Proc. Natl. Acad. Sci. USA Vol. 92, pp. 4537-4541, May 1995 Cell Biology Functional expression of low density lipoprotein receptor-related protein is controlled by receptor-associated protein in vivo THOMAS E. WILLNOW*, SCOTr A. ARMSTRONG*, ROBERT E. HAMMERt, AND JOACHIM HERZ* Departments of *Molecular Genetics and tBiochemistry and Howard Hughes Medical Institute, University of Texas Southwestern Medical Center, Dallas, TX 75235 Communicated by Joseph L. Goldstein, The University of Texas Southwestern Medical Center, Dallas, TX, February 15, 1995 ABSTRACT The 39-kDa receptor-associated protein terminal HNEL tetraamino acid motif has prompted Strick- (RAP) associates with the multifunctional low density li- land et al (32) to propose a possible role of the KDEL receptor poprotein (LDL) receptor-related protein (LRP) and thereby in the process. When RAP is overexpressed the retention prevents the binding of all known ligands, including a2- system becomes saturated and RAP is secreted from the cell, macroglobulin and chylomicron remnants. RAP is predomi- resulting in autocrine/paracrine inhibition of LRP function in nantly localized in the endoplasmic reticulum, raising the vitro and in vivo. We have used this effect in a previous study possibility that it functions as a chaperone or escort protein (33) to provide evidence that LRP participates in the clearance in the biosynthesis or intracellular transport of LRP. Here we of remnant lipoproteins by the liver. have used gene targeting to show that RAP promotes the Based on its biochemical properties RAP has been proposed expression of functional LRP in vivo. The amount of mature, to function as a chaperone during biosynthesis, as an escort processed LRP is reduced in liver and brain of RAP-deficient protein within the secretory pathway, and as a short-acting mice. As a result, hepatic clearance of a2-macroglobulin is modulator of LRP activity (2, 25, 26, 31, 34). In this study we impaired and remnant lipoproteins accumulate in the plasma have used gene targeting to disrupt the RAP gene in mice to ofRAP-deficient mice that also lack functional LDL receptors. determine whether RAP is required for normal LRP expres- These results are consistent with the hypothesis that RAP sion and activity. We found that homozygous RAP-deficient stabilizes LRP within the secretory pathway. They also suggest mice are viable and superficially normal. LRP expression in a further mechanism by which the activity of an endocytic liver and brain is significantly reduced in these animals. This receptor may be modulated in vivo. results in impaired clearance of LRP-specific circulating li- gands by the liver. Our results indicate that RAP controls LRP The low density lipoprotein (LDL) receptor-related protein expression, possibly by stabilizing the receptor during biosyn- (LRP) (1) is a ubiquitously expressed endocytic receptor thesis or transport along the secretory pathway. (reviewed in ref. 2) that binds a diverse group of ligands, including lipoproteins (3-6), lipoprotein lipase (7), proteases EXPERIMENTAL PROCEDURES (8, 9), protease inhibitors and protease:inhibitor complexes Methylamine-activated human a2-macroglobulin was a gift (10-15), bacterial toxins (16, 17), viruses (18), and lactoferrin from Dudley Strickland (American Red Cross, Rockville, (19). Some of these macromolecules compete with each other MD); antibodies directed against synaptophysin were provided for a common site on LRP, while others bind to independent by T. Rosahl and T. C. Siidhof (Howard Hughes Medical sites (20). LRP is an essential gene (21, 22), probably because Institute, University of Texas Southwestern Medical Center, it participates in such a large number of diverse biological Dallas). Bovine asialofetuin was purchased from Sigma. Glu- processes. tathione S-transferase/RAP fusion protein was produced as LRP is synthesized as a 600-kDa precursor (LRP 600) that described (25). Mice used in the experiments were C57BL/6J is cleaved to generate an amino-terminal 515-kDa (LRP 515) x 129SvJ hybrid male mice (age, 12-20 weeks) bred in house fragment and a carboxyl-terminal 85-kDa (LRP 85) fragment. and fed ad libitum throughout the course of the experiment This proteolytic processing takes place in a post-Golgi secre- (Teklad 6% mouse/rat diet 7001; Teklad, Madison, WI). tory compartment (23). LRP 515 harbors all known ligand Animal care and experimental procedures involving animals binding sites and remains noncovalently associated with LRP were conducted in accordance with institutional guidelines. 