Surface Expression of Human CD14 in Chinese Hamster Ovary ...

Vol. 268,No. 29,Issue of October 15,pp. 22055-22059, 1993 Printed in U.S.A.

THEJOURNAL OF BIOUYXCAL CHEMISTRY Q 1993 by The American Society for Biochemistry

and Molecular Biology, Inc.

Surface Expressionof Human CD14 in Chinese Hamster Ovary Fibroblasts- Imparts Macrophage-like Responsiveness to Bacterial Endotoxin* (Received for publication, February 10, 1993, and in revised form, May 11, 1993)

Douglas T. GolenbockSQ,Yannan LiuS,Frederick H. Millhaml, Mason W. Freemanll, and Raphael A. Zoeller** From the $Department of Internal Medicine, Division of Infectious Diseases, the llSection of Critical Care, Department of Surgery, Boston City Hospital and the Boston University School of Medicine, Boston, Massachusetts 02118, the ILipid Metabolism Unit, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114, and the **Department of Biophysics, Boston University School of Medicine, Boston, Massachusetts 02118

While an increasing number of membrane-bound LPS-bindCardiovascular collapse associated with Gram-negative septicemia is believed to result from the stimulation ing proteins havebeen identified by ligand binding techniques of phagocytes by bacteriallipopolysaccharide (endo- (7-121, none of these molecules has been definitively demontoxin, LPS).It remains unclear how endotoxin activates strated to transduce the LPS activation signals directly. Curphagocytes, but recent evidence suggests the involve- rent evidence suggests that the cellular glycosyl phosphatiment of the glycosyl phosphatidylinositol-linked myelo- dylinositol-linked protein, CD14, is the best candidate for an cyte antigen, CD14. We report that transfection of hu- LPS-signaling receptor. LPS forms complexes with lipopolysacman CD14 into Chinesehamsterovaryfibroblasts charide binding protein and “septin” proteins, both constitutransfers macrophage-likeresponsiveness to otherwise ents of serum, and these LPS complexes bind CD14 (13, 14). LPS-unresponsive cells. Thesedatademonstratethat Furthermore, CD14-deficient monocytes (from human patients LPS-inducedresponsiveness can be transferred to a hetwith paroxysmal nocturnalhemoglobinuria) also failed to bind erologous non-responder cell type by expression of a LPS-serum complexes (15). Consistentwiththehypothesis single leukocyte-specificgene product. that CD14 functions as anLPS-signaling receptor were observations that LPS signaling was serum-enhanced (8, 16, 17). The outermost leafletof the outer membraneof Gram-nega- Monoclonal antibodies to CD14 partially interfered with the tive bacteria is composed almost entirely of lipopolysaccharide signaling process (8, 15, 17-20), and transfection of the human (LPS,l endotoxin (1, 2)). During the course of serious Gram- CD14 cDNA intomurine 70W3 B-lymphocyte tumor cells, negative bacterial infections, LPS-induced stimulation of which are less sensitive than macrophages to LPS, enhanced phagocytes(polymorphonuclearleukocytes, monocytes, and the sensitivityof this lineto LPS (21). Otherobservations have macrophages) results in the cascadingproduction and release led to the argument thatCD14 may not act alone as a signal transducing protein but may work in conjunction with an as yet of inflammatory mediators. The activity of these mediators, which include arachidonicacid metabolites (e.g. prostaglandins unidentified signalingreceptor. First, CD14 lacks a cytoplasmic domain or structural motifs knownto be responsiblefor cellular PGEz and PGDz) and cytokines (e.g. tumor necrosis factor-a and interleukin-lp) is, in turn, responsible for hypotension, end signaling. Second, a number of lipid A analogs have been deinhibition maynot be organ failure, andoften death (3). Recent epidemiological sur- scribed as LPS antagonists, yet theofsite (18, 22). CD14 veys of bacteremia in the United States estimate that between The evaluation of CD14 as a signaling protein is difficult in 40,000 and 70,000 Americans die asa result of Gram-negative leukocytic phagocytes because CD14 cannot be isolated easily septicemia per year (3, 4). from other LPS-binding and -signaling proteins present on LPS is composed of a complex and variable carbohydrate these cell types. For this reason, we engineered a CD14-exmoiety covalently linked to a highly conserved anionic phospholipid knownas lipid A, which anchors LPS into the bacterialpressing cell line in Chinese hamster ovary fibroblast (CH0)outer membrane (5).The toxicity of LPS is mediated through K1 cells by gene transfection. Our results demonstrate that the its lipid Amoiety (1,2,5).Although membrane-bound leukocyte addition of CD14 to the genetic background of an otherwise receptors for lipid A have been hypothesized to exist for many entirely unresponsive cell type is suficient to alter sensitivity years (61, the fundamental detailsof how LPS activates leuko- to bacteriallipopolysaccharide to a degree that approaches that of macrophages. CHO/CD14 will be useful in examining the cytes are only beginning to emerge. of other LPS-signaling proteins, in role of CD14, in the absence * This work was supported by National Institutes of Health Grants cellular responses to Gram-negativeendotoxin. R29GM47127 and Pol-AI33087 (to D. T.G.), R 0 1 HL45098 (to M. W. F.), and R29 DK40192 (to R. A. Z.). The costs of publication of this EXPERIMENTALPROCEDURES article were defrayed in part by the payment of page charges. This Reagents article must thereforebe hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Unless otherwise stated, all chemical and immunologic reagents 0 To whomcorrespondenceshould be addressed: Maxwell Finland Laboratory for InfectiousDiseases, 774 Albany St., Boston, MA 02118. were purchased from Sigma. All reagents used in LPSexperiments The abbreviationsused are: LPS, lipopolysaccharide; PMA, phorbol were determined to be pyrogen-free either by the manufacturer or by Limulus amebocyte lysate testing (MA Bioproducts, Walkersville, MD). 12-myristate13-acetate; HS, human serum; FCS, fetal calfserum; acid (3H-AA) and [cx-~~PI~CTP SFM, serum-free medium; 3H-AA, [5,6,8,9,11,12,14,15-3H]arachidonic[5,6,8,9,11,12,14,15,-3HlArachidonic acid; CHO-K1, Chinesehamster ovary fibroblasts-Kl;PBS, phosphate- were from DuPont NEN. Phosphate-buffered saline (PBS), Ham’s F-12, buffered saline; mAb, monoclonal antibody; LDL, low density lipopro- and RPMI 1640 were from MA Bioproducts. Defined fetal calf serum tein. was from Hyclone (Logan, UT). Human serum (HS) was derived from

