Protein transfer across microsomal membranes reassembled from separated membrane components

Nature Vol. 273 15 June 1978

569

Larger uncertainties, perhaps o f the order o f 0.1 A , should be expected f o r the hydrogen a t o m positions. Several structural features o f the bilirubin bis anion d o not differ substantially f r o m those o f the free acid in the crystal 1 1 1 2 . T h u s , the ridge tile shape is confirmed with a dihedral angle o f 98° between the two oxodipyrromethene moieties which are both planar in a cisoid arrangement ( s y n - Z configuration). Values o f b o n d lengths (Fig. 3) indicate an essentially double and single b o n d order for C 4 - C 5 (and C I S C I 6 ) and C 5 - C 6 (and C 1 4 - C 1 5 ) respectively. Bond distances and bond angles are in g o o d agreement with those determined in a single o x o d i p y r r o m e t h e n e compound 2 ' 1 . In the chloroform molecule and in the i s o p r o p y l a m m o n i u m ion values o f bond lengths and angles are also in the normal range. T h e c a r b o n - o x y g e n bond lengths in the carboxyl group are approximately equal, corresponding to a carboxylate form. T h e nature o f l a c t i m - l a c t a m tautomerism in bilirubin is contro­ versial 7-25 . In this study the values o f carbon nitrogen and c a r b o n - o x y g e n bond lengths in the terminal rings, the experi­ mental evidence in these rings o f a hydrogen bound to the nitrogen a t o m , and the large value observed for the C 4 C 5 - C 6 b o n d a n g l e — a s a consequence of H • • H repulsion in the oxodipyrromethene portion (see Fig. 3) allow the lactam form to be assigned.

O C ® N © 0 oH.

Fig. 3 The oxodipyrromethene moiety in the bilirubin-isopropylamine-chloroform complex. T h e hydrogen b o n d pattern is o f particular relevance (see Fig. 2). In contrast to bilirubin 11 - 12 , in the present c o m p l e x the c o n f o r m a t i o n o f the pigment molecule is stabilised by only two pairs o f N H • • • O intramolecular bonds, each pair involving one carboxylate oxygen. Each o f the remaining carboxylate oxygens is engaged not only in strong hydrogen bonds with two i s o p r o p y l a m m o n i u m ions but also in a weak C H • • O hydrogen b o n d with c h l o r o f o r m . T h e shortest H bond of the complex involves the lactam oxygen a t o m and the third hydrogen o f the i s o p r o p y l a m m o n i u m ion. T h u s the structure of the d i - i s o p r o p y l a m m o n i u m bilirubinate might be viewed as a model for bilirubin-protein associations. W e thank the University o f G e n o v a and the Italian C N R for financial support. A N G E L O M U G NOLI

Istituto di Chimica Fisica, Universita di Genova, Palazzo delle Scienze, Corso Europa, 16132 Genova, Italy PAOLO MANITTO DIEGO MONTI

Istituto di Chimica Organica, Universita di Milano, and Centro CNR per le Sostanze Organiche Naturali, Via Saldini 50, 20133 Milano, Italy

Received 24 February; accepted 23 March 1978. 1. Schmid, R . in Jaundice (eds G o r e s k y , C . A . & Fischer, M . M . ) (Plenum, N e w 2.

