The structure of human lymphotoxin (tumor necrosis factor-beta) at 1.9-A resolution.

The three-dimensional structure of recombinant human lymphotoxin (residues 24-171 of the mature protein) has been determined by x-ray crystallography at 1.9-A resolution (Rcryst = 0.215 for I greater than 3 sigma (I)). Phases were derived by molecular replacement using tumor necrosis factor (TNF-alpha) as a search model. Like TNF-alpha, lymphotoxin (LT) folds to form a "jellyroll" beta-sheet sandwich. Three-fold related LT subunits form a trimer stabilized primarily by hydrophobic interactions. A cluster of 6 basic residues around the 3-fold axis may account for the acid lability of the trimer. Although the structural cores of TNF-alpha and LT are similar, insertions and deletions relative to TNF-alpha occur in loops at the "top" of the LT trimer and significantly alter the local structure and the overall shape trimer is highly conserved. The sites of two mutations (Asp-50 and Tyr-108) that abolish the cytotoxicity of LT are contained within poorly ordered loops of polypeptide chain that flank the cleft between neighboring subunits at the base of the molecule, suggesting that the receptor recognizes an intersubunit binding site.

X-ray structure of human relaxin at 1.5 A. Comparison to insulin and implications for receptor binding determinants.

The X-ray crystal structure of relaxin at 1.5 A resolution is reported for the physiologically active form of the human hormone. Relaxin is a small, two-chain polypeptide that is a member of the protein hormone family that also includes insulin and the insulin-like growth factors IGF-I and IGF-II. These hormones have biologically diverse activities but are structurally similar, sharing a distinctive pattern of cysteine and glycine residues. The predicted structural homology of relaxin to insulin is confirmed by this structural analysis; however, there are significant differences in the terminal regions of the b-chain. Although relaxin, like insulin, crystallizes as a dimer, the orientation of the molecules in the respective dimers is completely different. The region of the relaxin molecule proposed to be involved in receptor binding is part of the dimer interface, suggesting that some of the other residues contained in the dimer contact surface might be receptor binding determinants as well. The proposed receptor binding determinants for insulin likewise include residues at its dimer interface. However, because the dimer contacts of relaxin and insulin are quite different, it appears that these two structurally related hormones have evolved somewhat dissimilar mechanisms for receptor binding.

Self-association of human and porcine relaxin as assessed by analytical ultracentrifugation and circular dichroism.

The self-association properties of recombinant DNA derived human relaxin, and porcine relaxin isolated from porcine ovaries, have been studied by sedimentation equilibrium analytical ultracentrifugation and circular dichroism (CD). The human relaxin ultracentrifuge data were adequately defined by a monomer-dimer self-association model with an association constant of approximately 6 x 10(5) M-1, whereas porcine relaxin was essentially monomeric in solution. An approximate 5-fold increase in weight fraction of human relaxin monomer elicited by dilution of the protein resulted in no change in the far-UV CD spectrum at 220 nm. In contrast, after the same increase in weight fraction of monomer, the near-UV circular dichroism spectra for human relaxin exhibited a significant decrease in the amplitude for the CD bands near 277 and 284 nm. These CD bands, which may be assigned to the lone tyrosine in human relaxin, are superimposed on a broad envelope that is probably due to the three disulfide chromophores. Although both the human and porcine proteins contain two tryptophan residues, the near-UV CD spectra exhibit only a broad shoulder near 295 nm rather than the strong CD bands often found for tryptophan. Moreover, there is little change in this broad band after dilution of human relaxin to concentrations that resulted in a 4-fold increase in monomer weight fraction. These data suggest that dissociation of the human relaxin dimer to monomer is not accompanied by large overall changes in secondary structure or alteration in the average tryptophan environment, whereas there is a significant change in the tyrosine environment.(ABSTRACT TRUNCATED AT 250 WORDS)

Use of recombinant DNA derived human relaxin to probe the structure of the native protein.

