Strategies for high-throughput gene cloning and expression.

High-throughput approaches for gene cloning and expression require the development of new, nonstandard tools for use by molecular biologists and biochemists. We have developed and implemented a series of methods that enable the production of expression constructs in 96-well plate format. A screening process is described that facilitates the identification of bacterial clones expressing soluble protein. Application of the solubility screen then provides a plate map that identifies the location of wells containing clones producing soluble proteins. A series of semi-automated methods can then be applied for validation of solubility and production of freezer stocks for the protein production group. This process provides an 80% success rate for the identification of clones producing soluble protein and results in a significant decrease in the level of effort required for the labor-intensive components of validation and preparation of freezer stocks. This process is customized for large-scale structural genomics programs that rely on the production of large amounts of soluble proteins for crystallization trials.

Serologic cross-reactions among rabbit secretory IgA molecules: evidence for multiple subclasses of secretory IgA-f molecules.

Serologic cross-reactions among allotypes of rabbit secretory IgA (sIgA) of the f-subclass were examined by quantitative radiobinding assays with various allo-anti-alpha chain reagents. Numerous cross-reactions were observed which demonstrated the complexity of the C alpha f allotypes. One of the reagents used in these studies reacted not only with all sIgA molecules of the immunogen C alpha f allotype but also with all sIgA molecules of the other C alpha f allotypes. Aliquots of the antiserum were each adsorbed with IgA molecules of these C alpha f allotypes and then used in radiobinding studies with sIgA-f molecules of the various C alpha f allotypes. The adsorbed reagents reacted with some but not all sIgA-f molecules, thus indicating that the C alpha f allotypes each comprise more than one serologically distinguishable subset. Results from these radiobinding experiments were used to develop a model in which each of the five IgA-f allotypes comprises at least two serologically distinguishable subsets. Each of these subsets expresses a unique pattern of C alpha determinants. These C alpha determinants appear to be protein in nature rather than carbohydrate as periodate oxidation of the sIgA-f glycoproteins does not affect the reactions of the molecules with the cross-reactive alloreagents. IgA molecules of the serologically distinguishable subsets presumably differ in amino acid sequence in the constant region and, thus, would be products of distinct C alpha genes. Therefore, it is probable that the cross-reactions of these molecules, which were previously thought to comprise a single subclass, sIgA-f, may reflect the presence of more than one subclass of sIgA-f molecules, i.e., a third subclass of rabbit sIgA.

Covalent binding to alpha-macroglobulins of a protein with free SH groups produced by activated B cells: blocking by D-penicillamine and gold compounds.

alpha 2-Macroglobulin (alpha 2M) complexed with proteinases or modified by the action of amines has been shown to affect immune responses in vitro, though as yet the mechanisms are poorly understood. Supernates from rabbit lymphoid cells cultured in medium with normal rabbit serum and 35S-methionine (or 14C-leucine) were found to contain intensely radiolabeled alpha-macroglobulins (alpha M) (alpha 1 and alpha 2) on electrophoresis. When human alpha 2 M, instead of rabbit serum, was added to cultures, it also appeared radiolabeled, suggesting that lymphocyte-produced proteins (LyP) formed complexes with serum alpha M. These alpha M-associated LyP were produced in greater quantity when lymphocytes were cultured in the presence of mitogens; they were not produced by cells cultured in the presence of cycloheximide; they were produced primarily by B cells rather than T cells or macrophages. Pretreatment of serum or alpha M with methylamine, enhanced rather than inhibited the formation of LyP-alpha M complexes, a finding which is contrary to that expected if the LyP were a proteinase. Since this methylamine treatment of alpha M also results in the generation of free SH groups from the internal thioester bonds of alpha M, the formation of disulfide bonds between LyP and alpha M was considered. Indeed, (a) the LyP-alpha M complex formation was inhibited by N-ethylmaleimide, aurothiomalate, sodium aurothioglucose or D-penicillamine; (b) blocking the SH groups with NEM, of either culture fluid supernates or serum, had an inhibitory effect on the formation of these complexes; (c) the LyP-alpha M complexes were dissociated by sodium dodecyl sulfate (SDS) only after their reduction with 2-mercaptoethanol (2-ME). Thus, a disulfide bond was formed between alpha M and LyP with free SH groups (SH-LyP). Molecular sieving by high performance liquid chromatography (HPLC) of the serum-free radiolabeled supernates indicated that SH-LyP eluted at a position corresponding to a polypeptide of mol. wt of about 22,000. However, SDS-PAGE of the 22,000 mol. wt HPLC fraction showed that the major protein was approximately mol. wt 11,000 under both reducing and non-reducing conditions. In addition, the SH-LyP reduced by 2-ME from its binding site on alpha 2M had a mol. wt of about 11,000 in SDS-PAGE, suggesting that it was a non-covalent homodimer of mol. wt 11,000 polypeptides. We suggest that alpha 2M as well as SH-LyP may affect the immune system by functioning as SH-reactive agents.