22055

22056

CD14 Expression Imparts

LPS Responsiveness to

clotted whole blood from healthy volunteers and heat-inactivated a t 56 "C for 45 min. Tissue culture plastic, centrifuge tubes and culture tubes were from Falcon (Becton Dickinson, San Jose, CA). When necessary, glassware and plasticware were rendered pyrogen-free by baking at 270 and 125 "C, respectively, overnight. Salmonella minnesota R595 LPS (ReLPS)was the gift of N. Qureshi and K. "akayama (University of Wisconsin, Madison, WI).Synthetic lipid IVa was from ICN Biomedicals (Costa Mesa, CA). All lipids were prepared as stock suspensions at 1mg/ml in PBS and stored at -20 "C. The suspensions were defrosted and sonicated in a waterbath sonicator (Laboratory Supplies Inc., Hicksville, W, this sonicator has a single setting) prior to use. Phorbol 12-myristate 13-acetate (PMA) was stored at 0.3 mglml in dimethyl sulfoxide at -70 "C and diluted 1:10,000 in medium immediately prior to use. 1,25-Dihydroxycholecalciferol(vitamin D3, Biomole, Plymouth Meeting, PA) was stored a t 10 mM in ethanol at -20 "C.

CHO Fibroblasts

supplemented with 10% fetal calf serum (F-l2/FCS) a t lo7cells/ml, and a 25-pl portion of the cell suspension was stained by direct or indirect immunofluorescence and analyzed byflow microfluorimetry as described in Ref. 17. The final concentrations of the monoclonal antibodies used to label the cells were 10 pg/ml for 3C10, My4, and the myeloma protein control antibodies and -3 pdml for leu-M3. Northern Blot Determination of CD14 mRNA Levels