This report describes the physical, chemical, and biological characterization of recombinant human relaxin (rhRlx) used as a probe to establish the disulfide pairing in native human relaxin. This strategy is necessary since native human relaxin is only available in the nanogram range. The relaxin molecule is composed of two nonidentical peptide chains, an A-chain 24 amino acids in length and a B-chain of 29 amino acids, linked by two disulfide bridges with an additional disulfide linkage in the A-chain. Native relaxin isolated from human corpora lutea was compared to rhRlx by reversed-phase chromatography, partial sequence analysis, mass spectroscopy, and bioassay. The potency of rhRlx was established by its ability to stimulate cAMP from primary human uterine endometrial cells. Native relaxin isolated from human corpora lutea was equipotent to chemically synthesized relaxin, which in turn was equipotent to rhRlx. A tryptic map was developed for rhRlx to confirm the complete amino acid sequence and assignment of the disulfide bonds. The three disulfide bonds (CysA10-CysA15, CysA11-CysB11, and CysA24-CysB23) were assigned by mass spectrometric analysis of the tryptic peptides and by comparison to chemically synthesized peptides disulfide linked in the two most probable configurations. In addition, the observed amino acid composition and sequence of rhRlx was in agreement with that predicted from the cDNA sequence with the exception that the A-chain amino terminal was pyroglutamic acid. The migration of rhRlx upon sodium dodecyl sulfate-polyacrylamide gel electrophoresis was consistent with a monomeric structure, and the identity of the band was demonstrated by immunoblotting.

Structural characterization by mass spectrometry of native and recombinant human relaxin.

Mass spectrometry has played a key role in characterizing the primary structure of native and recombinant relaxin, a peptide hormone that induces ripening of the cervix prior to childbirth. The peptide is composed of two chains, A and B, and is formed from a single-chain prohormone, as is insulin. Aside from conserved cysteines, though, it has little sequence homology with insulin. Due to the small amounts of native peptide initially available (less than 10 pmol), traditional techniques could not provide information on the blocked A-chain sequence, on the carboxyterminal sequences, nor on other possible post-translational modifications. Mass measurements by fast atom bombardment (FAB) were made on reduced human relaxin isolated from corpora lutea. The detection limit by FAB for reduced relaxin was 500 fmol. The B-chain was four amino acids shorter than expected from comparison of the previously known cDNA sequence with homologous rat and porcine sequences. The A-chain, as predicted, was 24 amino acids in length and had a pyroglutamic acid residue on the amino-terminus. The purified samples were homogeneous with no other post-translational modifications. The recombinant relaxin molecule was also extensively characterized by mass spectrometry. In addition to the intact molecule, all tryptic peptides were characterized by FAB. A capillary high-performance liquid chromatography continuous-flow FAB system, developed for high-sensitivity peptide mapping, aided in these analyses. Finally, the three disulfide bonds were shown by tandem mass spectrometry to match those of insulin.

Sequence-specific 1H-NMR assignments and identification of two small antiparallel beta-sheets in the solution structure of recombinant human transforming growth factor alpha.

Transforming growth factor alpha (TGF alpha) is a small mitogenic protein with about 35% sequence identity with epidermal growth factor (EGF). TGF alpha-like proteins have been proposed to play a role in oncogenesis and wound healing. This report describes sequence-specific 1H-NMR resonance assignments for recombinant human TGF alpha (hTGF alpha). These assignments provide the basis for interpreting NMR data which demonstrate that the solution structure of hTGF alpha includes an antiparallel beta-sheet involving residues Gly-19 to Leu-24 and Lys-29 to Cys-34 and a second, smaller, antiparallel beta-sheet involving residues Tyr-38 and Val-39 and His-45 and Ala-46. These data, together with constraints imposed by the disulfide bonds, are combined to construct a molecular model of the polypeptide chain fold for residues Cys-8 to Ala-46. The resulting structure is similar to that of mouse and human EGF. Human TGF alpha and mouse EGF, however, differ with respect to their structural dynamics, since amide proton/deuteron exchange is much faster for hTGF alpha than for mouse EGF at pH 3.5.

