THE JOURNAL OF BIOLOGICAL CHEMISTRY (0 1987 by The American Society of Biological Chemists, Inc.

Vol. 262, No. 21, Issue of July 25, pp. 10189-10194,1987 Printed in U.S.A.

A Functional Interaction between the Signal Peptide and the Translation Apparatus Is Detected by the Use of a Single Point Mutation Which Blocks Translocation across Mammalian Endoplasmic Reticulum*

(Received for publication, December 17,1986)

Ibrahim IbrahimiSS and Reiner Gentzn From the $European Molecular Biology Laboratory, Postfach 10.2209, Meyerhofstrasse 1, 0-6900 Heidelberg, Federal Republic of Germany and the ITF. Hoffmann-La Roche and Co. A.G., ZFE, Grenzacher Strasse 124, CH-4002 Basel, Switzerland

A functional interaction between the signal sequence and the translation apparatus which may serve as a first step in chain targeting to the membrane is de- scribed. To this end, we exploited the powerful tech- nique of molecular cloning in a procaryotic system and the well characterized translocation system of mam- malian endoplasmic reticulum. The signal peptide of subunit B of the heat labile enterotoxin of Escherichia coli (EltB) was fused to several proteins. Single base substitutions were introduced in the signal peptide and their effect on protein synthesis and translocation was studied. We sought a single amino acid substitution which may define certain steps in the coordinated reg- ulation of chain synthesis and targeting to the mem- brane. The substitution of proline for leucine at residue -8 in the signal peptide abolished all known functions of the signal peptide. In contrast to wild type signal peptide, the mutant signal peptide did not lead to arrest of nascent chain synthesis by signal recognition parti- cle or translocation of the precursor protein across the membrane of the endoplasmic reticulum. Furthermore, the mutant signal peptide was not cleaved by purified E. coli signal peptidase. Interestingly, the mutation resulted in about a 2-fold increase in the rate of syn- thesis of the precursor protein, suggesting a role for the signal peptide in regulating the synthesis of the nascent secretory chain as a means of ensuring early and efficient targeting of this chain to the membrane. This role might involve interaction of the signal pep- tide with components of the translation apparatus and/ or endogenous signal recognition particle. These re- sults were obtained with three different fusion proteins carrying the signal peptide of EltB thus leading to the conclusion that the effect of the mutation on the struc- ture and function of the signal peptide is independent of the succeeding sequence to which the signal peptide is attached.

Typical secretory proteins are synthesized with a variable extension of amino acids at the amino terminus called the signal peptide. This peptide directs the translocation of the protein across membrane barriers by serving as a ligand which is recognized by receptors in the membrane. To understand

* The costs of Publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

5 Supported by the Alexander von Humboldt-Stiftung. To whom correspondence should be addressed.

the interactions between the signal peptide and its receptor(s), a process which leads to chain translocation across the mem- brane, structural and functional analysis of the signal peptide and the receptor(s) is under investigation in many laborato- ries.

In mammalian ER,’ a receptor for the signal peptide, called SRP, was isolated (Walter and Blobel, 1980) and detailed analysis of its mechanism of action is progressing rapidly (reviewed in Hortsch and Meyer, 1986). Signal peptidase, the enzyme responsible for cleavage of the signal peptide, has also been isolated as a complex and some of its subunits were suggested to function as a receptor for the signal peptide that is involved in the completion of the translocation process initiated by SRP (Evans et al., 1986).

In prokaryotes, mutation analysis (reviewed in Oliver, 1985) suggested the presence of signal peptide receptors, but only preliminary biochemical evidence was obtained (Muller and Blobel, 1984).

The structural features of a signal peptide which specify recognition by the translocation apparatus and cleavage by signal peptidase have been widely investigated but not pre- cisely identified (von Heijne, 1986; Duffaud et al., 1985). Comparative analysis of the several hundred signal peptides with known amino acid sequence failed to pinpoint the com- mon structural features for a consensus signal peptide. The signal peptides seemed so diverse (Von Heijne, 1985b), and as new signal peptides were reported some of the rules deduced from previous signal peptides had to be modified (see “Dis- cussion”).

In an alternative approach to this problem, mutation anal- ysis was used to delineate the structural features that are essential for the function of a signal peptide. Most mutations in the signal peptide resulted in partial loss of function; only few mutations blocked the secretion process completely (Ken- dall et al., 1986; Iida et al., 1985; Ryan and Bassford, 1985; Kadonaga et al., 1985; Emr and Silhavy, 1983; Inouye et al., 1982; Koshland et al., 1982). These studies pointed to the importance of the overall conformation of a particular signal peptide rather than specific sequence features. The effect of

The abbreviations used are: ER, endoplasmic reticulum; EltB, subunit B of the heat labile enterotoxin of E. coli; SRP, signal recognition particle; RM-KN, salt-extracted and nuclease-treated microsomes; Ifn-y, human y-interferon; Pdgf, platelet-derived growth factor (v-sis); Cat+ and Dhfr+, bacterial chloramphenicol acetyltrans- ferase (Cat) and dihydrofolate reductase (Dhfr) with 18 and 5 extra amino acids at the amino terminus, respectively; Sp-Cat+, Sp-Dhfr’, Sp-Ifn-y, and Sp-Pdgf, fusions between the EltB signal peptide and each of Cat’, Dhfr+, Ifn-7, and Pdgf, respectively; SDS-PAGE, so- dium dodecyl sulfate-polyacrylamide gel electrophoresis.

