James L. Baber

Structure and dynamics of KH domains from FBP bound to single-stranded DNA

Gene regulation can be tightly controlled by recognition of DNA deformations that are induced by stress generated during transcription. The KH domains of the FUSE-binding protein (FBP), a regulator of c-myc expression, bind in vivo and in vitro to the single-stranded far-upstream element (FUSE), 1,500 base pairs upstream from the c-myc promoter. FBP bound to FUSE acts through TFIIH at the promoter. Here we report the solution structure of a complex between the KH3 and KH4 domains of FBP and a 29-base single-stranded DNA from FUSE. The KH domains recognize two sites, 9–10 bases in length, separated by 5 bases, with KH4 bound to the 5′ site and KH3 to the 3′ site. The central portion of each site comprises a tetrad of sequence 5′d-ATTC for KH4 and 5′d-TTTT for KH3. Dynamics measurements show that the two KH domains bind as articulated modules to single-stranded DNA, providing a flexible framework with which to recognize transient, moving targets.

Molecular basis of sequence-specific single-stranded DNA recognition by KH domains: solution structure of a complex between hnRNP K KH3 and single-stranded DNA

To elucidate the basis of sequence‐specific single‐stranded (ss) DNA recognition by K homology (KH) domains, we have solved the solution structure of a complex between the KH3 domain of the transcriptional regulator heterogeneous nuclear ribonucleoprotein K (hnRNP K) and a 10mer ssDNA. We show that hnRNP K KH3 specifically recognizes a tetrad of sequence 5′d‐TCCC. The complex is stabilized by a dense network of methyl‐oxygen hydrogen bonds involving the methyl groups of three isoleucine residues and the O2 and N3 atoms of the two central cytosine bases. Comparison with the recently solved structure of a specific protein–ssDNA complex involving the KH3 and KH4 domains of the far upstream element (FUSE) binding protein FBP suggests that the amino acid located five residues N‐terminal of the invariant GXXG motif, which is characteristic of all KH domains, plays a crucial role in discrimination of the first two bases of the tetrad.

Structure and Dynamics of Full-Length HIV-1 Capsid Protein in Solution.

The HIV-1 capsid protein plays a crucial role in viral infectivity, assembling into a cone that encloses the viral RNA. In the mature virion, the N-terminal domain of the capsid protein forms hexameric and pentameric rings, while C-terminal domain homodimers connect adjacent N-terminal domain rings to one another. Structures of disulfide-linked hexamer and pentamer assemblies, as well as structures of the isolated domains, have been solved previously. The dimer configuration in C-terminal domain constructs differs in solution (residues 144-231) and crystal (residues 146-231) structures by ∼30°, and it has been postulated that the former connects the hexamers while the latter links pentamers to hexamers. Here we study the structure and dynamics of full-length capsid protein in solution, comprising a mixture of monomeric and dimeric forms in dynamic equilibrium, using ensemble simulated annealing driven by experimental NMR residual dipolar couplings and X-ray scattering data. The complexity of the system necessitated the development of a novel computational framework that should be generally applicable to many other challenging systems that currently escape structural characterization by standard application of mainstream techniques of structural biology. We show that the orientation of the C-terminal domains in dimeric full-length capsid and isolated C-terminal domain constructs is the same in solution, and we obtain a quantitative description of the conformational space sampled by the N-terminal domain relative to the C-terminal domain on the nano- to millisecond time scale. The positional distribution of the N-terminal domain relative to the C-terminal domain is large and modulated by the oligomerization state of the C-terminal domain. We also show that a model of the hexamer/pentamer assembly can be readily generated with a single configuration of the C-terminal domain dimer, and that capsid assembly likely proceeds via conformational selection of sparsely populated configurations of the N-terminal domain within the capsid protein dimer.

