Inhibiting protein-protein interactions: a model for antagonist design.

Protein-protein interactions (PPI) are a ubiquitous mode of transmitting signals in cells and tissues. We are testing a stepwise, generic, structure-driven approach for finding low molecular weight inhibitors of protein-protein interactions. The approach requires development of a high-affinity, single chain antibody directed specifically against the interaction surface of one of the proteins to obtain structural information on the interface. To this end, we developed a single chain antibody (sc1E3) against hIL-1beta that exhibited the equivalent affinity of the soluble IL-1 receptor type I (sIL-1R) for hIL-1beta and competitively blocked the sIL-1R from binding to the cytokine. The antibody proved to be more specific for hIL-1beta than the sIL-1R in that it failed to bind to either murine IL-1beta or human/murine IL-1alpha proteins. Additionally, failure of sc1E3 to bind to several hIL-1beta mutant proteins, altered at receptor site B, indicated that the antibody interacted preferentially with this site. This, coupled with other surface plasmon resonance and isothermal titration calorimetry measurements, shows that sc1E3 can achieve comparable affinity of binding hIL-1beta as the receptor through interactions at a smaller interface. This stable single chain antibody based heterodimer has simplified the complexity of the IL-1/IL-1R PPI system and will facilitate the design of the low molecular weight inhibitors of this interaction.

Determining selectivity of drugs by quantitative two-dimensional gel analysis. A study of tenidap, piroxicam, and dexamethasone.

In vitro pharmacologic measures of drug specificity are well established, i.e. drug interaction with a specific target such as an enzyme, receptor, or ion channel. However, in vitro measures of drug selectivity, defined as effects on secondary targets, are lacking. Two-dimensional gel electrophoresis (2-D gel) was examined as a measure of drug selectivity by comparing the effects of three drugs, tenidap, piroxicam, and dexamethasone, on the synthesis of intracellular proteins in lipopolysaccharide (LPS)-stimulated murine macrophages. A set of 902 35S-methionine-labeled proteins were separated consistently, identified by their coordinates of apparent isoelectric point and molecular weight, and quantified. LPS altered the concentrations of 45 proteins. Tenidap, at 10 microM, affected a total of five proteins (suppressed three; stimulated two), whereas piroxicam, at 10 microM, suppressed two proteins. Dexamethasone at 0.01 microM suppressed eight proteins and stimulated one. Thus, none of the drugs reversed the LPS-induced changes. Two of the eight proteins suppressed by dexamethasone were also suppressed by tenidap and were identified as proIL-1 alpha and proIL-1 beta. Since the subset of affected proteins provided a unique protein "fingerprint" for each drug, the three drugs were mechanistically differentiated by 2-D gel analysis. Compared to LPS (5% affected proteins), all three drugs were selective (< or = 1% affected) with piroxicam > tenidap > dexamethasone. With identification of affected proteins, this technique can provide a useful in vitro assessment of drug selectivity.

Purification and preliminary crystallographic studies of penicillin G acylase from Providencia rettgeri.

Two isoforms of the heterodimeric enzyme penicillin G acylase (EC 3.5.1.11) from Providencia rettgeri ATCC 31052 (strain Bro1) were purified to near homogeneity. The isoforms exhibited comparable enzymatic activities but differed slightly in the molecular weight and pI of their respective alpha-subunit. The origin of this difference was traced to the partial conversion of the N-terminal Gln of the alpha-subunit to pyrrolidonecarboxylic acid (pyro-Glu). The boundaries of the mature enzyme within the translated DNA sequence of the wild-type propeptide (GenBank M86533) were determined. The results conclusively identified the length of the signal peptide and the position of the spacer cleaved from the propeptide to form the active heterodimer. The molecular weights of the alpha- and beta-subunits, based on these termini, were 23.7 and 62.2 kDa, respectively. Both isoforms were crystallized independently as hexagonal bipyramids up to 0.60 mm in diameter in either space group P6(1)22 or P6(5)22 (a = b = 140.5 A and c = 209.5 A) from ammonium sulfate solutions buffered by 50 mM potassium phosphate at pH 7.5. The presence of glycerol, although not required, facilitated crystal growth. Native and heavy atom derivative data were collected to 3.0 A resolution, and the calculation of isomorphous replacement phases is under way.