10189

10190 Signal Sequence Recognition in Protein Synthesis and Translocation

mutations on the function of signal peptides has been studied almost exclusively in uiuo and mainly in prokaryotic systems. The analysis of protein translocation in such systems is complicated by the inherent problems of fractionating cellular compartments and the many unwanted side reactions (Tom- massen et al., 1985).

In this report we describe the effect of a single amino acid substitution in the signal peptide on translocation across ER membranes in an eucaryotic in vitro translation-translocation system. The substitution of histidine for glutamine at residue -20 gave a phenotype which was indistinguishable from wild type, whereas the additional substitution of proline for leucine at residue -8 resulted in a mutant signal peptide which failed to display any of its expected functions. We describe, for the first time, a possible functional interaction of the signal peptide with components of the translation apparatus in addition to its well established interaction with components of the translocation apparatus (Walter et al., 1981). This interaction might play a role in regulating the synthesis of nascent secretory polypeptides in a manner which ensures efficient and early targeting to the membrane.

MATERIALS AND METHODS

Escherichia coli RNA polymerase and 7-mGpppA were from Phar- macia P-L Biochemicals, [35S]methionine (1000 Ci/mmol) was from Amersham Corp., England and placental ribonuclease inhibitor was from Bethesda Research Laboratory. Cell-free lysate from wheat germ, rough microsomes, salt-washed and nuclease-treated micro- somes (RM-KN), and SRP were prepared and used as previously described (Walter et al., 1981). Purified E. coli leader peptidase was a generous gift from R. Zimmerman, Miinchen.

Plasmid Construction-Plasmids pSP-CAT1, pSP-DHFR1, and pSP-IFN-yl carry the nucleotide sequence coding for the signal peptide of the heat labile enterotoxin of E. coli (EltB) fused to the coding regions of each of bacterial chloramphenicol acetyltransferase, mouse dihydrofolate reductase, and human y-interferon (Ifn-y), re- spectively. These plasmids were constructed using synthetic oligo- nucleotide coding for the signal peptide of EltB, cDNA coding for each of the proteins, and derivatives of plasmid pDS5/3 (Stueber et al., 1984, Bujard et al., 1987). Two mutations in the signal peptide were isolated during the construction of pSP-DHFR1. The first mutation was designed to create an Sphl site at the beginning of the signal peptide and the second mutation was spontaneous. The mutant signal peptides are referred to as Sp2 and Sp3 and the wild type as Spl. Spl and Sp2 were phenotypically identical. The coding regions of bacterial chloramphenicol acetyltransferase and Ifn-y were cloned in fusion to the coding region of the Sp3 mutant signal peptide in addition to Spl. Diagramatic illustration of the relevant regions of the plasmids (Fig. 1A) and the sequence of the wild type and mutant signal peptides (Fig. 1 8 ) are shown. Details of the construction

A pSP-CAT1 _ _ _ P/O RES sp 1' rat ".

pSP-OHFR1 P/O RBS sp )r dhfr

P I 0 RES sp I I , .... ."

p S P - I F N - g l : Ifn-5

... , I ".

B S p l A T G A A T A i A G T A A i A TGT T A T GTT T T A T T T ACG GCG T T A C T A TCC TCT C T A T A T & C A T eL)GCT-.~ M N K V K C Y V L F T A L L S S L Y P H G A

sp2. (*T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H

Sp3 C A T ' ' ' ' ' ' ' . . . . . ..c[A . . . . . . . . . . . . . . . H P

FIG. 1. Diagramatic representation of the relevant regions of the plasmids ( A ) and the sequence of the wild type and mutant signal peptides ( B ) . The plasmids were constructed and the nucleotide sequence determined as described under "Materials and Methods." The plasmids called pSP-CAT1, pSP-DHFRl, and pSP-IFN-yl carry the cat, dhfr, and ifn-y genes fused to the coding region of the signal peptide of EltB, respectively. The symbols stand for promoter-operator region (P/O), ribosomal binding site (RBS), signal peptide (sp) , and joining region ( j r ) . The sequence of the wild type (spl ) and mutant (sp2 and sp3) signal peptides are shown in B. The standard one-letter abbreviations of amino acids are used.

procedures and isolation of mutants in the signal sequence will be described elsewhere?

