In vitro scanning saturation mutagenesis of all the specificity determining residues in an antibody binding site.

For the first time, each specificity determining residue (SDR) in the binding site of an antibody has been replaced with every other possible single amino acid substitution, and the resulting mutants analyzed for binding affinity and specificity. The studies were conducted on a variant of the 26-10 antidigoxin single chain Fv (scFv) using in vitro scanning saturation mutagenesis, a new process that allows the high throughput production and characterization of antibody mutants [Burks,E.A., Chen,G., Georgiou,G. and Iverson,B.L. (1997) Proc. Natl Acad. Sci. USA, 94, 412-417]. Single amino acid mutants of 26-10 scFv were identified that modulated specificity in dramatic fashion. The overall plasticity of the antibody binding site with respect to amino acid replacement was also evaluated, revealing that 86% of all mutants retained measurable binding activity. Finally, by analyzing the physical properties of amino acid substitutions with respect to their effect on hapten binding, conclusions were drawn regarding the functional role played by the wild-type residue at each SDR position. The reported results highlight the value of in vitro scanning saturation mutagenesis for engineering antibody binding specificity, for evaluating the plasticity of proteins, and for comprehensive structure-function studies and analysis.     

Amino acid substitutions outside a preselected antigenic region in hemoglobin affect the binding to monoclonal antibodies obtained by immunization with the synthetic region

It is often assumed that amino acid substitutions outside a protein antigenic site have no effect on the reactivity of a protein variant with antibodies, especially monoclonal antibodies (mAbs). Substitutions that exert an effect on the reactivity of a protein variant with mAbs are frequently considered to be within the antigenic site of the mAb. To test this assumption, two mAbs [IgG_l(k) and IgG_2a (k)] were prepared by immunization with a synthetic peptide corresponding to region 63–78 of the chain of human hemoglobin (Hb). The peptide was used as an immunogen in its free form (i.e., without conjugation to a carrier), so that the results will not be made ambiguous by peptide modification nor by an immune response to sites spanning peptide and protein carrier. In addition to their reaction with human Hb, the mAbs were also studied with four primate Hbs which had no substitutions within region 63–78 and only a limited number of substitutions which occurred outside of, and at considerable distances in three-dimensional (3D) structure from, this region. Inhibition studies revealed substantial differences in the binding affinities of some of the primate Hbs, relative to human Hb. Some of the substitutions caused major decreases in binding, although they were at considerable distances in the 3D structure from the indicated site residues. It is concluded that substitutions in a protein, even when distant from an antigenic site, can exert major influences on the protein's reactivity with anti-site mAbs.     

Analogues of oxytocin antagonists bearing a ureido group in the amino acid side chain at position 4 or 5.

Substitution of the side chain carboxamido group at position 4 in the potent oxytocin antagonist (OTA) [ThiaPmp(1), D-Trp(2), Cys(6), Arg(8)]-OT, PA, in which ThiaPmp = beta,beta-(3-thiapentamethylene)-beta-mercaptopropionic acid, led to [Orn(Car)(4)]-PA, ([Cit(4)]-PA), which had uterotonic antagonistic activity equal to that of PA. The same modification at position 5, leading to [Cit(5)]-PA, resulted in antagonistic potency more than 10 times lower than that of PA. This paper also describes the same substitutions introduced in the highly potent OTA [Pen(6)]-PA (antioxytocic in vitro pA(2) = 8.72). Analogues of the general formula [U(4)-X(5)-Pen(6)]-PA, in which U = Lys, Orn, Dab, Dap or X = Orn, Dab or Dap, were synthesized by SPPS. Each of these analogues was carbamoylated by treatment with KCNO in DMF-H(2)O, yielding the corresponding U(Car)(4) or X(Car)(5) derivatives. In the uterotonic assay, the substitution with the ureido group at Gln(4) results in retention of high antagonistic potency, albeit somewhat lower than that of PA, e.g. [Orn(Car)(4), Pen(6)]-PA and [Dab(Car)(4), Pen(6)]-PA having pA(2) = 8.52 and pA(2) = 8.42 respectively. In the pressor assay, [Lys(Car)(4), Pen(6)]-PA and [Dab(Car)(4), Pen(6)]-PA were somewhat weaker antagonists of arginine vasopressin than [Pen(6)]-PA; [Dap(Car)(4), Pen(6)]-PA showed only a faint trace of pressor agonistic activity. The substitution with the ureido group at position 5 leads to a significant loss of OTA potency in the in vitro uterotonic assay. The [Orn(Car)(5), Pen(6)]-PA was the most potent of the series (pA(2) = 8.05). An interesting finding is that [Dap(Car)(5), Pen(6)]-PA is equipotent with its precursor [Dap(5), Pen(6)]-PA (potency in the uterotonic test in vitro, pA(2) = 7.71 and pA(2) = 7.68, respectively). Furthermore, neither [Dap(5), Pen(6)]-PA nor [Dap(5), Pen(6), Gly(9)]-PA exhibited activity in the antidiuretic or pressor assays. Although these last two analogues show some decrease in antioxytocin potency, they behave as pure oxytocin antagonists, which makes them attractive candidates for further studies on the development of potent and specific OTAs.     