The antichlamydial, antiviral, and antiproliferative activities of human gamma interferon are dependent on the integrity of the C terminus of the interferon molecule.

The effects of recombinant human gamma interferon (rHuIFN-gamma; two identical monomers of 140 residues in length) and of two re-engineered C-terminal variants, rHuIFN-gamma Tetra-Ser (residues 129 to 132 replaced by serine) and rHuIFN-gamma 125 (two identical monomers of 125 residues each with the last 14 residues plus an additional alanine from the C terminus deleted), were compared in terms of several in vitro biological activities. By using three different human cell lines (HeLa 229, HEp-2, and A549), the interferons were tested for their ability to inhibit: (i) growth of Chlamydia trachomatis; (ii) replication of encephalomyocarditis virus; and (iii) cell growth. rHuIFN-gamma restricted the growth of chlamydiae to 50% of the non-IFN-treated control at concentrations ranging from 0.01 to 0.05 ng/ml, depending on the cell type assayed. One of the modified proteins, rHuIFN-gamma Tetra-Ser, also decreased the growth of chlamydiae, but it required a concentration of approximately 0.5 ng/ml to produce 50% inhibition. rHuIFN-gamma 125 had the lowest antichlamydial activity of the three IFN-gamma variants tested; concentrations of 1 to 20 ng/ml were needed to reduce the growth of C. trachomatis to 50% of that of the control. The relative antiviral and antiproliferative activities of the three IFN-gamma preparations paralleled their antichlamydial activities in these three cell lines. The antiencephalomyocarditis virus activities of rHuIFN-gamma Tetra-Ser and rHuIFN-gamma 125 were reduced by approximately 10-fold and 10(2)- to 10(3)-fold, respectively, compared with the antiviral activity of rHuIFN-gamma. Proliferation of the three cell lines was restricted to approximately 50% of the control with 0.5 to 10 ng of rHuIFN-gamma per ml. Inhibition of cell growth by rHuIFN-gamma Tetra-Ser was significant only at concentrations equal to or greater than 30 ng/ml, and the rHuIFN-gamma 125 variant did not significantly decrease the growth of any of the three cell lines at the concentrations tested. These results suggest that the C-terminal portion of rHuIFN-gamma is critical for maintaining the conformation necessary for inducing the antichlamydial, antiviral, and antiproliferative activities of the molecule.

Activation of human polymorphonuclear neutrophil functions by interferon-gamma and tumor necrosis factors.

Recombinant human interferon-gamma (rHuIFN-gamma) and natural human tumor necrosis factor beta (nHuTNF-beta) (previously called lymphotoxin), purified to homogeneity, were used to assess their effects on certain functions of human polymorphonuclear neutrophils (PMN) in vitro. The treatment of PMN with 100 U of either rHuIFN-gamma or nHuTNF-beta for 20 min significantly increased their ability to phagocytize 1.5-microns latex beads as detected by flow cytometry. Preparations of recombinant human TNF-beta (rHuTNF-beta) showed activities similar to those of its natural counterpart in activating phagocytosis. In addition, a significant enhancement in PMN-mediated antibody-dependent cellular cytotoxicity was observed after treatment for 2 hr with IFN gamma and both TNF-alpha and TNF-beta. The enhancement by treatment with a combination of rHuIFN-gamma and nHuTNF-beta exceeded the enhancement caused by either agent alone. We also show that although lipopolysaccharide (LPS) is a potent stimulator of PMN function, polymyxin B can block LPS-induced but not lymphokine-induced activation. These data demonstrate new activities for both TNF-alpha and TNF-beta in augmenting the phagocytic and cytotoxic activities of PMN.

The synergistic anti-proliferative effect of gamma-interferon and human lymphotoxin.

