Dengue virus NS4A cytoplasmic domain binding to liposomes is sensitive to membrane curvature.

Dengue virus (DENV) infection is a growing public health threat with more than one-third of the world's population at risk. Non-structural protein 4A (NS4A), one of the least characterized viral proteins, is a highly hydrophobic transmembrane protein thought to induce the membrane alterations that harbor the viral replication complex. The NS4A N-terminal (amino acids 1-48), has been proposed to contain an amphipathic α-helix (AH). Mutations (L6E; M10E) designed to reduce the amphipathic character of the predicted AH, abolished viral replication and reduced NS4A oligomerization. Nuclear magnetic resonance (NMR) spectroscopy was used to characterize the N-terminal cytoplasmic region (amino acids 1-48) of both wild type and mutant NS4A in the presence of SDS micelles. Binding of the two N-terminal NS4A peptides to liposomes was studied as a function of membrane curvature and lipid composition. The NS4A N-terminal was found to contain two AHs separated by a non-helical linker. The above mentioned mutations did not significantly affect the helical secondary structure of this domain. However, they reduced the affinity of the N-terminal NS4A domain for lipid membranes. Binding of wild type NS4A(1-48) to liposomes is highly dependent on membrane curvature.

A Toolbox of Genetically Encoded FRET-Based Biosensors for Rapid l-Lysine Analysis.

Background: The fast development of microbial production strains for basic and fine chemicals is increasingly carried out in small scale cultivation systems to allow for higher throughput. Such parallelized systems create a need for new rapid online detection systems to quantify the respective target compound. In this regard, biosensors, especially genetically encoded Förster resonance energy transfer (FRET)-based biosensors, offer tremendous opportunities. As a proof-of-concept, we have created a toolbox of FRET-based biosensors for the ratiometric determination of l-lysine in fermentation broth. Methods: The sensor toolbox was constructed based on a sensor that consists of an optimized central lysine-/arginine-/ornithine-binding protein (LAO-BP) flanked by two fluorescent proteins (enhanced cyan fluorescent protein (ECFP), Citrine). Further sensor variants with altered affinity and sensitivity were obtained by circular permutation of the binding protein as well as the introduction of flexible and rigid linkers between the fluorescent proteins and the LAO-BP, respectively. Results: The sensor prototype was applied to monitor the extracellular l-lysine concentration of the l-lysine producing Corynebacterium glutamicum (C. glutamicum) strain DM1933 in a BioLector® microscale cultivation device. The results matched well with data obtained by HPLC analysis and the Ninhydrin assay, demonstrating the high potential of FRET-based biosensors for high-throughput microbial bioprocess optimization.

Structural features of antiviral DNA cytidine deaminases.

The APOBEC3 (A3) family of cytidine deaminases plays a vital role for innate defense against retroviruses. Lentiviruses such as HIV-1 evolved the Vif protein that triggers A3 protein degradation. There are seven A3 proteins, A3A-A3H, found in humans. All A3 proteins can deaminate cytidines to uridines in single-stranded DNA (ssDNA), generated during viral reverse transcription. A3 proteins have either one or two cytidine deaminase domains (CD). The CDs coordinate a zinc ion, and their amino acid specificity classifies the A3s into A3Z1, A3Z2, and A3Z3. A3 proteins occur as monomers, dimers, and large oligomeric complexes. Studies on the nature of A3 oligomerization, as well as the mode of interaction of A3s with RNA and ssDNA are partially controversial. High-resolution structures of the catalytic CD2 of A3G and A3F as well as of the single CD proteins A3A and A3C have been published recently. The NMR and X-ray crystal structures show globular proteins with six α-helices and five ß sheets arranged in a characteristic motif (α1-ß1-ß2/2'-α2-ß3-α3-ß4-α4-ß5-α5-α6). However, the detailed arrangement and extension of individual structure elements and their relevance for A3 complex formation and activity remains a matter of debate and will be highlighted in this review.

Full-length Vpu and human CD4(372-433) in phospholipid bilayers as seen by magic angle spinning NMR.

