SumlnaryCD14 is a 55-kD protein found as a glycosylphosphatidylinositol (GPI)-anchored protein on the surface of monocytes, macrophages, and polymorphonuclear leukocytes, and as a soluble protein in the blood. Both forms of CD14 participate in the serum-dependent responses of cells to bacterial lipopolysaccharide (LPS). While CD14 has been described as a receptor for complexes of LPS with LPS-binding protein (LBP), there has been no direct evidence showing whether a ternary complex of LPS, LBP, and CD14 is formed, or whether CD14 binds LPS directly. Using nondenaturing polyacrylamide gel electrophoresis (native PAGE), we show that recombinant soluble CD14 (rsCD14) binds LPS in the absence of LBP or other proteins. Binding of LPS to CD14 is stable and of low stoichiometry (one or two molecules of LPS per rsCD14). Recombinant LBP (rLBP) does not form detectable ternary complexes with rsCD14 and LPS, but it does accelerate the binding of LPS to rsCD14, rLBP facilitates the interaction of LPS with rsCD14 at substoichiometric concentrations, suggesting that LBP functions catalytically, as a lipid transfer protein. Complexes of LPS and rsCD14 formed in the absence of LBP or other serum proteins strongly stimulate integrin function on PMN and expression of E-selectin on endothelial cells, demonstrating that LBP is not necessary for CD14-dependent stimulation of cells. These results suggest that CD14 acts as a soluble and cell surface receptor for LPS, and that LBP may function primarily to accelerate the binding of LPS to CD14. R cent work has described several serum and cell surface proteins that are necessary for responses of leukocytes to low concentrations of bacterial LPS (endotoxin) (1). LPSbinding protein (LBP),I an acute phase reactant, binds LPS (2) and greatly enhances the sensitivity of cells to LPS (3). Normal serum and plasma also enhance responses to LPS, and a multicomponent factor termed septin has been proposed to serve this function (4). LBP (5) and septin (4) each bind to LPS-coated particles and promote the interaction of these particles with CD14 (6, 7), a glycosylphosphatidylinositol (GPI)-anchored protein of monocytes, macrophages, and PMN (8, 9, 10). CD14 is necessary for serum-or LBPmediated responses of cells to LPS, such as the production 1 Abbreviations used in this paper: CHO, Chinese hamster ovary; ELPS, sheep erythrocytes coated with LPS; ELPS-LBP, ELPS opsonized with LBP; GPI, glycosylphosphatidylinositol; HAP, Dulbecco's PBS with 0.5 U/ml aprotinin, 0.05% human serum albumin, 3 mM D-glucose; HUVEC, human umbilical vein endothelial cells; LBP, LPS-binding protein; NHP, normal human plasma; PD-EDTA, Dulbecco's PBS lacking Ca 2+ and Mg 2+ with 1 mM EDTA; Ra, strain R60; Re, strain R595; rLBP, recombinant LBP; rsCD14, recombinant soluble CD14. of TNF by monocytes (6) and an increase in the adhesive properties of 132-integrins on PMN (7).Cells that do not express CD14, such as endothelial cells, also respond to low concentrations of LPS in the presence of serum. We have shown that these resp...

The p38 mitogen-activated protein kinases (MAPK) are activated by cellular stresses and play an important role in regulating gene expression. We have isolated a cDNA encoding a novel protein kinase that has significant homology (57% amino acid identity) to human p38␣/ CSBP. The novel kinase, p38␦, has a nucleotide sequence encoding a protein of 365 amino acids with a putative TGY dual phosphorylation motif. Dot-blot analysis of p38␦ mRNA in 50 human tissues revealed a distribution profile of p38␦ that differs from p38␣. p38␦ is highly expressed in salivary gland, pituitary gland, and adrenal gland, whereas p38␣ is highly expressed in placenta, cerebellum, bone marrow, thyroid gland, peripheral leukocytes, liver, and spleen. Like p38␣, p38␦ is activated by cellular stress and proinflammatory cytokines. p38␦ phosphorylates ATF-2 and PHAS-I, but not MAPK-activated protein kinase-2 and -3, known in vivo and in vitro substrates of p38␣. We also observed that p38␦ was strongly activated by MKK3 and MKK6, while p38␣ was preferentially activated by MKK6. Other experiments showed that a potent p38␣ kinase inhibitor AMG 2372 minimally inhibited the kinase activity of p38␦. Taken together, these data indicate that p38␦ is a new member of the p38 MAPK family and that p38␦ likely has functions distinct from that of p38␣.