Long Distance Measurements up to 160 Å in the GroEL Tetradecamer Using Q-Band DEER EPR Spectroscopy

Current distance measurements between spin-labels on multimeric protonated proteins using double electron-electron resonance (DEER) EPR spectroscopy are generally limited to the 15-60 Å range. Here we show how DEER experiments can be extended to dipolar evolution times of ca. 80 μs, permitting distances up to 170 Å to be accessed in multimeric proteins. The method relies on sparse spin-labeling, supplemented by deuteration of protein and solvent, to minimize the deleterious impact of multispin effects and substantially increase the apparent spin-label phase memory relaxation time, complemented by high sensitivity afforded by measurements at Q-band. We demonstrate the approach using the tetradecameric molecular machine GroEL as an example. Two engineered surface-exposed mutants, R268C and E315C, are used to measure pairwise distance distributions with mean values ranging from 20 to 100 Å and from 30 to 160 Å, respectively, both within and between the two heptameric rings of GroEL. The measured distance distributions are consistent with the known crystal structure of apo GroEL. The methodology presented here should significantly expand the use of DEER for the structural characterization of conformational changes in higher order oligomers.

Dependence of Distance Distributions Derived from Double Electron–Electron Resonance Pulsed EPR Spectroscopy on Pulse-Sequence Time†

Pulsed double electron-electron resonance (DEER) provides pairwise P(r) distance distributions in doubly spin labeled proteins. We report that in protonated proteins, P(r) is dependent on the length of the second echo period T owing to local environmental effects on the spin-label phase memory relaxation time Tm . For the protein ABD, this effect results in a 1.4 Å increase in the P(r) maximum from T=6 to 20 μs. Protein A has a bimodal P(r) distribution, and the relative height of the shorter distance peak at T=10 μs, the shortest value required to obtain a reliable P(r), is reduced by 40 % relative to that found by extrapolation to T=0. Our results indicate that data at a series of T values are essential for quantitative interpretation of DEER to determine the extent of the T dependence and to extrapolate the results to T=0. Complete deuteration (99 %) of the protein was accompanied by a significant increase in Tm and effectively abolished the P(r) dependence on T.

Tilted, Uninterrupted, Monomeric HIV-1 gp41 Transmembrane Helix from Residual Dipolar Couplings.

Cryo-electron microscopy and X-ray crystallography have shown that the pre- and postfusion states of the HIV-1 gp41 viral coat protein, although very different from one another, each adopt C3 symmetric structures. A stable homotrimeric structure for the transmembrane domain (TM) also was modeled and supported by experimental data. For a C3 symmetric structure, alignment in an anisotropic medium must be axially symmetric, with the unique axis of the alignment tensor coinciding with the C3 axis. However, NMR residual dipolar couplings (RDCs) measured under three different alignment conditions were found to be incompatible with C3 symmetry. Subsequent measurements by paramagnetic relaxation enhancement, analytical ultracentrifugation, and DEER EPR, indicate that the transmembrane domain is monomeric. 15N NMR relaxation data and RDCs show that TM is highly ordered and uninterrupted for a total length of 32 residues, extending well into the membrane proximal external region.

Insights into the Conformation of the Membrane Proximal Regions Critical to the Trimerization of the HIV-1 gp41 Ectodomain Bound to Dodecyl Phosphocholine Micelles.