85, which contains the membrane anchor and the cytoplasmic Protein iodinations and DNA manipulations were performed domain. as described (25). Protein content of membrane preparations The binding of ligands to LRP can be blocked in vitro by a was determined using the Coomassie plus protein assay re- 39-kDa polypeptide designated receptor-associated protein agent (Pierce). (RAP) (24-26) that copurifies with LRP on a2-macroglobulin Disruption of the RAP Gene. A 14-kb fragment of the affinity columns (10, 11). RAP also binds tightly to gp330, an murine RAP gene was isolated from a commercial genomic endocytic receptor that resembles LRP and is abundantly library (Stratagene) by hybridization screening with a rat expressed in the kidney (19, 27, 28), and to the very low density cDNA probe (25). The fragment was subcloned into lipoprotein (VLDL) receptor (29). Both proteins are also pGEM5Zf(+) and the pol2sneobpA expression cassette (35) members of the LDL receptor gene family. RAP also interacts was inserted into an Nco I site located within the leader peptide weakly with the LDL receptor (30). sequence. The neo expression cassette was flanked on the 5' Under normal circumstances RAP resides predominantly in side by a 1-kb Nco I fragment (short arm) and on the 3' end the endoplasmic reticulum (ER) and in the Golgi complex, by an adjacent 10-kb Sca I fragment (long arm). Two copies of while only a small percentage is present on the cell surface or the herpes simplex virus thymidine kinase gene were inserted in endosomes (27, 30, 31). The mechanism by which RAP is in tandem at the 5' end of the short homology region. retained in the ER is unknown, but the presence of a carboxyl- Abbreviations: LDL, low density lipoprotein; LRP, LDL receptor- The publication costs of this article were defrayed in part by page charge related protein; RAP, receptor-associated protein; CR, chylomicron payment. This article must therefore be hereby marked "advertisement" in remnant(s); apo, apolipoprotein(s); VLDL, very low density lipopro- accordance with 18 U.S.C. §1734 solely to indicate this fact. tein; ER, endoplasmic reticulum. 4537 Downloaded by guest on September 28, 2021 4538 Cell Biology: Willnow et at Proc. Natl. Acad. Sci. USA 92 (1995) of the linearized vector into Electroporation replacement A B LIVER BRAIN murine embryonic stem cells (JH1 and AB1) and derivation of I RAP|+ germ-line chimeras from four independent stem cell clones | RAP .+I|-/-|+/- ,+,/+|--,-l-,-1 12kb- S '6*7 8RAP was performed according to standard protocols (36). Targeted --hko RAP- ", .;.1 stem cell clones were identified PCR under the conditions by 8kb- -wt LRP 515- '." .:.:..'. :' described by Soriano et al (35) using a primer located within the 3' untranslated region of the neo expression cassette LRP 85-*- (5'-GATTGGGAAGACAATAGCAGGCATGC-3') and an- LDLR-- ,l1 -Synapt other primer located outside the targeting construct and 12 3 4 upstream of the short homology segment (5'-TGATTGG- TACCATCTCTGGGCTGG-3'). FIG. 2. Southern and immunoblot analysis of wild-type and RAP- Immunoblot and Blot of Membrane Pro- deficient mice. (A) Twenty micrograms of genomic tail DNA from Ligand Analysis and RAP- teins. Membrane were from mouse tissues wild-type (lane 4), heterozygous (lane 1), homozygous proteins prepared deficient mice 2 and was with Sca I and and 50 of lane was on (lanes 3) digested analyzed by ,ug protein per separated 4-15% Southern the rat RAP cDNA as a under conditions with- blotting using hybridization probe. SDS/polyacrylamide gels nonreducing Diagnostic fragments for wild-type (wt) and disrupted allele (ko) are out boiling (final concentration of SDS, 2.3%) and transferred indicated. (B) Fifty micrograms of membrane proteins prepared from to nitrocellulose paper at 4°C as described (5). Filters were liver (lanes 1-4) or brain (lanes 5-8) of wild-type (lanes 1, 2, 5, and 6) incubated with polyclonal rabbit IgGs (5 ,tg/ml). Bound IgG or RAP-deficient mice (lanes 3, 4, 7, and 8) was separated by SDS gel was detected using 125I-labeled goat anti-rabbit IgG (1 x 106 electrophoresis on 4-15% polyacrylamide gels under nonreducing cpm/ml) (5) or by enhanced chemoluminescence using horse- conditions, and the proteins were transferred to nitrocellulose filters. radish peroxidase-coupled donkey anti-rabbit IgG and the Filters were incubated with polyclonal rabbit IgGs (5 ,g/ml) directed ECL system (Amersham) according to the manufacturer's against the extracellular portion of rat LRP (LRP515), the cytoplasmic recommendations. tail of human LRP (LRP 85), rat RAP, bovine LDL receptor (LDLR), or rat synaptophysin (Synapt.). Bound IgG was detected using the Turnover Experiments. 125I-labeled human methylamine- enhanced chemoluminescence system (ECL, Amersham). activated a2-macroglobulin or bovine asialofetuin was injected into the external jugular vein of anesthetized mice (5 ,ug per Mendelian frequency (-1:2:1). We did not observe any vari- animal). At designated time points