Proc. Natl. Acad. Sci. USA
`Vol. 77, No. 7, pp. 3988-3992, July 1980 '
`Biochemistry
`
`Bacteria mature preproinsulin to proinsulin
`
`(hybrid signal sequences/secretion/signal peptidase/immunoprecipitation/protein processing)
`
`KAREN TALMADGE, JIM KAUFMAN, AND WALTER GILBERT
`
`Department of Biochemistry and Molecular Biology, Harvard University, Cambridge, Massachusetts 02138
`
`Contributed by Walter Gilbert, April 25, 1980
`
`reproinsulin gene into the
`By inserting the rat
`ABSTRACT
`bacterial prepenicillinase ene, we ormed a variety of hybrid
`bacterial-eukaryotic si
`a sefluences attached to proinsulin.
`Among these were the our to owing constructions: rat proin-
`sulin attached to the entire penicillinase signal se$;ence and
`rat preproinsulin fused to all of, to half of, or only to
`first four
`amino acids of the bacterial signal sequence. In all four cases,
`more than 90% of the rat insulin antigen ap ared in the peri-
`plasmic space. By immuno recipitation an determination of
`the amino acid sequences 0 the radiolabeled roducts, we show
`that the bacteria correctly process both the acterial and the
`eukaryotic signal sequences of these hybrid proteins. The
`cleavage of the eukaryotic signal by bacterial peptidase, in this
`case, generates proinsulin.
`
`Secretion is an essential feature of cells The precursors of almost
`all secreted proteins, both eukaryotic and prokaryotic, contain
`an amino-terminal extension (ref. 1; see ref. 2 for review). The
`signal hypothesis (1, 3) proposes that this peptide, the signal
`sequence, serves to bind the protein to the membrane and then
`to lead it acres. Sometime during transport, the signal sequence
`is removed and the preprotein is thereby processed to the ma-
`ture form.
`In bacteria, direct evidence establishes that the signal se-
`quence is essential for transport. Mutations have been described
`for two proteins (4, 5) that lead to the accumulation of the
`mutant product in the cytoplasm as the preprotein. In each case,
`the mutation results in an amino acid replacement in the signal
`sequence. Furthermore, rat proinsulin, attached to a complete
`bacterial signal sequence,
`is efficiently transported (6, 7);
`lacking a signal, it is not (7).
`The mechanism of secretion is quite general. Shields and
`Blobel (8) have used dog pancreas microsomes to segregate and
`process fish preproinsulin. Moreover, Fraser and Bruce (9)
`showed that chicken ovalbumin is secreted (50%) from bacterial
`cells when that gene is cloned in bacteria. Ovalburnin is unique
`among secreted proteins studied so far: it does not have an
`amino-terminal extension (10), although it may contain an in-
`ternal signal sequence (11). We have recently shown (7) that
`a normal eukaryotic
`sequence, the rat preproinsulin signal
`sequence, directs the efficient secretion of rat insulin antigen
`in bacteria. Is this eukaryotic presequence processed?
`
`MATERIALS AND METIIODS
`
`Materials. Chicken lysozyme, chicken ovalbumin, sperm
`whale myoglobin, and iodoacetamide were from Sigma; bovine
`proinsulin was a gift of Donald Steiner; human [32-microglo-
`bulin was a gift of Cox Terhorst; H235SO4 (carrier-free) and
`L-[4,5-3H(N)]leucine (50 Ci/mmol; 1 Ci = 3.7 X 101° bec-
`querels) was purchased from New England Nuclear. An IgG
`fraction of anti-insulin antiserum (from Miles) was prepared
`as described by Broome and Gilbert (12). Staphylococcus au-
`
`The publication costs of this article were defrayed in part by page
`charge payment. This article must therefore be hereby marked “ad-
`oertisement" in accordance with 18 U. S. C. §l734 solely to indicate
`this fact.
`
`reus strain Cowan I was heat-killed and formalin-treated by
`the method of Kessler (13) and resuspended (10% volume/
`volume) in NET buffer (50 mM Tris-HCI, pH 7.5/5 mM
`EDTA/0.15 M NaCl) (13).