Total cellular RNA was extracted from the transfectanta and vitamin D,-treated (0.1 PM for 48 h) THP-1 cells using a commercial reagent (Tri-reagentTM,Medical Research Center, Cincinnati, OH) per the manufacturer's instructions. RNA was fractionated on a denaturing 1.2% agarose, formaldehyde gel (29), transferred to Zeta-probe membranes (Bio-Rad),air-dried, and fixed by baking for 1h a t 80 "C. Blob were then prehybridized in 10 mlof hybridization buffer (5 x Denhardt's solution, 50% formamide, 0.2% SDS, 5 x SSPE, and 200 pglml boiled Antibodies and Plasmids salmon sperm DNA) for 4 h at 42 "C. A BstI-Not1 fragment of pCD14 Anti-CDl4 dbs-Purified 3C10 (IgG2b) was the gift of Dr. S. D. (100 ng of cDNA/blot) was labeled with [c~-~,P]~CTP using a random Wright (Rockefeller University,N Y ) , My4 (IgG2b) was purchased from priming kit (Boehringer Mannheim) and the labeled probe purified by Coulter (Hialeah, FL), and fluorescein isothiocyanate-conjugated centrifugation over a Bio-Rad spin column per the manufacturer's inleu-M3 (IgG2b) from Becton Dickinson. structions. Purified cDNA probes contained 55% unincorporated 32P Other Antibodies-Unconjugated mouse myeloma IgG2b was pur- with a specific activity of approximately lo8 c p d p g DNA. The radiolachased from Bionetics(Charleston, SC). Fluorescein isothiocyanate-la- beled cDNAprobewas denatured by boiling for5 min and added tothe beled goat anti-mouse IgG was from Boehringer Mannheim. hybridization buffer. After overnight hybridization at 42 "C, the blots Plasmids-A full-length cDNA plasmid for f-actin was the gift of Dr. were washed sequentially as follows: 2 x SSC + 0.5% SDS for15 min at Bradley Schwartz (University of Wisconsin Medical School, Madison, 22 "C, 0.5 x SSC + 0.5% SDS for15 min a t 22 "C; 0.1x SSC + 0.5% SDS WI). The cDNA for human CD14, the gift ofDr. Simon Law (Merck for 15 min at 22 "C, and 0.1 x SSC + 0.5% SDS for 30 minat 65 "C. The Research Laboratories, Rahway, NJ), was provided in aBstI-Not1site in blots were then placed in plastic wrap and autoradiographed at -70 "C the mammalian expression vector pcDNAI (Invitrogen, San Diego, CAI; for 1-3 days using Kodak X-AFt film. this plasmid is referred to as pCD14. Plasmid pKoNeo was from Stratagene (La Jolla, CA) and encodes the gene neomycin phosphotransferase, Assay for Arachidonic Acid Release which confers resistance to G418 (this plasmid can be made available Cells were plated in 24-well tissue culture dishes (50,000 cells/well) through the authors upon request). A plasmid encoding the cDNA for the type I1 rabbit scavenger receptor was as previously described (23). and allowed to attach overnight in F-l2/FCS. 3H-AAwas dried under a nitrogen stream, resuspended in a small volume in F-l2/FCS in a 12 x 75-mm pyrogen-free borosilicateglass tube, and subjected to 15-30 s of Cell Lines a n d Culture Conditions sonic irradiation in awater bath sonicator. Medium was removed from RAW 264.7, an LPS-responsive murine macrophage-like cell line the cell monolayers and replaced with 0.25 ml of F-l2/FCS containing (24), THP-1, a human promonomyelocytic line (25), and the CHO-K1 0.4 pCi/wellof 3H-AAfor 6 or 16 h. After labeling, the cells werewashed fibroblast line were purchased from the American Type Culture Collec- three times with 0.75 ml of F-12 supplemented with 1%human serum tion (Rockville, MD). The transfected lines CHO/CD14 and CHO/Neo (F-l2/HS).After a final aspiration, either 0.75 ml of F-l2/HS or F-12/HS were derived from CHO-K1 as described below. RAW 264.7, CHO-K1, plus a stimulant (LPS or PMA) were added. Dishes were returned t o a and the transfected