The lipocalin protein family: structural and sequence overview.

Lipocalins are remarkably diverse at the sequence level yet have highly conserved structures. Most lipocalins share three characteristic conserved sequence motifs - the kernel lipocalins - while others are more divergent family members - the outlier lipocalins - typically sharing only one or two. This classification is a useful tool for analysing the family, and within these large sets are smaller groups sharing much higher levels of sequence similarity. The lipocalins are also part of a larger protein superfamily: the calycins, which includes the fatty acid binding proteins, avidins, a group of metalloproteinase inhibitors, and triabin. The superfamily is characterised by a similar structure (a repeated +1 topology beta-barrel) and by the conservation of a remarkable structural signature.

Molecular modelling and site-directed mutagenesis of the active site of endothelin-converting enzyme.

Mammalian endothelin-converting enzyme is a membrane-bound metalloprotease; its C-terminal domain contains sequence motifs characteristic of zinc metalloproteases. We examined residues expected from molecular modelling to be important for substrate binding using selectively mutated recombinant rat ECE-1alpha expressed in CHO cells. A conserved N-A-Ar-Ar (Ar = aromatic) motif is likely to be important for substrate binding. Mutating N550 to Gln or Y552 to Phe reduces Vmax/Km by 8- and 18-fold, respectively. The equivalent residue to Y553 in thermolysin binds the inhibitor through its NH group. Removing this putative interaction by mutating Tyr to Pro destroys activity, but mutating it to Ala or Phe also removes most activity. Mutating G583 (in a conserved GGI motif N-terminal of the zinc-binding helix) to Ala has no measurable effect, but mutating G584 to Ala destroys activity. Changing V583 in the zinc-binding helix to Met, to mimic the sequence pattern in bovine ECE-2, increases Vmax/Km to 1.7-fold that of the wild-type. Assays of phosphoramidon binding follow the pattern of those of substrate binding, but the IC50 of the more potent ECE inhibitor CGS 26303 was not significantly altered by any of these mutations, suggesting that this compound may bind to ECE in a different mode from phosphoramidon.

Mutagenesis of Glu403 to Cys in rabbit neutral endopeptidase-24.11 (neprilysin) creates a disulphide-linked homodimer: analogy with endothelin-converting enzyme.

Neutral endopeptidase-24.11 (NEP; neprilysin; EC 3.4.24.11) and endothelin-converting enzyme (ECE) are related zinc metallopeptidases involved in the processing of biologically active peptides. Only ECE, however, exists as a disulphide-linked homodimer. The covalent linkage in rat ECE is between Cys412 in each subunit, which is equivalent to Glu403 in rabbit NEP. Here we report that directed mutagenesis of Glu403 to cysteine in rabbit NEP creates a disulphide-linked homodimer, as revealed by transient transfection in COS-1 cells and SDS/PAGE of a membrane fraction. Under reducing conditions, both the mutant (E403C) and the wild-type NEP migrate as a polypeptide of 92 kDa. However, under non-reducing conditions, the Mr of the wild type remains unchanged, whereas that of the mutant is doubled. Co-transfection of wild-type ECE and E403C NEP cDNA did not result in the production of a NEP-ECE heterodimer. Comparison of the kinetic constants for wild-type and E403C mutant NEP with either [D-Ala2,Leu5]enkephalin or 3-carb oxypropanoyl-alanyl-alanyl- leucine-4-nitroanilide(Suc-Ala-Ala-Leu-NH-Np) as substrate show a decrease of approx. 50% in Vmax/Km for the mutant form. The IC50 value for inhibition of the mutant by phosphoramidon or thiorphan is increased 3-fold and 5-fold respectively. Although NEP and ECE exhibit only about 40% identity and differ substantially in substrate specificity and some other characteristics, these data indicate that they have considerable similarity in three-dimensional structure, allowing dimer formation in the mutant NEP with the disulphide link probably occurring in a hydrophilic surface loop.

Studies of the membrane topology of the rat erythrocyte H+/lactate cotransporter (MCT1).

