Vol. 149, No. 2JOURNAL OF BACTERIOLOGY, Feb. 1982, p. 789-7920021-9193/82/020789-04$02.00/0

Secretion and Processing of Ribose-Binding Protein inEscherichia coli

JEFFREY L. GARWIN AND JON BECKWITH*Department ofMicrobiology and Molecular Genetics, Harvard Medical School, Boston, Massachusetts 02115

Received 14 August 1981/Accepted 5 October 1981

The periplasmic D-ribose-binding protein of Escherichia coli K-12 is madeinitially as a larger precursor form. This precursor was observed in wild-type cellsand more stably in cells inhibited for protein secretion. The precursor could beprocessed to the mature D-ribose-binding protein either co- or posttranslationally.The secretion pathway of the D-ribose-binding protein and that of the maltose-binding protein have many characteristics in common.

The periplasmic space of Escherichia coli,located between the inner and outer membranes,contains two major classes of proteins: (i) degra-dative enzymes that perform scavenger func-tions and (ii) binding proteins that function insmall molecule transport or chemotaxis (1). Tothe latter class belong binding proteins for thesugars D-ribose, maltose, L-arabinose, and D-galactose. The transport systems for maltoseand ribose appear to be absolutely dependent onthe respective binding proteins, whereas signifi-cant binding-protein-independent transport oc-curs for arabinose and galactose (for a review,see reference 9).To study the mechanism by which proteins,

such as these binding proteins, are secretedacross membranes, we wish ultimately to isolatemutants of E. coli that are pleiotropically defec-tive in the localization of proteins to the peri-plasm. Since the transport and subsequent me-tabolism of both ribose and maltose depend onbinding protein function, a potential procedurefor detecting such mutants would screen forsimultaneous loss of the ability to utilize eithersugar. Although many aspects of the secretion ofmaltose-binding protein (MBP) have been wellcharacterized, ribose-binding protein (RBP) hasbeen less well studied. In this note we show thatRBP, like MBP, follows a secretion pathwaythat involves a larger precursor as a kineticintermediate. Furthermore, cellular perturba-tions that cause accumulation of precursor MBPalso caused accumulation of precursor RBP, to asimilar extent and in the same subcellular com-partments.Antibody to RBP was used to detect mature

and precursor forms of the protein from cellextracts. The RBP used as antigen was preparedby SP-Sephadex chromatography of periplasmicextracts (11), followed by a subtractive adsorp-tion against QAE-Sephadex (pH 8.0) and prepar-ative sodium dodecyl sulfate-polyacrylamide gel

electrophoresis (further details of antigen prepa-ration will be published elsewhere). The proce-dure for rabbit immunization with rehydrated,lyophilized gel slices has been described previ-ously (10).The specificity of the antiserum was demon-

strated by precipitations of extracts from a strainwild type for the ribose pathway and strainscarrying mutant alleles. Strain MRiO (araDaraC Aliac-169 trp(Am) rpsL), which is induc-ible, and strain MC4100 (araD Alac-169 thirpsL), which is constitutive for the ribose sys-tem (J. Lopilato and J. Beckwith, unpublisheddata), showed the expected patterns for RBPsynthesis (Fig. 1, lanes 1, 2, 4, and 5). StrainMRilO6 (MC4100 derivative), carrying a muta-tion (DG19) that reduces RBP synthesis (3),showed little corresponding material in antibodyprecipitates (Fig. 1, lane 3).To detect the precursor of RBP, we used two

bacterial strains which, under the appropriateconditions, accumulate the precursor of MBP.The first strain, MM18 (MC4100, A p7247), pro-duces a hybrid protein composed of an N-terminalportion ofMBP and the C-terminal portion of P-galactosidase. When the synthesis of this hybridprotein is induced by adding maltose to thegrowth medium of MM18, proper localizationand processing of a number of exported pro-teins, including MPB, is inhibited (4). Here weshow the same to be true for RBP. When strainMM18 was grown in minimal glycerol mediumand its protein was labeled for 1 min, predomi-nantly normal-molecular-weight RBP was ob-served on gels (Fig. 1, lane 7). However, whenMM18 was grown for 2 h in the presence of0.2%maltose and then pulse-labeled, the predominantspecies precipitated by anti-RBP antibody was aprotein of apparent molecular weight 2,000 to3,000 greater than that of mature RBP (Fig. 1,lane 8). Qualitatively identical results were ob-served for MBP in strain MM18 (4).