PECAM-1 (CD31) is a 130-kDa member of the immunoglobulin (Ig) gene superfamily that is constitutively expressed at high concentration at endothelial cell intercellular junctions and at moderate density on the surface of circulating leukocytes and platelets. Recent in vitro and in vivo studies have shown the PECAM-1 plays a central role in mediating the extravasation of leukocytes from the vessel wall in response to inflammatory mediators. To study the binding characteristics of PECAM-1, phospholipid vesicles were prepared and examined by flow cytometry and immunofluorescence microscopy for their ability to associate with each other and with cells. Proteoliposomes containing high concentrations of PECAM-1 interacted homophilically with each other, forming large self-aggregates. PECAM-1 proteoliposomes, as well as soluble bivalent PECAM-1 in the form of a PECAM-1/IgG immunoadhesin, associated homophilically with cells expressing human, but not murine, PECAM-1. This binding could be completely inhibited by monoclonal antibody Fab fragments specific for Ig homology Domain 1 or Domains 1 + 2. Binding studies using cells expressing human PECAM-1 deletion mutants and murine/human chimeras confirmed that both Ig Domains 1 and 2 were both necessary and sufficient for homophilic binding. In contrast, engagement of membrane-proximal Domain 6 with monoclonal antibody Fab fragments had the opposite effect and augmented the binding of PECAM-1 proteoliposomes to cells. Thus, PECAM-1, like certain integrins, appears to be capable of antibody-induced conformational changes that alter affinity for its ligand. Similar changes induced by physiologic stimuli could be important in regulating the function of PECAM-1 in vascular cells.

A method to isolate fragments of DNA that promote gene expression in Bacillus subtilis is described. The system is based on production ofcatechol 2,3-dioxygenase [CatO2ase; catechol:oxygen 2,3-oxidoreductase (decyclizing), EC 1. 13 Bacillus subtilis is an attractive alternative to Escherichia coli as a host for expression of cloned genes. The Gram-positive organism is nonpathogenic, free ofendotoxins, and an important producer of extracellular enzymes on a large industrial scale. Critical to the development ofthe microorganism as a host-vector system for recombinant DNA technology is the efficient expression of heterospecific genes. To express plasmid-borne genes in B. subtilis, transcriptional or translational signals that differ from those of E. coli are required (1). Plasmid vectors suitable for cloning fragments of DNA that carry transcriptional promoter or termination signals for Gram-negative bacteria into E. coli have been characterized (2-6). Detection in these systems is based on expression ofgenes that encode P-galactosidase (4-6) or confer antibiotic resistance to host cells (2, 3). An approach similar to the latter has been successful in B. subtilis using chloramphenicol acetyltransferase genes originating from Bacillus pumilus (7) or the transposable genetic element Tn9 (8). In the E. coli /-galactosidase system, selection ofDNA fragments that promote expression of the lacZ gene is based on an easily visualized color change ofbacterial colonies grown on indicator plates containing a chromogenic substrate (4). An analogous system that functions in B. subtilis would greatly facilitate the effort to decipher problems of heterospecific gene expression in Gram-positive bacteria.In this report, we present a method whereby fragments of DNA that promote expression of a foreign gene in B. subtilis are detected by a change ofcolor of bacterial colonies. The system is based on the cloning and expression, in B. subtilis, ofthe xylE gene, which originated from the TOL plasmid pWWO (9) of Pseudomonas putida mt-2. The assay is rapid and inexpensive, does not require special indicator plates but offers the advantages ofa genetic indicator test (10), and can be used for the development of efficient plasmid gene expression vectors. MATERIAL AND METHODSBacterial Strains and Plasmids. The B. subtilis strains used are derivatives of Marburg strain 168. Strains BZ2 cysB3 recE4 and TGB1 trpC2 recE4 spo331 were constructed by transformation (11). MI112 argl5 leuB thr5 r-mM recE4 was from T. Tanaka. Bacillus licheniformis 9945A and Bacillus pumilus BP1 were obtained from the Bacillus Genetic Stock Center (Ohio State University, Columbus). E. coli strain BZ18 was from W. Arber; C600 rj m' was from J. W. Little; Pseudomonas putida mt-2 was donated by K. Timmis. The bifunctional E. coli/B. subtilis plasmid pHV33 (12) was obtained from R. Dedonder. Plasmid DNA was prepared by an alkaline extraction procedure (13) or a cleared lysate method (14) followed by cesium chloride/ ethidium bromide density gradient centrifugati...

The c-Jun N-terminal kinase (JNK), or stressactivated protein kinase plays a crucial role in cellular responses stimulated by environmental stress and proinflammatory cytokines. However, the mechanisms that lead to the activation of the JNK pathway have not been elucidated. We have isolated a cDNA encoding a novel protein kinase that has significant sequence similarities to human germinal center kinase (GCK) and human hematopoietic progenitor kinase 1. The novel GCK-like kinase (GLK) has a nucleotide sequence that encodes an ORF of 885 amino acids with 11 kinase subdomains. Endogenous GLK could be activated by UV radiation and proinflammatory cytokine tumor necrosis factor ␣. When transiently expressed in 293 cells, GLK specifically activated the JNK, but not the p42͞44 MAPK ͞ extracellular signal-regulated kinase or p38 kinase signaling pathways. Interestingly, deletion of amino acids 353-835 in the putative C-terminal regulatory region, or mutation of Lys-35 in the putative ATP-binding domain, markedly reduced the ability of GLK to activate JNK. This result indicates that both kinase activity and the C-terminal region of GLK are required for maximal activation of JNK. Furthermore, GLK-induced JNK activation could be inhibited by a dominant-negative mutant of mitogen-activated protein kinase kinase kinase 1 (MEKK1) or mitogen-activated protein kinase kinase 4͞SAPK͞ERK kinase 1 (SEK1), suggesting that GLK may function upstream of MEKK1 in the JNK signaling pathway.