HIV-1 Vpu and CD4(372-433), a peptide comprising the transmembrane and cytoplasmic domain of human CD4, were recombinantly expressed in Escherichia coli, uniformly labeled with 13C and 15N isotopes, and separately reconstituted into phospholipid bilayers. Highly resolved dipolar cross-polarization (CP)-based solid-state NMR spectra of the two transmembrane proteins were recorded under magic angle sample spinning. Partial assignment of 13C resonances was achieved. Site-specific assignments were obtained for 13 amino acid residues of CD4(372-433) and two Vpu residues. Additional amino acid type-specific assignments were achieved for 10 amino acid spin systems for both CD4(372-433) and Vpu. Further, structural flexibility was probed with different dipolar recoupling techniques, and the correct insertion of the transmembrane domains into the lipid bilayers was confirmed by proton spin diffusion experiments.

NMR structure of the transmembrane and cytoplasmic domains of human CD4 in micelles.

The human cluster determinant 4 (CD4) is a type I transmembrane glycoprotein involved in T-cell signalling. It is expressed primarily on the surface of T helper cells but also on subsets of memory and regulatory T lymphocytes (CD4(+) cells). It serves as a coreceptor in T-cell receptor recognition of MHC II antigen complexes. Besides its cellular functions, CD4 serves as the main receptor for human immunodeficiency virus type I (HIV-1). During T-cell infection, the CD4 extracellular domain is bound by HIV-1 gp120, the viral surface glycoprotein, which triggers a number of conformational changes ultimately resulting in virion entry of the cell. Subsequently, CD4 is downregulated in infected cells by multiple strategies that involve direct interactions of the HIV-1 proteins VpU and Nef with the cytoplasmic part of CD4. In the present work, we describe the NOE-based solution structure of the transmembrane and cytoplasmic domains of the cystein-free variant of CD4 (CD4mut) in dodecylphosphocholine (DPC) micelles. Furthermore, we have characterized micelle-inserted CD4mut by paramagentic relaxation enhancement (PRE) agents and (1)H-(15)N heteronuclear NOE data. CD4mut features a stable and well-defined transmembrane helix from M372 to V395 buried in the micellar core and a cytoplasmic helix ranging from A404 to L413. Experimental data suggest the amphipathic cytoplasmic helix to be in close contact with the micellar surface. The role of the amphipathic helix and its interaction with the micellar surface is discussed with respect to the biological function of the full-length CD4 protein.

Nanodiscs allow the use of integral membrane proteins as analytes in surface plasmon resonance studies.

Nanodiscs are small-sized and flat model membranes that provide a close to native environment for reconstitution of integral membrane proteins. Incorporation of membrane proteins into nanodiscs results in water-soluble proteolipid particles making the membrane proteins amenable to a multitude of bioanalytical techniques originally developed for soluble proteins. The transmembrane domain of the human CD4 receptor was fused to ubiquitin with a preceding N-terminal decahistidine tag. The resulting integral membrane protein was incorporated into nanodiscs. Binding of the nanodisc-inserted histidine-tagged protein to a monoclonal anti-pentahistidine antibody was quantified using surface plasmon resonance (SPR) experiments. For the first time, a membrane-inserted transmembrane protein was employed as analyte while the antibody served as ligand immobilized on the sensor chip surface. SPR experiments were conducted in single-cycle mode. We demonstrate that the nanodisc-incorporated membrane protein showed nearly identical affinity toward the antibody as did the soluble decahistidine-tagged ubiquitin studied in a comparative experiment. Advantages of the new experimental setup and potential applications are discussed.

Thermophoresis: The Case of Streptavidin and Biotin.

Thermophoretic behavior of a free protein changes upon ligand binding and gives access to information on the binding constants. The Soret effect has also been proven to be a promising tool to gain information on the hydration layer, as the temperature dependence of the thermodiffusion behavior is sensitive to solute-solvent interactions. In this work, we perform systematic thermophoretic measurements of the protein streptavidin (STV) and of the complex STV with biotin (B) using thermal diffusion forced Rayleigh scattering (TDFRS). Our experiments show that the temperature sensitivity of the Soret coefficient is reduced for the complex compared to the free protein. We discuss our data in comparison with recent quasi-elastic neutron scattering (QENS) measurements. As the QENS measurement has been performed in heavy water, we perform additional measurements in water/heavy water mixtures. Finally, we also elucidate the challenges arising from the quantiative thermophoretic study of complex multicomponent systems such as protein solutions.