Single-channel studies on linear gramicidins with altered amino acid side chains. Effects of altering the polarity of the side chain at position 1 in gramicidin A.

The modulation of gramicidin A single-channel characteristics by the amino acid side chains was investigated using gramicidin A analogues in which the NH2 terminal valine was chemically replaced by other amino acids. The replacements were chosen such that pairs of analogues would have essentially isosteric side chains of different polarities at position 1 (valine vs. trifluorovaline or hexafluorovaline; norvaline vs. S-methyl-cysteine; and norleucine vs. methionine). Even though the side chains are not in direct contact with the permeating ions, the single-channel conductances for Na+ and Cs+ are markedly affected by the changes in the physico-chemical characteristics of the side chains. The maximum single-channel conductance for Na+ is decreased by as much as 10-fold in channels formed by analogues with polar side chains at position 1 compared with their counterparts with nonpolar side chains, while the Na+ affinity is fairly insensitive to these changes. The relative conductance changes seen with Cs+ were less than those seen with Na+; the ion selectivity of the channels with polar side chains at position 1 was increased. Hybrid channels could form between compounds with a polar side chain at position 1 and either valine gramicidin A or their counterparts with a nonpolar side chain at position 1. The structure of channels formed by the modified gramicidins is thus essentially identical to the structure of channels formed by valine gramicidin A. The polarity of the side chain at position 1 is an important determinant of the permeability characteristics of the gramicidin A channel. We discuss the importance of having structural information when interpreting the functional consequences of site-directed amino acid modifications.     

Analogues of a potent oxytocin antagonist with truncated C-terminus or shorter amino acid side chain of the basic amino acid at position 8.

Twelve analogues were synthesized, their structure derived from modifications of [(S)Pmp1, D-Trp2, Pen6, Arg8]oxytocin, PA, in which (S)Pmp = beta,beta-(3-thiapentamethylene-beta-mercaptopropionic acid). PA is a potent antagonist of the uterotonic effect of oxytocin in the rat (uterotonic test in vitro, pA2 = 8.86) and in the baboon. Truncated analogues of PA from the C-terminus were systematically prepared ending in either the free acid or the amide, i.e. PA1-9 acid, PA1-8 acid, PA1-7 acid, PA1-6 acid, PA1-8 amide, PA1-7 amide and PA1-6 amide. PA1-8 amide was roughly as potent as PA in the rat uterotonic assay in vitro, and the shorter amides were only somewhat weaker antagonists. All four acid analogues were weaker antagonists than PA but still maintained rather high antagonistic potency. These findings suggest that, if these truncated acids form as metabolites in vivo, they may contribute to the overall biological effect of PA and their contribution should be taken into account. Furthermore, using these analogues, the radioimmunoassay measurements of PA may be standardized, as they may cross react with PA antibodies and interfere with the determination. In addition, five analogues were made by substituting Arg8 of PA with Lys, Orn8, Dab8, Dap8 and Cit8. All of these analogues maintained high potency as OTAs in the uterotonic assay, although their activity was only about 1.5-3 times lower than PA. The most potent analogue in the uterotonic assay, [Dap8]PA, pA2 = 8.53, had weak pressor activity (pA2 = 6.90) and no antidiuretic effect. The pressor activity was lower for all tested acids, and for PA1-6 acid it was even below the detection limit. Additionally, PA1-9 acid, PA1-7 acid and PA1-6 acid showed no antidiuretic activity. Hence, the PA1-6 acid is a potent OTA with pA2 = 8.27 and no measurable effect in the pressor or antidiuretic tests and thus it is a pure oxytocin antagonist. This fact makes it an attractive candidate for further studies on inhibition of OT biological effects and on preterm labour.     

