Oligonucleotide-directed mutagenesis using M13-derived vectors: an efficient and general procedure for the production of point mutations in any fragment of DNA

Membrane assembly of the triple-spanning coronavirus M protein. Individual transmembrane domains show preferred orientation

The M protein of mouse hepatitis virus strain A59 is a triple-spanning membrane protein which assembles with an uncleaved internal signal sequence, adopting an NexoCcyt orientation. To study the insertion mechanism of this protein, domains potentially involved in topogenesis were deleted and the effects analyzed in topogenesis were deleted and the effects analyzed in several ways. Mutant proteins were synthesized in a cell-free translation system in the presence of microsomal membranes, and their integration and topology were determined by alkaline extraction and by protease-protection experiments. By expression in COS-1 and Madin-Darby canine kidney-II cells, the topology of the mutant proteins was also analyzed in vivo. Glycosylation was used as a biochemical marker to assess the disposition of the NH2 terminus. An indirect immunofluorescence assay on semi-intact Madin-Darby canine kidney-II cells using domain-specific antibodies served to identify the cytoplasmically exposed domains. The results show that each membrane-spanning domain acts independently as an insertion and anchor signal and adopts an intrinsic preferred orientation in the lipid bilayer which corresponds to the disposition of the transmembrane domain in the wild-type assembled protein. These observations provide further insight into the mechanism of membrane integration of multispanning proteins. A model for the insertion of the coronavirus M protein is proposed.

Ectopic recombination within homologous immunoglobulin mu gene constant regions in a mouse hybridoma cell line

We have transferred a pSV2neo vector containing the wild-type constant region of the immunoglobulin mu gene (C mu) into the mutant hybridoma igm482, which bears a 2-bp deletion in the third constant-region exon of its haploid chromosomal mu gene (C mu 3). Independent igm482 transformants contain the wild-type immunoglobulin C mu region stably integrated in ectopic chromosomal positions. We report here that the wild-type immunoglobulin C mu region can function as the donor sequence in a gene conversion event which corrects the 2-bp deletion in the mutant igm482 chromosomal C mu 3 exon. The homologous recombination event restores normal immunoglobulin M production in the mutant cell.

Membrane assembly of the triple-spanning coronavirus M protein. Individual transmembrane domains show preferred orientation

The M protein of mouse hepatitis virus strain A59 is a triple-spanning membrane protein which assembles with an uncleaved internal signal sequence, adopting an NexoCcyt orientation. To study the insertion mechanism of this protein, domains potentially involved in topogenesis were deleted and the effects analyzed in topogenesis were deleted and the effects analyzed in several ways. Mutant proteins were synthesized in a cell-free translation system in the presence of microsomal membranes, and their integration and topology were determined by alkaline extraction and by protease-protection experiments. By expression in COS-1 and Madin-Darby canine kidney-II cells, the topology of the mutant proteins was also analyzed in vivo. Glycosylation was used as a biochemical marker to assess the disposition of the NH2 terminus. An indirect immunofluorescence assay on semi-intact Madin-Darby canine kidney-II cells using domain-specific antibodies served to identify the cytoplasmically exposed domains. The results show that each membrane-spanning domain acts independently as an insertion and anchor signal and adopts an intrinsic preferred orientation in the lipid bilayer which corresponds to the disposition of the transmembrane domain in the wild-type assembled protein. These observations provide further insight into the mechanism of membrane integration of multispanning proteins. A model for the insertion of the coronavirus M protein is proposed.

DNA inversions between short inverted repeats in Escherichia coli

Using site-specific mutagenesis in vitro, we have constructed Escherichia coli strains that allow the detection of the inversion of an 800-bp segment in the lac region. The invertible segment is bounded by inverted repeats of either 12 or 23 bp. Inversions occurring at these inverted repeats will restore the Lac+ phenotype. Inversions can be detected at both short homologies at frequencies ranging from 0.5 x 10(-8) to 1 x 10(-7). These events, which have been verified by DNA sequence analysis, are reduced up to 1000-fold in strains deficient for either RecA, RecB or RecC. They are not reduced in strains deficient in the RecF, J pathway. These results show that the RecB,C,D system can mediate rearrangements at short sequence repeats, and probably plays a major role in cellular rearrangements.

cis-active sequences near the 5'-termini of beet necrotic yellow vein virus RNAs 3 and 4

RNAs 3 and 4 of the multicomponent genome of beet necrotic yellow vein virus are dispensable for infection of Chenopodium quinoa leaves. We have used mutagenesis of biologically active RNA 3 transcripts to identify 5'-proximal sequences essential in cis for RNA 3 amplification. One such element, Box I, (nucleotides 283-292) was complementary to the first 10 residues (Box I') following the 5'-terminal cap. A second cis-active element (Box II) was identified between nucleotides 237-244 and was complementary to nucleotides 16-23 (Box II'). Other cis-active sequences exist between Box II' and II but have not been mapped to fine scale. Most sequence substitutions in Boxes I and II or in the 5'-proximal complementary sequences were lethal but compensatory mutations designed to restore Box I/I' or Box II/II' base pairing restored viability, suggesting that secondary structure involving these elements rather than their exact sequence is the critical feature. Transcripts bearing short deletions near residue 200 were replicated but did not assemble into virions, indicating that this region contains or contributes to a cis-active encapsidation signal. Similar experiments with RNA 4 transcript have shown that 5'-proximal cis-essential elements are limited to the first 400 residues of this RNA. Essential subdomains within this region have not been mapped but there are no structures obviously homologous to Boxes I/I' and II/II' of RNA 3.

A ribosomal ambiguity mutation in the 530 loop of E. coli 16S rRNA

A series of base substitution and deletion mutations were constructed in the highly conserved 530 stem and loop region of E. coli 16S rRNA involved in binding of tRNA to the ribosomal A site. Base substitution and deletion of G517 produced significant effects on cell growth rate and translational fidelity, permitting readthrough of UGA, UAG and UAA stop codons as well as stimulating +1 and -1 frameshifting in vivo. By contrast, mutations at position 534 had little or no effect on growth rate or translational fidelity. The results demonstrate the importance of G517 in maintaining translational fidelity but do not support a base pairing interaction between G517 and U534.