I

i&fl?* M c D o n a g h , A . F . Ann. N Y.Acad. Sci. 244, 533-552 (1975). £

,

e

^

r S S K S & ' o r , W . ) 309-335 (P.enum,

5. E t t e r - K j e l s a a s , H . & K u c n z l c . C . C . Biochim. biophvs. Acta 400. 83-94 (1975). 6. Brodersen. R . & T h e i l « a a r d , J . Scand. J. din. Lab. Invest. 24. 3 9 5 - 3 9 8 (1969). 7. K u e n z l e . C . O , W e i b e l , M . N.. P e l l o n i . R . R . & Hemmcrich, P. Biochem. J. 133, 3 6 4 - 3 6 8 (1973). 8. M a n i t t o , P . . S e v e r i n i - R i c c a . G . & M o n t i , D . Gazzetta Chim. Ital. 104, 633-637 (1974), 9. K n e l l , A . J . , H a n c o c k , F. & H u t c h i n s o n , D . W . in Metabolism and Chemistry of Bilirubin and Related Tetrapyrrds (eds Bakkan, A . F. & F o g , J . ) 234-240 (Pediatric Research Institute, O s l o , 1975). SO, M a n i t t o , P. & M o n t i , D . J. chem. Sue. chem. Commttn. 1 2 2 - 1 2 3 (1976). 11. B o n n e t t , R . D a v i c s . J . E. & H u r s t h o u s e . M . B. Nature 262. 3 2 6 - 3 2 8 (1976). 12. Le Bas, G . , Allegret, A . & de R u n g o . C. 4th European Crvstallographic Meeting Oxford (1977)/ 13. Billing, B. H . in The Liver (cd. Gall. A . ) ch. 1 (Williams and W i l k i n s , Baltimore, 1973). 14. K u e n z l e . C . C , G i t z e l m a n n - C u m a r a s a m y , N . & W i l s o n , K . )...!. bid. Chem. 251, 8 0 1 - 8 0 7 (1976). N is. 1. M & J a k o b y . W , B. 15. K a m i s a k a . J . K... H a b i g , W . H . . Ketley, J . Eur. J. Biochem. 60, 155 161 (1975). Z . & Arias, t. M . Biochemistry 14, 16. Kamisaka. J , K . . Listowsky. I., Gatmaitan 2 1 7 5 - 2 1 8 0 (1975). 17. T i p p i n g , E . , K e t t e r e r , B. & Christodoulides. L . . Biochem..I. 157,211-216(1976). 18. J a c o b s e n . J . Int. J. Peptide Protein Res. 9. 2 3 5 - 2 3 9 (1977). 19. Jacobsen. C . Eur. J. Biochem. 27, 5 1 3 - 5 1 9 (1972). 20. J a c o b s e n , C . Int. J. Peptide Protein Res. 7. 159-163 (1975). 21. M c D o n a g h . A , F. & A s s i s i , F . FEBS Lett. 18, 315 317 (1971). A 27. 3 6 8 - 3 7 6 22. G e r m a i n , G . , M a i n , P. & W o o i f s o n . M . M . Acta crystallogr. (1971). 23. Shcldrick. W . S. J. C. S. Perkin II 1457-1462 (1976). 24. C u l l e n , D . L , et at. Tetrahedron 33, 4 7 7 - 4 8 3 (1977). 25. S e v e r i n i - R i c c a , G... M a n i t t o , P., M o n t i , D . & Randall, E. V. Gazzetta Chim. Ital. 105, 1 2 7 3 - 1 2 7 7 (1975).

Protein transfer across microsomal membranes reassembled from separated membrane components THE synthesis o f secretory proteins is initiated o n ribosomes in the cell cytoplasm and the first part o f the polypeptide chain to appear is a short sequence termed the signal sequence 1 . T h i s signal is thought to direct the ribosomal complex to the endoplasmic reticulum ( E R ) membrane where the synthesis o f the rest o f the secretory protein is tightly coupled to its transfer across the membrane into the cisternal space. In all but one case, it is clear that the polypeptide chain is processed during transfer to remove the signal sequence 2 - 3 . Little is k n o w n ab o ut the m e m b r a n e proteins involved in conveying the growing polypeptide chain across the membrane. A classical approach to learning m o r e about these proteins would be to dissect the microsomal membrane into inactive c o m p o n e n t s that can be reassembled subsequently into a functional entity. T h i s w o u l d allow the purification o f c o m p o n e n t s involved in protein transfer. Here we report the first such reassembly. Messenger R N A coding for light-chain i m m u n o g l o b u l i n (Ig) was translated in a wheat germ cell-free system in the presence o f RNase-treated rough microsomes ( R M R ) f r o m canine pancreas 5 . T h e authentic Ig light chain ( L i ) was synthesised, together with a small a m o u n t o f the light-chain precursor ( p - L i ) containing the signal peptide (Fig. 1, track 1). T h e authentic light chain was resistant to added proteases and its precursor was completely degraded (Fig. 1, track 2). Since added proteases degrade all but those proteins inside the vesicles, the transfer o f the light-chain precursor into the vesicles is tightly coupled to the removal o f the signal peptide. I n the absence o f membranes, n o authentic light chain was synthesised (Fig. 1, track 3) and the precursor was completely degraded by added proteases (Fig. 1, track 4). T h e precursor synthesised in the presence o f R M R membranes was < 1 0 % o f the total Ig Li and could be increased by limiting the a m o u n t o f membrane, and hence the number o f sites, available for protein transfer. In all the experiments described the total membrane surface area was adjusted (using the phospholipid content o f the membrane) to the same level, which, for R M R membranes, was sufficient to transfer a n d process about 9 0 % o f the total Ig Li synthesised. W h e n R M R membranes were washed with 0.5 M KC1, the resulting R M R K membranes had almost lost the capacity to transfer and process the Ig Li. Precursor was synthesised together with o n l y a small a m o u n t o f the authentic light chain (Fig. 1, track 5), and it was completely degraded b y added proteases (Fig. 1, track 6). T h e small a m o u n t o f authentic light