Site-directed double fluorescent tagging of human renin and collagenase (MMP-1) substrate peptides using the periodate oxidation of N-terminal serine. An apparently general strategy for provision of energy-transfer substrates for proteases.

Periodate in neutral aqueous solution rapidly converts N-terminal Ser or Thr to an alpha-N-glyoxylyl moiety that can serve as the locus for incorporation of a modifying group [Geoghegan, K. F., and Stroh, J. G. (1992) Bioconjugate Chem. 3, 138-146. Gaertner, H. F. et al. (1992) Bioconjugate Chem. 3, 262-268]. The usefulness of this procedure has been further illuminated in a route to "energy-transfer" substrates for endoproteases. Each such substrate is an oligopeptide cleavable by a proteinase, but modified (usually at its termini) with two chromophores that form an energy donor-acceptor pair. Production of these substrates is an exercise in double site-directed peptide modification. The new route is composed of three steps, beginning from an unprotected peptide in which a sequence recognized by the pertinent enzyme is placed between N-terminal Ser and C-terminal Lys. Lys may not occur elsewhere in the peptide. Periodate oxidation converts the N-terminal Ser to an alpha-N-glyoxylyl group, which is then allowed to form a hydrazone with the carbohydrazide derivative Lucifer Yellow CH, a hydrophilic fluor with a large Stokes shift (excitation maximum, 425 nm; emission maximum, 525 nm). Finally, the modified peptide is allowed to react with 5-carboxytetramethylrhodamine succinimidyl ester. This reaction selectively modifies the epsilon-amino group of C-terminal Lys, the only amino group remaining in the peptide. 5-Carboxytetramethylrhodamine strongly (> 90%) quenches Lucifer Yellow fluorescence by resonance energy transfer in the intact substrate, but enzyme-catalyzed cleavage eliminates the quenching. The resulting increase in fluorescence may be used to follow the hydrolytic reaction.(ABSTRACT TRUNCATED AT 250 WORDS)

Secondary structural analysis of two recombinant murine proteins, interleukins 1 alpha and 1 beta: is infrared spectroscopy sufficient to assign structure?

The secondary structure for two murine recombinant proteins, interleukins 1 alpha and 1 beta (rmIL-1 alpha and -1 beta), has been analyzed by Fourier transform infrared (IR) spectroscopy and then compared to results obtained by X-ray diffraction, circular dichroism (CD), and nuclear magnetic resonance (NMR) spectroscopy. The IR results obtained here for rmIL-1 alpha and -1 beta suggested that their secondary structures consisted predominantly of beta-sheets or strands. However, the analysis also revealed a significant absorption band near 1656 cm-1, which is typically assigned to alpha-helical or random structures. When these same murine polypeptides were analyzed by CD, no evidence of alpha-helical structures was observed. Further, published X-ray diffraction and NMR studies characterizing the human forms of IL-1 alpha and -1 beta indicate the absence of alpha-helices and that the human proteins are composed mainly of beta-strands (i.e., greater than 55%), with approximately 24% of the amino acids involved in large loops connecting the strands. The murine IL-1 proteins, when compared to their respective human counterparts, each show greater than 80% sequence homology. Given this fact, the CD analyses, and the result that this IR band amounted to 21% of the overall integrated area, the absorption peak at 1656 cm-1 was attributed to the presence of large loops rather than to alpha-helical or random structures. Such a structural assignment appears reasonable and is totally consistent with the established existence of large loops in the human forms as well as in other proteins found to fold similarly (viz., human bFGF).(ABSTRACT TRUNCATED AT 250 WORDS)

Role of prostaglandins in the behavioral changes induced by murine interleukin 1 alpha in the rat.