Covalent disulfide binding of human IL-1 beta to alpha 2-macroglobulin: inhibition by D-penicillamine.

Secreted human IL-1 beta is known to have two free SH groups due to unpaired cysteines (positions 8 and 71). Alpha 2-Macroglobulin (alpha 2-M) has internal thioester bonds between cysteine and glutamate residues. Free SH groups may be generated at these alpha 2M residues through the action of proteinases, amines such as methylamine, or at a slow rate, by H2O ("aging" of alpha 2M). Thus, the possibility that IL-1 beta forms a disulfide bond with alpha 2M was investigated. 125I-labeled human rIL-1 beta (15 kDa) was incubated with fresh normal human serum or with purified alpha 2M, treated or not with methylamine. The mixtures were submitted to nondenaturing and denaturing polyacrylamide gel electrophoresis (PAGE) followed by autoradiography. IL-1 beta bound to commercially purified "aged" alpha 2M and to alpha 2M in methylamine-treated serum but not to native serum alpha 2M. It did not bind detectably to any other serum proteins. The addition of D-penicillamine (D-pen) during the reaction of [125I]rIL-1 beta with serum or purified alpha 2M blocked the covalent binding of rIL-1 beta to alpha 2M. [125I]rIL-1 beta was removed from alpha 2M by 2-mercaptoethanol in SDS. Thus, disulfide bonds were formed between the free SH groups on [125I]rIL-1 beta and those resulting from the cleavage of the internal thioester bonds of alpha 2M. "Cold" rIL-1 beta and a Cys71----Ser71 rIL-1 beta mutant effectively competed with [125I]rIL.1 beta for binding sites on alpha 2M. When complexes of rIL-1 beta or the mutant rIL-1 beta and alpha 2M were subjected to nonreducing SDS-PAGE and subsequent Western blot analysis, the rIL-1 beta molecules were found to be present in the alpha 2M bands in a dose-dependent manner. rIL-1 beta attached to alpha 2M in the presence or absence of D-pen showed similar biological activity in the mouse thymocyte-assay. Thus, rIL-1 beta attached noncovalently to alpha 2M is biologically active. The lack of inhibition of rIL-1 beta activity by binding to methylamine-treated alpha 2M in the absence of D-pen suggests, but does not prove, that the covalently bound rIL-1 beta is also active. We concluded that human rIL-1 beta binds to alpha 2M through the Cys at position 8 and that D-pen inhibits this binding. We speculate that this inhibitory effect may contribute to the therapeutic benefits of D-pen in patients with rheumatoid arthritis.

Identifying immunoglobulin isotypes, allotypes and idiotypes by a hemolytic assay in gel.

We developed a hemolytic radial immunodiffusion assay for identifying immunoglobulin (Ig) isotypes, allotypes and idiotypes by using gels containing erythrocytes coated with anti-Ig antibody or erythrocytes coated with Staphylococcal protein A. These indicator cells lysed specifically when treated sequentially with Ig antigen, the appropriate anti-Ig antiserum (developer) and complement. To identify these Ig subpopulations, we used monospecific indicator cells, e.g. erythrocytes coated with antibody specific for an Ig isotype, and developers with broader specificities ('multispecific'), e.g. antiserum to Fab. Alternatively, we used 'multispecific' indicator cells, e.g. erythrocytes coated with antibody to Fab and monospecific developers, e.g. antiserum to Ig idiotype. To identify Ig subpopulations specifically, either the indicator cells or the developer need to be monospecific. When both the indicator cells and the developer were monospecific, e.g. to allotype and to isotype, the specificity was determined by both reagents and ultimately restricted by the reagent with the narrower specificity, that is, reacting with the smallest Ig subpopulation. This sensitive hemolytic assay may be used to quantitate subpopulations of Ig molecules and may be modified into a reverse plaque forming cell assay to count lymphocytes secreting a given Ig class, type, allotype and idiotype.

