(Received for publication, March 16, 1992). Maria Z. KounnasS, Randal E. Morris$,. Michael R. ThompsonT, David J. FitzGeraldII**,. Dudley K. Strickland+**, and.
Communication
THEJOURNAL OF BIOLOGICAL CHEMISTRY Vol. 267, No. 18, Issue of June 25, pp. 12420-12423,1992 Printed in U.S. A.
The crs-Macroglobulin Receptor/ Low Density Lipoprotein Receptor-related Protein Binds and Internalizes Pseudomonas Exotoxin A*
receptor (1, 2) which mediates the plasma clearance of aZM activated by reaction with proteinases (designated azM*). This receptor is foundinmanytissues and cells and is expressed in high levels in fibroblasts (3-5) and hepatocytes (6, 7). Amino acid sequencing studies of the purified receptor (8,9) revealed this molecule to be identical to the low density lipoprotein receptor-related protein (LRP) (10). Herz et al. (Received for publication, March 16, 1992) (11)demonstrated that this receptor is synthesized as a 600Maria Z. KounnasS, Randal E. Morris$, kDa single-chain precursor, which is cleaved in the transMichael R. ThompsonT, David J. FitzGeraldII**, Golgi to a 515-kDa heavy chain and an 85-kDa light chain Dudley K. Strickland+**, and that remain noncovalently associated on the cell surface. The Catharine B. Saelinger**$$ 515-kDa chain contains 15 epidermal growth factor repeats From the $Biochemistry Laboratoy, American Red Cross, and 31 low density lipoprotein receptor ligand binding repeats Rockville, Maryland 20855, the TDivision of Digestive and is responsible for ligand binding, while the 85-kDa polyDisease, the §Department of Anatomy and Cell Biology, and the $$Department of Molecular Genetics, Biochemistry peptide contains seven epidermal growth factor repeats,a and Microbiology, University of Cincinnati College of transmembrane domain, and NPXY consensus internalizaMedicine, Cincinnati, Ohio 45267-0524, and the tion sequences within the cytoplasmic domain. IILaboratory of Molecular Biology, National Cancer In addition to its role in removal of azM*, azMR/LRP also Institute, Division of Cancer Biology, Diagnosis and Centers, Bethesda, Maryland 20892 has been proposed to be responsible for hepatic clearance of chylomicron remnants. This proposal is supported by a numThe a2-macroglobulin receptor/low density lipopro- ber of studies which have documented the in vitro interaction tein receptor-related protein (a2MR/LRP) is a large of apoE (E’), apoE-enriched P-VLDL particles (13-16), or cell-surface glycoprotein consisting of a 515-kDa and lipoprotein lipase (17) with this receptor. a2MR/LRP also an 85-kDa polypeptide; this receptor is thought to be binds 2 mol of a 39-kDa protein, termed the receptor-associresponsible for the binding and endocytosis of acti- ated protein (RAP), with high affinity (KO= 14 nM) (18). vated a2-macroglobulinand apoE-enriched &very low density lipoprotein. A similar high molecular weight While the primary structure of RAP has been determined glycoprotein has been identified as a potential receptor (19), the function of this molecule is not fully understood. for Pseudomonas exotoxin A (PE). We demonstrate The fact that RAP is able to inhibit ligand binding by this that the aZMR/LRP and the PE-binding glycoprotein receptor (18,20, 21) has led to theproposal that thismolecule have a similar mobility upon sodium dodecyl sulfate- may modulate ligand binding i n uiuo. polyacrylamide gel electrophoresis and are immunoRecent studies have identified another large cell surface logically indistinguishable. Furthermore, affinity-pu- glycoprotein isolated from mouse fibroblast cell lines and rified a2MR/LRP binds specifically to PE but not to a mouse liver (22, 23) that specifically binds to PE. PE was mutant toxin defective in its ability to bind cells. The originally described as one of several virulence factors pro39-kDa receptor-associated protein, which blocks binding of ligands to aZMR/LRP, also preventsbinding duced by Pseudomonas aeruginosa (24). More recently, this and subsequent toxicity of PE for mouse fibroblasts. toxin has been employed as a component of recombinant The concentration of receptor-associated protein that toxins developed as novel therapeutics (25). PE, which is a was required to reduce binding and toxicity to 50% three-domain bacterial toxin (26), is lethal for cells because was approximately 14 nM, a value virtually