Human immunodeficiency virus type 1 gag-pol frameshifting is dependent on downstream mRNA secondary structure: demonstration by expression in vivo

The human immunodeficiency virus type 1 (HIV-1) Gag-Pol fusion polyprotein is produced via ribosomal frameshifting. Previous studies in vitro and in Saccharomyces cerevisiae have argued against a significant role for RNA secondary structure 3' of the shift site, in contrast with other systems, in which such structure has been shown to be required. Here we show, by expressing the HIV-1 gag-pol domain in cultured vertebrate cells, that a stem-loop structure 3' of the HIV-1 shift site is indeed important for wild-type levels of frameshifting in vivo.

Novel mutations in the pheA gene of Escherichia coli K-12 which result in highly feedback inhibition-resistant variants of chorismate mutase/prephenate dehydratase

The bifunctional enzyme chorismate mutase/prephenate dehydratase (EC 5.4.99.5/4.2.1.51), which is encoded by the pheA gene of Escherichia coli K-12, is subject to strong feedback inhibition by L-phenylalanine. Inhibition of the prephenate dehydratase activity is almost complete at concentrations of L-phenylalanine greater than 1 mM. The pheA gene was cloned, and the promoter region was modified to enable constitutive expression of the gene on plasmid pJN302. As a preliminary to sequence analysis, a small DNA insertion at codon 338 of the pheA gene unexpectedly resulted in a partial loss of prephenate dehydratase feedback inhibition. Four other mutations in the pheA gene were identified following nitrous acid treatment of pJN302 and selection of E. coli transformants that were resistant to the toxic phenylalanine analog beta-2-thienylalanine. Each of the four mutations was located within codons 304 to 310 of the pheA gene and generated either a substitution or an in-frame deletion. The mutations led to activation of both enzymatic activities at low phenylalanine concentrations, and three of the resulting enzyme variants displayed almost complete resistance to feedback inhibition of prephenate dehydratase by phenylalanine concentrations up to 200 mM. In all four cases the mutations mapped in a region of the enzyme that has not been implicated previously in feedback inhibition sensitivity of the enzyme.

Envelope glycoprotein and CD4 independence of vpu-facilitated human immunodeficiency virus type 1 capsid export

The effect of vpu on the release of human immunodeficiency type 1 capsid proteins was examined in the presence or absence of virus-encoded envelope glycoproteins as well as in cells which constitutively express either the CD4 or CD8 protein. The results show that vpu-mediated facilitated export of capsid proteins from HeLa cells does not require expression of the envelope glycoprotein. The experiments also show that export of virus capsid proteins from HeLa cells facilitated by vpu is not affected by coexpression of either the CD4 or CD8 protein. The vpu protein acts in trans to facilitate export of virus capsid proteins from HeLa cells.

Mutational analysis of the human immunodeficiency virus type 1 Eli Nef function

The studies presented here define an internally consistent experimental system that permits systematic analysis of the effect of nef on the rate of the human immunodeficiency virus type 1 (HIV-1) replication in a CD4+ tumor T-cell line and in primary peripheral blood mononuclear cells. The parental full-length Nef protein, derived from the Eli strain of HIV-1, accelerates virus replication in both cell types. Mutations that destabilize or alter the intracellular location of the protein affect the ability of the Nef protein to accelerate virus replication. A set of mutants was made in amino acids proposed to be required for Nef function, including threonine and serine residues proposed to be targets for phosphorylation, and in sequences thought to resemble the G-1, G-3, and G-4 sites of the family of G proteins. In most cases alterations of the critical amino acids yield stable Nef proteins of parental phenotype. These results challenge the existing theories for the mechanism of Nef function. The results also identify two residues in the carboxyl half of the protein that are important for Nef function.

Factors affecting homologous overexpression of the Saccharomyces cerevisiae lanosterol 14 alpha-demethylase gene

The Saccharomyces cerevisiae Lanosterol 14 alpha-demethylase (14DM) gene was overexpressed in S. cerevisiae using promoter sequences of the highly expressed S. cerevisiae glyceraldehyde-3-phosphate dehydrogenase (TDH3) gene. To investigate factors affecting 14DM overproduction, the levels of 14DM-specific RNAs, apoprotein, and heme protein, respectively, were determined and the 14DM-specific RNA levels compared with the RNA levels originating from the endogenous TDH gene(s). The quantitative measurements revealed that the 14DM steady-state RNA levels reached were some three- to five-fold below the theoretically expected values. With a view towards further improving expression of the 14DM gene, the spacing between the TDH3 promoter and the AUG was adjusted precisely and to rule out possible toxic effects exerted by the 14DM protein, the TDH3 promoter was placed under galactose regulation by introducing an UASG segment. Furthermore, the effects of the gene copy number on 14DM overproduction were investigated. From the analysis of the improved expression constructs five conclusions could be reached: (1) expression from the native 14DM gene is comparable to the expression driven by the TDH3 promoter-14DM fusion construct on single copy plasmid vectors; (2) expression from the TDH3 promoter-14DM construct on single-copy vectors is nearly as efficient as expression from the corresponding endogenous TDH3 gene; (3) the gene copy number has an effect on the relative expression levels of the TDH3 promoter-14DM constructs; (4) the steady-state amounts of protein produced are very nearly proportional to gene dosage; and (5) protein toxicity does not have a major impact on 14DM production. The maximum yield of 14DM was in the order of 7% of the total yeast protein and the maximum production of functional 14DM heme protein appears to be limited by the availability of heme.