lines were maintained in Ham's F-12 while THP-1 37"C 5% CO, incubator, and after 45 min, a 0.2-ml portion of the cells were maintained in RPMI 1640 (M.A. Bioproducts, Walkersville, supernatant was removed and counted. For the experiment described in MD). Tissue culture medium was supplemented with 10%defined fetal Fig. 4, a serum-free medium (SFM)was used, which consisted of Hams calf serum andciprofloxacinat 10 pg/ml. Celllines were grownat 37 "C F-12 mixed with an equal volume of macrophage serum-free medium in a humidified 5% CO, environment a t a density of 2-5 x lo5 celldml. (Life Technologies, Inc.; this commercial medium contains human seAlthough CHO transfectants were selected and maintained in medium rum albumin, transferrin, andinsulin. Macrophage serum-free medium supplemented with G418 (400 pg of active druglml), cells were main- was assayed by S. D. Wright (RockefellerUniversity) for "septin" activtained in theabsence of G418 forat least 48 h prior to experimentation. ity (14) and byP. Tobias (Scripps Institute, La Jolla, CA) for lipopolysaccharide binding protein (16); neither serum protein was detectable. Generation of CHO Cell Lines Expressing CD14 a n d the This mixture of medium was used becausethe release of 3H-AAcannot Scavenger (Acetylated LDL) Receptor be measured without some extracellular protein present t o bind the released fatty acid; thus themacrophage serum-free medium served as CHO-K1 cells were plated at lo6 celld100-mm diameter tissue cul- compatiblemedium and a source of LPS-free human albumin. The ture dish overnight. The cells were then washed free of medium with remainder of the cell monolayer was washed and lysed in 0.5 N NaOH, PBS and either 1)co-transfected with pCD14 (2 &dish) and pKoNeo and the cell-associated countdmin were determined. The maximal re(0.7 pg/dish) or 2) pKoNeo alone (0.7 pg/dish) using CaPO, precipitation sponse of CHO/CD14 was usually 2-&fold higher (range, 1.2-15-fold) as described in Ref. 26.After 48h, the medium was replaced with fresh than theresponse of unstimulated or CHO/Neo control cells, representmedium containing G418 (400 pg of active drug/ml). After 4 weeks ing release of - 3 4 % of the totalincorporated label. All data areshown under selective conditions, isolates were generated from the G418-re- as thetotal released 3H-AAin 0.2 mlof supernatant. Each figure shown sistant population using limiting dilution. Cell surface expression of is representative of at least three separate experiments. CD14 in each clonal strain was determined by flow microfluorimetry (17) following incubation of the cells with phycoerythrin-labeled monoRESULTS AND DISCUSSION clonal antibody leu-M3. A subclone expressing high levels of CD14 is To assess the potential of CD14 to serve as a signaling rereferred to here as CHO/CD14; CHO cells transfected with pKoNeo ceptor, a fibroblast-like cell line that is unresponsive to LPS, alone are referred to as CHO/Neo. To generate cells expressing the scavenger (acetylated LDL) receptor (27), CHO cells were transfected CHO-K1, was co-transfected with an expression plasmid enwith the full-length cDNA forthe type I1 rabbit scavenger receptor (23) coding the cDNA for CD14 and plasmid pKoNeo, encoding reby Capo4 precipitation followed by G418 selection as previously de- sistance to the amino glycoside, G418. Afterselection in G418, scribed (28). A clone capable of binding and metabolizing acetylated LDL (28), referred to here as CHO/rSR, was chosen for further experi- a clonal isolate, CHO/CD14, expressing cell surface CD14(Fig. l),was expanded and used for further experiments. A control ments.