1. Hydrophobicity analysis of the monocarboxylate/proton cotransporter MCT1 (lactate transporter) suggests a structure with 12 transmembrane (TM) segments, presumed to be alpha-helical. 2. A series of anti-peptide antibodies have been raised against regions of the MCT1 sequence, which each recognize a polypeptide of approx. 40 kDa in rat erythrocytes. The topology of rat MCT1 was investigated by studying the immunoreactive fragments derived from proteolytic digestion of the protein in intact rat erythrocytes and leaky membranes. 3. Reactivity with an anti-(C-terminus) antibody was prevented on treatment of leaky membranes, but not intact cells, with carboxypeptidase Y, indicating that the C-terminus of the protein is cytoplasmically disposed. 4. Treatment of intact cells in saline buffer with trypsin, chymotrypsin, bromelain and protease K (up to 1 mg/ml) resulted in no degradation of MCT1, indicating the absence of any large exposed extracellular loop. In a buffer of low ionic strength (containing sucrose), cleavage was observed with bromelain at an extracellular site, probably TM9/10.5. Treatment of leaky membranes with low (less than 100 micrograms/ml) concentrations of several proteases resulted in fragmentation of MCT1, reflecting cleavage at the cytoplasmic face of the membrane. These treatments generated N-terminal fragments of apparent molecular mass approx. 17-19 kDa that were resistant to further degradation. The epitopes for the TM6/7 and C-terminal antibodies were either lost from the membrane or destroyed under most of these conditions, indicating that these regions of the protein are located in the cytoplasm. 6. More detailed structural prediction analysis of MCT-related sequences was made assuming the constraints placed upon the possible arrangements by the experimental data outlined above. This analysis provided additional strong evidence for the 12-TM-segment model, with cytoplasmic N- and C-terminal ends and a large internal loop between TM6 and TM7. The predicted helices were assigned moments of hydrophobicity and residue substitution; for a number of TM segments this permitted the prediction of the sides of the helix that faced membrane lipid and the interior of the protein.

Mutational analysis and molecular modelling of an amino acid sequence motif conserved in antiporters but not symporters in a transporter superfamily.

Elements of a 'G X8 G X3 G P X2 G G' amino acid sequence motif were conserved in the fifth predicted membrane-spanning domains of 31 antiporters, but none of 27 symporters or uniporters that together comprise a 'superfamily' of structurally, related transport proteins. Molecular modelling and mechanics predicted that the GP dipeptide of this motif bends the antiporters' fifth transmembrane helices, and that the repeating pattern of glycine residues forms a pocket, devoid of side chains, on the surface of these helices. The glycine residue in the motif's GP dipeptide was conserved in 90% of these antiporters with alanine being the only observed substitution. Replacement of the glycine residue of the GP dipeptide with alanine and serine reduced the level of tetracycline resistance conferred by TetA(C), a tetracycline/H+ antiporter, by 74 and 81%, respectively. All other substitutions totally abolished resistance to tetracycline. In contrast, replacement of the glycine residue of the GP dipeptide did not abolish increased susceptibility to cadmium, another phenotype conferred by TetA(C) independent of resistance to tetracycline. These results suggest that the glycine of the GP dipeptide is necessary for the tetracycline/H+ antiport activity of TetA(C), rather than its expression, stability, or general three-dimensional structure.

Topological analyses of the L-fucose-H+ symport protein, FucP, from Escherichia coli.

The transport of L-fucose into Escherichia coli is mediated by the L-fucose-H+ symport protein (FucP). The fucP gene has been sequenced and encodes a hydrophobic protein that contains 438 amino acid residues, with a predicted M(r) of 47,773. The hydropathic profile of FucP indicates 10 to 12 hydrophobic regions that could span the membrane as alpha-helices. A 12-helix model with the N- and C-termini located in the cytoplasm was derived from the hydropathic profile and from application of the 'positive inside' rule. This model was tested using beta-lactamase fusion technology. Analyses of 62 different FucP-beta-lactamase fusions suggested that the FucP protein crosses the cytoplasmic membrane of E. coli 12 times, with the N- and C-termini in the cytoplasm. From measurements of [14C]-L-fucose uptake, it was deduced that the last putative transmembrane region must be complete for transport activity to be retained and that the four C-terminal residues were unnecessary for transport activity. Fourier transform analyses show that all the predicted helices contain a periodicity that enables hydrophobic/hydrophilic faces to be identified; these were particularly evident in putative helices 1, 3, 4, 5, 6, 10 and 11.

Molecular modeling of the active site of endothelin-converting enzyme.

