Structural models of DYNLL1 with interacting partners: African swine fever virus protein p54 and postsynaptic scaffolding protein gephyrin.

DYNLL1, the smallest dynein light chain, interacts with different cargos facilitating their cellular transport. Usually the sequence recognized in the targets is homologous to the GIQVD or the KXTQT motifs with a glutamine that is important for binding. Here we add two new examples of DYNLL1 targets that can be classified into these two groups: ASFV p54 and gephyrin. Using NMR we demonstrate the direct interaction between DYNLL1 and two peptides derived from their interacting sequences. We model the structure of both complexes and show that the overall binding mode is preserved as in other complexes despite differences at the residue-specific interactions.

African swine fever virus protein p54 interacts with the microtubular motor complex through direct binding to light-chain dynein.

Dynein is a minus-end-directed microtubule-associated motor protein involved in cargo transport in the cytoplasm. African swine fever virus (ASFV), a large DNA virus, hijacks the microtubule motor complex cellular transport machinery during virus infection of the cell through direct binding of virus protein p54 to the light chain of cytoplasmic dynein (LC8). Interaction of p54 and LC8 occurs both in vitro and in cells, and the two proteins colocalize at the microtubular organizing center during viral infection. p50/dynamitin, a dominant-negative inhibitor of dynein-dynactin function, impeded ASFV infection, suggesting an essential role for dynein during virus infection. A 13-amino-acid domain of p54 was sufficient for binding to LC8, an SQT motif within this domain being critical for this binding. Direct binding of a viral structural protein to LC8, a small molecule of the dynein motor complex, could constitute a molecular mechanism for microtubule-mediated virus transport.

Identification of novel cellular proteins that bind to the LC8 dynein light chain using a pepscan technique.

Dynein is a minus end-directed microtubule motor that serves multiple cellular functions. We have performed a fine mapping of the 8 kDa dynein light chain (LC8) binding sites throughout the development of a library of consecutive synthetic dodecapeptides covering the amino acid sequences of the various proteins known to interact with this dynein member according to the yeast two hybrid system. Two different consensus sequences were identified: GIQVD present in nNOS, in DNA cytosine methyl transferase and also in GKAP, where it is present twice in the protein sequence. The other LC8 binding motif is KSTQT, present in Bim, dynein heavy chain, Kid-1, protein 4 and also in swallow. Interestingly, this KSTQT motif is also present in several viruses known to associate with microtubules during retrograde transport from the plasma membrane to the nucleus during viral infection.

Cloning, expression, and purification of histidine-tagged preS domains of hepatitis B virus.

The preS domains of the hepatitis B virus are hydrophilic polypeptides that have been implicated, among other functions, in the binding of the virus to hepatocytes and in the induction of virus-neutralizing antibodies. A method of overproducing the preS domains of two different subtypes, adw and ayw, has been developed by adding a 6x His tag at the carboxy-terminal end of the polypeptides. Codons for the 6x His were added in reverse primers used to amplify the corresponding cDNAs. The polymerase chain reaction products were cloned into the expression vectors pET-3d (subtype ayw) and pT7-7 (subtype adw), under the control of the inducible bacteriophage T7 RNA polymerase promoter. Upon induction with isopropyl-beta-d-thiogalactopyranoside, proteins were overexpressed and purified by affinity chromatography on a Ni-nitrilotriacetic acid agarose column. This method yielded 20-40 mg of highly pure and very stable proteins per liter of cell culture. Circular dichroism and fluorescence spectroscopy of isolated preS-his-ayw and preS-his-adw, as well as their ability to bind polymerized human serum albumin, indicate that the 6x His tag does not modify the native-like conformation and, therefore, they may be considered as useful tools to study the function of these viral polypeptide regions.

The Akt kinase signals directly to endothelial nitric oxide synthase.