Highly purified preparations of rHuIFN-gamma and HLT synergistically inhibit the in vitro proliferation of WI38VA13 cells, a SV40 virus-transformed derivative of the human cell line W138. The growth of nontumorigenic WI38 cells, established from normal diploid embryonic lung tissue, was not inhibited under identical incubation conditions; rMuIFN-gamma and HLT have the same in vitro effects on the murine melanoma B16. Thus, the combination of rMuIFN-gamma and HLT synergistically inhibits the growth of B16 cells, although incubation with either lymphokine alone has no significant anti-proliferative effect. Flow cytometric analysis of the cell cycle of asynchronous cells indicates that when B16 melanoma cells are treated with rMuIFN-gamma and HLT in combination, more cells accumulate in the G0/G1 phase than when the cells are treated with rMuIFN-gamma alone. It appears that populations of rapidly proliferating tumor cells, which normally contain a large percentage in the S+G2+M phases of the cell cycle, become arrested in the quiescent G0/G1 phase by the combination of IFN-gamma and HLT, resulting in the inhibition of cell growth.

Evidence for a gamma-interferon receptor that regulates macrophage tumoricidal activity.

Gamma-interferon (IFN-gamma) is the macrophage-activating factor (MAF) produced by normal murine splenic cells and the murine T cell hybridoma 24/G1 that induces nonspecific tumoricidal activity in macrophages. Incubation of 24/G1 supernatants diluted to 8.3 IRU IFN-gamma/ml with 6 X 10(6) elicited peritoneal macrophages or bone marrow-derived macrophages for 4 h at 37 degrees C, resulted in removal of 80% of the MAF activity from the lymphokine preparation. Loss of activity appeared to result from absorption and not consumption because (a) 40% of the activity was removed after exposure to macrophage for 30 min at 4 degrees C, (b) no reduction of MAF activity was detected when the 24/G1 supernatant was incubated with macrophage culture supernatants, and (c) macrophage-treated supernatants showed a selective loss of MAF activity but not interleukin 2 (IL-2) activity. Absorption was dependent on the input of either IFN-gamma or macrophages and was time dependent at 37 degrees C but not at 4 degrees C. With four rodent species tested, absorption of murine IFN-gamma displayed species specificity. However, cultured human peripheral blood monocytes and the human histiocytic lymphoma cell line U937 were able to absorb the murine lymphokine. Although the majority of murine cell lines tested absorbed 24/G1 MAF activity, two murine macrophage cell lines, P388D1 and J774, were identified which absorbed significantly reduced amounts of natural IFN-gamma. Purified murine recombinant IFN-gamma was absorbed by elicited macrophages but not by P388D1. Normal macrophages but not P388D1 bound fluoresceinated microspheres coated with recombinant IFN-gamma and binding was inhibited by pretreatment of the normal cells with 24/G1 supernatants. Scatchard plot analysis showed that 12,000 molecules of soluble 125I-recombinant IFN-gamma bound per bone marrow macrophage with a Ka of 0.9 X 10(8) M-1. Binding was quantitatively inhibitable by natural IFN-gamma but not by murine IFN alpha. IFN-beta competed only weakly. Monoclonal antibodies against IFN-gamma either inhibited or enhanced MAF activity by blocking or increasing IFN-gamma binding to macrophages, respectively. These results indicate that IFN-gamma reacts with a receptor on macrophage in a specific and saturable manner and this interaction initiates macrophage activation.

Natural human interferon-gamma. Complete amino acid sequence and determination of sites of glycosylation.

Fresh human peripheral blood lymphocytes were induced with desacetylthymosin -alpha 1 and staphylococcal enterotoxin B. The induced gamma interferon (or IFN-gamma, immune interferon, type II interferon) was purified to homogeneity utilizing controlled-pore glass, concanavalin A-Sepharose, Bio-Gel P100, or Sephacryl S-200, and reversed phase high performance liquid chromatography. This procedure resulted in two active species with apparent Mr = 20,000 and 25,000 as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Both species were found to have identical amino acid sequences with a pyroglutamate residue as NH2-terminus. In both cases six different COOH termini were found. They are, at least qualitatively, identical in both species. There are two possible Asn-X-Ser/Thr glycosylation sites. Both carry carbohydrates in the Mr = 25,000 species whereas in the Mr = 20,000 species only one site is glycosylated. This likely explains the difference in apparent molecular weight between the two species and the expected molecular weight based upon the amino acid sequence.