Serologic and structural comparisons of rabbit IgA allotypes.

Serologic and structural comparisons of the rabbit IgA-g allotypes revealed that 1) the IgA-g allotypes have multiple allotypic determinant sites, 2) the g74, g76 and g77 allotypic specificities have several allotypic determinants in common whereas g75 molecules do not appear to have allotypic determinants in common with g74, g76 and g77 molecules, 3) the partial amino acid sequence of alpha chain from g75 and g76 Fc2alpha fragments differ by at least one amino acid residue, and 4) the g74 alpha-chains may have the "extra" intradomain disulfide bond in the Calpha2 domain whereas the g75 and g76 alpha chains lack this disulfide bond. Thus, multiple mutational events must have occurred during the evolution of g74, g75 and g77 genes.

The Production of Polyclonal Antibodies in Laboratory Animals. The Report and Recommendations of ECVAM Workshop 35.

This is the report of the thirty-fifth of a series of workshops organised by the European Centre for the Validation of Alternative Methods (ECVAM). ECVAM's main goal, as defined in 1993 by its Scientific Advisory Committee, is to promote the scientific and regulatory acceptance of alternative methods which are of importance to the biosciences and which reduce, refine or replace the use of laboratory animals. One of the first priorities set by ECVAM was the implementation of procedures which would enable it to become well informed about the state-of-the-art of non-animal test development and validation, and the potential for the possible incorporation of alternative tests into regulatory procedures. It was decided that this would be best achieved by the organisation of ECVAM workshops on specific topics, at which small groups of invited experts would review the current status of various types of in vitro tests and their potential uses, and make recommendations about the best ways forward (1). This joint ECVAM/FELASA (Federation of European Laboratory Animal Science Associations) workshop on The Immunisation of Laboratory Animals for the Production of Polyclonal Antibodies was held in Utrecht (The Netherlands), on 20-22 March 1998, under the co-chairmanship of Coenraad Hendriksen (RIVM, Bilthoven, The Netherlands) and Wim de Leeuw (Inspectorate for Health Protection, The Netherlands). The participants, all experts in the fields of immunology, laboratory animal science, or regulation, came from universities, industry and regulatory bodies. The aims of the workshop were: a) to discuss and evaluate current immunisation procedures for the production of polyclonal antibodies (including route of injection, animal species and adjuvant ); and b) to draft recommendations and guidelines to improve the immunisation procedures, with regard both to animal welfare and to the optimisation of immunisation protocols. This report summarises the outcome of the discussions and includes a number of recommendations and a set of draft guidelines (included in Appendix 1).

Physiologic and behavioral assessment of rabbits immunized with Freund's complete adjuvant.

Although the use of Freund's Complete Adjuvant (FCA) has been discouraged for the production of polyclonal antibodies, little clinical evidence supports the belief that FCA necessarily affects the well-being of immunized rabbits. We designed the present study to determine whether immunization at multiple sites with small volumes of Freund's adjuvant affects rabbit well-being. We injected 18 female New Zealand White rabbits (six animals per group) with antigen in FCA, Freund's Incomplete Adjuvant, or physiologic saline in the following volumes and routes: 0.02 to 0.03 mL intradermally in each of 30 to 40 sites and 0.1 mL subcutaneously in each of two sites. The body weight, temperature, complete blood count, and behavior of the rabbits in the home cage, upon handling, and in an open field did not differ significantly among the immunization groups during the 7-week assessment period. Only the degree of induration around injection sites differed: as expected, FCA induced the greatest response at the injection sites, but the sites were neither ulcerative nor necrotic, nor did palpation of the sites induce any apparent discomfort to the rabbits. We conclude that FCA may be used safely and humanely in rabbits if small volumes are injected intradermally or subcutaneously in multiple sites.