cell line, CHO/Neo, was transfected with a pKoNeoalone. CHO/ Neo cells expressed no CD14 as assayed by immunofluoresCHO transfectants were detached from the surface of tissue culture cence staining with an anti-CD14 antibody (Fig. 1). Northern dishes after washing monolayers once with PBS and incubating cells for blot analysis (30) confirmed an extremely high level of CD14 10 min with 1mM EDTA in PBS. Cells were resuspended in Ham's F-12 expression in CHO/CD14; steady state mRNA levels were conFlow Microfluorimetry Analysis of CHO Transfectants

CD14 Expression Imparts LPS Responsiveness to CHO Fibroblasts A.

’O01

I 28s -

1

CHO/NEO.

+

LEU-M3

CHOINEO.

+

IS0

10

18s

100

-I

1000

FIG.1. CHO cells transfected with the human CD14 gene bind monoclonalantibody to CD14 and produce abundant CD14 mRNA. CHO cells wereco-transfected with pCD14andpKoNeo or transfected with pKoNeo alone. Shownabove is a fluorescence-activated cell sorter histogram (see “Experimental Procedures”for details) of both cell lines analyzed after staining with phycoerythrin-conjugated leu-M3 or a directlyconjugatedmatchedcontrol mAb (n = 10,000 events). Similar results were obtained with the anti-CDl4 mAbs 3C10 and mAb My4 using indirect immunofluorescence.Inset, CD14 mRNA levels in vitamin D,-treated THP-1 cells (10 pg/lune), CHO/Neo (10 pg/lum), and CHO/CD14 (1 pgllune); identical samples probed with 32P-labeledf-actin cDNA demonstratedthat the quantity and qualityof the transcript in lunes I and 2 were similar (data not shown). See “Experimental Procedures”for details.

sistently almost 10-fold greater than the expression of transcripts obtained from the cell type from which the CD14 cDNA was originally cloned (vitamin D3-treated THP-1 monocytes (25)); CD14 mRNA could not be detected in CHO/Neo (Fig. 1, inset 1. One well established response to cellular activators suchas LPS is the release of arachidonic acid from phospholipid pools (24). Both CHO/Neo and CHOICD14 were capable of releasing arachidonic acid when stimulated withPMA (Fig. 2 A ) and the calcium ionophore A23187 (lov6M, data notshown). When LPS was used as a stimulus, only CHOICD14 released arachidonic acid whileCHO/Neo was unresponsive, even at extremely high levels of LPS (Fig. 2B). The LPS dose-response curve showed that CHOICD14 was almostas sensitive as a highly responsive macrophage-like cell line, the RAW 264.7 (Fig. 3). Both specific monoclonal antibody to CD14 (mAb My4, 100 pg/ml) and the endotoxin antagonist lipid IVa (100 ng/ml) (17,31-33) inhibited the effects of LPS (10 ng/ml) in CHOICD14 (data not shown). These data demonstrate that expression the of CD14 on the cell surface was sufficient to impart macrophage-like LPS responsiveness to theCHO cell line. CD14 is considered to be a receptor for LPS only when endotoxin is complexed to serum LPS accessory proteins (8, 14, 16,211. We sought to determine if LPS responsiveness in CHOI Neo could occur in the absence of serum. Such a question is difficult to pose in isolated monocytes because serum accessory proteins are presentat low concentrations even after washing (171, and CD14-independent activation of monocytes can occur under what appear be to rigorously serum-free conditions (22). CHOlCD14 cells, in contrast toisolated leukocytes, could withstand extensive washing and growth in serum-free medium without apparent loss of viability or responsiveness, and the response to LPS could not be attributed to other putative LPS binding proteins. CHOICD14 cells were consistently hyporesponsive to LPS in the absence of serum (Fig. 4). Only the highest concentrations of LPS tested succeeded in activating the CDlCbearing CHO cells without serum present. Require-