`Radiolabeling of Proteins. Escherichia coli K-12 strains
`PR13 bearing insulin plasmids p287.47 (which produces protein
`i27/+4), p2-11.1947 (protein il2/-21), p2l8.CB6 (protein
`i4/-21), or pKT4l (a control plasmid with no insulin insert),
`and FMA10/M1357 bearing insulin plasmid p280. 1947 (protein
`i25/-21) [all described by Talmadge et al. (7)] were grown
`overnight in 2YT medium (14) supplemented with thymidine
`at 40 ug/ml for FMAl0. Fifty microliters was inoculated into
`10 ml of S medium (15) supplemented either with thiamine at
`10 pg/ml and thymidine at 40 ug/ml for FMAl0 or with L-
`leucine and L-threonine at 40 ug/ml each for PR13, and then
`grown to OD55o of 0.3. Five millicuries of H235SO4 was added
`to all cells (except PR13/p287.47) and incubation was continued
`1 hr with shaking at 37°C (PRl3) or 34°C (FMAIO). PR1-‘3
`bearing p287.47 was harvested, resuspended in 10 ml of S
`medium supplemented with L-threonine at 40 ;.¢g/ml and in-
`cubated for 1 hr at 37°C with shaking with 5 mCi of H235SO4
`and 2.5 mCi of [3H]leucine.
`Immunoprecipitations. Labeled cells were harvested,
`resuspended in 100 ul of Tris-HCI, pH 8/20% sucrose and in-
`cubated l5 min with 1(1) ul of lysozyme at 20mg/ml in 20mM
`EDTA, pH 8. The cells were pelleted by centrifugation for 5
`min at 10,000 rpm in a Sorvall SS-34 rotor, and the supernatant
`was diluted with 800 ul of 150 mM Tris-HCl, pH 8/2% Triton
`X-l00/0.2 M EDTA. Alternately, labeled cells were harvested,
`resuspended in 100 pl of the Tris/sucrose buffer as above, in-
`cubated with 100 ul of lysozyme in EDTA as above, and lysed
`with 800 pl of Triton buffer as above. The cell debris was pel-
`leted at 16,500 rpm for 1 hr in a Sorvall SA600 rotor. A 200- to
`1000-fold excess [as determined by radioimmunoassay (7)] of
`an IgG fraction of guinea pig anti-insulin serum was added to
`each supernatant, and the mixture was held for 1 hr at 37°C and
`then 1 hr on ice. One hundred microliters of heat-killed, for-
`malin-treated S. aureus (10% vol/vol) was added, and the
`mixture was incubated for 30 min on ice and washed by the
`method of Kessler (13).
`Polyacrylamide Gel Electrophoresis. S. aureus bacteria
`complexed to proteins to be analyzed by polyacrylamide gel
`electrophoresis were resuspended in 100 pl of sample buffer
`[200 mM Tris-HCl, pH 6.8/ 10% (vol/vol) glycerol/0.01%
`(wt/vol) bromophenol blue/5 mM EDTA/2% NaDodSO4/
`dithiothreitol (freshly added to 10 mM)], boiled 3 min, allowed
`to cool to room temperature, and incubated for 20 min with 20
`ul of 0.5 M iodoacetamide. Thirty microliters of sample buffer
`made 200 mM in dithiothreitol was added, the room "temper-
`ature incubation was continued another l0 min, and the bac-
`teria were removed by centrifugation. Aliquots (10-50 ul) were
`loaded onto a 15% Laemmli NaDodSO4/polyacrylamide gel
`
`Abbreviation: HPLC, hi h-
`
`3988
`
`rformance liquid chromatography.
`Mylan v. Genentech
`ylan V. Genentech
`IPR2016-00710
`Genentech Exhibit 2133
`
`IPR2016-00710
`
`Genentech Exhibit 2133
`
`

`
`Biochemistry: Talmadge et al.
`
`Proc. Natl. Acad. Sci. USA 77 (1980)
`
`3989
`
`(16) with 7 M urea in the bottom gel and 2 mM EDTA added
`to all buffers. One microgram each of sperm whale myoglobin,
`chicken lysozyme, human 52-microglobulin, and bovine
`proinsulin were nm as molecular weight markers. The stained
`dried gel was autoradiographed on Kodak XR-5 film. S. aureus
`bacteria complexed to proteins whose sequences were to be
`determined were resuspended in 100 pl of Maizel gel buffer
`(17), boiled 3 min, and centrifuged 5 min in a Sorvall SS-34 rotor
`at 10,000 rpm. The supernatant was run on a 15% Maizel gel
`(17), the wet gel was autoradiographed on Kodak XR-5 film for
`1 hr at 4°C, and the protein was eluted from a cmshed gel slice
`for 8 hr at room temperature with shaking in 1-2 ml of 50 mM
`ammonium bicarbonate, pH 7.5/0.2 mg of ovalbumin per
`ml/0.2 mM dithiothreitol/0.1% NaDodSO4. The crushed gel
`was removed by filtration through silicone-treated glass wool
`and the protein was lyophilized.