Continuous infusion of murine recombinant interleukin 1 alpha (rIL-1 alpha) produces weight loss, appetite suppression, reduction in horizontal locomotor activity (crossovers) and vertical locomotor activity (rears), and an increase in drinking behavior in the rat. The role of prostaglandins (PG) in the elicitation of these effects was studied. Infusion of rIL-1 alpha produced a transient increase in serum (PGs) which peaked at 24 to 48 h. This increase was completely inhibited by piroxicam. However, inhibition of circulating PG by piroxicam did not block the reductions in appetite, crossover, and rears induced by rIL-1 alpha; it restored normal drinking behavior and only partially restored body weight. Continuous intraperitoneal infusion of PGE2 at 24 micrograms/day exposed the animals to serum levels of PGE2 comparable to those produced by infusion with rIL-1 alpha. Yet, at the point of maximum weight loss induced by rIL-1 alpha (72 h), PGE2 infusion resulted in only a quarter of the weight loss. Compared with rIL-1 alpha, continuously infused PGE2 produced significantly smaller reductions in appetite, crossovers, and rears, and had no effect on drinking behavior. From these observations, we conclude that the rIL-1 alpha-induced increase in drinking behavior was fully dependent on products of the cyclooxygenase pathway, but not necessarily PGE2. However, because of the failure of piroxicam to fully reverse rIL-1 alpha effects on eating, mobility, and weight loss, there must also be a significant PG-independent component to account for the full range of rIL-1 alpha effects.

Reduction of biological activity of murine recombinant interleukin-1 beta by selective deamidation at asparagine-149.

A biologically active preparation of murine recombinant interleukin-1 beta (mIL-1 beta) from Escherichia coli cell lysates contained tow forms of mIL-1 beta with pI 8.7 and pI 8.1, respectively. Treatment with 0.1 M Tris, pH 8.5, at 37 degrees C for 35 h converted the pI 8.7 form to the pI 8.1 form by the selective deamidation of an asparagine residue (Asn149) in the mIL-1 beta molecule. Deamidated mIL-1 beta had 3- to 5-fold lower co-mitogenic activity and receptor affinity than the unmodified form.

Isolation and characterization of biologically active murine interleukin-1 alpha derived from expression of a synthetic gene in Escherichia coli.

A murine interleukin-1 alpha (mIL-1 alpha) gene coding for amino acids 115 to 270 of the precursor protein (Lomedico, P.T., Gubler, U., Hellmann, C.P., Dukovich, M., Giri, J.G., Pan, Y.E., Collier, K., Semionow, R., Chua, A.O. and Mizel, S.B. (1984) Nature 312, 458-462) was chemically synthesized and expressed in Escherichia coli. mIL-1 alpha, in the form of insoluble inclusion bodies, accounted for approx. 30% of total cellular protein produced by the recombinant strain. A simple isolation protocol was developed in which inclusion body material was first solubilized in 3 M guanidine hydrochloride, and the mIL-1 alpha was then simultaneously purified and allowed to fold to its active conformation by dialysis against distilled water. This procedure yielded pure, biologically active mIL-1 alpha with 41% recovery of the mIL-1 alpha present in the guanidine hydrochloride extract. The purified preparation had the expected amino acid composition, a molar absorptivity of 28,200 M-1.cm-1 and a pI of 5.2. No methionyl-mIL-1 alpha was detected by N-terminal sequence analysis, and the endotoxin level was less than 10 pg per micrograms of mIL-1 alpha. The specific biological activity was 3.10(7) units/mg in a co-mitogenic thymocyte proliferation assay. In addition to full-length mIL-1 alpha, the preparation contained N-terminally truncated mIL-1 alpha species (mainly des-4 and des-6 amino acid forms). The truncated species were isolated and found to have the same biological activity as the complete polypeptide. Thus, the active fragment of mIL-1 alpha appears to consist of a proteinase-sensitive N-terminal region which is not essential for activity, and a proteinase-resistant core which harbors the essential determinants of its cytokine function.