Protein Synthesis ana' Translocation Assays-Plasmid DNA was transcribed by purified E. coli polymerase in the presence of the capping structure 7-mGpppA as previously described (Stueber et al., 1984). The resulting mRNA was translated in the wheat germ lysate (Bujard et al., 1987). [36S]methionine was used at 1 pCi/pl of trans- lation mixture. For synchronized translation assays, all components of the translation except mRNA were mixed. The mixture and the mRNA were warmed at 25 "C for 2 min and then mixed together. Translation was allowed to proceed for 2 min at which time further initiation was blocked by the addition of 7-mGp to a final concentra- tion of 8 mM. The incubation was continued and aliquots were taken at the indicated intervals and precipitated with cold 10% trichloroa- cetic acid. For protein translocation, SRP and/or RM-KN were included in the translation mixtures a t a concentration of 0.08 Am/ ml for SRP and 0.8 Am/ml for RM-KN. This amount was considered as 1 unit in each case. Blanks containing all compensating buffers and ions were added to control translations as indicated in figure legends.

Posttranslational Treatments-Where indicated, translation prod- ucts were treated with proteinase K at a concentration of 200 pg/ml in the presence of 10% sucrose. Digestion was allowed to proceed for 60 min on ice after which it was stopped by the addition of phenyl- methylsulfonyl fluoride to a final concentration of 2 mM. Processing by purified E. coli signal peptidase was carried out on a sample of the translation mixture in the presence of 25 mM Tris-HC1, pH 8.0, and 0.025% SDS. 1 pl of the enzyme preparation which was suspended in 0.7% octyl glucoside was used in 10 p1 of digestion assay and incu- bation was carried out for 30 min at 37 "C.

Analysis of Translation Products-All samples were precipitated with 10% trichloroacetic acid on ice. The pellets were solubilized in sample buffer and subjected to SDS-PAGE on slab gels containing 10-15% acrylamide gradient as previously described (Laemmli, 1970). Radioactive protein bands were visualized by fluorography using EN3- HANCE (New England Nuclear).

RESULTS

We have used molecular cloning to investigate the role of the signal peptide in secretory protein synthesis and translo- cation across ER membranes. We sought a single point mu- tation which might define new steps in the targeting of the nascent secretory chain to the membrane. The coding region of the signal peptide of EltB was constructed from overlapping synthetic oligonucleotides and fused to the N25 promotor- operator-ribosomal binding site region of a derivative of plas- mid pDS5/3 (Stueber et al., 1984). This construction was used to fuse several cytoplasmic and secretory proteins to the signal peptide. We have studied the synthesis and translocation of these proteins in both prokaryotic and eukaryotic systems. We found that this signal peptide can direct the translocation of mouse dihydrofolate reductase (Bujard et al., 1987), bacte- rial chloramphenicol acetyltransferase (Ibrahimi and Fuchs, 1987), human Ifn-7 and Pdgf (v-s~s)~ across ER membranes.

Two mutant signal peptides were isolated and their effect on chain synthesis and translocation was studied. One of these mutants involved the replacement of asparagine residue -20 by histidine and was functionally indistinguishable from wild type, and the second mutant, which involved the addi- tional replacement of leucine residue -8 by proline, was completely nonfunctional. Thereafter the term mutant will refer to this nonfunctional signal peptide unless otherwise indicated. We studied the influence of this mutant signal peptide on the polypeptide chain synthesis in the wheat germ cell free translation system and on translocation across ER membranes. Diagramatic representations of the relevant re- gions of the constructions (Fig. 1A) and the sequence of the wild type and mutant signal peptides (Fig 1B) are shown.

B. Gentz, W. Bannwarth, D. Stueber, and H. Bujard, manuscript

Ibrahimi and Gentz, manuscript in preparation. in preparation.

Signal Sequence Recognition in Protein Synthesis and Translocation 10191

Detailed description of the construction procedures and anal- ysis of the constructs by nucleotide sequencing will be pre- sented elsewhere.'

The Replacement of Proline for Leucine in the Signal Peptide Prevents Nascent Chain Arrest by SRP and Processing by the E R Membrane-SRP which was isolated from mammalian E R arrests the synthesis of nascent secretory polypeptides in wheat germ cell-free translation system (Walter et al., 1981). The arrest serves to target the nascent chain to the membrane where the translocation process is initiated (Walter and Blo- bel, 1981). The synthesis of four fusion proteins composed of the EltB signal peptide fused to each of bacterial chloram- phenicol acetyltransferase (Sp-Cat'), (Ibrahimi and Fuchs, 1987), mouse dihydrofolate reductase (Sp-Dhfr'), (Bujard et al., 1987), human Ifn-y (Sp-Ifn-y), and human Pdgf (Sp- Pdgf)3 was found to be arrested by SRP. Here we show that the mutant signal peptide was fused to bacterial chloram- phenicol acetyltransferase, dihydrofolate reductase, and Ifn- y and in all three cases it failed to cause arrest in nascent chain synthesis by S R P (Figs. 2 and 3).