Strong Adverse Contribution of Conformational Dynamics to Streptavidin-Biotin Binding.

Molecular dynamics plays an important role for the biological function of proteins. For protein ligand interactions, changes of conformational entropy of protein and hydration layer are relevant for the binding process. Quasielastic neutron scattering (QENS) was used to investigate differences in protein dynamics and conformational entropy of ligand-bound and ligand-free streptavidin. Protein dynamics were probed both on the fast picosecond time scale using neutron time-of-flight spectroscopy and on the slower nanosecond time scale using high-resolution neutron backscattering spectroscopy. We found the internal equilibrium motions of streptavidin and the corresponding mean square displacements (MSDs) to be greatly reduced upon biotin binding. On the basis of the observed MSDs, we calculated the difference of conformational entropy ΔSconf of the protein component between ligand-bound and ligand-free streptavidin. The rather large negative ΔSconf value (-2 kJ mol-1 K-1 on the nanosecond time scale) obtained for the streptavidin tetramer seems to be counterintuitive, given the exceptionally high affinity of streptavidin-biotin binding. Literature data on the total entropy change ΔS observed upon biotin binding to streptavidin, which includes contributions from both the protein and the hydration water, suggest partial compensation of the unfavorable ΔSconf by a large positive entropy gain of the surrounding hydration layer and water molecules that are displaced during ligand binding.

Helix formation in arrestin accompanies recognition of photoactivated rhodopsin.

Binding of arrestin to photoactivated phosphorylated rhodopsin terminates the amplification of visual signals in photoreceptor cells. Currently, there is no crystal structure of a rhodopsin-arrestin complex available, although structures of unbound rhodopsin and arrestin have been determined. High-affinity receptor binding is dependent on distinct arrestin sites responsible for recognition of rhodopsin activation and phosphorylation. The loop connecting beta-strands V and VI in rod arrestin has been implicated in the recognition of active rhodopsin. We report the structure of receptor-bound arrestin peptide Arr(67-77) mimicking this loop based on solution NMR data. The peptide binds photoactivated rhodopsin in the unphosphorylated and phosphorylated form with similar affinities and stabilizes the metarhodopsin II photointermediate. A largely alpha-helical conformation of the receptor-bound peptide is observed.

Integral membrane proteins in nanodiscs can be studied by solution NMR spectroscopy.

We present a two-dimensional solution NMR spectrum of an integral membrane protein (IMP) in a nanodisc. Solution NMR relies on rapid isotropic tumbling of the analyte with correlation times in the nanosecond range. IMPs in a cellular membrane do not satisfy this condition. Previous liquid-state NMR studies on IMPs were conducted in organic solvent or artificial membrane mimicking particles like detergent micelles. Nanodiscs are relatively small (150 kDa), detergent-free model membranes that are suitable for functional reconstitution of IMPs. Nanodiscs allow solubilization of integral membrane proteins in a nearly native lipid bilayer environment. The 70 residue polypeptide CD4mut was incorporated into nanodiscs. CD4mut features one transmembrane helix. The aliphatic (1)H-(13)C HSQC spectrum of nanodiscs with inserted, ((13)C, (15)N)-labeled CD4mut exhibits reasonably dispersed protein and lipid NMR signals. Our results demonstrate that IMPs in nanodiscs are amenable to liquid-state NMR methodology.

Structure of the SLy1 SAM homodimer reveals a new interface for SAM domain self-association.