Cellular distribution of rabbit IgA allotypes.

Intestinal tissue from rabbits heterozygous at one or both of the alpha-chain subclass loci, i.e., f or g, was examined with mixtures of rhodamine-labeled and fluorescein-labeled antibodies each directed against one of the allelic products of the f or g locus. Individual plasma cells were stained by only one fluorescent reagent and thus exhibited both allelic exclusion and sublcass exclusion. In contrast, the epithelial cells of the Lieberkükhn's glands sid not exhibit allelic exclusion. Studies with rabbits heterozygous at the f and g loci revealed that the ratio of the number of cells expressing maternal-type f cells to paternal-type f is equal to the ratio of the number of maternal-type g cells to paternal-type g cells. These data are used to suggest that the expression of each of the genes in the heavy chain chromosomal region of a particular allogroup is under the same control.

Expression of 12 rabbit IgA C alpha genes as chimeric rabbit-mouse IgA antibodies.

Serologic analysis of rabbit secretory IgA initially identified two subclasses of IgA, IgA-f and IgA-g. Recent molecular genetic studies have resulted in the identification and cloning of 13 genes encoding the constant region (C) of rabbit IgA heavy chains. Each of these 13 C alpha genes, C alpha 1-C alpha 13, was subcloned into an expression vector containing the VDJ (V, variable; D, diversity; J, joining) gene of a dansyl (DNS)-binding hybridoma antibody. The alpha heavy-chain constructs were transfected into SP2/0 cells producing murine light chains with specificity for DNS. Of the 13 resulting transfectomas, 12 were shown by ELISA to secrete DNS-binding chimeric rabbit-mouse IgA molecules. By immunoblot analysis, the 12 IgA-producing transfectomas were shown to secrete alpha chains ranging in size from 60 to 72 kDa. These data suggest that rabbit IgA may be composed of as many as 12 IgA isotypes. This is in marked contrast to mouse and human, in which only 1 and 2 IgA isotypes, respectively, are found. Serologic analyses, using anti-IgA-f and anti-IgA-g alloantisera, revealed that 11 of the 12 transfectoma IgAs reacted with anti-IgA-f and not with anti-IgA-g antibodies and that one reacted with anti-IgA-g and not with anti-IgA-f antibodies. Each of the IgA-producing transfectomas was cocultured with a Madin-Darby canine kidney cell line expressing the rabbit polymeric immunoglobulin receptor, and the transcytosed IgA antibodies were analyzed by immunoblots to determine whether they associated with secretory component (SC) through covalent or noncovalent interactions. Each of the 11 IgA-f isotypes was shown to bind SC by a disulfide linkage, whereas the single IgA-g isotype appeared to bind SC through noncovalent interactions only.

Recombinant rabbit secretory immunoglobulin molecules: alpha chains with maternal (paternal) variable-region allotypes and paternal (maternal) constant-region allotypes.

A population of IgA molecules having heavy chains coded by two parental chromosomes in trans position has been identified in rabbits heterozygous at both the V(H)a locus, which controls allotypic specificities on the variable part of heavy chains, and the C(alpha)g locus, which controls allotypic specificities on the constant part of alpha chains. These recombinant molecules have alpha-chain allotypic specificities controlled by both the maternal V(H)a gene and the paternal C(alpha)g gene or conversely, the paternal V(H)a gene and the maternal C(alpha)g gene. These recombinant molecules were found in F(ab)(2alpha) fractions obtained after passage of F(ab)(2alpha) preparations through immunosorbent columns designed to remove one population of F(ab)(2alpha) molecules, i.e., g74- or g75-type molecules. The effluent F(ab)(2alpha) fractions were then examined by radioprecipitation methods for allotypic specificities controlled by the V(H)a and C(alpha)g loci. About 40% of the g75 F(ab)(2alpha) molecules from each of three rabbits with the a(1)g(74) and a(2)g(75) allogroups were alg75 recombinants. These alg75 recombinant molecules represented from 2.5-5.6% of the total unfractionated F(ab)(2alpha) sample. The F(ab)(2alpha) fractions from two rabbits with the a(1)g(75) and a(3)g(74) allogroups had from 1.8-8.2% recombinant molecules: some were alg74 recombinants and some were a3g75 recombinants. Somatic recombination as a mechanism responsible for the synthesis of polypeptide chains in which part of the information is obtained from one chromosome and part from the homologous chromosome is discussed.