Completion of the translocation process results in segrega- tion of the nascent chain inside the E R microsomes and cleavage of the signal peptide by signal peptidase, to generate the authentic protein. Whereas this was the case for all fusion proteins containing the wild type signal peptide, the three fusions containing the mutant signal peptide were not pro- cessed by the E R microsomes (Figs. 2 and 3). The arrest by S R P and processing by the microsomes for Sp-Dhfr+ with a wild type signal peptide is reproduced here as a positive control (Fig. 2C).

The Mutant Signal Peptide Does Not Direct Translocation across E R Microsomal Membranes nor Is It Cleaved by Signal Peptidase-The process of secretory protein translocation across E R membranes is invisioned to consist of several stages. These stages include recognition of the nascent chain by S R P , targeting to the membrane, integration into the membrane, transfer across the membrane, and release and

I pSP-DHFR3

1 1 - " 1 1 - " 1 1 RM-KN - - - - 1 2 2 1 - 1 2 2 1 - 1 2 2 1 SRP C PSP-DHFR1 B pSP-CAT3

1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 k D

a -30 -sp-cat*

"Sp-Dhfr'

FIG. 2. The mutant signal peptide does not lead to arrest of nascent chain synthesis by SRP or processing by the micro- somes. DNA of the two plasmids pSP-DHFR3 and pSP-CAT3 car- rying the mutant signal peptide, and the plasmid pSP-DHFRI car- rying the wild type signal peptide, was transcribed, and the resulting mRNA was translated in wheat germ lysate. Translations were sup- plemented with no components ( l a n e I) , 1 unit of SRP (lane 2). 2 units of SRP ( l a n e 3), 2 units of SRP plus 1 unit of RM-KN ( l a n e 4 ) , or 1 unit of SRP plus 1 unit of RM-KN ( l a n e 5). Numbers on the right side indicate molecular mass standards in kilodalton (IZD). The arrows point to the expected fusion proteins.

A pSP-DHFR3

- - + + - + + R " K N - - + + - + + - - + + - + + SRP -

c pSP-IFb3 B pSP-CAT3 + - + - + + - + - + - + + - + - + - + +

1 2 3 4 5 6 7 1 2 3 h 5 6 7 1 2 3 4 5 6 7

"" I " "

FIG. 3. The mutant signal peptide does not direct translo- cation across ER microsomes. DNA of the plasmids pSP-DHFR3, pSP-CAT3, andpSP-IFN-y3 carrying the mutant signal peptide was transcribed, and the resulting mRNA was translated in wheat germ lysate. Translations were supplemented with no components (lane I); SRP (lane 2) ; RM-KN (lane 3), SRP plus RM-KN (lane 4 ) . Translation products as in lanes 1 and 4 were treated with proteinase K in the absence (lanes 5 and 6 ) and presence (lane 7) of 0.5% Triton x-100.

segregation of the completed chain inside the microsome. At some undefined stage during this process, the signal peptide is cleaved by signal peptidase which is believed to reside on the luminal side of the membrane. The whole process leads to generation of the authentic protein inside the microsome and as a consequence it becomes resistant to digestion by externally added proteases.

The cleavage of the signal peptide is not an obligatory consequence of or requirement for translocation. Mutations which block signal sequence cleavage but do not interfere with the translocation process have been identified (Russel and Model, 1981; Inouye et al., 1986). In this case, the precursor protein would be protected from protease digestion. We tested this possibility for the mutant signal peptide under investi- gation. Fig. 3 shows that the mutant signal peptide was unable to direct translocation of any of the fusion proteins across the microsomal membrane as evidenced by their sensitivity to digestion by proteinase K. The wild type signal peptide was shown to direct translocation of the fusion proteins across the membrane as evidenced by the resistance of the authentic products to proteinase K digestion in addition to its glycosyl- ation in the case of Sp-Cat (Ibrahimi and Fuchs, 1987) and S p - I f n - ~ . ~ We further investigated the cleavage of the wild type and mutant signal peptides by purified E. coli signal peptidase in a posttranslational and translocation-independ- ent assay. Fig. 4 shows that whereas the wild type signal peptide was cleaved by the signal peptidase, the mutant signal peptide was not. Thus the mutation must have abolished the recognition specificity for cleavage by signal peptidase as well.

The Mutation in the Signal Peptide Results in a Specific Increase i n the Rate of Nascent Chain Synthesis-In the course of studying the translocation of the fusion proteins having the wild type and mutant signal peptides, we observed that the mutant signal peptide resulted in a higher level of protein synthesis compared to wild type signal peptide. This effect was most pronounced in the case of the Sp-Cat' protein which could be due to the slow rate of synthesis of this protein compared to the other fusion proteins (see "Discussion"). Systematic analysis of this effect revealed that the Sp-Cat' protein with the mutant signal peptide was synthesized twice as fast as the Sp-Cat' with the wild type signal peptide (Fig.