Efficient cell-to-cell movement of beet necrotic yellow vein virus requires 3' proximal genes located on RNA 2

RNA 2 of beet necrotic yellow vein virus (BNYVV) carries six open reading frames. The four 3' proximal frames encode the proteins P42, P13, P15, and P14. The first three species present homologies to proteins encoded by three overlapping open reading frames (the triple gene block) in potexviruses, carlaviruses, and barley stripe mosaic virus. P14 does not display homology with other known plant viral proteins. The functions of P42, P13, P15, and P14 were investigated by site-directed mutagenesis. Full-length transcripts of wild-type BNYVV RNAs 1 and 2 were infectious when coinoculated to protoplasts or leaves of Chenopodium quinoa. RNA 2 transcripts in which P42, P13, and P15 were prematurely terminated by frameshift mutations replicated in protoplasts (when inoculated with wild-type RNA 1) but were not infectious to leaves, indicating that the triple gene block proteins of BNYVV are essential for viral cell-to-cell spread. Mutations in P14 were not lethal in leaf infections but smaller local lesions and lesser amounts of viral RNA were produced. RNA 2-related subgenomic RNA species of 2.6, 1.4, and 0.7 kb were detected; they presumably direct synthesis of P42, P13, and P14. No species of the length predicted for a P15-specific subgenomic RNA was detected.

In vitro mutagenesis of biologically active transcripts of beet necrotic yellow vein virus RNA 2: evidence that a domain of the 75-kDa readthrough protein is important for efficient virus assembly

RNA 2 of the multipartite genome of beet necrotic yellow vein virus carries the cistron for 21-kDa viral coat protein at its 5' extremity. The amber termination codon of the coat protein cistron undergoes suppression approximately 10% of the time so that translation continues into an adjacent 54-kDa open reading frame, yielding a 75-kDa readthrough protein. The roles of coat protein and the readthrough protein in infection were investigated with biologically active transcripts of RNA 2. Much of the coat protein cistron of the RNA 2 transcript could be deleted without interfering with viral replication and local lesion formation on leaves, although formation of the rod-shaped virions did not occur. Mutants in which the amber coat protein termination codon was replaced with an ochre codon or a tyrosine codon were also viable. The ochre codon was suppressed both in vitro and in planta. The mutant containing the tyrosine substitution produced only the 75-kDa read-through protein and was deficient in viral assembly. Deletions in the 54-kDa readthrough domain were also viable in planta but had different effects on virus assembly. A deletion in the C-terminal portion of the readthrough domain did not interfere with RNA packaging but, unexpectedly, deletions in the N-terminal portion were assembly deficient, although 21-kDa coat protein was produced in planta. Thus, the 75-kDa protein can apparently intervene in virion assembly even though it has not been detected in purified virions.

Global regulation of mitochondrial biogenesis in Saccharomyces cerevisiae: ABF1 and CPF1 play opposite roles in regulating expression of the QCR8 gene, which encodes subunit VIII of the mitochondrial ubiquinol-cytochrome c oxidoreductase

The multifunctional DNA-binding proteins ABF1 and CPF1 bind in a mutually exclusive manner to the promoter region of the QCR8 gene, which encodes 11-kDa subunit VIII of the Saccharomyces cerevisiae mitochondrial ubiquinol-cytochrome c oxidoreductase (QCR). We investigated the roles that the two factors play in transcriptional regulation of this gene. To this end, the overlapping binding sites for ABF1 and CPF1 were mutated and placed in the chromosomal context of the QCR8 promoter. The effects on transcription of the QCR8 gene were analyzed both under steady-state conditions and during nutritional shifts. We found that ABF1 is required for repressed and derepressed transcription levels and for efficient induction of transcription upon escape from catabolite repression, independently of DNA replication. CPF1 acts as a negative regulator, modulating the overall induction response. Alleviation of repression through CPF1 requires passage through the S phase. Implications of these findings for the roles played by ABF1 and CPF1 in global regulation of mitochondrial biogenesis are discussed.

Cytoplasmic domain truncation enhances fusion activity by the exterior glycoprotein complex of human immunodeficiency virus type 2 in selected cell types

To investigate the glycoprotein determinants of viral cytopathology, we constructed chimeric env genes between a noncytopathic strain of human immunodeficiency virus type 2 (HIV-2), designated HIV-2/ST, and a highly fusogenic and cytopathic variant derived from this virus. Expression of the resulting chimeric glycoproteins indicated that efficient syncytium formation in the human T-cell line Sup T1 mapped to the C-terminal region of the transmembrane (TM) glycoprotein subunit. In this region, the wild-type and cytopathic ST glycoproteins differed by only four amino acids and by the presence of a premature termination codon in the cytopathic variant. Subsequent site-directed mutagenesis indicated that the cytoplasmic domain truncation was responsible for the enhanced fusion activity. This modification, however, increased the fusion activity of the glycoprotein only in Sup T1 cells (in which the ST variant arose) but not in Molt 4 clone 8 or peripheral blood mononuclear cells. These observations indicate that the length of the cytoplasmic domain of the HIV-2 glycoprotein modulates the fusion activity of the exterior glycoprotein complex in a cell-specific manner. Such adaptability appears to permit the emergence of fusogenic variants during HIV-2 passage in vitro and may also regulate viral growth or cytopathic effects in selected cell types during natural infection in vivo.