CHOA’eo

22057

CD14 Expression Imparts LPS Responsiveness to CHO Fibroblasts

22058

104

I

CHOICD14 ---.O----

I /1

a a jl

rn

............A ............................................

0

II

0.1

I

1

*............................................ I

10

I

100

2500

1

A I

1000

RAW 264.7 CHOIrSR CHO/CD14 FIG. 5. CHO cells transf‘ected with the macrophage scavenger [ L P S J , ng/ml receptor failto respond to LPS.RAW 264.7 macrophage-like tumor FIG.3. LPS responsiveness in transfected CHO fibroblastsand cells, CHO/rSR and CHO/CD14, were plated, labeled with3H-AA overa murine macrophage cell line.The response of the murine macro- night, and assayed for LPS responsiveness as described under “Experiphage cell line, RAW 264.7, was compared withthat of CHO/CD14 and mental Procedures.”CHO/rSR were also unresponsive tothe tetraacyl CHO/Neo. Cells were labeled for 6 h with 3H-AA prior to stimulation lipid A precursor, lipid IVa (100 ng/ml, data not shown). All values with medium containing 1%human serum and varying concentrations represent the mean * S.D. of three separate determinations. of LPS. Stimulation and determination of 3H-AA release wereperformed as described under “Experimental Procedures.”All values repLPS signaling has remained a complex field, and the identiresent the mean * the S.D. of three separate determinations. fication of putative LPS receptors has failed to elucidate how this important toxin activates monocytes, polymorphonuclear leukocytes, and other LPS-sensitive cells. The data presented here indicate that the expression of CD14 is sufficient to impart endotoxin sensitivity to an otherwise non-responsive cell, an effect that may be ubiquitous in mammalian cells. Although expressed in a heterologous cell line, CD14 expression resulted in monocyte-like attributes with respect to endotoxin responsiveness in the transfectants,i.e. LPS activity was serum-enhanced, demonstrable at very low concentrations of LPS and 20000 inhibited by known LPS antagonists. The ability of CHO fibroblasts to respond to LPS was not limited to arachidonate metabolism. In macrophages, the transcriptional activator, NF-KB is translocated from the cytosol to the nucleus in response to 10000 LPS and, in part, serves to regulate LPS-induced expression of ......5”-.-”5! important proinflammatory cytokines such as tumor necrosis factor-a! (35).Although Northern blot analysis of RNA extracted from LPS-stimulated CHO/CD14 and probed withhuman O ! I1 I I I I cDNA probes under low stringency conditions failed to reveal 0 0.1 1 10 100 1000 LPS-induced production of tumor necrosis factor-a! and inter[LPS], nglml leukin-l/3 transcript (data not shown), LPS-induced translocation of an NF-KB-like protein occurred in response to LPS.’ FIG.4. Stimulation of CHO/CD14 by LPS is enhanced by the Cells were plated in 24-well tissue culture This finding suggests that the diverse responses presence of human serum. LPSto seen in dishes andwashed five times ina serum-freemedium (see “Experimen- macrophages could be triggeredinitiallythrough a single tal Procedures”) immediately prior to overnight labeling 3H-AA with in SFM. After labeling, cellmonolayers were washed three final times CD14/NF-KB-linked pathway but that other factors, missing with SFM and then stimulated with graded amounts of LPS in SFM from CHO cells, must also be required. aloneorSFMsupplementedwith 1%humanserum.After45min, Recently, two hypothetical models have been proposed to supernatants were collected and counted for released 3H-AA. Experi- explain how LPSmightinduce a transmembranesignal ments performed with cells, which were grown for several days in serum-free medium priorto stimulation with LPS, yielded similar results through CD14 binding (2). 1)CD14 may directly activate cells complexes; 2 ) (data not shown). All values represent the mean * the S.D. of three after binding by LPS and LPS-serum protein separate determinations. CD14 may function as a binding receptor and activate cells only in collaboration with a second signaling receptor (analogous to the interleukin-6 receptor (36)). Our data cannot distinguish lishing the ability of LPS to bind to the transfected scavenger between these two possibilities. The CD14 transfectants dereceptor (data not shown). However, when these cells were scribed in this paper may, however, prove to be useful in studies assayed for responsiveness, LPS failed to stimulate 3H-AA re- designed to explore these models. Because CHO cells are widely lease (Fig. 5). Synthetic lipid IVa (100ng/ml) also failed to employed in somatic mutagenesis studies, theCHOlCD14 line activatescavengerreceptor-bearingtransfectants,although is amenable toa genetic complementation analysisdesigned to the cell linewascapable of respondingto PMA (datanot shown). Thus, the presence of an LPS-binding protein on the * R. DeLude, M. Fenton, and D. Golenbock, manuscript in preparation. surface of cells does not guarantee endotoxin responsiveness.