Effects of bacteria-produced human alpha, beta, and gamma interferons on in vitro immune functions.

The effects of bacteria-produced human interferons (HuIFN) alpha, beta, and gamma on in vitro immune functions of human peripheral blood mononuclear cells (PBMC) were studied. Proliferative response to phytohemagglutinin was significantly inhibited by the addition of HuIFN-alpha 2 or HuIFN-beta at 10, 100, or 1000 U/ml. In contrast, HuIFN-gamma showed suppressive activities only when added at 1000 U/ml. HuIFN-alpha 2 or HuIFN-beta caused significant inhibition of human mixed-lymphocyte reaction (MLR) as measured by [3H]thymidine incorporation. Similar inhibition was caused by HuIFN-gamma when it was added only at very low concentrations (1 U/ml); 10, 100, or 1000 U/ml resulted in no or only a modest increase in MLR. All three interferons exhibited dose-related effects on PWM-induced immunoglobulin synthesis in cultures of PBMC. These data demonstrate that purified interferons produced by recombinant DNA technology can significantly alter in vitro immune functions and that HuIFN-gamma has properties which are different from those of HuIFN-alpha 2 or HuIFN-beta.

Interrelationships of human interferon-gamma with lymphotoxin and monocyte cytotoxin.

Crude preparations of interferon (IFN)-gamma derived from human peripheral blood leukocyte (PBL) cultures induced with 12-O-tetra-decanoylphorbol-13-acetate (TPA) and phytohemagglutinin (PHA) were more cytotoxic to HeLa cells than partially purified nautral or highly purified recombinant human IFN-gamma preparations. Conditioned media from PBL cultures contained, in addition to IFN-gamma, a mixture of cytotoxins, including classic lymphocyte-derived lymphotoxin (LT), and a TPA-induced cytotoxic activity produced by the adherent cell population (presumably monocytes). These two types of cytotoxins, indistinguishable in the mouse L929 cell LT assay, could be differentiated by an antiserum prepared against LT derived from the B lymphoblastoid cell line RPMI 1788. This antiserum neutralized lymphocyte-derived classic LT but failed to neutralize the activity of the monocyte-derived cytotoxin. Processing of conditioned media by sequential chromatography on silicic acid, Con A-Sepharose, and DEAE-Sephacel failed to separate IFN-gamma from the LT activity. However, this procedure did remove the monocyte-derived cytotoxic activity present in the original starting material, leaving predominantly classic LT. This LT showed a slightly basic isoelectric point (pI 7.6) which partially overlapped the more basic pI range of IFN-gamma. The two lymphokine activities also could not be completely separated by fast protein liquid chromatography or molecular sieve chromatography. LT in these partially purified preparations was associated with a protein having an apparent molecular weight of 58,000 on gel filtration. This form dissociated partially into a 20,000 mol wt species after denaturation with 0.1% NaDodSO4. IFN-gamma could be selectively removed from preparations containing both IFN-gamma and LT with the aid of monoclonal antibody to IFN-gamma. The addition of purified LT to purified E. coli-derived recombinant human IFN-gamma resulted in a marked synergistic enhancement of cytotoxicity for HeLa cells.

Biological and antigenic similarities of murine interferon-gamma and macrophage-activating factor.