Disulfide-bonded discontinuous epitopes on the glycoprotein of vesicular stomatitis virus (New Jersey serotype)

Intrachain disulfide bonds between paired cysteines in the glycoprotein (G) of vesicular stomatitis virus (VSV) are required for the recognition of discontinuous epitopes by specific monoclonal antibodies (MAbs) (W. Keil and R. R. Wagner, Virology 170:392-407, 1989). Cleavage by Staphylococcus aureus V8 protease of the 517-amino-acid VSV-New Jersey G protein, limited to the glutamic acid at residue 110, resulted in a protein (designated GV8) with greatly retarded migration by polyacrylamide gel electrophoresis (PAGE) under nonreducing conditions. By Western blot (immunoblot) analysis, protein GV8 was found to lose discontinuous epitope IV, which maps within the first 193 NH2-terminal amino acids. These data, coupled with those obtained by PAGE migration of a vector-expressed, truncated protein (G1-193) under reducing and nonreducing conditions, lead us to postulate the existence of a major loop structure within the first 193 NH2-terminal amino acids of the G protein, possibly anchored by a disulfide bond between cysteine 108 and cysteine 169, encompassing epitope IV. Site-directed mutants in which 10 of the 12 cysteines were individually converted to serines in vaccinia virus-based vectors expressing these single-site mutant G proteins were also constructed, each of which was then tested by immunoprecipitation for its capacity to recognize epitope-specific MAbs. These results showed that mutations in NH2-terminal cysteines 130, 174, and, to a lesser extent, 193 all resulted in the loss of neutralization epitope VIII. A mutation at NH2-terminal cysteine 130 also resulted in the loss of neutralization epitope VII, as did a mutation at cysteine 108 to a lesser extent. Both epitopes VII and VIII disappeared when mutations were made in COOH-distal cysteine 235, 240, or 273, the general map locations of epitopes VII and VIII. These studies also reveal that distal, as well as proximal, cysteine residues markedly influence the disulfide-bond secondary structure, which ostensibly determines the conformational structure of the VSV-New Jersey G protein required for presentation of the major discontinuous epitopes recognized by neutralizing MAbs.

Global regulation of mitochondrial biogenesis in Saccharomyces cerevisiae: ABF1 and CPF1 play opposite roles in regulating expression of the QCR8 gene, which encodes subunit VIII of the mitochondrial ubiquinol-cytochrome c oxidoreductase

The multifunctional DNA-binding proteins ABF1 and CPF1 bind in a mutually exclusive manner to the promoter region of the QCR8 gene, which encodes 11-kDa subunit VIII of the Saccharomyces cerevisiae mitochondrial ubiquinol-cytochrome c oxidoreductase (QCR). We investigated the roles that the two factors play in transcriptional regulation of this gene. To this end, the overlapping binding sites for ABF1 and CPF1 were mutated and placed in the chromosomal context of the QCR8 promoter. The effects on transcription of the QCR8 gene were analyzed both under steady-state conditions and during nutritional shifts. We found that ABF1 is required for repressed and derepressed transcription levels and for efficient induction of transcription upon escape from catabolite repression, independently of DNA replication. CPF1 acts as a negative regulator, modulating the overall induction response. Alleviation of repression through CPF1 requires passage through the S phase. Implications of these findings for the roles played by ABF1 and CPF1 in global regulation of mitochondrial biogenesis are discussed.

Distinct phosphotyrosines on a growth factor receptor bind to specific molecules that mediate different signaling pathways

The receptor for platelet-derived growth factor (PDGF) binds two proteins containing SH2 domains, GTPase activating protein (GAP) and phosphatidylinositol 3-kinase (PI3-kinase). The sites on the receptor that mediate this interaction were identified by using phosphotyrosine-containing peptides representing receptor sequences to block specifically binding of either PI3-kinase or GAP. These results suggested that PI3-kinase binds two phosphotyrosine residues, each located in a 5 aa motif with an essential methionine at the fourth position C-terminal to the tyrosine. Point mutations at these sites caused a selective elimination of PI3-kinase binding and loss of PDGF-stimulated DNA synthesis. Mutation of the binding site for GAP prevented the receptor from associating with or phosphorylating GAP, but had no effect on PI3-kinase binding and little effect on DNA synthesis. Therefore, GAP and PI3-kinase interact with the receptor by binding to different phosphotyrosine-containing sequence motifs.

Overexpression of the E1B 55-kilodalton (482R) protein of human adenovirus type 12 appears to permit efficient transformation of primary baby rat kidney cells in the absence of the E1B 19-kilodalton protein

To analyze the structure and function of the E1B 19,000-molecular-weight protein (19K protein) (163R) of human adenovirus type 12, mutants were produced at various positions across the 163R-coding sequence. Viruses bearing mutations within the first 100 or so amino acids yielded unstable 163R-related products, induced DNA degradation and enhanced cytopathic effect (cyt/deg phenotype) in KB cells, and transformed primary rodent cells at much lower efficiencies than wild-type (wt) virus. Deletion of the final 16 residues at the carboxy terminus had no phenotypic effect. Alteration of residue 105 reduced transforming efficiency significantly, suggesting that this region of 163R is functionally important. Disruption of the AUG initiation codon at nucleotide 1542 blocked production of 163R completely but resulted in higher levels of E1B 55K-482R protein synthesis and a transforming efficiency similar to that of wt virus. These data suggested that while 163R is of some importance, normal transforming efficiencies can be obtained in its absence if 482R is overexpressed.

Correlation of co-ordinated amino acid changes at the two-domain interface of cysteine proteases with protein stability

The engineering of a protein containing an alternative local residue packing for a set of side-chains has proven to be a major challenge because compositional, volumetric and steric constraints must be respected. Homologous proteins should provide examples of alternative groups of residues leading to a similar functional result. The functional significance of a pair of co-ordinated changes that are observed in the cysteine proteases family has been investigated by comparing the effect of individual or double changes on secretion, stability and activity of papain. The two changes are not independent. Detrimental effects of single mutations at one of the two positions can be partly suppressed by the co-ordinated mutation that reproduces naturally occurring contacts, indicating that these changes are concerted. Single mutations at the other position produce milder effects, suggesting a pathway for evolution.