New Y o r k , 1976). Macmillan Journals L t d 1978

Nature Vol. 273 15 June 1978

570 c h a i n w a s protected against digestion by protease a n d represented the residual transfer activity o f the R M R K membranes. P r o t e i n transfer a n d processing c o u l d be fully restored to R M R K membranes b y readdition o f the salt extract ( S E ) ; the authentic light c h a i n was synthesised together with s o m e precursor (Fig. 1, track 7) but o n l y the former was resistant to a d d e d proteases (Fig. 1, track 8). T r e a t m e n t with high salt thus removes, reversibly, m e m b r a n e c o m p o n e n t s essential for protein transfer and processing. T h e active c o m p o n e n t s in the S E were characterised using a rapid, quantitative assay (see legend t o T a b l e 1) devised to facilitate their eventual purification. T h e assay determines the percentage o f total protein synthesised that is resistant t o a d d e d proteases because o f transfer into vesicles. W e used canine pancreatic m R N A in this assay since it codes m a i n l y for secretory proteins a n d 4 0 - 5 5 % o f the protein synthesised is resistant to protease digestion. A t least 8 0 % o f the pancreatic m R N A used in these experiments codes for secretory proteins ( u n p u b l i s h e d observations), so this level o f resistance suggests that a fraction o f the microsomal vesicles, containing newlysynthesised secretory proteins, are leaky to added proteases

Protease Membrane R M R

+

+

+ RM

S E protein (/ug)

Fig. 2 Titration of the salt-washed microsomal membranes with the salt-extracted membrane proteins. Incubations were carried out as described in Fig. 1 and the quantitative assays as described in Table 1. The salt extract was concentrated by ultrafiltration using a Minicon with a B15 membrane (molecular weight cutoff = 15,000) and varying amounts were added to R M R K membranes (50 ug membrane phospholipid per 100 ul assay).

78

56

34

12

0.1 0.5 1.0 5.0 10.0 50.0

R K

+ RM

R K +

SE

p-Ti-

Fig. 1 Effect of KC1 treatment of canine rough microsomes on the transfer and processing of light chain immunoglobulin. Rough microsomes ( R M ) from canine pancreas were prepared as described previously 1 , and suspended to I mg membrane phospholipid ml 1 in 20 m M H E PES PH 7.5 (at 20 °C), 110 m M potassium acetate, 3 m M magnesium acetate and 2 mM dithiothreitol (microsomal buffer). CaCl 2 was added to 1 m M and micrococcal nuclease (Boehringer) to l O u g m l " 1 followed by incubation at 20 °C for 15 min to remove endogenous m R N A . T h e nuclease was inhibited by chelating Ca 2 + with E G T A added to 2 m M , the membranes washed by pelleting through 0.5 M sucrose, 20 m M H E PES pH 7.5 at 60,000^ for 20 min at 4 °C and resuspended in microsomal buffer containing 1 mM E G T A . Aliquots of these R M R membranes containing 50 ug of membrane phospholipid (90 ug protein) were used in each 100-w" assay. For the salt treatment, R M R membranes were suspended in ice-cold 0.5 M KC1, 20 m M H E P E S pH 7.5 to 10 mg membrane phospholipid ml""1 and pelleted as above. The pellet ( R M R K membranes) was resuspended in microsomal buffer and 50 ug membrane phospholipid (80 ug protein) was used in each 100-ul assay. In a parallel experiment, the sucrose cushion was omitted and the mixture centrifuged at 100,000# for 1 h at 4 °C. T h e supernatant salt extract (SE) was retained and 8 ug of protein was used in every 100 ul assay. The R M R , R M R K and S E preparations were frozen in liquid nitrogen and stored at — 80 °C in small aliquots. The assay for transfer and processing has been described in detail previously 1 . Briefly, m R N A was isolated from a M O P C 41 tumour secreting Ig light chain, and translated in a cell-free wheat germ system containing 35Smethionine in the presence and absence of microsomal membranes. After incubation at 25 °C for 1 h, samples were either immunoprecipitated with antibody to light chain, or treated with proteases to determine the nature of the light chain transferred into the microsomal vesicles. Aliquots from both treatments were analysed by S D S - P A G E . The upper band is the precursor light chain (P-Li) and the lower band, derived from the precursor by proteolytic removal of the signal peptide, is the authentic light chain (Li) of immunoglobulin that would be secreted in vivo.