`Amino Acid Sequence Analysis. The protein isolated from
`a Maizel gel was resuspended in 100 pl of distilled water and
`3 mg of ovalbumin was added. The proteins were precipitated
`in 5 vol of acetone and resuspended in 200 pl of 70% (wt/vol)
`formic acid, and then 3 mg of Polybrene (Aldrich) in 200 pl of
`70% formic acid was added. Between 20,000 and 300,000 cpm
`was loaded onto a Beckman sequenator, updated model 890B,
`and successive steps of Edman degradation were performed,
`using a 0.1 M Quadrol program (18). The amino acid deriva-
`tives were collected after each cycle, dried under streaming
`nitrogen, and converted in 200 pl of 0.1 M HCl at 80°C for 10
`min. Then 20-100 pl was dried in a vacuum oven, resuspended
`in 100 pl of distilled water, and mixed with 2 ml of Aquasol, and
`radioactivity was measured by liquid scintillation counting.
`Fractions with radioactivity were extracted with ethyl acetate
`and the aqueous phase of the 35S-labeled fractions and the ethyl
`acetate phase of all the fractions were analyzed on a Waters
`high-performance liquid chromatography (HPLC) system,
`using an RCSS Radial Pak A (C13) column.
`
`RESULTS
`
`Description of Hybrid Proteins. In a set of plasmid con-
`structions designed to create a series of hybrid proteins, each
`containing a fusion of some portion of a bacterial signal se-
`quence (derived from penicillinase) to some part of a eukaryotic
`signal sequence (derived from rat preproinsulin), four con-
`structions transport more than 90% of the rat insulin antigen
`into the periplasmic space of E. coli (7). Fig. 1 shows the se-
`quences of these four hybrid proteins, named by a lower case
`and a pair of numbers: the first number referring to the last
`
`‘
`Prepenicillinase
`MSIQHFRVALIPFFAAFCLPVFA HPETLVK...
`
`prepenicillinase wild-type amino acid before the amino acids
`encoded by the insertion of the Pst restriction site, the second
`referring to the first amino acid of preproinsulin (negative
`numbers) or proinsulin (positive numbers). Either the complete
`bacterial signal sequence or the major part of the eukaryotic
`signal served to transport efficiently to the periplasm.
`Immunoprecipitation and Polyacrylamide Gel Electro-
`phoresis. We grew cells containing the insulin gene plasmids
`in a low-sulfate medium and labeled the proteins with H2"5SO4
`or with both H235SO4 and [3H]leucine. We isolated the labeled
`protein products from the periplasmic fraction by adding an
`excess of anti-insulin IgG and immunoprecipitating by incu-
`bating with formalin-treated, heat-killed S. aureus (13). Fig.
`2 shows an autoradiogram of the electrophoresis of the immu-
`noprecipitated proteins on a Laemmli NaDodSO4/polyacryl—
`amide gel (16) containing urea and EDTA. A dark new band
`appears in the four samples from insulin-antigen-producing
`cells (Fig. 2, lanes a—d) that is absent in the control precipitation
`(Fig. 2, lane e). Without processing, i25/-21 would have 142
`amino acids, il2/-21 would have 130, i27/+4 would have 121,
`and i4/-21 would have 118. Instead, i27/+4 (Fig. 2, lane a)
`is larger than the other three, which are all the same size and
`run close to, but slower than, the bovine proinsulin standard;
`bovine proinsulin is 5 amino acids shorter than rat proinsulin
`[the deletion is in the C peptide (22)]. The gel mobilities in
`comparison with those of molecular weight standards suggest
`that i27/ +4, which has a complete bacterial signal and no eu-
`karyotic signal, has also been processed. A similar pattern is
`obtained if the proteins are immunoprecipitated from a Triton
`lysate of whole cells (data not shown).