Rabbit secretory components of different allotypes vary in their carbohydrate content and their sites of N-linked glycosylation.

The asparagine-linked glycosylation sites in rabbit high and low Mr secretory components (SC) have been determined for the three known allotypes, t61, t62, and t63. Purified SC polypeptides were subjected to mild periodate oxidation of terminal nonreducing sugars followed by selective reduction with [3H]sodium borohydride, SC polypeptides were further proteolytically cleaved, and the 3H-labeled peptides were isolated and characterized. Both high and low Mr SCs of the three allotypes possess a common glycosylation site at the asparagine residue position 400, whereas the second site, in the amino-terminal domain of SC, was found to be variable: the t61 and t63 allotypes contained a glycosylation site at positions 70 and 90, respectively. Moreover, although the t62 allotype was found to contain a triplet acceptor site (N-X-S) at positions 90-92, analyses showed that less than 30% of the t62 allotype peptides encompassing this region were glycosylated at residue 90. Furthermore, the amino acid sequence of the t61 SC in the region of residues 69-90 varies by 8 and 10 amino acid substitutions when compared with the t62 and t63 allotype sequences, respectively. However, neither the variation in amino acid sequence nor the variation in degree or site of glycosylation measurably affected the non-covalent binding of domain 1 to dimeric IgA.

Structural variability of rabbit secretory components. Allotype-associated differences in the third, fourth, and fifth domains.

We have previously shown (Frutiger, S., Hughes, G. J., Hanly, W. C., Kingzette, M., and Jaton, J.-C. (1986) J. Biol. Chem. 261, 16673-16681) that limited tryptic digestion of the high Mr form of rabbit secretory component of allotypes t61, t62, and t63 generates two major fragments, the NH2-terminal domain and a 40-kDa fragment encompassing domains 3, 4, and 5. Similarly, from the low Mr form of secretory component, (SC) the NH2-terminal domain, together with a 30-kDa fragment containing domains 4 and 5, were released. These fragments were used as inhibitors in a sensitive competitive binding radioimmunoassay with noncross-reactive rabbit alloantisera to study the distribution and localization of the major allotype-specific allotopes within the SC polypeptide. The 40-kDa fragments were shown to inhibit the 125I-labeled intact SC/anti-SC allotype reaction to the extent of 90%, i.e. nearly as well as the intact homologous high Mr SC form. In contrast, the NH2-terminal fragments (domain 1) were not inhibitory. The low Mr SC of each allotype was less inhibitory on a molar basis than the homologous high Mr SC polypeptide, an observation compatible with the deletion of domains 2 and 3 in the smaller polypeptide (Deitcher, D. L., and Mostov, K. E. (1986) Mol. Cell. Biol. 6, 2712-2715; Frutiger, S., Hughes, G. J., Fonck, Ch., and Jaton, J.-C. (1987) J. Biol. Chem. 262, 1712-1715). The structural correlates of the allotypic specificities were evaluated by comparative peptide mapping of the 40-kDa fragments (allotypes t61, t62, and t63). The data suggest that the t61 allotype structure differs significantly from the t62 and t63 structures, the latter two being much more related to each other than to t61. These findings are in full agreement with the serological data. The inhibition results suggest that the major allotype-specific, noncross-reactive allotopes of SC are distributed throughout domains 3, 4, and 5, even though domain 4 appears to be more conserved than domains 3 and 5 between the allotypes t61 and t63. Seven amino acid substitutions between t61 and t63 have been detected within domains 3, 4, and 5.

The IgA heavy-chain gene family in rabbit: cloning and sequence analysis of 13 C alpha genes.