In vitro mutagenesis studies at the arginine residues of adenylate kinase. A revised binding site for AMP in the X-ray-deduced model

Although X-ray crystallographic and NMR studies have been made on the adenylate kinases, the substrate-binding sites are not unequivocally established. In an attempt to shed light on the binding sites for MgATP2- and for AMP2- in human cytosolic adenylate kinase (EC 2.7.4.3, hAK1), we have investigated the enzymic effects of replacement of the arginine residues (R44, R132, R138, and R149), which had been assumed by Pai et al. [Pai, E. F., Sachsenheimer, W., Schirmer, R. H., & Schulz, G. E. (1977) J. Mol. Biol. 114, 37-45] to interact with the phosphoryl groups of AMP2- and MgATP2-. With use of the site-directed mutagenesis method, point mutations were made in the artificial gene for hAK1 [Kim, H. J., Nishikawa, S., Tanaka, T., Uesugi, S., Takenaka, H., Hamada, M., & Kuby, S. A. (1989) Protein Eng. 2, 379-386] to replace these arginine residues with alanyl residues and yield the mutants R44A hAK1, R132A hAK1, R138A hAK1, and R149A hAK1. The resulting large increases in the Km,app values for AMP2- of the mutant enzymes, the relatively small increases in the Km,app values for MgATP2-, and the fact that the R132A, R138A, and R149A mutant enzymes proved to be very poor catalysts are consistent with the idea that the assigned substrate binding sites of Pai et al. (1977) have been reversed and that their ATP-binding site may be assigned as the AMP site.     

An analysis of the side chain requirement at position 177 within the lactose permease which confers the ability to recognize maltose.

In previous work (Brooker, R. J., and Wilson, T. H. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 3959-3963), lactose permease mutants were isolated which possessed an enhanced recognition for maltose. In some of these mutants, the wild-type alanine residue at position 177 was changed to valine or threonine. To gain further insight into the side chain requirement at position 177 that confers maltose recognition, further substitutions of isoleucine, leucine, phenylalanine, proline, and serine have been made via site-directed mutagenesis. Permeases containing alanine or serine exhibited poor maltose recognition whereas those containing isoleucine, leucine, phenylalanine, proline, or valine showed moderate or good recognition. As far as galactosides are concerned, the Val-177, Pro-177, and Ser-177 mutants were able to transport lactose as well as, or slightly better than, the wild-type strain. The other mutants displayed moderately reduced levels of lactose transport. For example, the Phe-177 mutant, which was the most defective, showed a level of downhill transport which was approximately 20% that of the wild-type strain. In uphill transport assays, all of the position 177 mutants were markedly defective in their ability to accumulate beta-D-thiomethylgalactopyranoside against a concentration gradient. Finally, the position 177 mutants were analyzed for their ability to catalyze an H+ leak. Interestingly, even though the wild-type permease does not leak H+ across the bacterial membrane, all of the position 177 mutants were shown to transport H+ in the absence of sugars. For most of the mutants, this H+ leak was blocked by the addition of beta-D-thiodigalactoside. Overall, these results are discussed with regard to the effects of position 177 substitutions on the sugar recognition site and H+ transport.     

Resistance of human immunodeficiency virus type 1 to neutralization by natural antisera occurs through single amino acid substitutions that cause changes in antibody binding at multiple sites.

The ability of human immunodeficiency virus type 1 (HIV-1) to replicate in the presence of strong immune responses to the virus may be due to its high mutation rate, which provides envelope gene variability for selection of neutralization-resistant variants. Understanding neutralization escape mechanisms is therefore important for the design of HIV-1 vaccines and our understanding of the disease process. In this report, we analyze mutations at amino acid positions 281 and 582 in the HIV-1 envelope, where substitutions confer resistance to broadly reactive neutralizing antisera from seropositive individuals. Neither of these mutations lies within an antibody-binding site, and therefore the mechanism of immune escape in both cases is by alteration of the shape of the envelope proteins. The conformation of the CD4-binding site is shown to be critical with regard to presentation of other discontinuous epitopes. From our analysis of the neutralization of these variants, we conclude that escape from polyclonal sera occurs through alterations at several different epitopes, generally resulting from single amino acid substitutions which influence envelope conformation. Experiments on a double mutant showed that the combination of both mutations is not additive, suggesting that these variants utilized alternate pathways to elicit similar alterations of the HIV-1 envelope structure.     

