Human-engineered monoclonal antibodies retain full specific binding activity by preserving non-CDR complementarity-modulating residues.

Humanization of murine monoclonal antibodies for human therapy has commonly been achieved by complementarity-determining region (CDR) grafting, in which murine CDR loops are grafted onto human framework regions. Difficulties with that method have revealed the importance of certain framework residues in determining both the 3-D structure of CDR loops and the overall affinity of the molecule for its specific ligand. In the general model of structure-function relationships presented here, each amino acid position in the variable region is classified according to the benefit of achieving a more human-like antibody versus the risk of decreasing or abolishing specific binding affinity. Substitutions of human residues at low-risk positions (exposed to solvent but not contributing to antigen binding or antibody structure) are likely to decrease immunogenicity with little or no effect on binding affinity. Changes at high-risk positions (directly involved in antigen binding, CDR stabilization or internal packing) are avoided to preserve the biological activity of the antibody. Moderate-risk changes are made with caution. This model has been tested experimentally using H65, an anti-CD5 murine monoclonal antibody, whose binding activity had been greatly reduced by two previous attempts at humanization by conventional CDR grafting. The new 'human-engineered' H65 antibody containing 20 low-risk human consensus substitutions (expressed as either IgG or Fab) retains the full binding avidity of parental murine and chimeric H65 antibodies. A human-engineered antibody with an additional 14 moderate-risk substitutions has unexpectedly enhanced avidity (3- to 7-fold). This method is generally applicable to the design of other human-engineered antibodies with therapeutic potential.

ELISA assay optimization using hyperbolic regression.

The enzyme-linked immunosorbent assay (ELISA) is routinely used for estimation of specific protein concentration in research and in manufacturing. To optimize both the sensitivity of the assay and its usable dynamic range, it is necessary to adjust the concentrations of reagents, as well as other experimental conditions. Hyperbolic regression is a novel mathematical technique which gives the investigator a quantitative analysis of the assay's kinetic response to variations in reagents and conditions. The equations for hyperbolic regression are easily programmed on a computer spreadsheet or even on a scientific pocket calculator. Hyperbolic regression can be used to facilitate the initial optimization of the assay conditions, or to estimate the concentrations of unknowns on a routine basis directly from the standard curve, or to monitor the long-term degradation in activity of antibody reagents and antigen standards.

Expression of foreign proteins in microorganisms.

The various alternative strategies for the expression of heterologous proteins in microorganisms are reviewed. To illustrate how these general considerations can be addressed in particular cases, the expression of chimeric human-mouse antibodies in Escherichia coli and the production of thaumatin, an intensely sweet plant protein, in yeast, are described.

Escherichia coli promoter -10 and -35 region homologies correlate with binding and isomerization kinetics.

The DNA promoter sequence at which gene transcription is initiated in Escherichia coli contains two distinct regions of conserved nucleotides. Coefficients of homology to the -10 region and the -35 region were computed for many different prokaryotic promoters. Linear equations were derived that relate the degree of homology for each promoter region to the two kinetic constants that describe the interaction of RNA polymerase with the promoter site on DNA: the strength KB of binding to form closed complex, and the rate kf of isomerization to form open complex. A graphical plot of -35 versus -10 promoter region homologies for many promoters suggest that certain classes of prokaryotic operons might utilize differential degrees of consensus homology in these two regions to achieve specific control over kf and KB, in addition to modulating overall promoter strength.

Hyperbolic regression analysis for kinetics, electrophoresis, ELISA, RIA, Bradford, Lowry, and other applications.

Hyperbolic regression analysis is an effective method for fitting experimental data points obtained from a variety of experiments in molecular biology, including enzyme kinetics, agarose gel electrophoresis of DNA fragments, SDS-polyacrylamide gel electrophoresis of proteins, enzyme-linked immunosorbent assays (ELISA), radioimmunoassays (RIA), Bradford protein quantitation assays, Lowry protein assays, and other applications. Hyperbolic regression yields excellent fitted curves without the biases that are introduced by carrying out linear regression on double reciprocal coordinates, and it produces one simple equation, encompassing all the data points, that can be used easily in a pocket calculator to estimate the values of unknown samples from the known standards.

Nucleotide sequence homologies in control regions of prokaryotic genomes.

Functional recognition sites for several regulatory factors, including RNA polymerase, cyclic adenosine monophosphate receptor protein and ribosomes, do not always have strong consensus nucleotide sequence homology, yet they are capable of biological activity. Using the computer, other nucleotide sequences can be found that have equal or significantly greater consensus homology, but whose biological function has not been characterized. This analysis shows that no arbitrary 'cutoff score' can successfully distinguish active recognition sites from uncharacterized homologies, due to the great natural diversity in the strength and conservation of functional sites. It also predicts that the strong 'cryptic' homologies presented here are of two types: some might already have a biological function which has so far not been detected, whereas certain single-point mutations might be able to confer activity upon the others by correcting a key structural defect.