Antibody framework residues affecting the conformation of the hypervariable loops

Rodent monoclonal antibodies have been "humanized" or "reshaped" for therapy by transplanting the antigen-binding loops from their variable domains onto the beta-sheet framework regions of human antibodies. However, additional substitutions in the human framework regions are sometimes required for high affinity antigen binding. Here we describe antigen binding by a reshaped antibody derived from the mouse anti-lysozyme antibody D1.3, and several variants in which point mutations had been introduced into framework positions to improve its affinity. The affinities were determined from the relaxation kinetics of reactant mixtures using quenching of fluorescence that occurs upon formation of the antibody-antigen complex. The dissociation constant of lysozyme ranged from 3.7 nM (for D1.3) to 260 nM. Measurement of antibody-antigen association kinetics using stopped-flow showed that D1.3 and most of the reshaped antibodies had bimolecular rate constants of 1.4 x 10(6) s-1 M-1, indicating that differences in equilibrium constant were predominantly due to different rates of dissociation of lysozyme from immune complexes. Mutations in a triad of heavy chain residues, 27, 29 and 71, contributed 0.9 kcal/mol in antigen binding free energy, and a Phe to Tyr substitution of light chain residue 71 contributed an additional 0.8 kcal/mol. The combined effect of all these mutations brought the affinity of the reshaped antibody to within a factor of 4 of D1.3. All of these substitutions were in the beta-sheet framework closely underlying the complementarity-determining regions, and do not participate in a direct interaction with antigen. The informed selection of residues in such positions may prove essential for the success of loop transplants in antibodies. Variation of these sites may also have a role in shaping the diversity of structures found in the primary repertoire, and in affinity maturation.

Hybrid myosin light chains containing a calcium-specific site from troponin C

Recombinant DNA approaches have allowed us to probe the mechanisms by which the regulatory light chains (RLCs) regulate myosin function by identifying the functional importance of specific regions of the RLC molecule. For example, we have demonstrated that the presence of high-affinity Ca2+/Mg(2+)-binding site in the N-terminal domain of the RLC is essential for the regulation of myosin-actin interaction [Reinach, F. C., Nagai, K. & Kendrick-Jones, J. (1986) Nature 322, 80-83]. To explore further the role of this metal-binding site in the RLC and generate an RLC with a Ca(2+)-specific site, we constructed four chicken skeletal muscle myosin regulatory light chain hybrid 'genes'. In these, the first domain containing the high-affinity Ca2+/Mg(2+)-binding site in the RLC was replaced with that containing the lower-affinity, Ca(2+)-specific, regulatory site from troponin C (TnC). In two of these hybrids, we replaced only the Ca(2+)-binding EF hand, while in the other two the EF hand and the N-terminal helix of TnC were transplanted. These hybrids were expressed in Escherichia coli in high yields and the purified proteins were used in calcium-binding experiments to assay the affinity and specificity of the sites and incorporated into scallop myosin to assay their regulatory behaviour. The results obtained show that the calcium-binding site from TnC, when transplanted into the RLC backbone, had a low affinity although most of its specificity appeared to be retained. As a result, although the TnC/RLC hybrids bound to scallop myosin and were able to activate the MgATPase activity of scallop acto-myosin, they were unable to regulate it. These results are in agreement with our previous findings that occupancy of the Ca2+/Mg2+ site in the RLC is essential for regulation. Our results suggest that the specificity and affinity of the calcium-binding site in troponin C is dependent on both intra- and inter-domain interactions within troponin C and that these latter interactions appear to be missing when this binding site is transplanted into the light chain backbone.

Expression and function of heterologous forms of malate dehydrogenase in yeast

The structure of the tricarboxylic acid cycle enzyme malate dehydrogenase is highly conserved in various organisms. To test the extent of functional conservation, the rat mitochondrial enzyme and the enzyme from Escherichia coli were expressed in a strain of Saccharomyces cerevisiae containing a disruption of the chromosomal MDH1 gene encoding yeast mitochondrial malate dehydrogenase. The authentic precursor form of the rat enzyme, expressed using a yeast promoter and a multicopy plasmid, was found to be efficiently targeted to yeast mitochondria and processed to a mature active form in vivo. Mitochondrial levels of the polypeptide and malate dehydrogenase activity were found to be similar to those for MDH1 in wild-type yeast cells. Efficient expression of the E. coli mdh gene was obtained with multicopy plasmids carrying gene fusions encoding either a mature form of the procaryotic enzyme or a precursor form with the amino terminal mitochondrial targeting sequence from yeast MDH1. Very low levels of mitochondrial import and processing of the precursor form were obtained in vivo and activity could be demonstrated for only the expressed precursor fusion protein. Results of in vitro import experiments suggest that the percursor form of the E. coli protein associates with yeast mitochondria but is not efficiently internalized. Respiratory rates measured for isolated yeast mitochondria containing the mammalian or procaryotic enzyme were, respectively, 83 and 62% of normal, suggesting efficient delivery of NADH to the respiratory chain. However, expression of the heterologous enzymes did not result in full complementation of growth phenotypes associated with disruption of the yeast MDH1 gene.

Defects in Moloney murine leukemia virus replication caused by a reverse transcriptase mutation modeled on the structure of Escherichia coli RNase H

We have studied a mutant Moloney murine leukemia virus with a deletion in reverse transcriptase (RT) which is predicted to make its RNase H domain resemble structurally that of human immunodeficiency virus RT. This deletion was based on improved RNase H homology alignments made possible by the recently solved three-dimensional structure for Escherichia coli RNase H. This mutant Moloney murine leukemia virus RT was fully active in the oligo(dT)-poly(rA) DNA polymerase assay and retained nearly all of wild-type RT's RNase H activity in an in situ RNase H gel assay. However, proviruses reconstructed to include this deletion were noninfectious. Minus-strand strong-stop DNA was made by the deletion mutant, but the amount of minus-strand translocation was intermediate to the very low level measured with RNase H-null virions and the high level seen with wild-type RT. The average length of translocated minus-strand DNA was shorter for the deletion mutant than for wild type, suggesting that mutations in the RNase H domain of RT also affect DNA polymerase activity.