( c o m p a r e ref. 6). T h e percentage o f leaky vesicles in any m e m b r a n e preparation, however, was unaffected b y the m e m b r a n e treatments we used. R M R m e m b r a n e s containing r a d i o l a b e l e d secretory proteins were subjected t o these various treatments a n d 4 0 - 5 0 % o f the proteins synthesised were f o u n d to be resistant to a d d e d proteases. M e m b r a n e disruption is therefore not the cause o f s o m e o f the results presented here. R M R membranes protected 3 9 % o f the total proteins synthesised a n d salt washing reduced this to 1 2 % . A p p r o x i mately half o f the protection afforded b y R M R K m e m b r a n e s to pancreatic secretory proteins can be attributed t o residual transfer activity ( c o m p a r e with Fig. 1, tracks 5,6 for the Ig L i ) ; the other half is a feature o f the assay a n d represents newlysynthesised proteins associated with m e m b r a n e vesicles so as to a l l o w o n l y partial digestion b y a d d e d proteases. T h i s small a m o u n t o f protein d i d not v a r y significantly f r o m experiment to experiment and d i d not affect interpretation o f the results. R e a d d i t i o n o f S E to R M R K m e m b r a n e s at the beginning o f the incubation restored the original level o f protection, whereas readdition at the end h a d n o effect. I n the absence o f a n y other membranes, S E alone was unable t o protect the secretory proteins ( T a b l e 1, experiments 1-5). Proteins are the active c o m p o n e n t s in the SE. D i a l y s i s to r e m o v e small molecules did not impair its ability t o restore transfer activity to R M R K membranes, whereas heat o r trypsin treatment destroyed it ( T a b l e 1, experiments 6 - 8 ) . R N A does not seem to be an active c o m p o n e n t because the extract was derived f r o m RNase-treated membranes. N o phospholipids were detected in the extract. T h e proteins which restore transfer activity t o R M R K vesicles are b o u n d tightly to the membrane. B y varying the a m o u n t o f S E a d d e d to R M R K m e m b r a n e s it was possible to increase the protection f r o m 12 to 55 % at saturating concentrations (Fig. 2). W e estimated that 3.4 ug o f S E protein h a d t o be a d d e d to R M R K membranes (containing 50jag p h o s p h o l i p i d ) t o saturate the m e m b r a n e sites that rebind the S E proteins. F r o m R M R m e m b r a n e s (containing 50 ug p h o s p h o l i p i d ) we obtained 3.1 ug o f S E protein. M o r e than 9 0 % o f S E protein a d d e d t o RMRK membranes must, therefore, h a v e been r e b o u n d . These data also indicate that R M R membranes, as isolated, are nearly saturated with S E proteins. T h e m e m b r a n e proteins extracted with high salt are not solely responsible f o r protein transfer across microsomal membranes because the S E will not restore transfer activity to R M R K membranes that have been trypsin treated, n o r will it confer transfer activity o n sonicated liposomes derived from

