key: cord-1023103-unssixx2
authors: Pease, Richard J.; Harrison, Georgina B.; Scott, James
title: Cotranslocational insertion of apolipoprotein B into the inner leaflet of the endoplasmic reticulum
date: 1991
journal: Nature
DOI: 10.1038/353448a0
sha: 6cf35329edb8d663a047ca4af4ffa28d84b351ee
doc_id: 1023103
cord_uid: unssixx2

APOLIPOPROTEIN(apo) B100 is required for the distribution of hepatic triglyceride to peripheral tissues as very-low-density lipoproteins. The translocation of apo B100 into the endoplasmic reticulum (ER) and its subsequent assembly into lipoprotein particles is of particular interest as the protein is both very large (relative molecular mass 512,000) and insoluble in water. It has been proposed that apo B translocation occurs in discrete stages and is completed post-translationally(1). Several sites of arrest of translocation were reported to be present in apo B15 (the N-terminal 15% of the protein). We have re-examined this question by in vitro translation coupled with translocation into microsomes, and find no evidence for transmembrane segments in truncated apo B proteins. Translocated apo B17 is strongly associated with the membrane-of the ER, being only partially releasable with alkaline carbonate, and remaining bound to the microsomes following disruption with saponin. The efficient binding of short segments of apo B, despite the absence of transmembrane domains, suggests that apo B is cotranslationally inserted into the inner leaflet of the ER. This will obviate problems caused by the size and insolubility of apo B100, because the growing hydrophobic protein chains will never exist in a lipid-free form during translocation. From the inner leaflet, apo B in association with membrane-derived lipid can bud into the lumen of the ER to form nascent lipoprotein particles.

Division of Molecular Medicine, MRC Clinical Research Centre, Watford Road, Harrow HA1 3UJ, UK APOLIPOPROTEIN (apo) B100 is required for the distribution of hepatic triglyceride to peripheral tissues as very-low-density lipoproteins. The translocation of apo B100 into the endoplasmic reticulum (ER) and its subsequent assembly into lipoprotein particles is of particular interest as the protein is both very large (relative molecular mass 512,000) and insoluble in water. It has been proposed that apo 8 translocation occurs in discrete stages and is completed post-translationally 1 • Several sites of arrest of translocation were reported to be present in apo B15 (the Nterminal 15% of the protein). We have re-examined this question by in vitro translation coupled with translocation into microsomes, and find no evidence for transmembrane segments in truncated apo 8 proteins. Translocated apo 817 is strongly associated with the membrane· of the ER, being only partially releasable with alkaline carbonate, and remaining bound to the microsomes following disruption with saponin. The efficient binding of short segments of apo 8, despite the absence of transmembrane domains, suggests that apo B is cotranslationally inserted into the inner leaflet of the ER. This will obviate problems caused by the size and insolubility of apo 8100, because the growing hydrophobic protein chains will never exist in a lipid-free form during translocation. From the inner leaflet, apo B in association with membrane-derived lipid can bud into the lumen of the ER to form nascent lipoprotein particles.

After biosynthesis apo BlOO is either assembled into lipoprotein particles and secreted, or degraded intracellularly 2 • 3 • Studies of cells in culture indicate that a proportion of the molecules are untranslocated and destined for degradation and that the remainder are translocated and quantitatively secreted 2 . Studies of apo B 15 translated in wheat-germ Iysates have suggested that translocation pauses while translation continues, imposing transient transmembrane configurations 1 ; we cannot confirm these observations.

To obtain efficient translation of apo B in reticulocyte lysates, a partial 5' untranslated region (UTR) of mouse encephalomyocarditis virus (EMCV) (ref. 4 ) was fused to apo B complemen-448 tary DNA. The EMCV-apo B transcripts encode MetAia before the apo B prepeptide sequence, but the extended signal still functions efficiently. Thus, apo B7 and B9 ( Fig. 1 a, b ) increase in size after translation in the presence of microsomes owing to glycosylation 5 , and apo B9 is protected from trypsin digestion when translated with microsomes ( Fig. la) . When apo B7 is translocated in the presence of the competitive glycosylation inhibitor N-acetyl-AsnTyrThr-carboxyamide6, the principal product (unglycosylated apo B7) migrates faster than pre-apo B7, consistent with cleavage of the signal sequence (Fig. lb) . Cleavage (-90%) was confirmed by radiosequencing of processed apo Bl5 (results not shown).

In cultured cells apo BlOO is carbonate-resistant 7 • 8 immediately after biosynthesis, which indicates that integration into the ER membrane has taken place 9 • We therefore investigated whether the truncated apo B proteins bind to microsomes. Under conditions in which soluble secretory proteins are quantitatively released but a transmembrane protein will remain associated with the membrane, membrane binding was significant with proteins in the range apo B9 to B 17, but less with apo B7 ( Fig.  2a ) . Similarly, intermediate carbonate releasability has been found with apo E in Golgi membranes 8 : we have confirmed the membrane association of apo Bl 7 following disruption with saponin ( Fig. 2b) .