Effects of continuously administered murine interleukin-1 alpha: tolerance development and granuloma formation.

Continuous infusion of murine recombinant interleukin-1 alpha (rIL-1 alpha) into rats by using intraperitoneally implanted osmotic pumps led to marked decreases in body weight, liver enzymes (serum glutamic oxalacetic transaminase, serum glutamic pyruvic transaminase, and sorbitol dehydrogenase), appetite, and mobility and increases in drinking, blood urea nitrogen, and total peripheral blood leukocytes within 3 days. Granuloma formation was found in the local area of rIL-1 alpha release. As early as day 3, a focal infiltrate of polymorphonuclear leukocytes, mononuclear leukocytes, and plasma cells filled the area; by day 6, extensive fibrosis was found. A loss of rIL-1 alpha-induced changes, with the exception of granuloma formation, occurred by day 10. A marked decrease in the response to rIL-1 alpha was also observed when animals were challenged by implantation of new pumps containing rIL-1 alpha, with monitoring of body weight, or by subcutaneous injection of rIL-1 alpha, with monitoring of serum colony-stimulating factor production. We propose that, even in the continuous presence of interleukin-1, replacement of the acute responses to interleukin-1 by restoration of more normal physiology may be advantageous upon acquisition of specific immunity.

The use of high field NMR to determine homogeneity in a preparation of human C5a produced by Escherichia coli.

The polypeptide backbone of human C5a was prepared by recombinant DNA techniques. Standard biochemical analysis guided the protein separation to give a sample of C5a which was deemed homogeneous. However, nuclear magnetic resonance (NMR) studies showed the material to be significantly heterogeneous. Reanalysis by high performance liquid chromatography (HPLC) corroborated the NMR results. Further separation by HPLC and analysis by NMR spectroscopy guided the isolation of rC5a to greater than 92% purity. NMR analysis, immunochemical and biological evaluation of the impurities showed them to be C5a structural variants. These results indicate that conventional methods of protein chemistry can fail to reveal heterogeneity in recombinant proteins, and in some circumstances NMR spectroscopy can aid in their purification.

The effects of continuous administration of murine interleukin-1 alpha in the rat.

Recombinant murine IL-1 alpha was administered continuously to rats by means of osmotic pumps implanted intraperitoneally. Continuous infusion of rIL-1 alpha in a range between 0.12 and 12.0 micrograms/day for four days was found to produce concentration-dependent weight loss. Behavioral parameters were continuously monitored and recorded at the 3.0 micrograms/day concentration in electronically-monitored activity cages during Days 2 through 5 of rIL-1 alpha administration. Parameters were separated into those affected during the dark phase (active period) or the light phase (resting period). Eating activity was found to be significantly reduced during each dark period through day 5, when compared with either untreated or PBS vehicle-infused animals. During the fourth and fifth days of infusion, however, eating behavior in animals infused with rIL-1 alpha began to increase toward control level in the latter, but not the earlier, half of the dark period. In contrast, drinking behavior was found to be significantly elevated only during the light periods. Continuous infusion of rIL-1 alpha also produced significant reductions in both horizontal locomotor activity (crossovers) and vertical locomotor activity (rears). However, in contrast to the trend toward a return of normal eating behavior, locomotor activity remained decreased through the fifth day of rIL-1 alpha infusion. These results suggest changes that could be produced by IL-1 in chronic inflammatory disease and infection.

Expression and regulation of the penicillin G acylase gene from Proteus rettgeri cloned in Escherichia coli.

The penicillin G acylase genes from the Proteus rettgeri wild type and from a hyperproducing mutant which is resistant to succinate repression were cloned in Escherichia coli K-12. Expression of both wild-type and mutant P. rettgeri acylase genes in E. coli K-12 was independent of orientation in the cloning vehicle and apparently resulted from recognition in E. coli of the P. rettgeri promoter sequences. The P. rettgeri acylase was secreted into the E. coli periplasmic space and was composed of subunits electrophoretically identical to those made in P. rettgeri. Expression of these genes in E. coli K-12 was not repressed by succinate as it is in P. rettgeri. Instead, expression of the enzymes was regulated by glucose catabolite repression.

