NMR analysis of the interaction of picornaviral proteinases Lb and 2A with their substrate eukaryotic initiation factor 4GII

Abstract

Messenger RNA is recruited to the eukaryotic ribosome by a complex including the eukaryotic initiation factor (eIF) 4E (the cap-binding protein), the scaffold protein eIF4G and the RNA helicase eIF4A. To shut-off host-cell protein synthesis, eIF4G is cleaved during picornaviral infection by a virally encoded proteinase; the structural basis of this reaction and its stimulation by eIF4E is unclear. We have structurally and biochemically investigated the interaction of purified foot-and-mouth disease virus (FMDV) leader proteinase (Lb(pro)), human rhinovirus 2 (HRV2) 2A proteinase (2A(pro)) and coxsackievirus B4 (CVB4) 2A(pro) with purified eIF4GII, eIF4E and the eIF4GII/eIF4E complex. Using nuclear magnetic resonance (NMR), we completed (13)C/(15) N sequential backbone assignment of human eIF4GII residues 551-745 and examined their binding to murine eIF4E. eIF4GII551-745 is intrinsically unstructured and remains so when bound to eIF4E. NMR and biophysical techniques for determining stoichiometry and binding constants revealed that the papain-like Lb(pro) only forms a stable complex with eIF4GII(551-745) in the presence of eIF4E, with KD values in the low nanomolar range; Lb(pro) contacts both eIF4GII and eIF4E. Furthermore, the unrelated chymotrypsin-like 2A(pro) from HRV2 and CVB4 also build a stable complex with eIF4GII/eIF4E, but with K(D) values in the low micromolar range. The HRV2 enzyme also forms a stable complex with eIF4E; however, none of the proteinases tested complex stably with eIF4GII alone. Thus, these three picornaviral proteinases have independently evolved to establish distinct triangular heterotrimeric protein complexes that may actively target ribosomes involved in mRNA recruitment to ensure efficient host cell shut-off.

Keywords: control of protein synthesis; protein-protein interactions; substrate binding; viral proteases; virus-host interactions.

Figure 1

Domain structure of human eIF4GII and cleavage of eIF4GII_551–745 by picornaviral proteases. (A) Schematic representation of human eIF4G with known structural domains indicated (the “ring” domain has only been shown in yeast) and the amino acid sequence of the fragment of eIF4GII_551–745 used. Protein binding sites are indicated as follows: eIF4E binding site, broken line; Lb^pro binding site, dotted line; 2A^pro binding region, straight line. Positions of the cleavage site of the Lb^pro and 2A^pro are marked by black and white triangles, respectively. The backbone NMR signals of 178 residues underlaid in black were assigned using triple resonance experiments, (B) In vitro cleavage of 5 µM eIF4GII_551–745 by 7 nM sLb^pro in the absence or the presence of 5 µM eIF4E, (C) In vitro cleavage of 5 µM eIF4GII_551–745L629A/L630A by 7 nM of sLb^pro in the absence or the presence of 25 µM eIF4E, (D) In vitro cleavage of 11 µM eIF4GII_653‐745 by 7 nM of sLb^pro in the absence or the presence of 5 µM eIF4E, (E) In vitro cleavage of 5 µM eIF4GII_551–745 by 80 nM HRV2 2A^pro in the absence or the presence of 5 µM eIF4E. The molar ratio of sLb^pro/HRV2 2A^pro to eIF4GII_551–745 is indicated below each cleavage figure.

Figure 2

eIF4GII_551–745 forms a ternary complex with eIF4E and sLb^proC51A. Complex formation was analysed via SEC on a HiLoad 16/60 Superdex 200 prep grade column together with aprotinin (6.5 kDa) as an internal standard. 0.5 mg of the individual or complexed proteins were analysed together with 1 mg of aprotinin. Complex formation of two or more proteins was performed by incubation at 4°C for 10 minutes. (A) eIF4GII_551–745, eIF4E and sLb^proC51A (B) eIF4GII_551–745 and sLb^proC51A (C) eIF4E and sLb^proC51A. Stable complexes (under conditions of SEC) were observed between eIF4GII_551–745 and eIF4E as well as between eIF4GII_551–745, eIF4E and sLb^proC51A (both A), but not between eIF4GII_551–745 and sLb^proC51A in the absence of eIF4E (B) and eIF4E and sLb^proC51A in the absence of eIF4GII_551–745 (C).

Figure 3

2D ¹H‐¹⁵N HSQC spectra of ¹⁵N labelled eIF4GII_551–745. Spectra are overlaid for comparison: native eIF4GII_551–745 (cyan), 1:1 complex of eIF4GII_551–745 and eIF4E (purple) and 1:1:1 complex eIF4GII_551–745, eIF4E and sLb^proC51A (red). Complexes were purified using SEC and concentrated to 0.7 mM. The narrow shift dispersion of cross‐peaks in the ¹H dimension indicates the lack of a defined tertiary structure for eIF4GII_551–745. Residues attenuated or disappearing upon the addition of the eIF4E or the sLb^proC51A are involved in or affected by complex formation. Signals discussed in the text are shown in insets.