Association of apo B with the ER membrane could either involve amphipathic binding to the inner leaflet, or the presence of transmembrane regions. We observed that after translation in reticulocyte lysates for either 40 min (when protein synthesis is ongoing) or for 90 min at 30 °C, both apo B 15 and apo BI 7 are protected from trypsin or proteinase K (Fig. 3a, b; and data not shown), thereby excluding the presence of transmembrane domains. To ensure reproducible inactivation of proteinase K , it was necessary to centrifuge the membranes out of the digestion mixture, but this protocol precludes the controls of post-translational addition of microsomes or total disruption with detergent to verify proteinase K susceptibility of apo BI 7. Instead it was shown that apo B 17 becomes susceptible to proteinase K after the membranes are permeabilized with saponin ( Fig. 3b ). Although we did not find transmembrane domains in apo B, we confirmed that E 1 glycoprotein is a transmembrane protein ( Fig. 3c ) 10 . Finally we determined whether generation of transmembrane domains resulted specifically from wheat-germ ribosomes. Reduced expression of some transmembrane domains in reticulocyte lysates relative to wheat-germ lysates has been reported 11 • 12 , although the topology of proteins expressed in mammalian cells is better predicted by translation in reticulocyte lysates 12 . But we find no evidence for transmembrane domains when apo B 15 is translated in wheat-germ lysates and the microsomes are digested with trypsin (results not shown) or proteinase K (Fig. 3d) . Using densitometry to quantitate full-length chains after digestion we find ( 1) 71 % ± 13 s.d . ( n = 9) recovery of apo B17 in reticulocyte lysates with trypsin and (2) 86% ±31 s.d. recovery ( n = 7) of apo BI 5 in wheat-germ lysates with proteinase K. In contrast, we find 0% (n = 4) recovery of apo 815 or B 17 in saponin-disrupted microsomes after proteinase K digestion. Although we routinely included the membrane stabilizer tetracaine 13 in digestion mixtures, we cannot detect transmembrane domains when tetracaine is absent but Ca2+ is present (results not shown)1; in contrast, substantially reduced levels of putative intermediates have been observed with tetracaine present (S. L. Chuck and V. R. Lingappa, unpublished), which suggests that the data in ref. 1 may be a result of overdigestion.

We conclude that apo Bis cotranslationally translocated into microsomes and binds to the inner leaflet of membrane. The binding of the amino terminus of apo B forms a focus for association of the remainder of the elongating protein with the membrane, circumventing the problems of size and hydrophobicity. It is probable that this amphipathic form of apo B NATURE · VOL 353 · 3 OCTOBER 1991 © 1991 Nature Publishing Group was translated in the presence (tracks 1, 2, 3, 6 ) or absence (4, 5) of microsames (m). To samples 4 and 5, microsomes were added post-translationally (pt). Samples were digested with trypsin (Trp) in the presence (track 6) or absence (tracks 2, 3, 4, 5) of detergent (Tx) or were undigested (track 1). Apo B9 is protected if cotranslationally imported into microsomes. Protected intermediates of synthesis (relative molecular masses (M,) 38,000 and 40,000 (38K and 40K)) are apparent in some translations, but do not affect the interpretation of the experiment. The yield of apo B9 is stimulated by microsomes (tracks 1 and 4) . When translation is carried out in the absence of microsomes (tracks 4, 5) peptidyl tRNA is apparent at -25K. B9.CHO, glycosylated apo B9. b, Apo B7 was translated in the presence (tracks 2, 3) or absence (1, 4) of microsomes, with (track 3) or without (tracks 1, 2, 4) 100 µM N-acetyl-AsnTyrThrcarboxyamide (NYT) 6 . The peptide inhibited glycosylation of apo B7 to generate a protein species (B7) of faster migration than pre-apo B7. indicating cleavage of the MetAla signal sequence. B7.CHO, glycosylated apo B7. METHODS. To generate the apo B9 transcript, construct EB9 was linearized with Sa/I and transcribed with T7 RNA polymerase (Boehringer). To generate the apo B7 transcript, construct EB9 was linearized at the Stul (1,199) subsequently buds into the lumen, carrying phospholipid and neutral lipids, to form the nascent lipoprotein particle 9 • D

Proc. natn. Acad. Sc

Goldstein and M. de Cuevas for the HD plasm id and 0. A. Deryabina for technical assistance

University of Cambridge) for advice on in vitro translation, J. Arms trong and G. Warren (ICRF, London) for the cDNA of the E1 glycoprotein, P. Hodges for his interest. T. Knott for synthesis and oligonucleotides, P. Byfield for protein sequencing, S. Pemberton for peptide synthesis. and I. Haywood for secretarial assistance

collected by centrifugation before (track 1) or after digestion wit h proteinase K for O min (track 2) or 90 min (track 3). Some partial synthesis products are present in all tracks but no new bands are generated by protease digestion. METHODS. Apo 815 capped transcripts were prepared by linearizing construct GB15 with Sa/I and transcribing with T7 polymerase in the presence of 0.5 mM GTP with 5 mM 7-methyl GpppG (Pharmacia), and fractionated on a Sephadex-G25 spin column (Pharmacia). Fractionated wheat germ lysate (Amersham) was reconstituted to 120 mM K+, with a 1 mM mixture of unlabelled amino acids excluding Met. Proteinase K digests were diluted with 2 volumes of 250 mg mI-1 BSA in 0.15 M NaCl with 5 mM PMSF. Membranes were collected in an airfuge and washed 3 times with NaCl solution before dissolving in SDS loading buffer plus 5 mM PMSF for SDS-PAGE. In some instances membranes were disrupted with 0.05% saponin before protease digestion. Proteins were resolved on 8-15% or, for c, on 7-17% SDS-PAGE polyacrylamide gradient gels. Subcloning: The 50-nucleotide 5' UTR of human /3-globin containing the natural Nea l site at the first A TG of the coding sequence 20 was synthesized as oligonucleotides with a 5' Sstl site and a 3' Sstll site to enable it to be cloned into the polylinker of plasmid pKS. Pre-apo 815 was constructed by ligation of Neal" Hindlll (2,279) fragments of cDNA to generate plasmid GB15.