10192 Signal Sequence Recognition in Protein Synthesis and Translocation

A B C 1 2 1 2 1 2

1 1 -0 - 1 1 t -

I FIG. 4. Processing of the fusion proteins carrying the wild

type and mutant signal peptides by purified E. coli signal peptidase. The processing reaction was carried out as described under "Materials and Methods." A-C show the results with the fusion proteins Sp-Cat', Sp-Dhfr', and Sp-Ifn-y, respectively. Lanes I and 2 display the results with the mutant and wild type signal peptides in each case.

100 - A 80

- / 60 LO-

-

2 c 2 0 - // 0 - dCO w

4 loo-

= 8 0 -

E B

60

2 0 -

/ 4 0 -

- /:

0 - 0'

0 20 4 0 60 80 Translatmn tlme Imm)

FIG. 5. The mutant signal peptide leads to a specific in- crease in rate of polypeptide chain synthesis. In A , DNA of the plasmid pSP-CAT1 carrying the wild type signal peptide and plasmid pSP-CAT3 carrying the mutant signal peptide were transcribed and the resulting mRNA was translated in wheat germ lysate. Samples of the translation mixture were taken a t the indicated intervals and precipitated with trichloroacetic acid on ice. The samples were sub- jected to SDS-PAGE and autoradiography. The bands corresponding to Sp-Cat' fusion protein were quantitated by densitometry and the normalized values plotted against translation time for the wild type (squares) and mutant type (circles). In B, the same analysis was done on a cytoplasmic polypeptide, ENV(80) (Certa et al., 1986). Plasmid DNA which encodes the polypeptide was transcribed, and the result- ing mRNA was included in the two translation reactions described above. The results are shown for the polypeptide synthesized in the wild type Sp-Cat' translation reaction (squares) and the mutant type Sp-Cat' translation reaction (circles).

5 A ) . The rate of synthesis of a cytoplasmic polypeptide, ENV(80) (Certa et al., 1986) as an internal standard in both translation reactions was the same (Fig. 5 B ) indicating that the effect on the apparent rate of synthesis of Sp-Cat' was specific and caused by the mutation in the region coding for the signal peptide.

The Effect of the Mutation on the Apparent Rate of Poly- peptide Chain Synthesis Is Not at the Level of Transcription or Initiation of Translation-The observed effect of the mu- tation on the rate of polypeptide chain synthesis was not due to a difference in DNA copy number used for transcription, since equal amounts of the two plasmids which have identical sizes were used. To check if the mutation has affected the transcription efficiency of the plasmid DNA, we measured the extent of incorporation of %-UTP in transcription reactions containing equal amounts of the two plasmids. The results indicated that the extent of incorporation was essentially the same (275,975 cpm for the wild type and 287,575 cpm for the mutant), and the profile of radioactive bands was identical (results not shown). These results indicate that the effect of the mutation on the rate of polypeptide chain synthesis is not at the level of transcription. The effect of the mutation on the initiation of protein synthesis was studied in synchronized translation assays. In such a translation, initiation is allowed to proceed for 2 min in the presence of excess mRNA, after which further initiation is prevented by the addition of 7- mGp. Fig. 6 shows a clear difference in the amount of Sp- Cat' protein synthesized at 10 min in the case of the wild type and mutant signal peptides but not in the case of globin whose mRNA was included in both translation reactions. This difference tends to disappear upon further incubation. Thus the effect of the mutation on the rate of polypeptide chain synthesis cannot be attributed to a difference in the initiation of chain synthesis. This leaves chain elongation as the most likely process which is affected by the mutation.

DISCUSSION

A functional interaction between the signal peptide and the translation apparatus which may serve as a first step in secretory chain targeting to the ER membrane is described. We have studied the influence of two amino acid substitutions, in the signal peptide on protein translocation across ER membranes and on secretory polypeptide chain synthesis. The replacement of asparagine residue -20 by histidine did not

A 1 2 3 L 5 6

=m "-

W M W M W M 5' 10' 70'

c

3 1 2

- -Sp-Cat'

"Globin

;ii

W M 10'

FIG. 6. The effect of the mutant signal peptide on the ap- parent rate of polypeptide chain synthesis is not at the level of chain initiation. In A , DNA of plasmids pSP-CAT1 and pSP- CAT3 was transcribed and synchronized translations were done as described under "Materials and Methods." Samples of the translation mixtures were precipitated with trichloroacetic acid at the indicated intervals and subjected to SDS-PAGE and autoradiography. The results are shown for the wild type ( W ) and mutant ( M ) signal peptides. In B, globin mRNA was included in the two translation reactions.