Figure 4

Effect of the addition of eIF4E, sLb^proC51A or HRV2 2A^proC106S on the signal intensities of the ¹⁵N HSQC spectrum of eIF4GII_551–745. A native spectrum of ¹⁵N eIF4GII_551–745 was recorded before every titration experiment and its intensities compared to those obtained after addition of the indicated protein(s). Normalized intensities of every assigned residue were calculated by dividing intensities by those of the control spectrum. Residues for which no amide backbone assignment value could be attributed are shown with a value of −0.1. The white boxes represent eIF4E binding site (BS), short black box the Lb^pro BS, long black box the HRV2 2A^pro BS and grey boxes represent the picornaviral cleavage sites (CS). (A) ¹⁵N eIF4GII_551–745 and eIF4E complex purified via SEC. (B) ¹⁵N eIF4GII_551–745/eIF4E/sLb^proC51A complex purified via SEC. (C) Titration experiment of ¹⁵N eIF4GII_551–745 and sLb^proC51A. (D) ¹⁵N eIF4GII_551–745/eIF4E/HRV2 2A^proC106S complex purified via SEC. (E) Titration experiment of ¹⁵N eIF4GII_551–745 and HRV2 2A^proC106S.

Figure 5

eIF4GII_551–745 also forms a complex with eIF4E and HRV2 2A^proC106S or CVB4 2A^proC110A. Complex formation was analysed via SEC on a HiLoad 16/60 Superdex 200 prep grade column (for HRV2 2A^pro) or on a HiLoad 16/60 Superdex 75 prep grade column (for CV 2A^pro) together with aprotinin (6.5 kDa) as an internal standard. 0.5 mg of the individual or complexed proteins were analysed together with 1 mg of aprotinin. Complex formation of two or more proteins was performed by incubation at 4°C for 10 minutes. (A) eIF4GII_551–745, eIF4E and HRV2 2A^proC106S (B) eIF4GII_551–745 and HRV2 2A^proC106S (C) eIF4E and HRV2 2A^proC106S (D) eIF4GII_551–745, eIF4E and CVB4 2A^proC110A (E) eIF4GII_551–745 and CVB4 2A^proC110A and (F) eIF4E and CVB4 2A^proC110A. Stable complexes were observed between eIF4GII_551–745, eIF4E and HRV2 2A^proC106S (A), eIF4E and 2A^proC106S (C) and eIF4GII_551–745, eIF4E and CVB4 2A^proC110A (D) but not between eIF4GII_551–745 and HRV2 2A^proC106S in the absence of eIF4E (B), eIF4GII_551–745 and CVB4 2A^proC110A in the absence of eIF4E (E) or eIF4E and CVB4 2A^proC110A in the absence of eIF4GII_551–745 (F).

Figure 6

Overlaid spectra from 2D ¹⁵N HSQC spectroscopy of native sLb^proC51A (grey), 1:1 sLb^proC51A and eIF4GII_551–745 (light blue), 1:1 sLb^proC51A and eIF4E (dark blue) and 1:1:1 sLb^proC51A, eIF4GII_551–745 and eIF4E (red). The eIF4GII_551–745/eIF4E/sLb^proC51A complex was purified using SEC and concentrated, whereas for binary complex formation the individual proteins were titrated into ¹⁵N sLb^proC51A. The assignments of the ¹⁵N signals have been assigned previously27 (BMRB Entry 15278). As a result of complex formation most residues of sLb^proC51A disappear upon the addition of the eIF4GII_551–745 or the eIF4E, since a globular, folded protein is more strongly affected by binding to an unstructured protein with a comparatively ‘puffed up’ shape. Due the resulting increased hydrodynamic drag of the complex the relaxation rates of the residues involved in the binding event are globally enhanced.

Figure 7

In vitro translation assays of the intermolecular processing of Lb^proC51A eIF4GI_619‐678 VP4/VP2 (A and C) or Lb^proC51A eIF4GI_599‐678 VP4/VP2 (B and D) by sLb^pro (A and B) and Lb^pro (C and D). Rabbit reticulocyte lysate (RRL) was programmed with 14 ng/µL RNA coding for Lb^proC51A eIF4GI_599‐678 VP4_5‐85/VP2_1‐23 (eIF4GI_599‐678 sequence with the eIF4E and the Lb^pro binding site) or Lb^proC51A eIF4GI_619‐678 VP4_5‐85/VP2_1‐23 (eIF4GI_619‐678 sequence lacking the eIF4E binding site) comprising the eIF4GI cleavage site SFANLG*RTTL. Precursor substrate protein was translated for 20 minutes in the presence of [³⁵S]‐Met before adding 14 ng/µL active sLb^pro or Lb^pro RNA in the presence of unlabelled Met to the reaction. 10 µL aliquots were taken at the indicated time points and cleavage of the substrate protein was analysed on 17.5% SDS gels followed by autoradiography (upper pictures). Uncleaved precursors Lb^proC51A eIF4GI_599‐678 VP4/VP2 or Lb^proC51A eIF4GI_619‐678 VP4/VP2 as well as cleavage products Lb^proC51A eIF4GI_599‐674, Lb^proC51A eIF4GI_675‐678 VP4_5‐85/VP2_1‐23 or Lb^proC51A eIF4GI_619‐674, Lb^proC51A eIF4GI_675‐678 VP4_5‐85/VP2_1‐23 are indicated. To observe cleavage on endogenous eIF4GI in the RRLs, 10 µL aliquots were analysed on 6% SDS gels followed by Western blotting using serum to detect the N‐terminal part of the eIF4GI protein (lower pictures). Uncleaved eIF4GI and cleavage products (CP_N) are indicated. Negative controls without any RNA (−sub, −prot) or comprising only RNA of the substrate protein (+sub, −prot) are shown at the right of each gel. Protein standards are shown on the left.

Figure 8

Model of the heterotrimeric triangular interactions of eIF4GII_551–745, eIF4E and sLb^proC51A. Binding sites for eIF4E and sLb^pro on the eIF4GII_551–745 are indicated in purple and red, respectively; a triangle indicates the Lb^pro cleavage site. Drawings were created in PyMol.55 PDB identifiers: 1QOL, Lb^proC51A; 1EJH, eIF4E.