Southern analysis has previously shown that the rabbit genome contains multiple genes coding for the constant regions of IgA heavy chains. In the present study, clones containing these C alpha genes have been isolated from cosmid and phage libraries. Restriction mapping and Southern analysis of the clones identified 13 non-allelic C alpha genes; 11 of the genes were clustered in individual or overlapping clones. The clustered genes are separated by 8-18 kb, and in total, the C alpha genes span a minimum of 160 kb of DNA. Southern analysis has shown that all genes within a cluster have the same transcriptional orientation, and that switch sequences are present 5' of at least 12 of the 13 genes. The nucleotide sequence of each C alpha gene was determined, and it appears that all genes are functional; thus, rabbit may have as many as 13 IgA isotypes. Comparisons of the protein sequences encoded by the 13 C alpha genes showed that the CH2 and CH3 domains of the alpha-chains are highly conserved, whereas the CH1 and hinge regions are highly diverse. Southern analysis of genomic DNA samples from other species within the order Lagomorpha showed that all samples had multiple C alpha hybridizing fragments. Thus, it is likely that all lagomorphs have multiple IgA isotypes and hence complex secretory immune systems.

Analysis of rabbit lung lavage immunoglobulins during the course of pulmonary inflammation induced with aerosolized antigen.

Lung lavage fluids (LLF) from rabbits with pigeon dropping extract (PDE)-induced granulomatous pulmonary inflammation were studied for protein and immunoglobulin (Ig) G and A levels. It was found that the protein levels of the lung fluids of rabbits increased to a maximum after 2-3 weeks of aerosol treatment with PDE during which time inflammation of the lung increases. This is followed by a gradual decrease in protein content as the inflammation wanes and the lung returns to normal. These variations primarily reflect changes in IgG and IgA levels. IgG and IgA levels follow different courses. IgA reaches a maximum in the first week of inflammation and then gradually decreases. In contrast, IgG reaches a maximum level (2-3 weeks) and stays at an elevated level throughout the 12 week period of aerosol treatment with PDE. Antibodies to PDE in these two classes of immunoglobulins do not entirely reflect the immunoglobulin class levels. IgA antibody levels reach a maximum after extended aerosol challenge while IgG antibody reaches a maximum early and then declines to background levels. The specificity of the non-PDE antibody IgG is unknown at present. The distribution of IgA subclass producing cells in the lung is different than in the gut. In the lung the major subclass is g while in the gut it is f. The distribution of subclasses of IgA in the LLF, however, does not appear to reflect the cellular distribution. The reason for this is not clear.

The amino-terminal domain of rabbit secretory component is responsible for noncovalent binding to immunoglobulin A dimers.

Rabbit secretory components (SC) constitute a family of markedly heterogeneous glycoproteins which are released in the secretions as free SC or as SC bound to polymeric immunoglobulins. The aim of this work was to determine the region of the SC polypeptides which is involved in IgA binding. The high and the low Mr forms of free SC (or IgA-dissociated bound SC) and the native secretory IgA complex were subjected to limited tryptic digestion. Chemically characterized peptides ranging in apparent size from 15 to 20 kDa, depending upon the allotype, were shown to be necessary and sufficient for efficient noncovalent binding to IgA dimers (subclass g). These fragments encompass the amino-terminal first domain of SC, i.e. residues 1-126, when aligned with the predicted amino acid sequence from a cDNA clone encoding the rabbit polymeric Ig receptor (Mostov, K.E., Friedlander, M., and Blobel, G. (1984) Nature 308, 37-43). The high and the low Mr forms of SC exhibited the same relative affinity for IgA dimers, suggesting that the postulated internal deletion in the smaller polypeptide (Kühn, L. C., Kocher, H.-P., Hanly, W.C., Cook, L., Jaton, J.-C., and Kraehenbuhl, J.-P. (1983) J. Biol. Chem. 258, 6653-6659) does not impair the IgA dimer recognition function.

Additional rabbit IgA allotypes, f69, f70, g76,g77: control by the C alphaf and C alphag loci.

Genetic and immunochemical studies have led to the identification of four additional rabbit IgA allotypes controlled by the Calphaf and Calphag loci. The folowing linkage combinations of the VHa and the 'new' alleles were observed among the populations of rabbits studied: a1f70g76, a1f69g77, and a2f69g77. Cross-reactions among g74, g76, and g77 molecules with various anti-g anti-allotype antisera indicate that the IgA-g allotypic specificities are comprised of multiple antigenic determinants. These studies provide a basis for further understanding of the evolution and gnetic control of the immunoglobulin heavy chain chromosomal region.