`Amino-Terminal Sequences of Radiolabeled Proteins. To
`verify that these proteins had been processed and to determine
`exactly where they had been clipped, we labeled il2/-21,
`i25/-21, and i4/-21 with H235SO4; i27/+4, with both H235SO4
`and [3H]leucine. After electrophoresis of the immunoprecipi-
`tates through a 15% Maizel gel (17), we autoradiographed the
`wet gel for an hour in the cold and then cut the samples directly
`out of the gel, using the autoradiograph as a template. Auto-
`mated, successive Edman degradations of the radioactive
`protein on a Beckman sequenator, using ovalbumin and Poly-
`brene as carriers, determined the positions of the sulfur-con-
`taining amino acids, methionine and cysteine, in the sequence.
`Fig. 1 shows the amino terminus of each protein. If the three
`candidates with most of the eukaryotic signal sequence (i25/
`-21, il2/-21, and i4/-21) are matured at
`the correct
`preproinsulin clipping site, radioactive cysteine should appear
`
`MSIQHFRVALIPFFAAFCLPVFA HPET
`
`127/+4
`125/-21 MSIQHFRVALIPFFAAFCLPVFA HP
`112/-21 MSIQHFRVALIP
`
`14/-21
`
`MSIQ
`
`AAGGGGGG
`
`ggscccc
`EQGGGGG
`gage
`
`QHLCGPHLVEALYLVCGE..
`WRMFLPLLALLVLWEPKPAQA FVKQHLCGPHLVEALYLVCGE...
`
`WRMFLPLLALLVLWEPKPAQA FVKQHLCGPHLVEALYLVCGE...
`WRMFLPLLALLVLWEPKPAQA FVKQHLCGPHLVEALYLVCGE...
`
`4
`MALWRMFLPLLALLVLWEPKPAQA FVKQHLCGPHLVEALYLVCGE..
`Preproinsulin
`
`FIG. 1. Amino acid sequences of hybrid proteins made as fusions between the prepenicillinase signal sequence and the preproinsulin signal
`sequence, constructed around Psi linkers, as described in ref. 7. Each sequence begins at the prepenicillinase fMet and ends at amino acid 21
`of proinsulin. Each line represents one continous sequence, which has been grouped from left to right to emphasize similarities and differences
`as follows: first group, prepenicillinase signal sequence amino acids; second group, matured penicillinase amino acids; third group, amino acids
`created by the inserted Pst linker (underlined) or by poly(G-C) tailing (glycines); fourth group, preproinsulin signal sequence amino acids; fifth
`group, matured proinsulin amino acids through amino acid 21. The sequence of prepenicillinase is from ref. 19, the sequence of preproinsulin
`is from refs. 6 and 20. The arrows indicate the sites of prepenicillinase and preproinsulin cleavage maturation. A = Ala, R = Arg, C = Cys, Q
`= Gln, E = Glu, G = Gly, H = His, I = Ile, L = Leu, K = Lys, M = Met, F = Phe, P = Pro, S = Ser, T = Thr, W = Trp, Y = Tyr, V = Val.
`
`

`
`3990
`
`Biochemistry: Talmadge et al.
`
`Proc. Natl. Acad. Sci. USA 77 (1980)
`
`:x
`
`ii
`
`3
`
`4i
`
`1
`
`160
`
`4000
`
`2000
`
`”S,cpm
`
`80
`
`350
`
`210
`
`70
`
`Immunoprecipitated rat insulin antigen from E. coli
`FIG. 2.
`strains bearing insulin plasmids, isolated and electrophoresed as
`described in the text. Lane a, i27/+4; b, i25/-21; c, i12/-21; d, i4/-21;
`e, PR13 bearing pKT41, a control plasmid without an insulin insert
`(7). The molecular weight markers, indicated by arrows, are, from top
`to bottom: sperm whale myoglobin (17,200), chicken lysozyme
`(14,400), human 32-microglobulin (11,600), and bovine proinsulin
`(8700). The molecular weight of authentic rat proinsulin is 9100 (21).
`The dye front is indicated, below the arrows, by a dot. The amount
`of material in each lane corresponds to an input of 05 mCi in the la-
`beling. The dry gel was exposed for 12 hr.