Sterile alpha motif (SAM) domains are protein interaction modules that are involved in a diverse range of biological functions such as transcriptional and translational regulation, cellular signalling, and regulation of developmental processes. SH3 domain-containing protein expressed in lymphocytes 1 (SLy1) is involved in immune regulation and contains a SAM domain of unknown function. In this report, the structure of the SLy1 SAM domain was solved and revealed that this SAM domain forms a symmetric homodimer through a novel interface. The interface consists primarily of the two long C-terminal helices, α5 and α5', of the domains packing against each other. The dimerization is characterized by a dissociation constant in the lower micromolar range. A SLy1 SAM domain construct with an extended N-terminus containing five additional amino acids of the SLy1 sequence further increases the stability of the homodimer, making the SLy1 SAM dimer two orders of magnitude more stable than previously studied SAM homodimers, suggesting that the SLy1 SAM dimerization is of functional significance. The SLy1 SAM homodimer contains an exposed mid-loop surface on each monomer, which may provide a scaffold for mediating interactions with other SAM domain-containing proteins via a typical mid-loop-end-helix interface.

Structural characterization of the transmembrane and cytoplasmic domains of human CD4.

Cluster determinant 4 (CD4) is a type I transmembrane glycoprotein of 58 kDa. It consists of an extracellular domain of 370 amino acids, a short transmembrane region, and a cytoplasmic domain of 40 amino acids at the C-terminal end. We investigated the structure of the 62 C-terminal residues of CD4, comprising its transmembrane and cytoplasmic domains. The five cysteine residues of this region have been replaced with serine and histidine residues in the polypeptide CD4mut. Uniformly 15N and 13C labeled protein was recombinantly expressed in E. coli and purified. Functional binding activity of CD4mut to protein VpU of the human immunodeficiency virus type 1 (HIV-1) was verified. Close to complete NMR resonance assignment of the 1H, 13C, and 15N spins of CD4mut was accomplished. The secondary structure of CD4mut in membrane simulating dodecylphosphocholine (DPC) micelles was characterized based on secondary chemical shift analysis, NOE-based proton-proton distances, and circular dichroism spectroscopy. A stable transmembrane helix and a short amphipathic helix in the cytoplasmic region were identified. The fractional helicity of the cytoplasmic helix appears to be stabilized in the presence of DPC micelles, although the extension of this helix is reduced in comparison to previous studies on synthetic peptides in aqueous solution. The role of the amphipathic helix and its potentially variable length is discussed with respect to the biological functions of CD4.

An indole-binding site is a major determinant of the ligand specificity of the GABA type A receptor-associated protein GABARAP.

The role of tryptophan as a key residue for ligand binding to the ubiquitin-like modifier GABA(A) receptor associated protein (GABARAP) was investigated. Two tryptophan-binding hydrophobic patches were identified on the conserved face of the GABARAP structure by NMR spectroscopy and molecular docking. GABARAP binding of indole and indole derivatives, including the free amino acid tryptophan was quantified. The two tryptophan binding sites can be clearly distinguished by mapping the NMR spectroscopy-derived residue-specific apparent dissociation constant, K(d), onto the three-dimensional structure of GABARAP. The biological relevance of tryptophan-binding pockets of GABARAP was supported by a highly conserved tryptophan residue in the GABARAP binding region of calreticulin, clathrin heavy chain, and the gamma2 subunit of the GABA(A) receptor. Replacement of tryptophan by alanine abolished ligand binding to GABARAP.

Enhancing the Catalytic Deamination Activity of APOBEC3C Is Insufficient to Inhibit Vif-Deficient HIV-1.

The retroviral restriction factors of the APOBEC3 (A3) cytidine deaminase family catalyze the deamination of cytidines in single-stranded viral DNA. APOBEC3C (A3C) is a strong antiviral factor against viral infectivity factor (vif)-deficient simian immunodeficiency virus Δvif, which is, however, a weak inhibitor against human immunodeficiency virus (HIV)-1 for reasons unknown. The precise link between the antiretroviral effect of A3C and its catalytic activity is incompletely understood. Here, we show that the S61P mutation in human A3C (A3C.S61P) boosted hypermutation in the viral genomes of simian immunodeficiency virus Δvif and murine leukemia virus but not in human immunodeficiency virus HIV-1Δvif. The enhanced antiviral activity of A3C.S61P correlated with enhanced in vitro cytidine deamination. Furthermore, the S61P mutation did not change the substrate specificity of A3C, ribonucleoprotein complex formation, self-association, Zinc coordination, or viral incorporation features. We propose that local structural changes induced by the serine-to-proline substitution are responsible for the gain of catalytic activity of A3C.S61P. Our results are a first step toward an understanding of A3C's DNA binding capacity, deamination-dependent editing, and antiviral functions at the molecular level. We conclude that the enhanced enzymatic activity of A3C is insufficient to restrict HIV-1, indicating an unknown escape mechanism of HIV-1.