Effects of substitution of aspartate-440 and tryptophan-487 in the thiamin diphosphate binding region of pyruvate decarboxylase from Zymomonas mobilis

A tryptophan residue at position 487 in Zymomonas mobilis pyruvate decarboxylase was altered to leucine by site-directed mutagenesis. This modified Z. mobilis pyruvate decarboxylase was active when expressed in Escherichia coli and had unchanged kinetics towards pyruvate. The enzyme showed a decreased affinity for the cofactors with the half-saturating concentrations increasing from 0.64 to 9.0 microM for thiamin diphosphate and from 4.21 to 45 microM for Mg2+. Unlike the wild-type enzyme, there was little quenching of tryptophan fluorescence upon adding cofactors to this modified form. The data suggest that tryptophan-487 is close to the cofactor binding site but is not required absolutely for pyruvate decarboxylase activity. Substitution of asparagine, threonine or glycine for aspartate-440, a residue which is conserved between many thiamin diphosphate-dependent enzymes, completely abolishes enzyme activity.

Protein engineering of antibodies

This article reviews the technical advances in antibody engineering and the clinical applications of these molecules. Recombinant DNA technology facilitates the construction and expression of engineered antibodies. These novel molecules are designed to meet specific applications. Although genomic and cDNA cloning have been used widely in the past to isolate the relevant antibody V domains, at present, the PCR-based cloning is the preferred system. Bacterial and mammalian expression systems are used commonly for the production of antibodies, antibody fragments, and antibody fusion proteins. A range of chimeric antibodies with murine V domains joined to C regions from human and other species have been produced and found to exhibit the expected binding characteristics and effector functions. Humanized antibodies have been developed to minimize the HAMA response, and bifunctional immunoglobulins are being used in tumor therapy and diagnosis. Single chain antibodies and fusion proteins with antibody specificities jointed to nonimmunoglobulin sequences provide a source of antibody-like molecules with novel properties. The potential applications of minimal recognition units and antigenized antibodies are described. Combinatorial libraries produced in bacteriophage present an alternative to hybridomas for the production of antibodies with the desired antigen binding specificities. Future developments in this field are discussed also.

In vitro mutational analysis of cis-acting RNA translational elements within the poliovirus type 2 5' untranslated region

Initiation of translation on poliovirus RNA occurs by internal binding of ribosomes to a region within the 5' untranslated region (UTR) of the mRNA. This region has been previously roughly mapped between nucleotides 140 and 631 of the 5' UTR and termed the ribosome landing pad. To identify cis-acting elements in the 5' UTR of poliovirus type 2 (Lansing strain) RNA that confer cap-independent internal initiation, we determined the in vitro translational efficiencies of a series of deletion and point mutations within the 5' UTR of the mRNA. The results demonstrate that the 3' border of the core poliovirus ribosome landing pad is located between nucleotides 556 and 585, whereas a region extending between nucleotides 585 and 612 confers enhanced translation. We studied two cis-acting elements within this region of the 5' UTR: a pyrimidine stretch which is critical for translation and an AUG (number 7 from the 5' end) that is located approximately 20 nucleotides downstream from the pyrimidine stretch and augments translation. We also show that the stem-loop structure which contains this AUG is not required for translation.

Expression of mammalian tyrosine aminotransferase in Saccharomyces cerevisiae and Escherichia coli. Purification to homogeneity and characterization of the enzyme overproduced in the bacteria

Rat liver tyrosine aminotransferase has been expressed in Saccharomyces cerevisiae and Escherichia coli. In yeast, the extent of production is 20-fold higher than that in rat liver after induction by dexamethasone, and reaches 250-fold higher in an E. coli strain carrying the T7 RNA polymerase transcription system. About 250 mg pure and homogeneous enzyme was obtained from 50 g transformed E. coli cells. Determination of Mr and pI, as well as analysis of N- and C-terminal amino acids, suggest that the isolated protein is native. The catalytic properties, similar to those of the enzyme from rat liver, confirm that it is fully active and that post-translational modifications in the mammalian cells are not essential for activity. Pyridoxal 5'-phosphate strongly protects the enzyme against thermal inactivation. After denaturation, 10 thiol groups, out of 16 in the polypeptide chain, react with 5,5'-dithiobis(2-nitrobenzoic acid) whereas only five or six are accessible under native conditions. Two thiols are rapidly modified with concomitant inactivation of the apoenzyme, but pyridoxal 5'-phosphate partially protects them in the holoenzyme. The results are interpreted in the light of the structure/function relationship in this enzyme.

Neither the endogenous nor a functional steroid hormone receptor binding site transactivate the ribosomal RNA gene promoter in vitro

The mammalian ribosomal RNA gene promoters exhibit a conserved sequence between positions +1 and +16 that shows a high degree of homology to the response element for glucocorticoids and progestins (GRE/PRE). These sequences bind specifically the glucocorticoid receptor and the progesterone receptor (PR) albeit with lower affinity than a canonical GRE/PRE. Because steroid hormones are known to affect expression of the ribosomal genes, we tested the influence of hormone receptors on the activity of the ribosomal RNA gene promoter in a cell-free transcription assay. Preparations of PR that induce transcription from the mouse mammary tumour virus (MMTV) promoter do not stimulate but slightly inhibit transcription from the ribosomal RNA gene promoter. This weak negative effect is not mediated through binding to the hypothetical GRE/PRE as a mutant promoter that does not bind receptor is equally repressed. Introduction of the functional MMTV GRE/PRE upstream of the basal ribosomal RNA gene promoter does not enhance its transcription in the presence of an active PR. Thus, RNA polymerase I transcription cannot be stimulated in vitro by cis elements and regulatory proteins that are active in RNA polymerase II transcription.

Resistance to ddI and sensitivity to AZT induced by a mutation in HIV-1 reverse transcriptase

Serial human immunodeficiency virus type-1 (HIV-1) isolates were obtained from five individuals with acquired immunodeficiency syndrome (AIDS) who changed therapy to 2',3'-dideoxyinosine (ddI) after at least 12 months of treatment with 3'-azido-3'-deoxythymidine (zidovudine, AZT). The in vitro sensitivity to ddI decreased during the 12 months following ddI initiation, whereas AZT sensitivity increased. Analysis of the reverse transcriptase coding region revealed a mutation associated with reduced sensitivity to ddI. When this mutation was present in the same genome as a mutation known to confer AZT resistance, the isolates showed increased sensitivity to AZT. Analysis of HIV-1 variants confirmed that the ddI resistance mutation alone conferred ddI and 2',3'-dideoxycytidine resistance, and suppressed the effect of the AZT resistance mutation. The use of combination therapy for HIV-1 disease may prevent drug-resistant isolates from emerging.