Endothelial nitric oxide synthase (eNOS) is an important modulator of angiogenesis and vascular tone [1]. It is stimulated by treatment of endothelial cells in a phosphatidylinositol 3-kinase (PI 3-kinase)-dependent fashion by insulin-like growth factor-1 (IGF-1) and vascular endothelial growth factor (VEGF) [2] [3] and is activated by phosphorylation at Ser1177 in the sequence RIRTQS(1177)F (in the single-letter amino acid code) [4]. The protein kinase Akt is an important downstream target of PI 3-kinase [5] [6], regulating VEGF-stimulated endothelial cell survival [7]. Akt phosphorylates substrates within a defined motif [8], which is present in the sequence surrounding Ser1177 in eNOS. Both Akt [5] [6] and eNOS [9] are localized to, and activated at, the plasma membrane. We found that purified Akt phosphorylated cardiac eNOS at Ser1177, resulting in activation of eNOS. Phosphorylation at this site was stimulated by treatment of bovine aortic endothelial cells (BAECs) with VEGF or IGF-1, and Akt was activated in parallel. Preincubation with wortmannin, an inhibitor of Akt signalling, reduced VEGF- or IGF-1-induced Akt activity and eNOS phosphorylation. Akt was detected in immunoprecipitates of eNOS from BAECs, and eNOS in immunoprecipitates of Akt, indicating that the two enzymes associate in vivo. It is thus apparent that Akt directly activates eNOS in endothelial cells. These results strongly suggest that Akt has an important role in the regulation of normal angiogenesis and raise the possibility that the enhanced activity of this kinase that occurs in carcinomas may contribute to tumor vascularization and survival.

Fusogenic activity of hepadnavirus peptides corresponding to sequences downstream of the putative cleavage site.

Sequence homology between the amino-terminal region of the S protein of hepatitis B Virus (HBV) and known fusion peptides from retroviruses and paramyxoviruses led us to propose that this region might be equally involved in the initial infective steps of hepadnaviruses. In fact, we showed that a synthetic peptide corresponding to the N-terminus region of the S protein of HBV had membrane-interacting properties and was able to induce liposome fusion adopting an extended (beta-sheet) conformation (Rodríguez-Crespo et al., 1996, 1995). We describe herein studies on the interaction of peptides derived from the N-terminal region of the S protein of duck (DHBV: Met-Ser-Gly-Thr-Phe-Gly-Gly-Ile-Leu-Ala-Gly-Leu-Ile-Gly-Leu-Leu) and woodchuck hepatitis B viruses (WHV: Met-Ser-Pro-Ser-Ser-Leu-Leu-Gly-Leu-Leu-Ala-Gly-Leu-Gln-Val-Val) with liposomes. These peptides were able to induce to a different extent aggregation, lipid mixing, and leakage of internal aqueous contents from both neutral and negatively charged phospholipid vesicles in a concentration-dependent and pH-independent manner. Fluorescence depolarization of 1,6-diphenyl-1,3,5-hexatriene-labeled vesicles indicated that both peptides become inserted into the hydrophobic core of the lipid bilayer. Circular dichroism studies indicated that the DHBV peptide adopts an extended conformation in the presence of lipids, whereas the WHV peptide displays a high content of alpha-helical conformation. Therefore, these results extend our previous findings obtained for human hepatitis B virus to other members of the hepadnavirus family and suggest that this region of the S protein is important in the initial steps of the infective cycle.

Nitric oxide synthases catalyze the activation of redox cycling and bioreductive anticancer agents.

Nitric oxide synthases (NOSs) play a crucial role in the control of blood flow, memory formation, and the immune response. These proteins can be structurally divided into oxygenase and reductase domains. The reductase domain shares a high degree of sequence homology with P450 reductase, which is thought to be the major enzyme responsible for the one-electron reduction of foreign compounds, including bioreductive antitumor agents currently undergoing clinical trials. In view of the structural similarities between NOS and P450 reductase, we investigated the capacity of NOS to reduce the hypoxic cytotoxin tirapazamine, the antitumor agent doxorubicin, and also the redox cycling compound menadione. All three isoforms exhibited high levels of activity toward these compounds. In the case of doxorubicin and menadione, the activity of NOS II was 5-10-fold higher than the other enzymes, whereas with tirapazamine, the activities were broadly similar. NOS-mediated metabolism of tirapazamine resulted in a large increase in plasmid DNA strand breaks, demonstrating that the reduction was a bioactivation process. In addition, tirapazamine inhibited NOS activity. Because nitric oxide is implicated in maintaining tumor vascular homeostasis, it is conceivable that tirapazamine could potentiate its own toxicity by increasing the degree of hypoxia. This study suggests that the NOSs could play a key role in the therapeutic effects of tirapazamine, particularly because NOS activity is markedly increased in several human tumors. In addition, the presence of NOS in the heart indicates that these enzymes may contribute to the cardiotoxicity of redox cycling drugs, such as doxorubicin.

Binding of dynein light chain (PIN) to neuronal nitric oxide synthase in the absence of inhibition.