Specific volumes of unsaturated phosphatidylcholines in the liquid crystalline lamellar phase.

The specific volumes of seven 1,2-diacyl-sn-glycero-3-phosphocholines with symmetric, unbranched acyl chains containing one, four, or six cis double bonds per chain, or with a saturated sn-1 chain and one, four, or six cis double bonds in the sn-2 chain were determined by the neutral buoyancy method. Experiments were conducted in the liquid crystalline lamellar phase over the temperature range from 5 to 35 degrees C. It is demonstrated that the molecular volume of phosphatidylcholines can be well approximated as the sum of a constant volume of the polar lipid head region and the temperature-dependent volumes of hydrocarbon chain CH2, CH, and terminal CH3 groups. A linear dependence of chain segment volumes on temperature was observed. A self-consistent set of partially temperature-dependent volumes is obtained that allows prediction of phosphatidylcholine molecular volumes within very tight error margins.

Structure and orientation of a G protein fragment in the receptor bound state from residual dipolar couplings.

Residual dipolar couplings for a ligand that is in fast exchange between a free state and a state where it is bound to a macroscopically ordered membrane protein carry precise information on the structure and orientation of the bound ligand. The couplings originate in the bound state but can be detected on the free ligand using standard high resolution NMR. This approach is used to study an analog of the C-terminal undecapeptide of the alpha-subunit of the heterotrimeric G protein transducin when bound to photo-activated rhodopsin. Rhodopsin is the major constituent of disk-shaped membrane vesicles from rod outer segments of bovine retinas, which align spontaneously in the NMR magnet. Photo-activation of rhodopsin triggers transient binding of the peptide, resulting in measurable dipolar contributions to 1J(NH) and 1J(CH) splittings. These dipolar couplings report on the time-averaged orientation of bond vectors in the bound peptide relative to the magnetic field, i.e. relative to the membrane normal. Approximate distance restraints of the bound conformation were derived from transferred NOEs, as measured from the difference of NOESY spectra recorded prior to and after photo-activation. The N-terminal eight residues of the bound undecapeptide adopt a near-ideal alpha-helical conformation. The helix is terminated by an alpha(L) type C-cap, with Gly9 at the C' position in the center of the reverse turn. The angle between the helix axis and the membrane normal is 40 degrees (+/-4) degrees. Peptide protons that make close contact with the receptor are identified by analysis of the NOESY cross-relaxation pattern and include the hydrophobic C terminus of the peptide.

Lipid-ethanol interaction studied by NMR on bicelles.

The interaction of ethanol with phospholipids was studied in bicelles at a physiologically relevant ethanol concentration of 20 mM and a lipid content of 14 wt % by high-resolution NMR. Transient association of ethanol with magnetically aligned bicelles imparts a small degree of anisotropy to the solute. This anisotropy allows detection of residual (1)H-(1)H and (1)H-(13)C dipolar couplings, which are superimposed on scalar couplings. Residual (2)H NMR quadrupole splittings of isotope-labeled ethanol were measured as well. The analysis of residual tensorial interactions yielded information on the orientation and motions of ethanol in the membrane-bound state. The fraction of phosphatidylcholine-bound ethanol was determined independently by gas chromatography and NMR. About 4% of ethanol is bound to phosphatidylcholine at a bicelle concentration of 14 wt % at 40 degrees C. Free and bound ethanol are in rapid exchange. The lifetime of ethanol association with phosphatidylcholine membranes is of the order of a few nanoseconds.

Recombinant production of the amino terminal cytoplasmic region of dengue virus non-structural protein 4A for structural studies.