Human immunodeficiency virus type 2 vpx protein augments viral infectivity

The genomes of HIV and SIV are complex and contain several accessory genes which modulate viral replication and pathogenicity. One of these genes, vpx, is unique to the HIV-2/SIV group of viruses and encodes a virion-associated protein of unknown function. To examine the function of vpx, we constructed a vpx-deficient HIV-2 proviral clone and characterized its in vitro biological properties. Following transfection into immortalized T-cell lines, vpx-mutant HIV-2 was fully replication competent and exhibited growth kinetics and cytopathic properties equivalent to wild-type HIV-2. In addition, vpx-deficient virions were indistinguishable from wild-type HIV-2 in ultrastructure, composition of major structural proteins, and reverse transcriptase activity. In PHA-stimulated normal peripheral blood mononuclear cells (PBMCs), however, vpx-deficient virus replicated at substantially lower titers and required a 100- to 1000-fold higher inoculum to establish a productive infection. This defect was localized to early events in the viral life cycle since vpx-deficient virus exhibited a 5- to 10-fold reduction in initial (single cycle) viral DNA synthesis following acute infection of primary PBMCs. Paradoxically, in long-term (9-23 months) cultures of immortalized T-cells (SupT1) continuous high level replication of vpx-deficient, but not wild-type, virus was observed, indicating less efficient viral spread and cell killing and a more attenuated phenotype of vpx-deficient HIV-2. Taken together, these results demonstrate that vpx is required for the production of fully infectious and cytopathic HIV-2 virions and that it functions early in the viral life cycle by facilitating viral entry and/or reverse transcription. The pronounced replicative defect of vpx-deficient HIV-2 in primary PBMCs but not in short-term cultures of immortalized T-cell lines emphasizes the need to characterize the properties of nonessential HIV accessory gene products in natural target cells.

Functional evidence that the Ser-195 residue of staphylococcal exfoliative toxin A is essential for biological activity

The substitution of the serine 195 residue of staphylococcal exfoliative toxin A by a cysteine residue led to a biologically inactive protein. This result is consistent with the hypothesis that exfoliative toxin A could be a protease or a lipase. However, no protease or lipase activity was detected with the native toxin.

A matched set of cat vectors for rapid mutational analysis of eukaryotic promoters and enhancers

The eukaryotic cat expression vectors, pBRAMScat1 and pBRAMScat2, were constructed to simplify the analysis of genomic fragments containing putative transcriptional regulatory elements. These vectors contain the f1 filamentous phage origin of replication for single-stranded DNA rescue, and permit site-directed mutagenesis, and dideoxy sequencing of nested deletion mutants using commercial T3, T7 and M13 universal forward/reverse primers. The above features eliminate the need to shuttle back and forth between a conventional cloning vector and the cat expression vector during the analysis of putative eukaryotic gene regulatory elements. Plasmid pBRAMScat1 contains the bacterial chloramphenicol acetyltransferase-encoding gene (cat) and no eukaryotic promoter and was designed for the analysis of eukaryotic promoters. Plasmid pBRAMScat2 contains the cat gene under the control of the Herpes simplex virus thymidine kinase promoter and was designed for the analysis of eukaryotic enhancers.

Effects of point mutations in a hinge region on the stability, folding, and enzymatic activity of Escherichia coli dihydrofolate reductase

The role of a hinge region in the folding, stability, and activity of Escherichia coli dihydrofolate reductase was investigated with three site-directed mutants at valine-88, the central residue of the hinge. The three mutants, V88A and V88I and a valine-88 deletion, were created to perturb the packing of hydrophobic residues in the interior of a loose turn formed by residues 85-91. Deleting the valine-88 residue destabilized the protein by 2.93 +/- 0.6 kcal/mol as determined by equilibrium unfolding transitions in urea monitored by circular dichroism at 20 degrees C. Substitution of alanine for valine-88 stabilized the protein by -0.20 +/- 0.02 kcal/mol, and the isoleucine substitution was mildly destabilizing by 1.73 +/- 0.2 kcal/mol. Although there was no clear correlation between side-chain volume and stability, these results suggest that side-chain interactions in the interior of the turn influence the folding and stability of dihydrofolate reductase. The specific activity of the valine deletion mutant was approximately twice that of the wild-type protein while the specific activities of the V88A and V88I proteins were only slightly greater than the wild type. The full time courses of the reactions catalyzed by the mutants were almost identical with that for the wild type, indicating no major changes in the kinetic mechanism. Additionally, the rate constants associated with interconversion between various forms of the apoenzyme were identical for the mutant and wild-type enzymes. The rate constants for refolding transitions were examined by dilution of urea-inactivated protein.(ABSTRACT TRUNCATED AT 250 WORDS)

The cloned RNA polymerase II transcription factor IID selects RNA polymerase III to transcribe the human U6 gene in vitro

Although the human U2 and U6 snRNA genes are transcribed by different RNA polymerases (i.e., RNA polymerases II and III, respectively), their promoters are very similar in structure. Both contain a proximal sequence element (PSE) and an octamer motif-containing enhancer, and these elements are interchangeable between the two promoters. The RNA polymerase III specificity of the U6 promoter is conferred by a single A/T-rich element located around position -25. Mutation of the A/T-rich region converts the U6 promoter into an RNA polymerase II promoter, whereas insertion of the A/T-rich region into the U2 promoter converts that promoter into an RNA polymerase III promoter. We show that this A/T-rich element can be replaced by a number of TATA boxes derived from mRNA promoters transcribed by RNA polymerase II with little effect on RNA polymerase III transcription. Furthermore, the cloned RNA polymerase II transcription factor TFIID both binds to the U6 A/T-rich region and directs accurate RNA polymerase III transcription in vitro. Mutations in the U6 A/T-rich region that convert the U6 promoter into an RNA polymerase II promoter also abolish TFIID binding. Together, these observations suggest that in the human snRNA promoters, unlike in mRNA promoters, binding of TFIID directs the assembly of RNA polymerase III transcription complexes, whereas the lack of TFIID binding results in the assembly of RNA polymerase II snRNA transcription complexes.