`
`at the 7th and 19th residues. Fig. 3 shows this unique pattern
`for all three proteins HPLC of the radioactive fractions proved
`that most of the 35$ radioactivity was originally in cysteine,
`except for the first fractions of il2/-21 (Fig. 3B) and -21
`(Fig. 3C), where the radioactivity was not in any amino acid
`and was probably the result of protein washing out of the se-
`quenator cup (data not shown). To test whether i27/+4 was
`matured at the end of the penicillinase signal sequence, we
`examined the positions of 358- and 31-I-labeled leucine. If the
`bacterial signal has been correctly removed, 358 should appear
`at residue 16, while 3H should appear at residues 15 and 20. Fig.
`4 shows this unique pattern, demonstrafing correct maturation
`of the bacterial signal when it is fused, four -amino acids away
`from the clipping site, to rat proinsulin. Again, HPLC proved
`that the 35S was originally in cysteine and the 3H was in leucine
`(data not shown).
`
`DISCUSSION
`
`These experiments test the ability of bacteria to mature a variety
`of signal sequences fused to rat proinsulin (Fig. 1). Three of
`these hybrid proteins have most of the eukaryotic signal se-
`quence attached to all of, half of, and only four amino acids of,
`the bacterial signal sequence. The sequencing data (Fig. 3)
`demonstrate that, in every case, the bacterial signal peptidase
`recognizes the eukaryotic clipping site and correctly matures
`
`--Q-Q. D--_I--—-_-Iu_
`
`
`
`FVKQHLCGPHLVEALYLVCGE
`20
`5
`10
`15
`Cycle
`FIG. 3. Location of 35S-containing residues in the amino-terminal
`region of the insulin products of three constructions containing the
`DNA encoding the preproinsulin signal sequence. The antigen was
`purified from H-235304-labeled cells by immunoprecipitation and
`NaDodS04/polyacrylamide gel electrophoresis and then subjected
`to automated Edman degradation. The amount of radioactivity re-
`leased by each cycle of degradation was determined by liquid scin-
`tillation counting. The amino-terminal sequence of authentic rat
`proinsulin is presented for comparison. (A) i25/-21: 20,000 cpm
`loaded, double-coupled at steps 1, 2, and 10, double-cleaved at step
`9, 10% of each cycle analyzed. (B) i12/-21: 150,000 cpm loaded,
`double-coupled at step 1, 50% each cycle analyzed. (C) i4/-21: 50,000
`cpm loaded, double-coupled at step 1, double-cleaved at step 9, 50%
`of each fraction analyzed.
`
`the hybrid preprotein to proinsulin. Furthermore, when the
`whole bacterial signal sequence is fused to rat proinsulin only
`four amino acids from the bacterial clipping site, this hybrid
`preprotein is also correctly matured (Fig. 4).
`The bacterial signal peptidase correctly matures the hybrid
`preproteins to proinsulin whether the site for clipping.is 29
`(i4/-21), 40 (i'12/ -21), or 52 (i25/-21) amino acids from the
`amino terminus. This clearly demonstrates that the infomiation
`to determine the site of cleavage is local and not dependent on
`the distance from the start of the signal peptide.
`A simple model to account for the site of clipping of any
`preprotein would be that the signal sequence extends outside
`the native folded protein and that the
`peptidase clips back
`to the protein surface. Our results do not support this model;
`it is unlikely that proinsulin protects the bacterial signal exactly
`as does penicillinase, yet we see precise, correct, clipping of
`
`

`
`Biochemistry: Talmadge et al.
`
`Proc. Natl. Acad. Sci. USA 77 (1980)
`
`3991
`
`among the generally hydrophobic amino acids found in signal
`sequences. The last, alanine, is a frequent delineator; about half
`of the known signal sequences (2) end with alanine. The third
`is proline, four amino acids from the clipping site. Schechter et
`al. (24) have isolated an immunoglobulin light chain variant
`that is matured by a cut three amino acids from an invariant
`glycine (at amino acid -4) despite replacement of the three
`intervening amino acids, and they propose that helix-breaking
`amino acids in this region create part of the structure that the
`signal peptidase recognizes in order to cut three amino acids
`toward the carboxyl-terminus. Although it is possible that the
`clipping of the eukaryotic signal by the bacterial signal pepti-
`dase is an artifact of this proline at -4, the ability of dog pan-
`creas rnicrosomes to segregate and to process preproteins from
`species as unrelated as dogs and fish (8), as well as the ability
`of bacteria to transport a eukaryotic preprotein (7), suggests
`that the mechanism of secretion is both general and ancient.