PIN, an 89-amino-acid polypeptide found in a rat hippocampal cDNA library using the yeast two-hybrid system and various neuronal nitric oxide synthase (nNOS) fragments as bait, was reported to be an inhibitor of nNOS (Science 274, 774-778, 1996). PIN reportedly inhibited nNOS selectively and did not interact with either the endothelial or inducible nitric oxide synthase isoforms. Inhibition was attributed to the ability of PIN to dissociate the catalytically active nNOS homodimer. PIN is a dynein light chain (J. Biol. Chem. 271, 19358-19366, 1996), which suggested that PIN may serve as an axonal transport protein for nNOS. We have synthesized a rat PIN cDNA by recursive polymerase chain reaction and have expressed the protein in Escherichia coli. Recombinant PIN is a folded dimeric, mostly alpha-helical protein with a single deeply buried tryptophan residue. We have also expressed and purified the nNOS fragment to which PIN reportedly binds (residues 163-245). This recombinant peptide has a disordered secondary structure. Gel-filtration experiments show that PIN binds to both the full-length nNOS and nNOS fragment. However, PIN neither inhibits nNOS activity nor dissociates the nNOS dimer into monomeric species. PIN thus possibly functions as a dynein light chain involved in nNOS axonal transport but is not an inhibitor of the enzyme. Our results agree with the proposal (Cell 82, 743-752, 1995) that the PIN recognition sequence in nNOS both lies outside the catalytic core and is not part of the monomer-monomer contact region.

Endothelial nitric oxide synthase: modulations of the distal heme site produced by progressive N-terminal deletions.

cDNAs coding for bovine endothelial nitric oxide synthase (eNOS) with N-terminal deletions of 52, 91, and 105 amino acids were constructed, and the proteins were expressed in Escherichia coli and purified by affinity chromatography. All three truncated proteins bind heme and exhibit the ferrous-CO absorption maximum at 444 nm characteristic of thiolate heme ligation. Deletion of the first 52 amino acids yields a fully active dimeric protein with the same spectroscopic properties as the wild-type. The myristoylation, palmitoylation, and polyproline domains of the enzyme located in the deleted region are therefore not required for full catalytic activity. The delta91 and delta105 proteins, which exhibit altered dimerization equilibria, retain 20 and 12%, respectively, of the maximal activity. Resonance Raman and UV-vis spectroscopy indicate that, in the absence of tetrahydrobiopterin (H4B) and l-Arg, the wild-type and delta52 proteins are predominantly five coordinate high spin, whereas the delta91 and delta105 proteins are six coordinate low spin. The delta91 and delta105 mutants bind H4B, as indicated by a concomitant decrease in the low-spin component of the UV-vis spectrum, but the binding of l-Arg is extremely slow ( approximately 15 min). Dithiothreitol readily coordinates as the sixth iron ligand in the delta91 and delta105 mutants but not in the delta52 or wild-type proteins. The dithiothreitol can be completely displaced by l-Arg but not by H4B. Resonance Raman comparison of wild-type eNOS and nNOS confirms that, in the absence of H4B and l-Arg, eNOS is primarily high spin whereas nNOS is predominantly six coordinate, low spin. The results indicate that Cys-101 is not critical for the binding of H4B and imply that some of the protein residues involved in dimer formation and in preservation of active site integrity are located, probably at the monomer-monomer interface, in the N-terminal end of the protein.

Cloning, expression, purification, and characterization of the major core protein (p26) from equine infectious anemia virus.

The gene coding for the major core protein (p26) of the lentivirus equine infectious anemia virus (EIAV) was cloned from EIAV infected serum, expressed in E. coli, and the resultant protein purified to electrophoretic homogeneity. The protein was expressed in a soluble form and was purified by conventional protein separation methods. When analyzed by SDS-PAGE, under both reducing and non-reducing conditions, the purified protein migrated as a 26 kDa monomer. Recombinant p26 (rp26), therefore, does not contain any intermolecular disulfide bond. Gel filtration chromatography also indicated that the protein occurs as a monomer in solution. Labeling of free sulphydryl groups with [1-14C]iodoacetamide suggests that none of the three cysteine residues of rp26 is involved in intramolecular disulfide bonds. The circular dichroism spectrum of rp26 was consistent with the following assignment of secondary structure elements: 51% a-helix, 15% beta-turn, and 34% aperiodic. Fluorescencespectroscopy revealed that the three tryptophan residues in rp26 occupy two different environments. These data support the conclusion that the recombinant protein is folded into an ordered and probably native conformation. Immunoblotting and enzyme immunoassay with EIAV infected sera demonstrated that recombinant p26 protein may be useful for diagnostic purposes.