Endothelin-converting enzyme (ECE) is a member of the zinc metalloproteinase family. It is much more specific in its protease activity than the bacterial metalloprotease thermolysin; we aim to construct a model of its active site to help to explain these differences. We aligned the sequence of human ECE with those of human neprilysin (which is 39% identical to ECE) and thermolysin. Residues believed to be important for inhibitor binding were assigned from the alignment and by analogy with structural and functional studies of these enzymes. These included a conserved IGG motif N-terminal to the zinc-binding HExxH motif, and a tyrosine residue that may be analogous to Y157 of thermolysin. We have used the program O to build a model of the active site of ECE based on the crystal structure of thermolysin.

Structural analysis and classification of lipocalins and related proteins using a profile-search method.

A three-dimensional profile method of detecting amino-acid sequences compatible with the tertiary structure of any protein has been applied to the lipocalin family of 8-stranded beta-barrels. Profiles derived from six well-resolved lipocalin crystal structures were used to search a comprehensive, non-redundant protein sequence database. Each profile identified a sub-group of lipocalin sequences although no single profile was sufficient to identify the whole family. The alpha-1-acid glycoprotein sub-family was not identified by any lipocalin profile, indicating that known sequence differences in otherwise well conserved regions of these proteins may be reflected in structural differences. The predicted similarity between the beta-lactoglobulin and alpha-2u-globulin structures was much more marked than the similarity between their sequences, and alpha-1-microglobulin sequences were found to be compatible with the structure of epididymal retinoic acid binding protein which has an additional long C-terminal helix. Proteins of unknown structure which were predicted to be compatible with the lipocalin fold include a human mucin. In cases where a large protein family of low overall sequence similarity contains a small number of known structures, this technique can be useful in determining or confirming subtle structural relationships between family members.

Molecular mechanics analysis of inhibitor binding to HIV-1 protease.

Crystallographic structures of HIV protease with three different peptide-mimetic inhibitors were subjected to energy minimization using molecular mechanics, the minimized structures analyzed and the inhibitor binding energies calculated. Partial charge assignment for the hydrogen bonded catalytic aspartic acids, Asp25 and -25', was in good agreement with charge calculations using semi-empirical molecular orbital methods. Root mean square deviations on minimization were small and similar for both subunits in the protease dimer. The surface loops, which had the largest B factors, changed most on minimization; the hydrophobic core and the inhibitor binding site showed little change. The distance-dependent dielectric of D(r) = 4r was found to be preferable to D(r) = r. Distance restraints were applied for the intermolecular hydrogen bonds to maintain the conformation of the inhibitor binding site. Using the dielectric of D(r) = 4r, the calculated interaction energy of the three inhibitors with the protease ranged from -53 to -56 kcal/mol. The psi groups of the inhibitors were changed to add or remove a 'transition state analogue' hydroxyl group, and the loss in energy on the removal of this group was calculated to be 0.9-1.7 kcal/mol. This would represent 19-36% of the total measured difference in binding energy between the inhibitors JG365 and MVT-101.

Antitumor imidazotetrazines. 25. Crystal structure of 8-carbamoyl-3-methylimidazo[5,1-d]-1,2,3,5-tetrazin-4(3H)-one (temozolomide) and structural comparisons with the related drugs mitozolomide and DTIC.

The antitumor imidazotetrazinone, temozolomide (5), C6H6N6O2, forms crystals with unit cell dimensions a = 17.332 (3), b = 7.351 (2), c = 13.247 (1), beta = 109.56 (1) degree and space group P21/c. A doubly hydrogen-bonded dimer constitutes the asymmetric unit. One carboxamide group forms an additional intermolecular NH...O hydrogen bond; in both molecules the carboxamide group is coplanar with the heterocycle and its NH2 group interacts with the imidazole nitrogen atom N(7). Molecular orbital calculations show the carbonyl carbon C(4) to be the most electron deficient atom, with relatively weak N(3)-C(4) and C(4)-N(5) bonds confirming that temozolomide should ring-open at this position in solution. The energy barrier to carboxamide group rotation of approximately 20 kJ mol-1 should permit interconversion between rotamers. In temozolomide and the related drug mitozolomide (4), N(7) is more negatively charged than N(1), which favors the formation of hydrogen bonds to the former atom in spite of their poor geometry. The relevance of these structural features to the action of temozolomide as a major-groove-directed prodrug of the alkylating agent MTIC (3) is discussed.