`Thus, we expect that all eukaryotic signals will be recognized
`by bacteria, that the preproteins will be secreted with some
`efficiency, and that the secreted protein will be correctly ma-
`tured.
`
`Inserting a eukaryotic gene into a bacterial gene normally
`produces a hybrid bacterial-eukaryotic protein. One method
`to eliminate the extraneous bacterial protein is to insert the
`codon for an unusual amino acid between the two genes and
`to subject the fused protein product to chemical cleavage (25).
`This only works for proteins that lack the unusual amino acid.
`An alternate strategy involves direct expression of the gene
`within the bacterium, where the entire eukaryotic gene (in-
`cluding its ATC initiation codon) is inserted downstream from
`a bacterial promotor. This produces a product with an extra
`formyl methionine at the amino terminus, arising from the
`bacterial initiator. The formyl group may be removed [reported
`for simian virus 40 tumor antigen (26) and rabbit /3—globin (27)],
`leaving an extra methionine in the case of rabbit [3-globin.
`Our results suggest a simple method for the production of the
`extract, native protein, applicable where the eukaryotic protein
`is normally secreted, in cases such as insulin, interferon, human
`growth hormone, and many other medically important pro-
`teins. If the gene for the preprotein is inserted downstream from
`a bacterial promoter, the mature protein, without extraneous
`bacterial amino acids, can be isolated from the bacterial peri-
`plasmic space.
`
`We gratefully acknowledge Peter Lomedico, Harry T. Orr, and
`Jeremy Knowles for helpful discussions and advice. This work was
`supported in part by National Institutes of Health Grant AM 21240
`and in part by Biogen.
`
`l. Milstein, C., Brownlee, G. G., Harrison, T. M. & Matthews, M.
`B. (1972) Nature (London) New Biol. 239, 117-120.
`2. Blobel, G., Walter, P., Change, C. N., Goldman, B. M., Erikson,
`A. H. & Lingappa, V. R.
`( 1979) Symp. Soc. Exp. Biol. 33,
`9-36.
`
`P93
`
`Blobel, G. & Dobberstein, B. (1975) j. Cell Biol. 67, 835-851.
`Emr, S. D., Hedgpeth, ]., Clement, J. M., Silhavy, T. I. &
`Hofnung, M. (1980) Nature (London) 285, 82-85.
`5. Bedoulle, H., Bassford, P., Fowler, A., Zabin, I., Beckwith, J. &
`Hofnung, M. (1980) Nature (London) 285, 78-81.
`6. Villa-Komaroff, L., Efstratiatis, A., Broome, S., Lomedico, P.,
`Tizard, R., Naber, S. P., Chick, W. L. & Gilbert, W. (1978) Proc.
`Natl. Acad. Sci. USA 75, 3727-3731.
`7. Talmadge, K., Stahl, S. & Gilbert, W. (1980) Pmc. Natl. Acad.
`Sci. USA 77, 3369-3373.
`Shields, D. & Blobel, C. (1977) Proc. Natl. Acad. Sci. USA 74,
`2059-2063.
`9. Fraser, T. & Bruce, B. J. (1978) Proc. Natl. Acad. Sci. USA 75,
`5936-5940.
`
`8.
`
`1 000 250
`
`120
`
`90
`
`"S,cpm 60
`
`30
`
`
`
`
`
`
`
`HPETAAGGGGGGQHLCGPHL
`5
`10
`15
`20
`Cycle
`FIG. 4. Location of the 35S-containing and [31-l]leucine residues
`in the amino-terminal sequence of i27/+4. The insulin antigen was
`purified from cells labeled with both H235SO4 and [31-I]leucine by
`immunoprecipitation and NaDodSO4/polyacrylamide gel electro-
`phoresis and then subjected to automated Edman degradation. (A)
`300,000 cpm loaded; (B) 85,000 cpm loaded. Double-coupling was
`done at step 1, double-cleaving at steps 2 and 18. The amount of 35S
`and 3H radioactivity released at each cycle of degradation was de-
`termined by liquid scintillation counting, with the crossover into the
`31-1 channel subtracted. Ten percent of each fraction was analyzed.