789

on Septem

ber 29, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

790 NOTES

a

.0

f ront

1 2 3 4 5 6 7 8 9 10 11

FIG. 1. Autoradiography of forms of RBP pulse-labeled with [3S]methionine in wild-type and mutant cells.Cells were labeled in minimal glycerol medium as described by Ito et al. (4). [35S]methionine (Amersham Corp.,Arlington Heights, Ill., specific activity, 1 Ci/mmol) was added to a final concentration of 20 ,uCi/ml for 1 min.Incorporation was stopped by adding trichloroacetic acid to 7% concentration (4). Subsequent immuneprecipitation was performed as described previously (4), by the wash modifications described by Oliver andBeckwith (in press) and with 0.5 mg of bovine serum albumin per ml in the immune precipitation mix. Samplesloaded on the gel represented immune precipitations from equivalent total counts of hot trichloroacetic acid-precipitable material (7). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis used the buffer system ofLaemmli (6), the separating gel containing 11% acrylamide and 0.29% N, N'-methylenebisacrylamide. The gelwas treated with 0.5 M salicylate (2), dried, and exposed to Kodak XR5 film at -70°C. For this experiment, allmedia contained glycerol, and the incubation temperature of the culture was 30°C, unless otherwise indicated.Ribose or maltose was added to a final concentration of 0.2% for 2 h when indicated. Lane 1, MRiO + ribose;lane 2, MRiO + glucose; lane 3, MRi1O6 + ribose; lane 4, MC4100; lane 5, MC4100 + maltose; lane 7, MM18;lane 8, MM18 + maltose; lane 9, MM52; lane 10, MM52 at 37°C for 2 h; lane 11, MC4100 at 37°C for 2 h. Lane 6contains molecular weight standards purchased from Sigma Chemical Co. (St. Louis, Mo.) and labeled with[3H]propionate (8): bovine serum albumin (molecular weight, 68,000); alkaline phosphatase (molecular weight,50,000); ovalbumin (molecular weight, 46,000); carbonic anhydrase (molecular weight, 30,000).

The second strain, MM52 (MC4100,secAI151), carries a conditional lethal mutation(secAI151) which has pleiotropic effects on se-cretion. The strain grows normally at 30°C, butafter 2 h of growth at 37°C, secretion and proc-essing of MBP are substantially inhibited andprecursor MBP accumulates (D. B. Oliver and J.Beckwith, Cell, in press). The inhibition wasspecific, affecting only a subset of periplasmicproteins. When MM52 was grown at 30°C, themajor labeled RBP species precipitated corre-sponded to the mature form (Fig. 1, lane 9).When MM52 was shifted to 37°C for 2 h, thepredominant species (Fig. 1, lane 10) comigratedwith the predominant species in lane 7, againsuggesting a precursor form of RBP.We also showed that the putative precursor

form of RBP can be detected in wild-type cellsafter a 20-s pulse of [35S]methionine, but that itdisappears after a "cold" chase of 40 s (Fig.2A). To show that the putative precursor is atrue kinetic precursor of periplasmic RBP, apulse-chase protocol was followed for strainsMM18 and MM52. When the RBP species waslabeled by a 2-min pulse followed by 0-, 2-, 6-,and 14-min chases, the precursor RBP appearedto be slowly processed to mature form in strainsMM18 and MM52. The result was consistentwith observations on the precursor of MBP instrain MM18 (4), whereas in strain MM52 pre-cursor MBP did not appear to be processed(Oliver and Beckwith, in press).

Further evidence that RBP and MBP aresecreted by similiar pathways was obtained by

J. BACTERIOL.

aw, Iwca :..

