Heterologous expression and characterization of alternatively spliced glutathione S-transferases from a single Anopheles gene.

Three cDNA sequences of glutathione S-transferase (GST), adgst1-2, adgst1-3 and adgst1-4, which are alternatively spliced products of the adgst1AS1 gene, were obtained from fourth instar larvae of Anopheles dirus mosquito by reverse transcriptase PCR reactions. The nucleotide sequences of these three cDNAs share >67% identity and the translated amino acid sequences share 61-64% identity. A comparison of the An. dirus to the An. gambiae enzymes shows that adGST1-2 versus agGST1-4, adGST1-3 versus agGST1-5 and adGST1-4 versus agGST1-3 have 85, 92 and 85% amino acid sequence identity, respectively, which confirms that orthologous isoenzymes occur across anopheline species. These three proteins were expressed at high levels, approximately 15-20 mg from 200 ml of E. coli culture. The recombinant enzymes were purified by affinity chromatography on an S-hexylglutathione agarose column. The subunit sizes of adGST1-2, adGST1-3 and adGST1-4 are 24.3, 23.9 and 25.1 kDa. The recombinant enzymes have high activities with 1-chloro-2,4-dinitrobenzene (CDNB), detectable activity with 1,2-dichloro-4-nitrobenzene but markedly low activity with ethacrynic acid and p-nitrophenethyl bromide. adGST1-3 was shown to be the most active enzyme from the kinetic studies. Permethrin inhibition of CDNB activity, at varying concentrations of CDNB, was significantly different, being uncompetitive for adGST1-2, noncompetitive for adGST1-3 and competitive for adGST1-4. In contrast, permethrin inhibition with varying glutathione concentrations was noncompetitive for all three GSTs. Despite the enzymes being splicing products of the same gene and sharing identical sequence in the N-terminal 45 amino acids, these GSTs show distinct substrate specificities, kinetic properties and inhibition properties modulated by the differences in the C-terminus.

The relative order of helical propensity of amino acids changes with solvent environment.

A model peptide of sequence Ac-Y-VAXAK-VAXAK-VAXAK-NH(2), where X is substituted with one of nineteen amino acids (P excluded), was synthesized and titrated with methanol to study helical propensity as a function of solvent environment. The CD spectra of these peptides are largely random coil in 2 mM sodium phosphate buffer (pH 5.5) and show a conformational change to alpha-helix with increasing methanol content. Singular value decomposition was used to correct the CD spectra for the absorbing side chains of W, Y, F, C, and M, and this correction can be substantial. With correction both W and F become good helix formers. The free energy for helix propagation was calculated using the Lifson-Roig statistical model for each of the nineteen amino acids at each point in their titration. The results show that the rank order of helical propensity for the nineteen amino acids changes with solvent environment. This result will be particularly important if proteins undergo hydrophobic collapse before secondary structures are formed, because amino acids can then see different solvent environments as the secondary structures are formed. Related amino acids are found to have interesting correlations in the shape of their titration curves. This finding provides one explanation for the limiting 70% accuracy in predicting secondary structure from sequence, since the helical propensities used are calculated for an average solvent environment. Proteins 2000;39:132-141.

Correcting the circular dichroism spectra of peptides for contributions of absorbing side chains.

The aromatic and sulfur-containing side chains Trp, Tyr, Phe, Cys, and Met contribute to the CD spectra of peptides and proteins in the amide region, interfering with the analysis for secondary structure. We propose a method to correct the CD spectra of peptides undergoing the helix-coil transition for contributions due to absorbing side chains using singular value decomposition. The method uses the common basis vectors obtained from an analysis of the CD spectra of related peptides without the aromatic and sulfur-containing amino acids. The common basis vectors are fitted to a portion of the CD spectrum of the peptide being corrected, in the range that is unaffected by its sidechain contributions. Then the resulting coefficients from the fitting are used along with the common basis vectors to regenerate the entire corrected spectrum. The method is illustrated for the CD spectra of the peptide sequence acetyl-Y-VAXAK-VAXAK-VAXAK-amide, where X is substituted with the 20 naturally occurring amino acids. This peptide model adopts a random-coil conformation in 2 mm sodium phosphate buffer, pH 5.5, and becomes an alpha helix in methanol/buffer solutions. The difference between the original and corrected spectra shows the contribution from the aromatic and sulfur-containing side chains.