New molecular biology methods for protein engineering

This review outlines recent advances in the application of molecular biological techniques to the study of protein structure and function. The chapter is divided into four main sections: methods for oligonucleotide-directed mutagenesis; mutational strategies for identifying functional residues and domains; systems for expression; and future developments. Few new methods were reported in 1990; however, a number of the papers that appeared represent refinements of previously reported strategies. This review is also published in Current Opinion in Structural Biology 1991, 1:605-610.

Absence of asparagine-linked oligosaccharides from glycoprotein D of herpes simplex virus type 1 results in a structurally altered but biologically active protein

Glycoprotein D (gD) of herpes simplex virus contains three utilized sites (Asn-X-Ser/Thr) for addition of asparagine-linked carbohydrates (N-CHO). Previously, we used oligonucleotide-directed mutagenesis to alter serine or threonine residues to alanine at each N-CHO addition site. Studies with monoclonal antibodies showed that a mutant protein lacking all three sites (now designated AAA) was structurally altered because of the amino acid change at residue 96 as well as the absence of the N-CHO. In this study, we constructed additional single mutations at site 1 (residues 94 and 96) and found that in most cases, the amino acid change itself adversely affected the conformation of gD. However, changing asparagine 94 to glutamine (Q) at site 1 had the least effect on gD. We constructed a second triple mutant, QAA, which lacked all three N-CHO signals. The antigenic conformation of QAA was similar to that of gD produced in the presence of tunicamycin (TM-gD). However, binding of MAbs to the AAA protein or to single mutants altered at site 1 was reduced compared with TM-gD. Wild-type gD and QAA proteins were equally susceptible to digestion by trypsin or Staphylococcus aureus V8 protease. In contrast, the AAA protein was more sensitive to trypsin but less sensitive to V8, again suggesting conformational alterations of the AAA protein. Despite what appeared to be large changes in structure, each mutant complemented the infectivity of a virus lacking gD (F-gD beta). We conclude that the N-CHO and amino acids at N-CHO site 1 play an important role in forming and/or maintaining gD structure, but none of the N-CHO are required for gD to function in the complementation assay.

A novel mechanism regulating growth factor association with the cell surface: identification of a PDGF retention domain

Platelet-derived growth factor (PDGF) chimeras were used to map a domain responsible for either efficient secretion of PDGF-A or the tight cell association of PDGF-B to their carboxy-terminal domains. Introduction of stop codons within PDGF-A or PDGF-B further dissected their respective carboxy-terminal domains. Although successive deletions of the PDGF-A carboxyl terminus did not impair its secretion, incremental deletions from the carboxyl terminus of PDGF-B abrogated its membrane retention properties and promoted secretion. By this approach, PDGF-B retention properties could be localized to PDGF-B residues 212-226. A processed form of PDGF-B, which contained this domain, was expressed at the cell surface but not released. Comparison of PDGF-B with PDGF-A revealed an analogous sequence located at the PDGF-A carboxyl terminus. We demonstrated that this PDGF-A domain also acts as a retention sequence under conditions that inhibit its proteolytic cleavage. Thus, differences in PDGF-A and PDGF-B secretion relate to differential proteolytic processing of analogous retention domains. All of these findings establish a new mechanism for stable growth factor presentation at the cell surface.

Altered DNA binding and replication activities of JC virus T-antigen mutants

Ten mutations were introduced into the JC virus (JCV) T antigen within a region corresponding to the SV40 T-antigen DNA binding domain (SV40 amino acids 131 to 220); nine of these increased homology between the two proteins in sequences critical for SV40 T antigen DNA binding. All mutant JCV T antigens bound to JCV and SV40 origins of DNA replication. Binding efficiency relative to the of wild-type JCV T antigen ranged from 83 to 301% for the JCV binding sites and from 44 to 240% for the SV40 binding sites. Nine mutant proteins promoted viral DNA replication in primary human fetal glial (PHFG) and CV-1 cells. In PHFG cells, promotion of DNA replication ranged from 26 to 220% relative to that of wild-type T antigen; in CV-1 cells it ranged from 14 to 522%. Coding sequences for five mutant proteins were transferred into the hybrid virus M1 (SV40) [M1(SV40) contains coding sequences from JCV and regulatory sequences from SV40]. Wild-type T antigen promoted replication weakly from the SV40 origin in these hybrid viruses in CV-1 cells (2% that from the JCV origin); replication driven by the mutant proteins ranged from 110 to 412% of that induced by the wild-type protein. Efficient specific DNA binding by a mutant T antigen was not a reliable indicator of that mutant protein's ability to promote DNA replication.

Optimization of the signal-sequence cleavage site for secretion from Bacillus subtilis of a 34-amino acid fragment of human parathyroid hormone

We have effected the secretion from Bacillus subtilis of a 34-amino acid (aa) fragment of human parathyroid hormone (PTH,1-34), using a Bacillus amyloliquefaciens neutral protease signal sequence. The secretion efficiency depended on the aa sequence near the signal-sequence cleavage site. We constructed a series of gene fusions encoding different pairs of aa between the signal sequence and PTH,1-34. There was a correlation between those polypeptides which were efficiently secreted and the potential for a beta-turn in the region just beyond the signal-sequence cleavage site. Based on this correlation, we constructed a gene fusion which specified Gly rather than Ala at the C terminus of the signal sequence, thus creating a beta-turn potential at the end of the signal sequence. The change provided a slight increase in secretion efficiency.