The binding of monoclonal antibodies to cell surface molecules. Quantitative analysis of the reactions and cross-reactions of an antibody (MB40.3) with four HLA-B molecules

The MB40.3 monoclonal antibody binds to four distinct HLA-B molecules; B7, B40, B40*, and B27. With Fab' fragments only the interaction with B7 and B40 was detected and the affinity for both was the same (1-2 X 10(8) M-1) suggesting the epitope is shared by the two molecules. Unlike many antibodies for which low affinity is due to a high-dissociation constant, that of MB40.3 results from a very low-association rate constant, coupled with a low-dissociation constant. In consequence, the affinity and avidity of Fab', F(ab')2, and IgG for B7 and B40 were found to be of a similar magnitude, soluble B7 was a more efficient competitor for antibody than cell surface B7 and in practice antibody bivalency was of little importance. The forward rate constant could be increased by removing Fc from the antibody or by removing sialic acid from the cells by treatment with neuraminidase. The neuraminidase treatment also produced an increase in the number of detectable cell surface HLA-A,B molecules. The affinity of MB40.3 for B40* and B27 was estimated to be less than 4 X 10(6) as no binding with Fab' was detected due to a high-dissociation rate. For these two HLA-B molecules bivalent attachment was critical, and it increased the strength of interaction with cell surface B40* and B27 to a point where the avidities were comparable to those obtained with B7 and B40, with B40* interacting more strongly than B27. The epitopes recognized by MB40.3 on B40* and B27 were thus shown to be structurally different from each other and from those on B7 and B40. The properties of this antibody contrast with those of other anti-HLA-A,B we have studied (Ways, J.P., and Parham, P. (1983) Biochem. J. 216, 423-432).     

Use of amber suppressors to investigate the thermostability of Bacillus licheniformis alpha-amylase. Amino acid replacements at 6 histidine residues reveal a critical position at His-133

A set of 12 Escherichia coli suppressor tRNAs, inserting different amino acids in response to an amber codon, has been used to create rapidly numerous protein variants of a thermostable amylase; by site-directed mutagenesis, amber mutations were first introduced into Bacillus licheniformis alpha-amylase gene at position His35, His133, His247, His293, His406, or His450; genes carrying one or two amber mutations were then expressed in the different suppressor strains, generating over 100 amylase variants with predicted amino acid changes that could be tested for thermostability. Within the detection limits of the assays, amino acid replacements at five histidine positions had no significant effect. In contrast, suppressed variants substituted at residue His133 clearly exhibited modified thermostability and could be either less stable or more stable than the wild-type amylase, depending on the amino acid inserted at this position; comparison of the variants indicates that the hydrophobicity of the substituting residue is an important but not a determinant factor of stabilization. The effect of the most stabilizing and destabilizing amino acid substitutions, His133 to Tyr and to Pro, respectively, were confirmed by introducing the corresponding missense mutations into the gene sequence. The advantages and limits of informational suppression in protein stability studies are discussed as well as structural features involved in the thermostability of B. licheniformis alpha-amylase.     

Amino acid substitutions at tryptophan 388 and tryptophan 412 of the HepG2 (Glut1) glucose transporter inhibit transport activity and targeting to the plasma membrane in Xenopus oocytes.