2

i

1 ...”.,,

CD14 Expression Imparts

LPS Responsiveness to

identify proteins, in addition to CD14, that are involved in the LPS-si@aling cascade. The Of the CD14-mediated signal transduction pathway, through the combined approaches of genetic transfer,mutantscreening, ligand binding, and the development of L p s receptor antagonists may be POtentially important in understanding the cellular biology associated with endotoxin-related disease. REFERENCES 1. Raetz, C. R. H. (1990)Annu. Reu. Biochem. 69, 129-170 2. Rietschel, E. T.,and Brade, H. (1992) Sci. Am. 267, 26-31 3. Bone, R. C. (1991)Ann. Intern. Med. 115,457469 4. Martin, M. A. (1991) Infect. Dis. Clin. North Am. 6,739-752 5. Rietschel, E. T. 11984) in Chemistry ofEndotoxins (Rietschel, E. T.,ed) Vol. 1, pp. 1A17, Elsevier Science Publishing Co., Inc., New York 6. Momson, D. C., and Rudbach, J. A. (1981) in Endotoxin-Cell Membrane Interaction Leading to Dansmembrane Signaling (Inman, F. P., and Mandy, W. J., ed) Vol. 8, p. 187, Plenum Press, New York 7. HamPton, R. y., GolenboCk, D. T., Penman, M.3 Krieger, M-7 and c. R. (1991)Nature 352,342444 8. Wright, s. D., h m w R. A., Tobias, p. s.,Ulevikh, R. J.. and Mathison, J. c . (1990) Science 249, 1431-1433 9. Wright, S . D., and Jong, M. T. C. (1986) J. Exp. Med. 164, 1876-1888 10. Kirkland, T. N., Virca, G. D., Kuus-Reichel, T., Multer, F.K., Kim, S . Y., m e v i k h , R. J., and Tobias, p. s. (1990) J . B i d . Chem. 2% 9520-9525 11. Lei, M., Stimpson, S . A., andMomson, D. C. (1991) J . Immunol. 147, 19251932 12. Parent, J. B. (1990) J . Biol. Chem. 268, 3345-3461 13. Wright, s. D., Tobias, P. s., Ulevitch, R. J., and Ramos, R. A. (1989) J . EXP. Med. 170,1231-1241 14. Wright, S . D., Ramos, R. A., Patel, M.,and Miller, D. S . (1992)J. Exp. Med. 176, 719-727 15. Couturier, C., Haeffner-Cavaillon,N.,Caroff, M., andKazatchkine, M. D. (1991) J . Immunol. 147, 1899-1904 16. Schumann, R. R., Leong, S. R., Flaggs, G. W., Gray, P.W., Wright, S. D.,

