Location of the glucuronosyltransferase domain in the heparan sulfate copolymerase EXT1 by analysis of Chinese hamster ovary cell mutants.

Heparan sulfate formation occurs by the copolymerization of glucuronic acid (GlcA) and N-acetylglucosamine (GlcNAc) residues. Recent studies have shown that these reactions are catalyzed by a copolymerase encoded by EXT1 and EXT2, members of the exostosin family of putative tumor suppressors linked to hereditary multiple exostoses. Previously, we identified a collection of Chinese hamster ovary cell mutants (pgsD) that failed to make heparan sulfate (Lidholt, K., Weinke, J. L., Kiser, C. S., Lugemwa, F. N., Bame, K. J., Cheifetz, S., Massagué, J., Lindahl, U., and Esko, J. D. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 2267-2271). Here, we show that pgsD mutants contain mutations that either alter GlcA transferase activity selectively or that affect both GlcNAc and GlcA transferase activities. Expression of EXT1 corrects the deficiencies in the mutants, whereas EXT2 and the related EXT-like cDNAs do not. Analysis of the EXT1 mutant alleles revealed clustered missense mutations in a domain that included a (D/E)X(D/E) motif thought to bind the nucleotide sugar from studies of other transferases. These findings provide insight into the location of the GlcA transferase subdomain of the enzyme and indicate that loss of the GlcA transferase domain may be sufficient to cause hereditary multiple exostoses.

Effects of mutating Asn-52 to isoleucine on the haem-linked properties of cytochrome c.

Asn-52 of rat cytochrome c and baker's yeast iso-1-cytochrome c was changed to isoleucine by site-directed mutagenesis and the mutated proteins expressed in and purified from cultures of transformed yeast. This mutation affected the affinity of the haem iron for the Met-80 sulphur in the ferric state and the reduction potential of the molecule. The yeast protein, in which the sulphur-iron bond is distinctly weaker than in vertebrate cytochromes c, became very similar to the latter: the pKa of the alkaline ionization rose from 8.3 to 9.4 and that of the acidic ionization decreased from 3.4 to 2.8. The rates of binding and dissociation of cyanide became markedly lower, and the affinity was lowered by half an order of magnitude. In the ferrous state the dissociation of cyanide from the variant yeast cytochrome c was three times slower than in the wild-type. The same mutation had analogous but less pronounced effects on rat cytochrome c: it did not alter the alkaline ionization pKa nor its affinity for cyanide, but it lowered its acidic ionization pKa from 2.8 to 2.2. These results indicate that the mutation of Asn-52 to isoleucine increases the stability of the cytochrome c closed-haem crevice as observed earlier for the mutation of Tyr-67 to phenylalanine [Luntz, Schejter, Garber and Margoliash (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 3524-3528], because of either its effects on the hydrogen-bonding of an interior water molecule or a general increase in the hydrophobicity of the protein in the domain occupied by the mutated residues. The reduction potentials were affected in different ways; the Eo of rat cytochrome c rose by 14 mV whereas that of the yeast iso-1 cychrome c was 30 mV lower as a result of the change of Asn-52 to isoleucine.

The significance of denaturant titrations of protein stability: a comparison of rat and baker's yeast cytochrome c and their site-directed asparagine-52-to-isoleucine mutants.

The residue asparagine-52 of rat cytochrome c and baker's yeast iso-1-cytochrome c was mutated to isoleucine by site-directed mutagenesis, and the unfolding of the wild-type and mutant proteins in urea or guanidinium chloride solutions was studied. Whereas the yeast mutant cytochrome unfolded in 4-7 M urea with a rate constant (k) of 1.7 x 10(-2) s-1, the rat mutant protein unfolded with k = 5.0 x 10(-2) s-1, followed by a slow partial refolding with k = 5.0 x 10(-4) s-1. Denaturant titrations indicated that the mutation increased the stability of the yeast cytochrome by 6.3 kJ (1.5 kcal)/mol, while it decreased that of the rat protein by 11.7 kJ (2.8 kcal)/mol. These results probably reflect structural differences between yeast iso-1 and vertebrate cytochromes c in the vicinity of the Asn-52 side chain.

Expression of recombinant cytochromes c from various species in Saccharomyces cerevisiae: post-translational modifications.

A complete protocol for the expression of recombinant cytochrome c genes from yeast, Drosophila melanogaster, and rat in a yeast strain, GM-3C-2, which does not express its own cytochromes c is described. The construction of the expression vectors, transformation and large-scale growth of the yeast, and preparation and purification of the recombinant cytochromes c are described. It was found that, contrary to the way yeast modifies its own cytochromes c, the recombinant proteins were partially acetylated at their N-terminus, except for the drosophila protein, which remained entirely unblocked. Furthermore, the yeast and rat proteins were close to fully trimethylated at lysine 72, while the drosophila protein could be separated chromatographically into forms containing tri-, di-, mono-, and unmethylated lysine 72 showing corresponding resonances in the NMR spectrum. These observations emphasize that, in employing expression procedures to obtain native or mutant forms of cytochrome c, it is essential to identify the variety and extent of post-translational modifications and to separate the preparation into pure monomolecular species. Otherwise, it may become impossible to distinguish between the influence of a site-directed mutation and unexamined post-translational modifications.

Relationship between local and global stabilities of proteins: site-directed mutants and chemically-modified derivatives of cytochrome c.