identical of its ability to irreversibly shut down protein synthesis (27). to the KO measured for the interaction of receptor- To accomplish this, thetoxin binds to a high molecular weight associated protein with thepurified receptor. Overall, cell surface glycoprotein (22) and enters cells by receptor the studies strongly suggest that the a2MR/LRP is remediated endocytosis (28-30). Following internalization, the sponsible for internalizing PE. toxin is cleaved by a cellular protease (31),reduced, and finally a 37-kDa fragment translocates to thecytosol where it ADPribosylates elongation factor 2 (31,32). Domain I of the toxin The az-macroglobulin receptor/low density lipoprotein mediates cell binding (33-35), while domain I1 has transloreceptor-related protein (azMR/LRP)’ is a large cell surface catingactivity and serves as the substrate for proteolytic * This work was supported in partby National Institutesof Health cleavage (31, 35). Domain I11 has theADP-ribosylating activGrants AI17529, GM42581, HL30200, and HL02113. The costs of ity (35-37) and also contains an endoplasmic reticulum retenpublication of this article were defrayed in part by the payment of tion sequence which may target the toxin to the endoplasmic page charges. This article must therefore be hereby marked “aduerreticulum (38, 39). In tissue culture, PE is most active against tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate fibroblastic cell lines (40, 41), while its primary target i n uiuo this fact. is the liver (42, 43). ** To whom correspondence should be addressed. The similarity in size between a2MR/LRP and the PEThe abbreviations used are: ol,MR/LRP, ol,-macroglobulin receptor/low density receptor-related protein; PE, Pseudomonas exotoxin binding proteinprompted us to examine the relationship A; RAP, receptor-associated protein; a2M*, activated a,-macroglobbetween these twomolecules. The results of the present ulin; GST, glutathione S-transferase; BSA, bovine serum albumin; binding TBS, Tris-buffered saline; SDS-PAGE, sodium dodecyl sulfate-poly- investigation reveal that the azMR/LRP and the PE glycoprotein are immunologically indistinguishable, that PE acrylamide gel electrophoresis; VLDL, very low density lipoprotein; binds specifically to the a2MR/LRP, and that the 39-kDa ELISA, enzyme-linked immunosorbent assay.
12420
n2MR/LRPInternalizes Rinds and Pseudomonas receptor-associated protein (RAP) also prevents binding and subsequent toxicityof PE for mouse fibroblasts. EXPERIMENTALPROCEDURES
/’roteins-u,M and tr,MR werepurified as descrihed ( 2 ) . PEhinding protein was isolated as descrihed ( 2 2 , 2 3 ) . RAP was prepared as a fusion protein with glutathione S-transferase using the p(;EX2 T vector(I’harmaciaLKRBiotechnology Inc.) as descrihed (18). 1’E was either purchased from LIST Riochemicals or prepared as descrihed ( 3 3 ) . I’EgIu57 was prepared as descrihed (M),while PEtransferrin was prepared according to the methods descrihed (44). An~ibodics-Polyclonal antihodies against the w,MR/I,RP heavy chain were prepared by immunizing rahhits with the 515-kIh human ,v,MR/I,RP heavy chain excised from SDS-gels. The resulting antibody WAS affinity-purified overimmohilized tr,MR/I,RP heavy chain. I’olyclonal anti-human u,MR/LRP light chainantisera were prepared by immunizing a rahhit with a synthetic peptide corresponding t o the carhoxyl terminus of the cytoplasmic domain (residues 45134525) conjugated to KLH as descrihed (10). Anti-I’E-hinding protein antisera were prepared hy immunizing rahhits with affinity-purified I,M cell-hinding protein. Immunohlolling-II-iereptor preparations were separated on 4 - 1 2 5 polyacrylamide gradient gels and electrophoretically transferred to as denitrocellulosememhranes.Immunohlotting wasperformed
scribed.' ( ’ d l Ilhxirity Exprriments-LM
