indicating that TF requires ribosome association to create high local concentrations of nascent polypeptide substrates for productive interaction in vivo. Binding to membrane-coupled peptides occurred through the central peptidyl-prolyl-cis/trans isomerase (PPIase) domain of TF, however, independently of prolyl residues. Crosslinking experiments showed that a TF fragment containing the PPIase domain linked to the ribosome via the N-terminal domain is sufficient for interaction with nascent polypeptide substrates. Homology modeling of the PPIase domain revealed a conserved FKBP(FK506-binding protein)-like binding pocket composed of exposed aromatic residues embedded in a groove with negative surface charge. The features of this groove complement well the determined substrate specificity of TF. Moreover, a mutation (E178V) in this putative substrate binding groove known to enhance PPIase activity also enhanced TF's association with a prolyl-free model peptide in solution and with nascent polypeptides. This result suggests that both prolyl-independent binding of peptide substrates and peptidyl-prolyl isomerization involve the same binding site.

Figure 1

TF binding to peptide libraries and localization of binding sites in native protein structures. (A) Binding of TF to 13meric peptide libraries derived from EF-Tu (elongation factor-Tu) and GlnRS (glutamine-tRNA-synthetase). Numbers at the right indicate the last peptide spot in the row. (B) Ribbon and (C) space filling representations (weblabviewer) of the structures of the corresponding native proteins [GlnRS in complex with its tRNA (green), dimer of EF-Tu]. Red segments in B correspond to the peptides bound with high affinity on the library in A. (C) Aromatic and positively charged residues within the binding sites are shown in yellow and blue, respectively.

Figure 2

Substrate specificity of TF. (A) For 2,842 peptides representing 20 protein sequences, the relative amino acid occurrence was determined. The relative occurrence of each amino acid in TF binding peptides with high affinity is normalized to its occurrence in the whole peptide libraries (set as 100%). Classifications of high, medium, low, and no affinity binders according to the occurrence of aromatic side chains (B) and the net charge (C). Positioning of enriched aromatic and basic residues within the 13meric peptides with high affinity for TF (D). TF recognition motif (E).

Figure 3

Peptide binding is mediated by TF's PPIase domain. (A) Domain organization of TF. (B) Comparative peptide scans (E. coli Lambda cI protein) after incubation with full-length TF (1) and the PPIase domain TF (145).

Figure 4

Interaction of TF with peptides in solution. (A) Fluorescence measurements of TF and F-pep1 (a, black solid line), of TF and F-pep1 quenched with a 60-fold excess of unlabeled pep1 (c, black dashed line), and of TF and F-pep1 with a 60-fold excess of the nonbinding pep2 (b, black dotted line). Fluorescence signals of F-pep1 alone (f, gray solid line) and of F-pep1 together with pep1 (e, gray dashed line) or pep2 (d, gray dotted line) are shown for control. The highest fluorescence amplitude was set as 1. (B) Titration of TF (triangles) and TF-E178V (circles) against F-pep1 for K_D determination of both protein-F-pep1 complexes (the calculated maximum fluorescence was set as 1 for both).

Figure 5

Interaction of TF with nascent polypeptides. Arrested ³⁵S-labeled nascent ICDH was synthesized in vitro in a TF-deficient transcription/translation system supplemented with TF fragments, TF (1) wild type, or TF-E178V. Minor bands visible in the autoradiography correspond to either residual full-length ICDH or endogenous background products of unknown identity. After crosslinking with disuccinimidyl suberate (lanes 2, 5, 8, and 11) and sucrose cushion centrifugation, ribosome-nascent chain complexes were coimmunoprecipitated to identify TF crosslinks (lanes 3, 6, 9, and 12). Brackets with stars indicate crosslinking products of TF variants. In lane 3, the bulky band of IgGs in the antisera comigrates with the radiolabeled crosslinking product and shifts part of it to lower molecular weight. Please note that TF (1) and TF-E178V are less efficiently recognized by TF antibodies as compared with wild-type TF.

Figure 6

Homology modeling of TF's PPIase domain TF (145) based on the yeast FKBP12 structure by using insight ii. (A) Model illustrated by weblabviewer. The stick representation shows the conserved aromatic residues of the FKBP-like binding pocket (yellow), numbered by their position in full-length TF. (B) Surface charge illustration of the model by using grasp. Red color represents negative surface charges and blue color positive surface charges (increasing intensity indicates increasing charge). Negatively charged residues at the border of the groove are labeled with their positions in the full-length protein. The dashed circle indicates the putative binding groove of this domain embedding the conserved aromatic residues of the binding pocket as shown in A.