The methionine 80 sulfur-heme iron bond of rat cytochrome c, whose stability is decreased by mutating the phylogenetically invariant residue proline 30 to alanine and increased when tyrosine 67 is changed to phenylalanine, recovers its wild-type characteristics when both substitutions are performed on the same molecule. Titrations with urea, analyzed according to the heteropolymer theory [Alonso, D. O. V., & Dill, K. A. (1991) Biochemistry 30, 5974-5985], indicate that both single mutations increase the solvent exposure of hydrophobic groups in the unfolded state, while in the double mutant this conformational perturbation disappears. Similar increases in solvent exposure of hydrophobic groups are observed when the sulfur-iron bond of the wild-type protein is broken by alkylation of the methionine sulfur, by high pH, or by binding the heme iron with cyanide. The compensatory effects of the two single mutations do not extend to the overall stability of the protein. The added loss of conformational stability due to the single mutations amounts to 7.3 kcal/mol out of the 9 kcal/mol representing the overall free energy of stabilization of the native conformation of the wild-type protein. The folded conformation of the doubly mutated protein is only 2 kcal/mol less stable than that of the wild type. These results indicate that the double mutant protein is able to retain the essential folding pattern of cytochrome c and the thermodynamic stability of the methionine sulfur-heme iron bond, in spite of structural differences that weaken the overall stability of the molecule.

Heparan sulfate glycosaminoglycans as primary cell surface receptors for herpes simplex virus.

Our current incomplete picture of the earliest events in HSV infection may be summarized as follows. The initial interaction of virus with cells is the binding of virion gC to heparan sulfate moieties of cell surface proteoglycans. Stable binding of virus to cells may require the interaction of other virion glycoproteins with other cell surface receptors as well (including the interaction of gB with heparan sulfate). Penetration of virus into the cell is mediated by fusion of the virion envelope with the cell plasma membrane. Events leading up to this fusion require the action of at least three viral glycoproteins (gB, gD and gH), one or more of which may interact with specific cell surface components. It seems likely that binding of gB to cell surface heparan sulfate may occur and may be important in the activation of some event required for virus penetration. Heparan sulfate is present not only as a constituent of cell surface proteoglycans but also as a component of the extracellular matrix and basement membranes in organized tissues. In addition, body fluids contain both heparin and heparin-binding proteins, either of which can prevent the binding of HSV to cells (WuDunn and Spear, 1989). As a consequence, the spread of HSV infection is probably influenced, not only by immune responses to the virus, but also by the probability that virus will be entrapped or inhibited from binding to cells by extracellular forms of heparin or heparan sulfate.

Changing the invariant proline-30 of rat and Drosophila melanogaster cytochromes c to alanine or valine destabilizes the heme crevice more than the overall conformation.

Drosophila melanogaster and rat cytochromes c in which proline-30 was converted to alanine or valine were expressed in a strain of baker's yeast, Saccharomyces cerevisiae, where they sustained aerobic growth. The mutations had no significant effect on the spectra or redox potentials but altered drastically the stability of the bond between the methionine-80 sulfur and the heme iron, as judged by four criteria: (i) the alkaline pKa values of the 695-nm band of the ferric form of the mutant proteins decreased by almost 1 pH unit as compared to the wild types; (ii) the acid pKa values increased by 0.5 to 1.2 pH units; (iii) the 695-nm band half-disappeared at temperatures 10-20 degrees C lower in the mutant proteins than in the wild types; and (iv) the 695-nm band of the mutant proteins was susceptible to concentrations of urea that had little influence on their overall structure. The valine-substituted rat cytochrome c had properties intermediate between those of the wild type and the alanine mutant. The destabilized coordinative bond is located in space a long distance from the mutation site. It is suggested that the mutations weaken the hydrogen bond between the carbonyl of residue 30 and the imino group of the imidazole of histidine-18, modifying the bonding of the heme iron by that imidazole, which, in turn, through a trans effect, weakens the bond between the heme iron and the other axial ligand, the sulfur of methionine-80. Alternatively, the effect of the mutations may be propagated allosterically along the peptide chain.

Amino acid sequence requirements for the association of apocytochrome c with mitochondria.

To examine the amino acid sequence requirements for the biphasic association of Drosophila melanogaster apocytochrome c with mouse liver mitochondria in vitro, recombinant constructs of the protein were prepared. Removal of the C-terminal sequence to residue 58 had little influence, but truncation to residue 50 decreased the association to low levels and removal to residue 36 was even more effective. However, a mutant missing the segment between residues 35 and 66 was fully functional, but, when the C-terminal segment from residue 36 was replaced with a noncytochrome c sequence, the high-affinity phase of the association was lost. A mutant in which residues 90, 91, 92, 96, and 100 were replaced by lysine, leucine, proline, proline, and proline, respectively, to prevent the possible formation of the C-terminal alpha-helix and another mutant in which the C-terminal segment from residue 90 to residue 120 was a noncytochrome c sequence had normal association. In contrast, replacing lysine-5, -7, and -8 by glutamine, glutamic acid, and asparagine, respectively, resulted in loss of the high-affinity phase. The same mutations in the apoprotein lacking the segment between residues 35 and 66 caused, in addition, a decrease of the low-affinity phase association. Thus, the N-terminal region is most critical for apocytochrome c association, but alternative segments of the central and/or C-terminal region can be utilized, where noncytochrome c sequences are ineffective. These results emphasize the wide disparity between the structural requirements for association with mitochondria and for the production of a functional holoprotein.