Quantitative computer analysis of signal sequence homologies in DNA.

Homologies to prokaryotic recognition sites for RNA polymerase, ribosomes, and cyclic-AMP receptor protein (CRP), are analyzed by a new computer program using weighting factors to account for the statistical variation at each position of the consensus. Known signal sequence sites are easily detected by this algorithm, and other sites with equally strong homology are found whose biological function is still unknown. Some sites are biologically active even though they have very weak homology. No arbitrary 'cutoff score' can distinguish active recognition sites from inactive homologies; experiments must determine why certain weak homologies are able to function while others are not.

A simple, computer-assisted assay to detect isotype-specific regulation of human immunoglobulin synthesis.

A simple, reliable, and computer-assisted assay has been developed to quantitate isotype-specific regulation of human immunoglobulin synthesis in vitro. The assay utilizes three separate human lymphoblast or myeloma cell lines, which secrete human immunoglobulins IgA, IgG, and IgE. Culture supernatants from 96-well tissue culture plates are then assayed for IgA, IgG, and IgE by a solid-phase enzyme-linked immunosorbent assay (ELISA) on a microtiter plate. Data collection and analysis is performed with the aid of computer programs designed for this assay. This assay has several advantages over other immunoglobulin regulation assays: no radioisotopes are used, thereby reducing cost and complexity; results are directly collected and quantified by computer analysis; the entire assay is completed in 3 days; reliability and reproducibility are increased by the use of established human cell lines; and co-culturing all three immunoglobulin-producing cell lines provides convenient internal controls for isotype specificity.

A unique secondary folding pattern for 5S RNA corresponds to the lowest energy homologous secondary structure in 17 different prokaryotes.

A general secondary structure is proposed for the 5S RNA of prokaryotic ribosomes, based on helical energy filtering calculations. We have considered all secondary structures that are common to 17 different prokaryotic 5S RNAs and for each 5S sequence calculated the (global) minimum energy secondary structure (300,000 common structures are possible for each sequence). The 17 different minimum energy secondary structures all correspond, with minor differences, to a single, secondary structure model. This is strong evidence that this general 5S folding pattern corresponds to the secondary structure of the functional 5S rRNA. The general 5S secondary structure is forked and in analogy with the cloverleaf of tRNA is named the "wishbone" model. It constant 8 double helical regions; one in the stem, four in the first, or constant arm, and three in the second arm. Four of these double helical regions are present in a model earlier proposed (1) and four additional regions not proposed by them are presented here. In the minimum energy general structure, the four helices in the constant arm are exactly 15 nucleotide pairs long. These helices are stacked in the sequences from gram-positive bacteria and probably stacked in gram-negative sequences as well. In sequences from gram-positive bacteria the length of the constant arm is maintained at 15 stacked pairs by an unusual minimum energy interaction involving a C26-G57 base pair intercalated between two adjacent helical regions.

The primary sequence of rabbit alpha-globin mRNA.

The rabbit alpha-globin DNA insertion in the chimeric plasmid pHb 72 (Liu et al., 1977) has been sequenced by the method of Maxam and Gilbert (1977). This has enabled us to determine the messenger RNA(mRNA) sequence beginning in the 5' untranslated region 9 nucleotides before the initiation codon and extending through the first 361 nucleotides of the translated region. The data reported here overlap and are in complete agreement with sequences determined by Baralle (1977) for the 5' end of the mRNA and by Proudfoot et al. (1977) for the 3' end. Our sequence is also in agreement with the partial complementary RNA (cRNA) sequencing data which we reported previously (Paddock et al., 1977), this work marks the completion of the primary sequence of the rabbit alpha-globin mRNA. These observations reaffirm the high fidelity with which gene copies can be synthesized in vitro, cloned in a bacterial plasmid and maintained in the host. The general features of the mRNA nucleotide sequence are duscussed with particular attention given to the base composition and codon preferences observed and to comparison of this sequence with other completed mRNA gene sequences. A new computer program has been used to search for the most stable base-pairing arrangement of the completed mRNA.

Computer method for predicting the secondary structure of single-stranded RNA.

We present a computer method utilizing published values for base pairing energies to compute the most energetically favorable secondary structure of an RNA from its primary nucleotide sequence. After listing all possible double-helical regions, every pair of mutally incompatible regions (whose nucleotides overlap) is examined to determine whether parts of those two regions can be combined by branch migration to form a pair of compatible new subregions which together are more stable than either of the original regions separately. These subregions are added to the list of base pairing regions which will compete to form the best overall structure. Then, a 'hyperstructure matrix' is generated, containing the unique topological relationship between every pair of regions. We have shown that the best structure can be chosen directly from this matrix, without the necessity of creating and examing every possible secondary structure. We have included the results from our solution of the 5S rRNA of the cyanobacterium Anacystis nidulans as an example of our program's capabilities.