The transitioning of the ectodomain of gp41 from a pre-hairpin to a six-helix bundle conformation is a crucial aspect of virus-cell fusion. To gain insight into the intermediary steps of the fusion process we have studied the pH and dodecyl phosphocholine (DPC) micelle dependent trimer association of gp41 by systematic deletion analysis of an optimized construct termed 17–172 (residues 528 to 683 of Env) that spans the fusion peptide proximal region (FPPR) to the membrane proximal external region (MPER) of gp41, by sedimentation velocity and double electron-electron resonance (DEER) EPR spectroscopy. Trimerization at pH 7 requires the presence of both the FPPR and MPER regions. However, at pH 4, the protein completely dissociates to monomers. DEER measurements reveal a partial fraying of the C-terminal MPER residues in the 17–172 trimer while the other regions, including the FPPR, remain compact. In accordance, truncating nine C-terminal MPER residues (675–683) in the 17–172 construct does not shift the trimer-monomer equilibrium significantly. Thus, in the context of the gp41 ectodomain spanning residues 17–172, trimerization is clearly dependent on FPPR and MPER regions even when the terminal residues of MPER unravel. The antibody Z13e1, which spans both the 2F5 and 4E10 epitopes in MPER, binds to 17–172 with a Kd of 1 ± 0.12 μM. Accordingly, individual antibodies 2F5 and 4E10 also recognize the 17–172 trimer/DPC complex. We propose that binding of the C-terminal residues of MPER to the surface of the DPC micelles models a correct positioning of the trimeric transmembrane domain anchored in the viral membrane.

Quantitative Resolution of Monomer-Dimer Populations by Inversion Modulated DEER EPR Spectroscopy.

A simple method, based on inversion modulated double electron-electron resonance electron paramagnetic resonance (DEER EPR) spectroscopy, is presented for determining populations of monomer and dimer in proteins (as well as any other biological macromolecules). The method is based on analysis of modulation depth versus electron double resonance (ELDOR) pulse flip angle. High accuracy is achieved by complete deuteration, extensive sampling of a large number of ELDOR pulse flip angle values, and combined analysis of differently labeled spin samples. We demonstrate the method using two different proteins: an obligate monomer exemplified by the small immunoglobulin binding B domain of protein A, and the p66 subunit of HIV-1 reverse transcriptase which exists as an equilibrium mixture of monomer and dimer species whose relative populations are affected by glycerol content. This information is crucial for quantitative analysis of distance distributions involving proteins that may exist as mixtures of monomer, dimer and high order multimers under the conditions of the DEER EPR experiment.

The C34 Peptide Fusion Inhibitor Binds to the Six-Helix Bundle Core Domain of HIV-1 gp41 by Displacement of the C-Terminal Helical Repeat Region.

The conformational transition of the core domain of HIV-1 gp41 from a prehairpin intermediate to a six-helix bundle is responsible for virus-cell fusion. Several inhibitors which target the N-heptad repeat helical coiled-coil trimer that is fully accessible in the prehairpin intermediate have been designed. One such inhibitor is the peptide C34 derived from the C-heptad repeat of gp41 that forms the exterior of the six-helix bundle. Here, using a variety of biophysical techniques, including dye tagging, size-exclusion chromatography combined with multiangle light scattering, double electron-electron resonance EPR spectroscopy, and circular dichroism, we investigate the binding of C34 to two six-helix bundle mimetics comprising N- and C-heptad repeats either without (core(SP)) or with (core(S)) a short spacer connecting the two. In the case of core(SP), C34 directly exchanges with the C-heptad repeat. For core(S), up to two molecules of C34 bind the six-helix bundle via displacement of the C-heptad repeat. These results suggest that fusion inhibitors such as C34 can target a continuum of transitioning conformational states from the prehairpin intermediate to the six-helix bundle prior to the occurrence of irreversible fusion of viral and target cell membranes.

Co‐Evolutionary Fitness Landscapes for Sequence Design

Efficient and accurate models to predict the fitness of a sequence would be extremely valuable in protein design. We have explored the use of statistical potentials for the coevolutionary fitness landscape, extracted from known protein sequences, in conjunction with Monte Carlo simulations, as a tool for design. As proof of principle, we created a series of predicted high-fitness sequences for three different protein folds, representative of different structural classes: the GA (all-α) and GB (α/β) binding domains of streptococcal protein G, and an SH3 (all-β) domain. We found that most of the designed proteins can fold stably to the target structure, and a structure for a representative of each for GA, GB and SH3 was determined. Several of our designed proteins were also able to bind to native ligands, in some cases with higher affinity than wild-type. Thus, a search using a statistical fitness landscape is a remarkably effective tool for finding novel stable protein sequences.