BACKGROUND: Dengue virus (DENV) is a mosquito-transmitted positive single strand RNA virus belonging to the Flaviviridae family. DENV causes dengue fever, currently the world's fastest-spreading tropical disease. Severe forms of the disease like dengue hemorrhagic fever and dengue shock syndrome are life-threatening. There is no specific treatment and no anti-DENV vaccines. Our recent data suggests that the amino terminal cytoplasmic region of the dengue virus non-structural protein 4A (NS4A) comprising amino acid residues 1 to 48 forms an amphipathic helix in the presence of membranes. Its amphipathic character was shown to be essential for viral replication. NMR-based structure-function analysis of the NS4A amino terminal region depends on its milligram-scale production and labeling with NMR active isotopes. METHODOLOGY/PRINCIPAL FINDINGS: This report describes the optimization of a uniform procedure for the expression and purification of the wild type NS4A(1-48) peptide and a peptide derived from a replication-deficient mutant NS4A(1-48; L6E, M10E) with disrupted amphipathic nature. A codon-optimized, synthetic gene for NS4A(1-48) was expressed as a fusion with a GST-GB1 dual tag in E. coli. Tobacco etch virus (TEV) protease mediated cleavage generated NS4A(1-48) peptides without any artificial overhang. Using the described protocol up to 4 milligrams of the wild type or up to 5 milligrams of the mutant peptide were obtained from a one-liter culture. Isotopic labeling of the peptides was achieved and initial NMR spectra were recorded. CONCLUSIONS/SIGNIFICANCE: Small molecules targeting amphipathic helices in the related Hepatitis C virus were shown to inhibit viral replication, representing a new class of antiviral drugs. These findings highlight the need for an efficient procedure that provides large quantities of the amphipathic helix containing NS4A peptides. The double tag strategy presented in this manuscript answers these needs yielding amounts that are sufficient for comprehensive biophysical and structural studies, which might reveal new drug targets.

Mapping the interaction between the cytoplasmic domains of HIV-1 viral protein U and human CD4 with NMR spectroscopy.

Viral protein U (VpU) of HIV-1 plays an important role in downregulation of the main HIV-1 receptor CD4 from the surface of infected cells. Physical binding of VpU to newly synthesized CD4 in the endoplasmic reticulum is an early step in a pathway leading to proteasomal degradation of CD4. In this study, regions in the cytoplasmic domain of VpU involved in CD4 binding were identified by NMR spectroscopy. Amino acids in both helices found in the cytoplasmic region of VpU in membrane-mimicking detergent micelles experience chemical shift perturbations upon binding to CD4, whereas amino acids between the two helices and at the C-terminus of VpU show no or only small changes, respectively. The topology of the complex was further studied with paramagnetic relaxation enhancement. Paramagnetic spin labels were attached at three sequence positions of a CD4 peptide comprising the transmembrane and cytosolic domains of the receptor. VpU binds to a membrane-proximal region in the cytoplasmic domain of CD4.

Single vector system for efficient N-myristoylation of recombinant proteins in E. coli.

BACKGROUND: N-myristoylation is a crucial covalent modification of numerous eukaryotic and viral proteins that is catalyzed by N-myristoyltransferase (NMT). Prokaryotes are lacking endogenous NMT activity. Recombinant production of N-myristoylated proteins in E. coli cells can be achieved by coexpression of heterologous NMT with the target protein. In the past, dual plasmid systems were used for this purpose. METHODOLOGY/PRINCIPAL FINDINGS: Here we describe a single vector system for efficient coexpression of substrate and enzyme suitable for production of co- or posttranslationally modified proteins. The approach was validated using the HIV-1 Nef protein as an example. A simple and efficient protocol for production of highly pure and completely N-myristoylated Nef is presented. The yield is about 20 mg myristoylated Nef per liter growth medium. CONCLUSIONS/SIGNIFICANCE: The single vector strategy allows diverse modifications of target proteins recombinantly coexpressed in E. coli with heterologous enzymes. The method is generally applicable and provides large amounts of quantitatively processed target protein that are sufficient for comprehensive biophysical and structural studies.