Murine peritoneal exudate cells (PEC) treated with murine recombinant interferon-gamma (IFN-gamma) (greater than 99% estimated purity), or concanavalin A-stimulated spleen cell supernatants developed tumoricidal properties (macrophage activation factor [MAF] activity). MAF activity was found to occur with treatments of 10 U/ml IFN-gamma, and at levels as low as 1 U/ml IFN-gamma if a second signal (5 ng/ml endotoxin) was present in the MAF assay. Endotoxin (lipopolysaccharide [LPS]) alone at these levels failed to induce MAF; induction of MAF was observed at 1,000-fold greater levels. The ability of IFN-gamma to stimulate murine PEC was species specific. Various sources of materials that displayed MAF activity, including supernatants from interleukin 2-dependent cloned cytotoxic murine T lymphocyte lines that did not display detectable antiviral activity, were neutralized by antibody raised and affinity purified against recombinant IFN-gamma. Thus, IFN-gamma, although never detectable by antiviral assays, appears to be present in many lymphokine preparations and has potent macrophage activation capability.

Importance of treatment regimen of interferon as an antitumor agent.

A highly purified hybrid human leukocyte interferon, IFN-alpha AD, produced in Escherichia coli, has been used to define optimum treatment conditions for L1210 leukemia in mice. Treatments prior to tumor inoculation were ineffective. Treatments from the third day post-tumor inoculation were most effective, and treatments every third day were more effective than were regimens involving more frequent treatments. IFN-alpha AD was effective in vivo against tumors formed from a line of L1210 cells resistant to IFN-alpha AD in cell cultures. These and other results indicate the importance and nature of indirect mechanisms of action for efficacy of interferons against tumors.

Antiviral effects of bacteria-derived human leukocyte interferons against encephalomyocarditis virus infection of squirrel monkeys. Brief report.

Two human leukocyte interferon sub-types (IFN-alpha A and -alpha D) produced in E. coli and a hybrid interferon (IFN-alpha AD[Bgl]) consisting of the N-terminal 61 residues of IFN-alpha A and the C-terminal 104 residues of IFN-alpha D were compared for antiviral activity against EMC virus infection of squirrel monkeys. Marked reduction in viremia and a decrease in the incidence of deaths occurred with highly purified preparations of the recombinant-DNA derived interferons. Dose response studies showed that IFN-alpha D and -alpha AD (Bgl) were more effective than IFN-alpha A.

Antiviral activity of bacteria-derived human alpha interferons against encephalomyocarditis virus infection of mice.

Bacteria-derived human leukocyte interferon (IFN) subtypes, IFN-alpha A, -alpha B, and -alpha D, and two hybrid IFNs, IFN-alpha AD and -alpha DA, were examined for both in vitro and in vivo antiviral activity. Two of these materials in highly purified form (IFN-alpha D and -alpha D) protect mice against lethal doses of encephalomyocarditis virus infection. A single dose of 1 microgram of protein of IFN-alpha D 3 h before infection conferred protection in both BDF1 and CD-1 mice against encephalomyocarditis virus infection, and multiple treatments with IFN-alpha D or IFN-alpha AD extend the mean survival time of infected mice. On a weight basis, IFN-alpha AD was approximately 100-fold more effective than IFN-alpha D. There is a direct correlation between the antiviral activity of the various human IFN species in L-929 cells and in mice for both single and multiple treatments before infection, but none of the cloned human IFN subtypes were effective when administered 24 h after infection. Mixtures of the two parental materials, IFN-alpha A and -alpha D, were not as protective as the hybrid molecule IFN-alpha AD, suggesting that IFNs with unique and altered species specificity can be produced by recombinant DNA methods.

Somatomedin-C shares the carboxy-terminal antigenic determinants with insulin-like growth factor-I.

Insulin-like growth factor I is a human plasma peptide with strong homology to human proinsulin. Many lines of evidence suggest that somatomedin-C and insulin-like growth factor I are identical molecules. However, no structural data regarding the carboxy-terminal end of somatomedin-C has yet been produced. We report a region-specific immunoassay directed against the carboxy-terminal end of insulin-like growth factor I. Somatomedin-C and insulin-like growth factor I are indistinguishable in this region-specific immunoassay. Insulin-like activity II and multiplication stimulating activity show partial cross reaction in this immunoassay. This study provides the first evidence about the carboxy-terminal portion of somatomedin-C and re-emphasizes that somatomedin-C and insulin-like growth factor I should be considered identical molecules.