`The amino-terminal sequence of i27/+4 matured at the correct bac-
`terial clipping site (see Fig. 1) is presented for comparison.
`
`i27/+4. The information for this clipping must therefore be
`contained in the signal sequence plus the first four amino acids
`of the mature protein. Lin et al. (23) characterized a mutant
`prelipoprotein in which a glycine, seven amino acids from the
`clipping site, was replaced with aspartic acid; the mutant pre-
`lipoprotein was transported but not cleaved. Thus, the car-
`boxyl-terrninal portion of the signal must participate in the
`processing.
`Does the processing of the eukaryotic signal sequence in
`bacteria indicate a general phenomenon, or is it a special case
`of fortuitous signal sequence similarities? If we align the pre-
`penicillinase (bacterial) and preproinsulin (eukaryotic) signal
`sequences, there are four amino acids which are the same dis-
`tance from the site of clipping, underlined below (see Fig. 1 for
`the one-letter amino acid code):
`
`Prepenicillinase
`MSIQHERVALIPFFAAFCLEVFA HPETLVK .
`
`.
`
`.
`
`Preproinsulin
`MALWMRELPLLALLVLWEPKEAQA FVKQHLC. .
`
`.
`
`The first two, phenylalanine and leucine, do not stand out
`
`

`
`3992
`
`Biochemistry: Talmadge et al.
`
`Proc. Natl. Acad. Sci. USA 77 (1980)
`
`10.
`
`11.
`
`12.
`
`13.
`14.
`
`15.
`
`16.
`17.
`
`& Walsh, K. (1978) Proc. Natl. Aoad.
`
`Palmiter, R. D., Gagnon,
`Sci. USA 75, 94-98.
`Lingappa, V. R., Lingappa, I. R. & Blobel, G. (1979) Nature
`(London) 281, 117-121.
`Broome, S. & Gilbert, W. (1978) Proc. Natl. Amd. Sci. USA 75,
`2746-2749.
`
`Immunol. 115, 1617-1624.
`Kessler, S. W. (1975)
`Miller, I. H. (1972) Experiments in Molecular Genetics (Cold
`Spring Harbor Laboratory, Cold Spring Harbor, NY).
`Roberts, R. B., Abelson, P. A., Cowie, D. B., Bolton, E. T. &
`Britten, R. J. (1957) Studies of Biosynthesis in Escherichia coli,
`(Carnegie Institution of Washington, Washington DC), Publ. No.
`607
`227, 680-685.
`Laemmli, U. K. (1970) Nature
`Maizel, ]. V., ]r. (1970) in Methods in Virology, eds. Mammo-
`rosch, K. & Koprowslti, H. (Academic, New York), Vol. 5, pp.
`179-246.
`
`Bauer, A. W., Margolies, M. N. & Haber, E. (1975) Biochemistry
`14, 3029-3035.
`
`C. (1978) Proc. Natl. Acad. Sci. USA 75, 3732-
`
`Sutcliffe,
`3736.
`Lomedico, P., Rosenthal, N., Efstratiadis, A., Gilbert, W., Kol-
`odner, R. & Tizard, R. (1979) Cell 18, 545-558.
`Dayhoff, M. 0. (1972) Atlas of Protein Sequences and Structure
`(National Biomedical Research Foundation, Washington, DC),
`Vol. 5.
`Nolan, C., Margoliash, E., Peterson, I. D. & Steiner, D. F. (1971)
`]. Biol. Chem. 246, 2780-2795.
`Lin, ]. I. C., Kanazawa, H., Ozols, ]. & Wu, H. C. (1978) Proc.
`Natl. Acad. Sci. USA 75, 4891-4895.
`Schechter, 1., Wolf, 0., Zemell, R. & Burstein, Y. (1979) Fed. Proc.
`Fed. Am. Soc. Exp. Biol. 38,1839-1845.
`_
`Itakura, K., Hirose, T., Crea, R., Riggs, A. D., I-Ieyneker, H.,
`Bolivar, F. & Boyer, H. F. (1977) Science I98, 1056-1063.
`Roberts, T., Bikel, 1., Yocum, R., Livingston, D. M. & Ptashne,
`M. (1979) Proc. Natl. Acad. Sci. USA 76, 5596-5600.
`Cuarante, L., Lauer, C., Roberts, T. M. & Ptashne, M. (1980) Cell
`20, 543-553.