Signal Sequence Recognition in Protein Synthesis and Translocation 10193

alter the function of the signal peptide, whereas the additional substitution of leucine residue -8 by proline resulted in a completely nonfunctional signal peptide and a faster rate of polypeptide chain synthesis. Although the comparison of the two mutant signal peptides correlates the loss of signal peptide function with the single amino acid substitution involving the replacement of leucine by proline, we cannot rule out the possibility that the second substitution contributes to this effect also.

The mutation seems to have abolished the functional rec- ognition of the signal peptide by SRP, the translocation apparatus, and signal peptidase as evidenced by the absence of chain synthesis arrest by SRP, translocation by the ER microsomes, or cleavage by signal peptidase. Furthermore, the mutation seems to have relaxed a constraint imposed by the signal peptide on the rate of nascent secretory chain elonga- tion. This effect was shown not to be at the level of transcrib- ing the DNA or initiation of protein synthesis. Therefore, we suggest that on the average the elongation of the polypeptide chain proceeds faster once the signal peptide is rendered nonfunctional by the mutation. Interaction of the signal pep- tide with the translation apparatus has been previously sug- gested, on theoretical grounds, to initiate the translocation process. In the relay helix hypothesis (Von Heijne, 1985a), it was proposed that the signal peptide may get compacted in the hydrophobic channel through the ribosome as an a-helix extending all the way up to the transferase site. This state is probably temporary and the compacted chain gets extruded from the ribosomal channel eventually. We suggest that this compacting process, resulting from the affinity of the hydro- phobic region of the signal peptide to the hydrophobic channel of the ribosome, slows down the process of elongation. The biding of SRP results in cessation of elongation in the heter- ologous wheat germ translation system or it may regulate secretory nascent chain elongation by an unknown mecha- nism in homologous mammalian translation systems. The nonfunctional mutant signal peptide would probably not get compacted in the ribosome and as a consequence, elongation of the nascent chain does not get hindered, and the process of targeting the nascent chain to the membrane does not take place.

The results presented in Fig. 5 and 6 show that the effect of the signal peptide on the rate of polypeptide chain synthesis is most pronounced during the early stages of chain elonga- tion, suggesting that once the signal peptide is extruded from the ribosome, the elongation process proceeds unhindered. To get direct evidence for this suggestion, the rate of incorpora- tion of radioactive amino acids in nascent chains has to be measured in a tightly synchronized translation system using single species of mRNA. To date such analysis has proved to be difficult. First, although the use of single species of mRNA in a translation system is possible, the translation system itself invariably contains endogenous mRNA which contrib- utes to the incorporation even after nuclease digestion. Fur- thermore, the appearance of polypeptides as a result of pre- mature termination or internal initiation poses a major prob- lem in the proposed analysis. Second, in a synchronized translation system, a certain time window is necessary to get a significant degree of chain initiation and subsequent elon- gation. The narrowest time window possible gives an appre- ciable degree of nonsynchrony which tends to expand with elongation time as a result of the nonuniformity of the elon- gation process compounded with the fact that initiation in- hibitors are not 100% effective. Finally, the quantitative pre- cipitation of short polypeptides for determining incorporation of radioactive amino acids is a major problem in this type of

analysis. One is left with the measurement of the rate of appearance of full-length polypeptides which is an average of the overall rate of elongation.

Whereas the interaction of the signal peptide with the ribosome is proposed to be the cause of the observed slowdown in the rate of chain synthesis, we cannot rule out possible interactions of the signal peptide with other components of the translation apparatus or the lysate such as indogenous SRP. However, the presence of either indogenous SRP or SRP arrested peptides in the wheat germ lysate has not been demonstrated. The system we described here could provide a useful assay for looking for indogenous SRP, if indeed the effect we reported here involved interaction between the signal peptide and indogenous SRP.

In view of the complexity of the translation process, only in cases involving minimal sequence differences, can the effect of a signal peptide on elongation be studied. Even in the case we presented here, we cannot rule out the possibility that the single base substitution in the coding region of the signal peptide has altered the folding of the mRNA in a manner that influences its translation efficiency. This possibility is, how- ever, considered to be unlikely since a single base substitution is not expected to cause significant changes in base pairing. A second requirement for observing a significant effect of the signal peptide on polypeptide chain elongation is a slow overall rate of chain elongation. This suggestion is based on the observation that the effect of the signal peptide on the rate of chain synthesis is most pronounced in the case of Sp- Cat+ compared to Sp-Dhfr+ and Sp-Ifn-y, respectively. This difference correlates with a difference in the overall rate of chain synthesis, Sp-Cat' being synthesized a t a rate which is an order of magnitude lower than the rate of synthesis of the other two proteins. Thus the faster the overall rate of chain synthesis is, the less significant is the retardation imposed on this rate by the signal peptide.

