Evaluation of affinity-tagged protein expression strategies using local and global isotope ratio measurements.

Elucidation of protein-protein interactions can provide new knowledge on protein function. Enrichments of affinity-tagged (or "bait") proteins with interaction partners generally include background, nonspecific protein artifacts. Furthermore, in vivo bait expression may introduce additional artifacts arising from altered physiology or metabolism. In this study, we compared these effects for chromosome and plasmid encoding strategies for bait proteins in two microbes: Escherichia coli and Rhodopseudomonas palustris. Differential metabolic labeling of strains expressing bait protein relative to the wild-type strain in each species allowed comparison by liquid chromatography tandem mass spectrometry (LC-MS-MS). At the local level of the protein complex, authentic interacting proteins of RNA polymerase (RNAP) were successfully discerned from artifactual proteins by the isotopic differentiation of interactions as random or targeted (I-DIRT, Tackett, A. J.; et al. J. Proteome Res. 2005, 4, 1752-1756). To investigate global effects of bait protein production, we compared proteomes from strains harboring a plasmid encoding an affinity-tagged subunit (RpoA) of RNAP with the corresponding wild-type strains. The RpoA abundance ratios of 0.8 for R. palustris and 1.7 for E. coli in plasmid strains versus wild-type indicated only slightly altered expression. While most other proteins also showed no appreciable difference in abundance, several that did show altered levels were involved in amino acid metabolism. Measurements at both local and global levels proved useful for evaluating in vitro and in vivo artifacts of plasmid-encoding strategies for bait protein expression.

A general system for studying protein-protein interactions in Gram-negative bacteria.

One of the most promising methods for large-scale studies of protein interactions is isolation of an affinity-tagged protein with its in vivo interaction partners, followed by mass spectrometric identification of the copurified proteins. Previous studies have generated affinity-tagged proteins using genetic tools or cloning systems that are specific to a particular organism. To enable protein-protein interaction studies across a wider range of Gram-negative bacteria, we have developed a methodology based on expression of affinity-tagged "bait" proteins from a medium copy-number plasmid. This construct is based on a broad-host-range vector backbone (pBBR1MCS5). The vector has been modified to incorporate the Gateway DEST vector recombination region, to facilitate cloning and expression of fusion proteins bearing a variety of affinity, fluorescent, or other tags. We demonstrate this methodology by characterizing interactions among subunits of the DNA-dependent RNA polymerase complex in two metabolically versatile Gram-negative microbial species of environmental interest, Rhodopseudomonas palustris CGA010 and Shewanella oneidensis MR-1. Results compared favorably with those for both plasmid and chromosomally encoded affinity-tagged fusion proteins expressed in a model organism, Escherichia coli.

Identification of an antilaminin-1 scFv that preferentially homes to vascular solid tumors.

The tumor vasculature and extracellular matrix make attractive targets for distinguishing solid tumors from normal cells. In solid tumors, the processes of angiogenesis and metastasis potentially give rise to unique epitopes not usually accessible in homeostatic organs. Specific targeting of solid tumors for radioimmunotherapy requires that the targeting agent accumulate rapidly and at high levels at the tumor site. This study involved the selection of scFvs that recognize laminin-1 in vitro from the Tomlinson I and J phage display libraries. Selected, purified scFvs were radioiodinated and injected in tumor-bearing mice. One of these, scFv 15-9, exhibited preferential accumulation at subcutaneous tumors when compared to other antilaminin scFvs or to a control scFv. Autoradiographic analysis indicated that scFv15- 9 also displayed a higher vessel:parenchyma ratio than did two other antilaminin scFvs, scFv 15-6 and scFv 15-1, indicating a preferential accumulation of scFv 15-9 around vessel structures. Immunohistochemistry confirmed that scFv 15-9 accumulated at sites of endothelial cells lining vessel structures where significant levels of laminin were present. These data demonstrate that scFv 15-9 binds to a specific epitope on laminin and has potential for tumor endoradiotherapy in subcutaneous tumors.

A diabody that dissociates to monomer forms at low concentration: effects on binding activity and tumor targeting.