Role of protein subunits in Proteus rettgeri penicillin G acylase.

Penicillin G acylase from Proteus rettgeri is an 80,000- to 90,000-dalton enzyme composed of two nonidentical subunits. Both subunits were required for enzymatic activity. The 65,000-dalton beta subunit contained a phenylmethylsulfonyl fluoride-sensitive residue required for enzymatic activity, and the 24,500-dalton alpha subunit contained the domain that imparts specificity for the penicillin side chain.

Experimental evolution of penicillin G acylases from Escherichia coli and Proteus rettgeri.

Proteus rettgeri and Escherichia coli W were shown to express structurally different penicillin G acylases. The enzymes had similar substrate specificity but differed in molecular weight, isoelectric point, and electrophoretic mobility in polyacrylamide gels and did not antigenically cross-react. When the organisms were subjected to environmental conditions which made expression of this enzyme essential for growth, spontaneous mutants were isolated that used different amides as the only source of nitrogen. These mutants acquired the ability to use amides for growth by deregulating the penicillin G acylase and by their evolution to novel substrate specificities. The enzymes expressed by mutants isolated from each genus appeared to have evolved in parallel since each acylase attained similar new substrate specificities when the organisms were subjected to identical selection pressure.

Repression of penicillin G acylase of Proteus rettgeri by tricarboxylic acid cycle intermediates.

The regulation of the penicillin acylase in proteus rettgeri ATCC 31052 was compared with that of the enzyme in Escherichia coli ATCC 9637. Unlike the E. coli acylase, the P. rettgeri enzyme was not induced by phenylacetic acid, nor was it subject to catabolite repression by glucose. The P. rettgeri acylase appears to be expressed constitutively but is subject to repression by the C4-dicarboxylic acids of the tricarboxylic acid cycle, succinate, fumarate, and malate.

Induction of 3-hydroxybenzoate 2-hydroxylase in a Pseudomonas testosteroni mutant.

Benzoate was established as the inducer of a unique 3-hydroxybenzoate 2-hydroxylase activity found in a Pseudomonas testosteroni mutant which is unable to grow on m-hydroxybenzoate as its sole source of carbon and energy.

Bioconversion of m-hydroxybenzoate to 2,3-dihydroxybenzoate by mutants of Pseudomonas testosteroni.

Mutans of Pseudomonas testosteroni were isolated for their inability to grow on m-hydroxybenzoate as sole carbon source. These mutants hydroxylated m-hydroxybenzoate for form 2,3-dihydroxybenzoate in high yeilds. The bioconversion described in this report represents the first reported example of 3-hydroxybenzoate 2-hydroxylase activity.

Involvement of threonine dehydratase in biosynthesis of the alpha-ketobutyrate prosthetic group of urocanase.

Seventeen mutants of Pseudomonas putida that were unable to grow on threonine as nitrogen source owing to a lack of threonine dehydratase were isolated, and all were found to be unable to synthesize active urocanase. Spontaneous revertants selected for urocanase production concomitantly regained threonine dehydratase. Mutants that were unable to utilize urocanate as carbon source were also isolated, and these were defective in urocanase formation but were normal in threonine dehydratase levels. Since alpha-ketobutyrate is the prosthetic group for urocanase, these results are consistent with the proposal that threonine dehydratase is necessary for urocanase prosthetic group biosynthesis. However, the lack of urocanase activity in threonine dehydratase-negative mutants was shown not to be the result of reduced levels of endogenous free alpha-ketobutyrate, nor to the participation of threonine dehydratase in the initiation of urocanase biosynthesis through the conversion of threonyl-tRNA(Thr) to alpha-ketobutyryl-tRNA(Thr). Other alternatives for the participation of threonine dehydratase in urocanase biosynthesis are discussed.