The structural features required for a functional signal peptide remain intriguing. Comparative and mutational anal- ysis of known signal peptides produced some rules which had to be changed in light of new data. Proline was assumed to be excluded from residues +1 to -3 (Von Heijne, 1983,1984) but some signal peptides turned out to have proline in this region (Lin and Gross, 1981; Davidson et al., 1982). It was also observed that the introduction of proline at residue -2 in one signal peptide caused a shift in the cleavage site which made proline the ultimate rather than the penultimate residue from the cleavage site (Weisman et al., 1986). In two other signal peptides, proline occurs at reside -2, but the cleavage site was assigned by homology to other signal peptides, in these two cases (Wetekam et al., 1982; Hudson et al., 1983). Amino acid sequence analysis of the amino-terminal ends of the authentic proteins in the latter cases would be required to establish the position of the cleavage site in order to determine whether this observation can be extended to other signal peptides.

Deletion mutants in the center of the hydrophobic region resulted in nonfunctional signal peptides. Analysis of the secondary structure of the wild type, mutant, and revertants which restore signal peptide function pointed to the impor- tance of a stable a-helical conformation in this region (Emr and Silhavy, 1983 and Bankaitis et al., 1984). The introduction of a proline residue in or around this region blocked signal peptide function in one case (Iida et al., 1985) but resulted in a moderately functional signal peptide in another case (Emr and Bassford, 1982).

The presence of proline a t residue -8 in the mutant signal peptide reported here is not exceptional. Proline at this resi- due occurs in several wild type signal peptides (Sherwood et

10194 Signal Sequence Recognition in Protein Synthesis and Translocation

\ * : \

I . ; \ . .

. . . *

V J

0 5 10 15 20 25 Leng th ( res idues)

FIG. 7. Scatter plot of total hydrophobicity versus length of the most hydrophobic segment for the EltB wild type (Spl) and mutant (Sp2 and Sp3) signal peptides, the signal peptides of human growth hormone (Denoto et al., 1981), human pla- cental lactogen (Barrera-Saldana, et al., 1983), rat apolipo- protein E (McLean et al., 1983), and bacteriorhodopsin (Dunn et al., 1981). The plot was done according to Von Heijne, 1986, and it includes, in addition to the above signal peptides (circles), the signal peptide of pre-P-lactamase (squares) as a typical prokaryotic signal peptide which is recognized by the translocation apparatus of both E. coli and ER, 170 eukaryotic signal peptides (dots), and 134 eukaryotic cytosolic proteins (40 amino-terminal residues) (X), taken from Von Heijne, 1986. The two lines define a region encompassing 92% of the signal peptides and 4% of the cytosolic sequences (1 kcal = 4.184 kJ). Note that the circled point which falls on the tower line corresponds to the EltB wild type (Spl) or mutant (Sp2) signal peptides. The circled point below the lower line corresponds to the EltB mutant (Sp3) signal peptide, and that to the extreme lower lejt end of the figure to the bacteriorhodopsin signal peptide.

al., 1979; Martial et aL, 1979, Dunn et aL, 1981; Denoto et al., 1981; Seeburg, 1982; Barrera-Saldana et al., 1983, and McLean et al., 1983). A scatter plot of total hydrophobicity uersus length of the most hydrophobic segment in the EltB wild type signal peptide, the two mutants, and the four signal peptides that have proline at residue -8 is shown in Fig. 7. It is clear that the EltB signal peptide falls on the border line which separates signal peptides from cytosolic proteins. The substi- tution of proline for leucine makes the mutant signal peptide fall well within the region of cytosolic proteins (Von Heijne, 1986). The very short signal peptide of bacteriorhodopsin seems to be an extreme exception. These observations em- phasize the fact that it is the overall folded structure of a particular signal peptide that specifies functionality rather than the nature or position of individual residues. Further- more, the results we reported here suggest that the signal peptide can act as a separate structural and functional unit independent of the type of mature sequence which is attached to it. Most interestingly, the evidence presented suggests that the signal peptide may act as a control domain which regulates the synthesis of the nascent secretory chain by interacting with the translation apparatus in a manner which would ensure efficient and early targeting of the chain to the ER membrane.

Acknowledgments-We are grateful to B. Dobberstein and H. Bu- jard for discussions and comments. We also than G. von Heijne for the help in the scatter plot, D. Stueber for various plasmid constructs, W. Bannwarth for oligonucleotide synthesis, H. Taatjes for excellent technical assistance, and A. Walter for typing the manuscript.