A human scFv, 15-9, was selected from a phage display library for binding to murine laminin-1. A diabody was made from the scFv by shortening the linker from 15 to 5 amino acids between the VH and VL sequence. Radioiodinated scFv and diabody were analyzed for size, binding to laminin, and biodistribution in tumor bearing mice. Diabody preparations at concentrations greater than 10 nM were largely dimer forms (approximately 60 kDa) as judged by gel filtration, but diluted diabody was eluted as a monomer (approximately 30 kDa). At low concentrations the radiolabeled diabody did not bind well to laminin. The (125)I diabody had significantly lower accumulation in tumors than did the scFv when injected at lower concentrations. These data indicate that the diabody dimer dissociates at concentrations of about 10nM resulting in monomers with no binding activity for laminin and poor tumor homing properties.

Genetic analysis of Pseudomonas aeruginosa by enterobacterial repetitive intergenic consensus polymerase chain reaction (PCR) and arbitrarily primed PCR: gel analysis compared with microchip gel electrophoresis.

OBJECTIVES: To assess the applicability of a newly emerging microchip gel electrophoresis for rapid strain differentiation among clinical isolates of Pseudomonas aeruginosa, and to compare this technique with the traditional gel method for DNA separation. METHODS: One hundred clinical strains of P. aeruginosa obtained from a hospital in northwestern Ohio were tested for reactivity to 3 serotype-specific monoclonal antibodies by enzyme-linked immunosorbent assay. Twelve strains (4 from each serogroup) were selected for DNA analysis by polymerase chain reaction (PCR)-based, single primer DNA fingerprinting methods with 3 different primers: 1 enterobacterial repetitive intergenic consensus PCR and 2 arbitrarily primed PCRs. The PCR products were analyzed by agarose slab gel and microchip gel electrophoresis. RESULTS: Of the 100 clinical isolates tested, 39% (4%, 14%, and 21%) were found to be serotypes 0:3, 0:6, and 0:11, respectively. Twelve strains were chosen for DNA analysis by PCR. The PCR products were analyzed by agarose slab gel electrophoresis and on microchips to determine interspecies diversity. Both methods demonstrated that different serotypes exhibited different electrophoretic patterns. Two strains (clinical strains 6 and 7, serotype 0:6) showed identical patterns, indicating a high degree of relatedness. CONCLUSION: In all cases, there was concordance between the electrophoretic patterns detected by the two methods. The capability of conducting both PCR and microchip gel electrophoresis offers an opportunity for an automated and rapid method for genetic analysis and differentiation among strains of P. aeruginosa and other microorganisms.

A monoclonal antibody to a carbohydrate epitope expressed on glycolipid and on alpha3beta1 integrin on human esophageal carcinoma.

A mouse monoclonal antibody (MAb-9) produced by immunization with a human esophageal carcinoma cell line, TE-2 (derived from undifferentiated squamous cell carcinoma) reacted specifically with about 30% of esophageal carcinoma cell lines and tissue sections from clinical samples. MAb-9 showed minimal reactivity with normal esophageal tissue. (125)I, fluorescent or gold particle labeled MAb-9 bound to TE-2 cell surfaces. (125)I-radiolabeled MAb-9 was used to detect reactive material from cell extracts in Western blot. Treatment of TE-2 membrane proteins with neuraminidase, N-glycanase or O-glycanase reduced antigen detection. Treatment of cells with periodic acid destroyed antibody binding in ELISA. Lipid extracts from cell membranes, containing glycolipids, also reacted with MAb-9. MAb-9 was used to purify target antigen from detergent solubilized membrane proteins and the prominent bands from subsequent gel electrophoresis were trypsin digested and analyzed by mass spectrometry. Peptides from alpha(3) and beta(1) integrin chains were identified. These data indicate that alpha3beta1integrin is prominently expressed on certain esophageal carcinomas and that a specific carbohydrate unit is selectively displayed on the alpha(3) integrin subunit as well as on glycolipid on the cell surface. The alpha3beta1 integrin expressed on A-431 carcinoma cells does not display this carbohydrate epitope and is not detected by MAb-9. Thus, expression of the carbohydrate epitope is the basis for the tumor selective reaction of MAb-9 with a subset of esophageal carcinomas.