cell monolayers were incuhated with increasing concentrations of RAP-GST or C S T for 90 min at 37 “C in serum free medium. Monolayers were cooled to 4 “C and incuhated with PE (40 ng) or transferrin-PE40 (500 ng) for 30 min. Control cells received no toxin. After 45 min at 4 ‘C, monolayers were washed and reincuhated in tissue culture medium for 4 h at 37 “C. Tritiated leucine (2 pCi/ml) was added for 30 min, and cells were processed as descrihed previously ( 2 8 ) . Binding of PF: to CdI.+”,M cell monolayers were incuhated with tissue culture medium containing either RAP-GST or GST. After incuhation for 18 h a t 4 “C. 2 pg/ml I’E was added for an additional 5 h a t 4 “C. The cells were then washed. harvested, homogenized, and the concentration of cell-associated toxin assayed hy EIJSA. Flatbottom micro-ELISA plates were washed with water, and sensitized with goatanti-I’E (1:500 dilution, List Biochemicals) in carhonnte buffer (15 mM Na,CO:!, 44 mM NaHCO:,, pH 9.6). Plates were washed with huffer A (2.25 mM KH,PO,, 7 mM Na,HPO,, 154 mM NaCI, 0.001”~thimerosal, 0.1”; Triton X-100,0.05°~ Tween 20, pH 7 . 2 ) and I h c k e d with 7% dry milk in phosphate buffered saline, pH 7.2, for 1 h at 25 “C. Known concentrations of PE or dilutions of homogenate were added, and plates were incuhated overnight at 4 “C. Plates were washed with huffer A, incuhated with hiotinyl sheep anti-PE (diluted 1:2500, Swiss Serum and Vaccine Institute; hiotinylated as descrihed in Ref. 46) in 3% RSA-phosphate-huffered saline for 1 h at 37 “C. washed. and incuhated wit.h streptavidin-horseradish peroxidase for 1 h. After a n additional washing, suhstrate (OPD) was added: after 30 min, color development WAS stopped and ahsorhance measured a t 492 nm. Lignnd Hlottinp--8 pg of human (r,MR/I,RP was separated on 515% polyacrylamide gels and transferred tonitrocellulose membranes. T h e memhrane was first hlocked with 3% milk in THS, pH 6.5.5 mM ( X I , , andthenincuhated for 1 hat 37 “C with 1 0 pg/ml PE or I’Eglu57. Strips were washed in T B S with 0.05% Tween 20 and inruhated for 1 h with goat anti-toxin. Next, memhranes were inrubated with rahhit anti-goat peroxidasefor 1 h, and reactive hands were visualized with the peroxidase suhstrate. 4-chloro-1-napthol. I.:/,/SA-Affinity-purified human placental tr,MR/LRPwas coated on microtiter plates (100 pI of 5 pg/ml) in THS, 5 mM CaCI,, pH 7.4. Following coating, the wells were blocked with 50 mg/ml RSA and inrreasing concentrations of PE or PEgIu57 were added. After 1 h at 37 “C, the wells were washed and a 1:200 dilution of goat anti-toxin added. Following a I-h incuhation, the wells were washed and incuhated with anti-goat IgG-HRP conjugate (1:2000). Binding was detected with 3,:3’,5,5‘-tetrarnethvlhenzidine(Kirkegaard X Perry I,ahs, lnr., Gaithershurg, MI)) suhstrate. M. Z.Kounnas, W. S. Argraves, and D. K. Strickland, suhmitted lnr puhlication.
Exotoxin A
12421
RESIJLTS
PE-binding Protein
Is
Immunologicnlly Rrlntrd to t r ( M R /
LRP-To examine the relationship between PE-binding protein and tu,MR/LRP, the mobility of these proteins on SIXPAGE were compared and immunoblotting experiments were performed. The heavy chain of affinity-purified f ~ & U W J W andtoxin-bindingproteinhadidenticalmohilityon SIX PAGE (Fig. 1A). T h e 85-kDa chain. although present, stains poorly with Coomassie Blue(2). Immunoblotting experiments showedthatantihodiesraisedagainsthumanplacental tr,MR/LRP reacted with the affinity-purified toxin-binding proteinderivedfrommouseliver(Fig. lH, l n n r 2 ) . As a positive control, the antibody was shown to react with the n,MR/LRP purified from human placenta (Fig. 1 N . lnnp / ). Similarly, antibodies against the murine toxin-binding protein recognized human tvJ4R/I,RI’ (Fig. lC, lnnr / ). In addition, both receptor preparations contained an 85-kDa light chainidentified by anantibodyraisedagainst a synthetic peptide corresponding to the carboxyl terminus of the human 85-kDa pol-ypeptide(Fig. 111). Thus, although the receptor preparations are derived from different tissues of different species,immunologicalcross-reactivitywasobserved, confirming that the I’E-binding protein is immunologically related to n2MR/LKP. T h e immunological similarity between these molecules suggests that these receptors are closely related, if not identical, and raises the possibility that tr:.“R/ LRP is responsible for transporting PI*: into the cell. This is further strengthened by the ohservation that dvlR/I,RP has been identified ona variety of established cell lines and mouse tissues which are susceptible to I’E, and is present in low or
A 1 ”.
2
D B C 1 2 1 2 1 2
-200 -116
-94
-66 (r,
-45
FIG. 1. S i m i l a r i t yb e t w e e n n2MR/I,RP a n dt a x i n - b i n d i n g protein. A , comparison of the mol)ility 0 1 t l w hravy chain ( 1 1 humin placental u2Ml