CHO Fibroblasts

22059

Mathison, J. C., Tobias, P. S . , andUlevitch, R. J. (1990)Science 249,14291431 17. Lynn, W. A,, Raetz, C. R. H., Qureshi, N., and Golenbock, T. D. (1991) J . Immunol. 147,3072-3079 18. Kitchens, R.L.9 Ulevitch, R. J.7 and Munford, R. S . (1992) J . ExP. Med. 176, 485494 19. Wright, S . D., Ramos, R. A,, Hermanowski-Vosatka, A. H., Rockwell, P., and Detmers, P. A. (1991) J . Exp. Med. 173,1281-1286 20. Heumann, D., Gallay, P., Barras, C., Zaech, P., Ulevitch, R. J., Tobias, P. S . , Glauser, M. P., and Baumgadner, J. D. (1992)J. Immunol. 148,3505-3512 21. Lee, J.-D., Kato, K., Tobias, P. S., Kirkland, T. N., and Ulevitch, R. J. (1992) J . Exp. Med. 175, 1697-1705 22. Lynn, W. A., Liu, Y., and Golenbock, D. T. (1992) submitted for publication. 23. Bickel, P. E., and Freeman, M. W. (1992) J . Clin. Inuest. BO, 1450-57 24. Zoeller, R. A., Wightman, P. D., Anderson, M. S., and Raetz, C. R. H. (1987)J . B i d . Chem. 262, 17212-17220 25. Tsuchiya, S., Yamabe, M., Yamaguchi, Y., Kobayashi, Y., Konno, T., and Tada, K.(1980) Int. J . Cancer 28, 171-176 26. Kingston, R. E. (1988)in Zntroduction ofDNA into Mammalian Cells(Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., and Smith, J. A., eds) Vol. 1, pp. 9.1.1-9.1.4, John Wiley and Sons, Inc., New York 27. Brown, M. S., and Goldstein, J . L. (1983)Annu. Reu. Biochem. 62, 223-261 28. Freeman, M., Ekkel, Y., Rohrer, L., Penman, M., Freedman, N. J.,Chisolm, G. M.,and Krieger, M. (1991)Proc. Natl. Acad. Sei. U. S. A . 88,49314935 29. Lehrach, H. D., Diamond, D., Wozney, J. M.,and Boedtker, H. (1977)Biochemistry 16,47434745 30. Davis, L.G., Dibner, M. D., and Battey,J. F. (1986)BasicMethods in Molecular Biology, Elsevier Science Publishing Co., Inc., New York 31. Golenbock, D. T., Hampton, R. Y., Qureshi, N., Takayama, K., and Raetz, C. R. H. (1991)J . Biol. Chem. 266, 19490-19498 R. S., Raetz, Harlan, C.and R. H., J. H. (1990) 32. Kovach, N. L., Yee, E., Munford, J. Exp. Med. 172, 77-84 33. Loppnow, H., Brade, H., Durrbaum, I., Dinarello, C. A,, Kusumoto, S . , RiE.-T., etschel, H. Flad, and D. (1989)J . Immunol. 142, 3229-3238 34. Rosenfeld, M. E.,and Lipton, B. A. (1992) Curr. Opin. Lipid. 3,318-323 Jongeneel, and S.A., C. 35. Shakhov,A. N., Collart, M. A., Vassalli, P., Nedospasov, V. (1990)J . Exp. Med. 171, 3 5 4 7 36. Hibi, M., Murakami, M.. Saito, M., Hirano, T., Taga, T., Kishimoto, and T. (1990) Cell 63, 1149-1157