Nature Vol. 273 15 June 1978

571

Table 1 Characterisation of the components in the salt extract that restore transfer activity to salt-washed microsomal membranes Experiment

Membrane or lipid*

Treatment

Proteaseresistant protein 0/

1 2 3

R M R membranes R M R K membranes R M R K membranes

A 4

RMHK membranes

5

—

6

R M R K membranes

7

8 9 10 11 12 13 14

J V i v i R K 11

icinui aiicc>

R M R K membranes Trypsin-treated RMRK membranes Trypsin-treated RMRK membranes Sonicated R M lipids Sonicated R M lipids Sonicated egg phosphatidylcholine Sonicated egg phosphatidylcholine

15

R M R K membranes

16 17

S M membranes S M membranes

— 8 jig SE protein added at beginning of incubation 8 j.ig SE protein added at end of incubation 8 ug SE protein 8 ug SE dialysed SE protein 8 jig heat-treated SE protein 8 ug trypsin-treated SE protein

39 12 43 4 1 42 11

11 5

8 ug SE protein

5

8 jig SE protein

1

1

3 8 jig SE protein 8 jig SE protein from SM membranes 8 jig SE protein

3

R M lipids or egg phosphatidylcholine ( T a b l e 1, experiments 9-14). T h e membrane proteins in the S E are not f o u n d in E R membranes devoid o f attached ribosomes in vivo. S m o o t h microsomal ( S M ) membranes, substantially freed o f R M membranes, were extracted with 0.5 M K C I a n d this salt extract did not confer transfer activity on R M R K membranes ( T a b l e 1, compare experiments 2 and 15). T h e S M vesicles were themselves unable to transfer secretory proteins ( T a b l e 1, experiment 16) and this inability was not due simply to a lack o f the m e m brane proteins present in the S E from rough microsomes. A d d i t i o n o f this S E to S M vesicles did not result in protein transfer (Table 1, experiment 17). T h e membrane proteins isolated by treating R M membranes with high salt are needed for the transfer o f secretory proteins across the E R membrane. T h e y can restore transfer activity to inactive, but intact, salt-washed vesicles and are normally b o u n d tightly to the membrane. T h e y are thus proteins b o u n d to the cytoplasmic side o f the microsomal membrane a n d probably d o not span it; they m a y , however, be bound to proteins that d o span the bilayer. Their location would suggest that they are involved in the binding o f the signal peptide a n d associated ribosomal c o m p l e x to the membrane. Other m e m brane proteins are necessary for protein transfer but they are more likely to be involved in events subsequent to binding. W e thank Hannelore Heinz for technical assistance a n d Henrik G a r o f f for critical reading o f the manuscript. Canine pancreas was kindly provided by Professor Siegfried Hoyer. GRAHAM WARREN