REFERENCES Bankaitis, V. A., Rasmussen, B. A., and Bassford, P. J., Jr. (1984)

Cell 37,243-252 Barrera-Saldaiia, H. A., Seeburg, P. H., and Saunders, G. F. (1983)

J. Biol. Chem. 258, 3787-3793 Bujard, H., Gentz, R., Lanzer, M., Stueber, D., Mueller, M., Ibrahimi,

I., Hauptle, M-T., and Dobberstein, B. (1987) Methods Enzymol., in press

Certa, U., Bannwarth, W., Stuber, D., Gentz, R., Lanzer, M., Le Grice, S., Guillot, F., Wendler, I., Hunsmann, G., Bujard, H., and

Davidson, J. M., Leslie, B., Wolt, T., Crystal, R. G. and Sandberg, L. B. (1982) Arch. Eiochern. Biophys. 281, 31-37

Denoto, F. M., Moore, D. D., and Goodman, H. M. (1981) Nucleic Acids Res. 9,3719-3730

Duffaud, G. D., Lehnhardt, S. K., March, P. E., and Inouye, M. (1985) Curr. Top. in Membr. and Tramp. 24, 65-104

Dunn, R., McCoy, J., Simsek, M., Majumdar, A., Chang, S. H., Rajbhandary, U. G., and Khorana, H. G. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 6744-6748

Emr, S. D., and Bassford, P. J., Jr. (1982) J . Biol. Clwm. 257,5852- 5860

Emr, S. D. and Silhavy, T. J. (1983) Proc. Natl. Acad. Sci. U. S. A. 80,4599-4603

Evans, E. A., Gilmore, R. and Blobel, G. (1986) Proc. Natl. Acad. Sci.

Hortsch, M., and Meyer, D. I. (1981) Znt. Reu. Cytol. 102, 215-242 Hudson, P., Haley, J., John, M., Cronk, M., Crawford, J., Haralam-

bidis, G., Tregear, G., Shine, J., and Niall, H. (1983) Nature 301,

MOUS, J. (1986) EMBO J. 5,3051-3056

U. S. A. 83, 581-585

628-631 Ibrahimi, I., and Fuchs, E. (1987) J . Bacteriol. 169, 1603-1610 Iida, A., Groarke, J. M., Park, S., Thom, J., Zabicky, J. H., Hazelbauer,

G. L. and Randall, L. L. (1985) EMBO J. 4, 1875-1880 Inouye, S., Soberon, X., Franceschini, T., Nakamura, K., Itakura, K.,

3441 and Inouye, M. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, 3438-

Inouye, S., Duffaud, G., and Inouye, M. (1986) J. Biol. Chem. 261,

Kadonaga, J. T., Pluckthun, A., and Knowles, J. R. (1985) J. Bwl.

Kendall, D. A., Bock, S. C., and Kaiser, E. T. (1986) Nature 231,

Koshland, D., Sauer, R. T., and Botstein, D. (1982) Cell 30,903-914 Laemmli, U. K. (1970) Nature 227,680-685 Lin, V., and Gross, J. K. (1981) Proc. Natl. Acad. Sci. U. S. A. 78,

Martial, J. A., Hallewell, R. A,, Baxter, J. D., and Good, H. M. (1979)

McClean, J. W., Fukazawa, C., and Taylor, J. M. (1983) J. Biol.

Mdler, M., and Blobel, G. (1984) Proc. Natl. Acad. Sci. U. S. A. 81,

Oliver, D. (1985) Annu. Reu. Microbiol. 39, 615-648 Russel, M., and Model, P. (1981) Proc. Natl. Acad. Sci. U. S. A. 78,

Ryan, J. P., and Bassford, P. J., Jr. (1985) J. Biol. Chem. 260, 14832-

Seeburg, P. H. (1982) DNA 1, 239-249 Shenvood, L. M., Burstein, Y., and Schechter, I. (1979) Proc. Natl.

Acad. Sci. U. S. A. 76,3819-3823 Stueber, D., Ibrahimi, I., Cutler, D., Dobberstein, B., and Bujard, H.

Tommassen, J., Leunissen, J., van Damme-Jongsten, M., and Over-

Von Heijne, G. (1983) Eur. J. Biochem. 133, 17-21 Von Heijne, G. (1984) J. Mol. Biol. 173, 243-251 Von Heijne, G. (1985a) FEBS Lett. 190, 1-5 Von Heijne, G. (1985b) Curr. Top. Membr. Tramp. 24, 151-179 Von Heijne, G. (1986) J. Mol. Biol. 189, 239-242 Walter, P., and Blobel, G. (1980) Proc. Natl. Acad. Sci. U. S. A . 77,

Walter, P., Ibrahimi, I., and Blobel, G. (1981) J . Cell Biol. 91, 545-

Weisman, L. S., Krummel, B. M., and Wilson, A. C. (1986) J . Biol.

Wetekam, W., Groneberg, J., Leine Weber, M., Wengenrnayer, F.,

10970-10975

Chem. 260,16192-16199

706-708

2825-2829

Science 205,602-606

Chem. 258,8993-9000

7737-7741

1717-1721

14837

(1984) EMEO J. 3 , 3143-3148

duin, P. (1985) EMEO J . 4, 1041-1047

7112-7116

550

C h m . 261,2309-2313

and Winnacker, F. (1982) Gene (Amst.) 19, 179-183