Structural studies on bio-active compounds. Part XIV. Molecular modelling of the interactions between pentamidine and DNA.

Molecular mechanics modelling was carried out on the antimicrobial aromatic diamidine, pentamidine, bound to the minor groove of several AT-rich DNA octamers. The pentamidine molecule was found to span four base pairs, with its highly charged amidinium groups forming hydrogen bonds to the O2 of thymine or the N3 of adenine, but not to the backbone phosphate groups. The mean binding energy of the pentamidine-DNA complexes was -52 kcal/mol, of the same order as that of berenil-DNA complexes. There was no significant energy difference between those models which contained GC base pairs and those which did not. The presence of bound pentamidine resulted in some distortion of helix geometry involving helix opening towards the minor groove in most cases and a decrease in the number of residues per turn.

Footprinting studies on the sequence-selective binding of pentamidine to DNA.

The sequence-selective binding of pentamidine, an antimicrobial aromatic diamidine, has been investigated by footprinting studies on two different DNA fragments using DNase I, micrococcal nuclease and hydroxyl radical as probes. Each probe reveals drug-induced protection from cleavage in AT-rich regions. The best binding sites consist of at least 5 consecutive AT base pairs. Three or less AT pairs do not constitute a pentamidine binding site.

Structural studies on bio-active compounds. Part 16. Synthesis and properties of sterically-constrained analogues of methylbenzoprim, a new dihydrofolate reductase inhibitor.

The new lipophilic dihydrofolate reductase inhibitor methylbenzoprim is a flexible substituted 2,4-diamino-5-arylpyrimidine bearing an ortho-disposed aromatic nitro group and an N-methylbenzylamine residue. Molecular modelling studies indicate that modifications which incorporate these structural features into a planar benzazole or benzazole-N-oxide arrangement could mimic the overall geometry of methylbenzoprim. Target compounds were synthesized by exploiting known reaction pathways involving ortho-interactions between diazonium-, azido- or nitro-substituents and the adjacent substituted amino group: these syntheses proceeded without detrimental effect on the diaminopyrimidine residue. None of the heterocyclic analogues were more potent than the lead molecule methylbenzoprim.

Anti-tumour imidazotetrazines. Part XXI. Mitozolomide and temozolomide: probes for the major groove of DNA.

The structural and electronic properties of the major groove-binding anti-tumour imidazotetrazinones, mitozolomide and temozolomide were studied using molecular orbital techniques. Structure-activity relationships of mitozolomide derivatives emphasized the importance of a hydrogen bond donor on the C8-substituent and showed that good activity would be expected for derivatives carrying a small C8-substituent with restricted negative potential. The coplanarity of the carboxamide substituent in mitozolomide is unlikely to be disrupted by hydrogen bonding in the major groove. Semi-empirical M.O. calculations gave a very high partial positive charge on C4, indicating an enhanced susceptibility to nucleophilic attack leading to ring opening at this position and the formation of a triazene alkylating agent.

Structural studies on bio-active compounds. Part XV. Structure-activity relationships for pyrimethamine and a series of diaminopyrimidine analogues versus bacterial dihydrofolate reductase.

The phenylpyrimidine derivative pyrimethamine and its congeners inhibit the enzyme dihydrofolate reductase (5,6,7,8-tetrahydrofolate:NADP+ oxidoreductase, EC 1.5.1.3) and are of interest as antiproliferative agents. In this study the equilibrium conformations of some pyrimethamine derivatives, and their interactions with Escherichia coli dihydrofolate reductase, were investigated using molecular modelling techniques. In each case the phenyl ring avoided coplanarity with the pyrimidine ring and attained a position approximately perpendicular to it, in agreement with crystal structures. A meta substituent could be placed either side of the pyrimidine plane, forming two non-equivalent, slowly interconverting solution conformations. Except for meta-azidopyrimethamine, both conformations of all the inhibitors were able to bind to the active site cleft of the enzyme with the diaminopyrimidine moiety, making the normal pattern of enzyme/inhibitor hydrogen bonds. One such conformation of the meta-azido compound failed to bind because of unacceptable steric clashes, whilst the other showed enhanced binding energy attributable to the occupation of a hydrophobic pocket by the azido group. The enhanced binding of 2,4-diamino-6-ethyl-5-phenylpyrimidine over its 6-methyl analogue was also related to attractive hydrophobic interactions.