Structural properties of the putative fusion peptide of hepatitis B virus upon interaction with phospholipids. Circular dichroism and Fourier-transform infrared spectroscopy studies.

A peptide corresponding to the N-terminal sequence of the S protein from hepatitis B virus (Met-Glu-Asn-Ile-Thr-Ser-Gly-Phe-Leu-Gly-Pro-Leu-Leu-Val-Leu-Gln) has been previously shown to interact with phospholipids and promote vesicle aggregation, phospholipid mixing, and liposome leakage, as well as erythrocyte lysis [Rodríguez-Crespo, I., Núñez, E., Gómez-Gutiérrez, J., Yélamos, B., Albar, J. P., Peterson, D. L. & Gavilanes, F. (1995) J. Gen. Virol. 76, 301-308]. The conformation of this putative fusion peptide has been studied, both at low and high peptide concentrations, by means of circular dichroism and Fourier-transform infrared spectroscopy, respectively. When the peptide is dissolved in trifluoroethanol, a significant population of alpha-helical structure is found in spite of the proline residue at position 11. In contrast, this hydrophobic oligopeptide has a high tendency to form large beta-sheet aggregates in aqueous buffers. Most of these aggregates can be eliminated by centrifugation. The peptide remaining in the supernatant adopts a non-ordered conformation. The aggregates can be dissociated by the anionic detergent sodium cholate, but the peptide still maintains an extended conformation. In the presence of acidic phospholipid vesicles, the putative fusion peptide adopts a highly stable beta-sheet conformation. Thus, unlike the fusion peptides of other viruses, an extended conformation seems to be the preferred structure when interacting with phospholipids. Such a conformation should be responsible for its membrane destabilization properties.

Phospholipid interactions of the putative fusion peptide of hepatitis B virus surface antigen S protein.

One of the first steps in the infective cycle of an enveloped virus consists of the fusion of the viral and cellular membranes. This process is usually achieved as a result of membrane destabilization brought about by a viral fusion peptide located at the amino terminus of one of the viral envelope glycoproteins. Previous sequence similarity studies by Rodríguez-Crespo et al. (Journal of General Virology 75, 637-639, 1994) have shown that a hydrophobic stretch in the amino-terminal sequence of the S protein of hepatitis B virus shares several characteristics with fusion peptides of retroviruses and paramyxoviruses. A 16 residue peptide with this sequence was synthesized and its interaction with liposomes characterized. This peptide was able to mediate vesicle aggregation, lipid mixing and liposome leakage in a pH dependent manner at concentrations ranging from 3.5 to 52.0 microM. These effects were specific for negatively charged phospholipid vesicles. The peptide was also able to haemolyse erythrocytes. This study supports the notion that the sequence might be important in the initial infective steps of this virus, interacting with the target membranes and bringing about their subsequent destabilization.

Site-directed mutagenesis of cysteine residues of hepatitis B surface antigen. Analysis of two single mutants and the double mutant.

The structure of hepatitis B surface antigen (HBsAg) is mainly maintained by an intricate disulfide network responsible for most of its structural and antigenic properties. Characterization of three cysteine-replacement mutants of HBsAg has been performed by both structural and immunological methods. Replacement of Cys121 or Cys124 with serine results in mutant proteins that show diminished binding titres to both monoclonal antibodies and to a polyclonal serum, indicating that a structural change has taken place. Circular dichroism analysis shows that the substitution of either of these two residues also diminishes the helical content of the protein. However, the double mutant, in which both cysteine residues have been simultaneously changed, reverts the properties of the single mutations, and shows similar behaviour to the wild-type protein. Both the single and double cysteine mutants are efficiently glycosylated and secreted from Chinese hamster ovary cells and, in all cases, the mutant proteins assemble into spherical particles of similar buoyant density to both the wild-type and serum derived HBsAg.

Prediction of a putative fusion peptide in the S protein of hepatitis B virus.

Sequence analysis of the S protein of hepatitis B virus (HBV) reveals a stretch of 23 hydrophobic amino acids in the amino-terminal region which shows a high degree of similarity with known fusogenic peptides from other viruses. Additionally, this sequence also appears to be highly conserved within the hepadnavirus family. Taken together, the different criteria used in this work suggest fusogenic activity in the amino-terminal region of the S protein of the envelope of HBV.