All 6 tryptophan residues in the human HepG2-type glucose transporter (Glut1) were individually altered by site-directed mutagenesis to investigate the role of these residues in transport function. Tryptophan residues in positions 48, 65, 186, 363, 388, and 412 of Glut1 were changed to either a glycine or leucine residue. Mutant mRNAs were synthesized and injected into Xenopus laevis oocytes. Transporter function as assessed by uptake of 2-deoxy-D-[3H]glucose or transport of 3-O-[3H]methylglucose was decreased in the 388 and 412 mutants but was unaltered in all other mutants. The amount of the mutant transporters expressed in total membrane and plasma membrane fractions was measured using Glut1-specific antibodies. Calculation of the intrinsic transport activity of each of the mutants using these data demonstrated that the reduced transport activity of the 412 mutants was caused entirely by a dramatic decrease in the intrinsic activity of the mutant proteins whereas the reduced activity of the 388 mutants was a result of a decreased level of the protein in oocytes, decreased targeting to the plasma membrane, and a modest decrease in the intrinsic activity. Protease/glycosidase mapping of in vitro translation products indicated that the effects of the 388 and 412 point mutations could not be attributed to a disruption in the ability of the mutant proteins to insert properly into the membrane. The ID50 for cytochalasin B inhibition of 2-deoxyglucose uptake was increased from 5 x 10(-7) M for the wild-type Glut1 to 4 x 10(-6) M in the 388 mutants but was unaltered in the 412 mutants. These observations suggest that 1) Trp-412 may comprise part of a hexose binding site or is involved in maintaining a local tertiary structure critical for transport function; 2) Trp-388 is involved in stabilizing the equilibrium binding of cytochalasin B to the transporter. Trp-388 may therefore lie near a substrate binding site and also appears to participate in stabilization of local tertiary structure important for full catalytic activity and efficient targeting to the Xenopus plasma membrane.     

Preparation and characterization of 3-monohydroxylated bile acids of different side chain length and configuration at C-3. Novel approach to the synthesis of 24-norlithocholic acid

A series of 3-monohydroxylated bile acids, in unlabeled and radioactive form, of varying side chain length and configuration at C-3 has been synthesized and rigorously characterized. They include: 3 alpha- and 3 beta-hydroxy-5 beta-androstane-17 beta-carboxylic acids (C20); 3 alpha- and 3 beta-hydroxy-5 beta-pregnan-21-oic acids (C21); 3 alpha- and 3 beta-hydroxy-23,24-bisnor-5 beta-cholan-22-oic acids (C22); 3 alpha- and 3 beta-hydroxy-24-nor-5 beta-cholan-23-oic acids (C23, norlithocholic and isonorlithocholic acids); and 3 beta-hydroxy-5 beta-cholan-24-oic acid (C24, isolithocholic acid). A novel approach to the degradation of lithocholic acid acetate to 24-norlithocholic acid is described. This degradation involves the photochemical modification of a Hunsdiecker reaction and Kornblum oxidation of the intermediate 23-bromide. The availability of these compounds makes it possible to study the metabolism and biological effects of short chain bile acids.     

Evidence that glutamic acid 49 of tryptophan synthase alpha subunit is a catalytic residue. Inactive mutant proteins substituted at position 49 bind ligands and transmit ligand-dependent to the beta subunit

Glutamic acid 49 of the alpha subunit of tryptophan synthase from Escherichia coli is an essential residue since 19 mutant proteins substituted at position 49 were found previously to be inactive. Our present findings that five mutants of the alpha subunit, substituted with Asp, Lys, Ala, Phe, or Gly at position 49, bind a substrate analog normally are further evidence that glutamic acid 49 is a catalytic base. Ligands of the alpha subunit also have similar effects on site-site interactions between the beta subunit and the wild type or mutant alpha subunits. These effects include inhibition of the activity of the beta subunit, reduction of the dissociation constant for D-tryptophan, and increase of the equilibrium concentration of a quinonoid intermediate formed with L-tryptophan.     

In vitro maximum binding of antigenic peptides to murine MHC class II molecules does not always take place at the acidic pH of the in vivo endosomal compartment.

Presentation of Ag to the T cell requires binding of specific peptide fragments of the Ag to MHC II molecules. The ability of a peptide to bind to MHC class II appears to be pH dependent. Recent reports indicate that the binding of peptide to MHC class II molecules takes place primarily within an endosomal compartment of the cell at around pH 5. In this study, we have explored the in vitro pH dependence of peptide binding to different haplotypes of murine MHC class II molecules. The binding of peptides to MHC II was analyzed and quantitated by silica gel TLC, using radiolabeled peptides. The MBP peptide fragments, MBP(1-14)A4 and MBP(88-101)Y88, bound maximally at pH 8 to IAk and IAs, respectively. The binding of PLP peptide fragment, PLP(138-151)Y138, to IAs was maximal at around neutral pH. The maximum binding of an OVA peptide fragment, OVA(323-340)Y340, to IAd, was found to occur at pH 6. Results presented in this report thus suggest that the in vitro maximum binding of peptide is pH dependent and does not always occur at pH 5. The optimum pH range for maximum binding may depend on the nature and net charge of the peptide and its interaction with MHC class II molecules.