AM&L .A.- is qm d pre RBP'a RBP

on Septem

ber 29, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 149, 1982

MRiO MM18 MM52

preRBP > _ _ ^_S_i . _ _ _= _ < n~~~~~~reRBP

RBP 4 ___Ii34I-._RBP,* _ _ _ < ~RBP

0 20 40 60 0 2 6 14 0 2 6 14sec min min

FIG. 2. Pulse-chase of precursor RBP in wild-type and mutant cells. An autoradiogram is shown of cellslabeled with [35S]methionine and processed as described in the legend to Fig. 1. Further experimental details arefound below. (A) Strain MRi0, grown at 30°C in the presence of 0.2% ribose, was labeled with 50 FCi of[35S]methionine per ml. After 20 s (pulse), nonradioactive methionine was added to 0.05%, and samples weresubjected to trichloroacetic acid precipitation after 0, 20, 40, and 60 s of chase. The lanes are labeled accordingly.Only the relevant portion of the gel is shown in the figure. (B) Strain MM18, grown for 90 min in the presence of0.2% maltose at 30°C, was pulse-labeled for 2 min with 20 FCi of [35S]methionine per ml. After nonradioactivemethionine was added to a final concentration of 0.05%, samples were subjected to trichloroacetic acidprecipitation after 0, 2, 6, and 14 min of chase. The lanes are labeled with the number of minutes of chase. (C)Strain MM52, grown for 120 min at 37°C in the presence of 0.2% maltose, was subjected to the same pulse-chaseprotocol as described for (B) above. The lanes are labeled 0, 2, 6, and 14 for the number of minutes of chase foreach sample.

ZMM18 zMM52\

preM8P,.MBP _

preRBP

RBP _ _:

1 2 3 4 5 6 7 8

FIG. 3. Subcellular localization of precursor and mature forms of RBP and MBP. A portion of anautoradiogram of sodium dodecyl sulfate-polyacrylamide gel is shown. Cultures prepared as indicated belowwere labeled with 20 pCi of [35S]methionine per ml for 2 min, followed by the addition of nonradioactivemethionine to 0.05% and a further 1-min chase. The incubations were terminated by the addition of the incubatedmaterial to an equal volume of crushed ice. To the ice had been added chloramphenicol sufficient for a finalconcentration of 100 ,ug/ml and NaN3 for a final concentration of 0.02%. All subsequent steps were performed at4 to 10°C, in the presence of chloramphenicol and NaN3. The subcellular fractionation into periplasmic,membrane, and cytoplasmic fractions was as described by Oliver and Beckwith (in press), except that 30 mMHEPES buffer (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) was used instead of Tris-chloride. Im-mune precipitation was performed as described in the legend to Fig. 1, except that both anti-RBP and anti-MBPrabbit serum were used in the immune precipitation. The anti-MBP serum was the gift of W. Boos. Lanes 1through 3, subcellular fractions from strain MM18, grown for 60 min, in the presence of 0.2% maltose and at30°C. Lanes 5 through 7, subcellular fractions from strain MM52, grown for 110 min at 37°C in the presence of0.2% maltose. Lanes 1 and 5, periplasmic fractions; lanes 2 and 6, membrane fractions; lanes 3 and 7,cytoplasmic fractions. Lane 4 contains radioactive standards labeled with [3H]propionate as described by Rockand Cronan (8). The visible band is carbonic anhydrase (molecular weight, 30,000). Lane 8 shows the matureRBP and MBP species immune-precipitated from the periplasmic fraction of MC4100.