13 11 14

Incubations were carried out in the presence of 35S-methionine as described in the legend to Fig. 1 except that 4 ug of canine pancreatic m R N A was added to each 100-jil assay instead of light chain m R N A . Unless otherwise stated, membranes, lipids and the salt extract were added at the beginning of the assay and the SE was derived from R M membranes. A t the end of the 1-h "incubation at 25 °C, the samples were cooled in ice and three aliquots taken for assay: In the first, 10-ul aliquot was sampled directly on to filter paper to determine the total protein synthesised; in the second, a 20-jil aliquot was treated with 80 ul of 2 m g m l ~ l proteinase K (Merck) in 50 m M HEPES pH 7.5 and incubated at 25 °C for 10 min, before a 50-ul aliquot was sampled on to filter paper; in the third, a 20-jil aliquot was treated with 80 ul of 2 mg ml" 1 proteinase K in 50 m M HEPES pH 7.5, 15 m M deoxycholate and incubated as in the second. All filter papers were treated according to the method of Mans and Novelli 8 . T h e proteinase K degrades nearly all of the proteins not inside the vesicle to such a small size that they do not precipitate on the filter paper when soaked in 10% T C A . The background ^Smethionine bound to the filter paper after proteinase K treatment in the presence of deoxycholate, was always < 5 % of the 35S in the total synthesised protein and was substracted from the total counts and those obtained in the presence of proteinase K alone. These corrected counts were used to determine the percentage of proteaseresistant protein[( 35 SCounts resistant to proteinase K/Total i a S counts) x 100]. In the absence of R M R , R M R K and SM membranes, the total 35 S -counts in the newly-synthesised proteins was typically I5i),uuu 4-20,000 c.p.m. In the presence of these membranes and irrespective of the various treatments, the total 35S counts were between 90 000 and 120,000 c.p.m. Smooth microsomes were prepared by the method of Adelman et aV and were extracted with 0.5 M K C as tor R M , membranes in Fig. I. Dialysis of the SE was earned out ;against 2 000 vol 0 5 M KCI, 20 m M HEPES pH 7.5 at 4 °C for 15 h. Heat treatment was carried out on a 50-ul aliquot of SE (0.8 mg prerfein ml" 1 ) at 70 °C for 5 min. Trypsin treatment of RMRK ^ ^ J * ™ carried out by treating 50-ul aliquots with 1 u o f 1 mg ml trypan and incubating at 25 °C for 20 min. Soybean trypsin m h . b t o ( 1 ul of 5 m g m l - 1 ) was then added and the samples used directly in the assay. Controls using a mixture of trypsin andsoybean [ trypsin inhibitor were shown to have no effect on any of t h « ^ " ^ t i o n s . R M lipids were extracted from R M membranes using 20 vol C H U r MeOH (2-1 v/v) and the organic phase removed under a stream ot N 2 a n d then TLuo. Egg phosphatidylcholine was purchased from Sigma Lipids were suspended to 5 mg ml m 50 m M HbKK> pH 7 5 by sonicadon for 5 x 2 min at < 10 °C using a^Branson sonifier with microtip set at 40 W . Phospholipids were assayed by the method of Bartlett9 and proteins by the method of Lowry• et a . . * In all cases the samples contained 50 ug phospholipids.

B E R N H A R D DOBBERSTEIN

European Molecular Biology Laboratory Postfach 10.2209 D6900 Heidelberg FRG

R e c e i v e d 12 D e c e m b e r 1977; a c c e p t e d 13 M a r c h 1978. 1. Blobel, G . & D o b b e r s t e i n , B. / . Cell Biol. 67, 8 3 5 - 8 5 1 , 8 5 2 - 8 6 7 (1975). 2. C a m p b e l l , P . N . & B l o b e l , G . FEBS Lett. 72, 2 1 5 - 2 2 6 (1976). 3. P a l m i t e r , R . D . , T h i b o d e a u , S. N . , G a g o n , J . & W a l s h , K . A . llth FEBS Mtg, Copenhagen, 1977 ( P e r g a m o n , L o n d o n , 1977). 4. Kreibich, G . . P e r e y r a , B. & Sabatini, D . D . Abstr. Tenth Int. Congr. Hiochem., Hamburg, 297 (1976). 5. D o b b e r s t e i n , B. & B l o b e l , G . Biochem. biophys. Res. Commun. 74, 1675 1782 (1977). 6. S a b a t i n i , D . D . & B l o b e l , G . / . Cell Biol. 45, 146-157 (1970). 7. A d e l m a n , M . R . , B l o b e l , G . & S a b a t i n i , D . D . / . Cell Biol. 56, 191-205 (1973) 8. M a n s , R . J . & N o v e l l i , G . D . Archs. Biochem. Biophys. 94, 48-53 9. B ( a n l e U , G . R . / . biol. Chem. 2-14, 466 471 (1959). , „ „ . , . . , 10 L o w r y . O . H . , R o s e b r o u g h , N . J . , F a r r , R . J . & R a n d a l l , R . J . J. hiol. Chem. 193, 265 275 (1951).

Nature Index and Binders T h e complete Index for 1977 is available, price £2.50 (U.K.),US$5.()0 (Rest of World). Copies of the 1976 index are still on sale, prices as above. Binders for the journal are also available at £8.00 ( U K ) , US$16.00 (Rest of World) for three: a year of Nature fits into three binders. A l l prices include postage. Payment may be made in any currency at the prevailing exchange rate. Orders should be sent, accompanied by remittance, to Macmillan Journals Ltd, Brunei Road, Basingstoke, Hampshire, R 0 2 1 zxs>. England.