NOTES 791

on Septem

ber 29, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

792 NOTES

subcellular fractionation experiments to deter-mine the location of their precursors. StrainsMM18 and MM52, which are derived fromMC4100 and are constitutive for RBP synthesis,were grown in the presence of maltose. Anti-RBP and anti-MBP were simultaneously includ-ed for immune precipitations of three subcellularfractions: periplasm, membranes, and cyto-plasm. In both MM18 and MM52, the matureand precursor forms of RBP and MBP cofrac-tionated despite the apparently different loca-tions of precursors in the two strains (Fig. 3).These results provide evidence that RBP is

made initially as a larger precursor form that issubsequently processed to mature size, as is thecase with all other periplasmic proteins studied.This higher-molecular-weight species was seenafter very short pulse-labeling periods in the wildtype and after even longer labeling periods incells in which protein secretion was inhibitedeither by an MPB-3-galactosidase hybrid proteinor by a conditional lethal mutation. In the lattercases, pulse-chase experiments indicated thatthis higher-molecular-weight species is a precur-sor of RBP. In wild-type cells the processing ofthe precursor took place so rapidly as to suggestpredominantly cotranslational cleavage. In con-trast, when secretion was inhibited to somedegree (strains MM18 and MM52), a much slow-er posttranslational processing occurred. Theseresults, in conjunction with those of others (4,5), raise the possibility that although cotransla-tional secretion is the normal mode of export formost proteins, a much slower posttranslationalmode is also possible. If this is the case, it isimportant in formulating models for the passageof normally secreted proteins through mem-branes.The results extend the range of known pleio-

tropic effects of both the hybrid protein (strainMM18) and the pleiotropic mutation of strainMM52. The latter is of particular interest sincethe secAI151 allele affects only a subset ofsecreted proteins (Oliver and Beckwith, in

press). The results support the conclusion thatRBP and MBP have many common features intheir export and, therefore, may share the samesecretion pathway. Thus, screens for mutantssimultaneously defective in secretion of bothRBP and MBP may well yield mutants in manysteps along the shared secretion pathway.

We thank Theresia Luna for technical assistance and AnnMcIntosh for assistance in the preparation of this manuscript.This work was supported by National Institutes of Health

,grant RO1-GM13017 to J.B. and a National Institute of Allergyand Infectious Diseases postdoctoral fellowship to J.L.G.

LITERATURE CITED1. Beacham, I. R. 1979. Periplasmic enzymes in gram-nega-

tive bacteria. Int. J. Biochem. 10:877-881.2. Chamberldn, J. P. 1979. Fluorographic detection of radio-

activity in polyacrylamide gels with water-soluble fluor,sodium salicylate. Anal. Biochem. 98:132-135.

3. Galloway, D. R., and C. E. Furlong. 1977. The role ofribose-binding protein in transport and chemotaxis inEscherichia coli K-12. Arch. Biochem. Biophys. 184:496-504.

4. Ito, K., P. J. Bassford, Jr., and J. Beckwith. 1981. Proteinlocalization in E. coli: is there a common step in thesecretion of periplasmic and outer membrane proteins?Cell 24:707-717.

5. Koshland, D., and D. Botstein. 1980. Secretion of beta-lactamase requires the carboxy end of the protein. Cell20:749-760.

6. Laemmll, U. K. 1970. Cleavage of structural proteinsduring the assembly of the head of bacteriophage T4.Nature (London) 227:680-685.

7. Mans, R. J., and G. D. Novelli. 1961. Measurement of theincorporation of radioactive amino acids into protein by afilter-paper disk method. Arch. Biochem. Biophys. 94:48-53.

8. Rock, C. O., and J. E. Cronan, Jr. 1979. Re-evaluation ofthe solution structure of acyl carrier protein. J. Biol.Chem. 254:9778-9785.

9. Silhavy, T. J., T. Ferencd, and W. Boos. 1978. Sugartransport systems in Escherichia coli, p. 127-169. In B. P.Rosen (ed.), Bacterial transport. Marcel Dekker, Inc.New York.

10. Tjian, R., D. Stinchcomb, and R. Losick. 1974. Antibodydirected against Bacillus subtilis a factor purified bysodium dodecyl sulfate slab gel electrophoresis. J. Biol.Chem. 250:8824-8828.

11. Wills, R. C., and C. E. Furlong. 1974. Purification andproperties of a ribose-binding protein from Escherichiacoli. J. Biol. Chem. 249:6926-6929.

J. BACTERIOL.

on Septem

ber 29, 2020 by guesthttp://jb.asm

.org/D

ownloaded from