Cell, Vol. 65, 367-368,
May 3, 1991, Copyright
0 1991 by Cell Press
Matters Arising
The Primary Pathway of Protein Export in E. coli Escherichia coli proteins (Ffh and F&Y) and 4.5s RNA species (Ffs) have been identified that exhibit sequence similarities to subunits of mammalian signal recognition particle (SRP) and docking protein (Poritz et al., 1988; Riimisch et al., 1989; Bernstein et al., 1989). Since the latter are known to be involved in targeting proteins for insertion into, or transit across; the membranes of the rough endoplasmic reticulum of mammalian cells, this raises the question as to whether these E. coli homologs are components of the protein export machinery. Both groups of researchers that have recently reported on this question (Ribes et al., 1990; Poritz et al., 1990). have proposed that the E. coli ribonucleoprotein particle containing 4.5s RNA does have a role in protein export. Most recently, it has been suggested (Rapoport, 1991) that this bacterial SRP homolog is a key component of the primary protein export pathway of E. coli. Such a finding would indicate that the signals for membrane translocation in mammalian and bacterial cells are recognized by molecules that are highly evolutionarily conserved. Although this is no doubt an attractive hypothesis, the current evidence does not justify the enthusiasm with which it has been set forth. In E. coli, the unparalleled ease of combining genetic, biochemical, and physiological approaches has already greatly facilitated studies of fundamental protein export mechanisms. Prior to the excitement concerning the discovery of bacterial SRP, a great deal had been learned concerning protein export in E. coli. Genetic approaches (Bieker et al., 1990) had led to the identification of various prl alleles selected on the basis of their ability to restore export of proteins with defective leader/signal peptides and conditional set alleles that resulted in cells exhibiting severe protein export defects under nonpermissive conditions. At least six previously unrecognized genes-secA (pr/D), secB, se&, se& (pr/G), secf, and secY (prlA)whose mutant phenotypes strongly indicated a role in protein export were identified in this manner. A variety of genetic and in vivo physiologic studies have provided a wealth of information concerning the function and order of the six Set proteins in the export pathway. Biochemical approaches began with in vitro translocation reactions (Rhoads et al., 1984; Muller and Blobel, 1984), subsequently established a requirement for most of the Set proteins for in vitro protein translocation, and, using strains engineered to overproduce Set proteins, have progressed to the isolation and functional reconstitution of most of the proteins required for translocation across the E. coli cytoplasmic membrane (Brundage et al., 1990). Current reviews of genetic and biochemical studies of this bacterial export pathway have appeared (Bieker et al., 1990; Schatz and Beckwith, 1990; Wickner et al., 1991).
In brief, translocation can be initiated either co- or posttranslationally; the former requires the nascent polypeptide to reach a certain critical length, but is not driven by translation (Randall, 1983). Many precursor proteins are “chaperoned” by forming a complex with the cytoplasmic SecB protein (Kumamoto, 1989). This prevents misfolding and binding of the precursor to nonproductive sites (Hart1 et al., 1990). The SecBlprecursor complex then binds to the membrane via an interaction with the peripheral cytoplasmic membrane protein SecA through the affinities of SecA for SecB (Hart1 et al., 1990) and for both the leader and mature domains of the precursor. Genetic analyses have strongly suggested direct interactions between the signal peptide and SecA (Fikes and Bassford, 1989), as well as with the integral membrane proteins SecE and SecY (Emr et al., 1981). SecY and SecE interact within the membrane (Brundage et al., 1990; Matsuyama et al., 1990; Bieker and Silhavy, 1990). The SecA protein is bound to the membrane by its affinities for SecY/E (Hart1 et al., 1990) and acidic phospholipids. Upon binding a precursor, SecA is activated as an ATPase, and this ATPase activity is required for translocation (Lill et al., 1989). Progress of the chain through the membrane (Tani et al., 1989) is driven by ATP hydrolysis (Chen and Tai, 1985) and the proton motive force (Schiebel et al., 1991). Integral cytoplasmic membrane proteins SecD and SecF may be involved in a late step in translocation, possibly in the release of newly processed proteins. There clearly is still much to be learned concerning this Set-mediated protein translocation pathway. What are the exact roles of each component in this process? How do ATP and the membrane potential drive translocation? Does the protein cross the membrane through the lipid phase or through a proteinaceous pore involving SecYlE? Nevertheless, the basic pathway is well established. Indeed, this is the only translocation reaction for which mutants are available in each relevant component, the genes have all been cloned and sequenced, and most of the functional proteins have been purified and reconstituted into translocation-competent proteoliposomes. How do the E. coli homologs of SRP (Ffh and Ffs) and docking protein (FtsY) fit into the Set-mediated protein export pathway? Rapoport (1991) has suggested that SRP-dependent, cotranslational translocation may constitute the primary export pathway of the bacterial cell, and that SecB chaperone function constitutes a posttranslational salvage pathway for those precursor proteins that fail to be cotranslationally targeted for export by SRP. This suggestion ignores in vivo studies (Randall, 1983) demonstrating that most precursor proteins normally display both cotranslational and posttranslational modes of export, as well as the documented secB null mutant defects (Kumamoto and Beckwith, 1985): the mutant accumulates a variety of precursor proteins, is defective in cotranslational export of proteins (Kumamoto and Gannon, 1988), the proposed function of the bacterial SRP, and is considerably delayed in posttranslational protein export.
Both Ribes et al. (1990) and Poritz et al. (1990) have argued strongly for a primary role of bacterial SRP in protein export. However, such a role would appear to be ruled out by the current data: First, bacterial SRP and docking protein homologs have not been identified in geneticselections for export mutants or in the biochemical dissection of in vitro translocation reactions. Second, direct tests of protein export in cells that have been depleted of functional 4.5s RNA showed normal export of maltose-binding protein, alkaline phosphatase, outer membrane protein A (OmpA), and ribose-binding protein. Secretion of ribosebinding protein and alkaline phosphatase does not require SecB, and therefore might have been proposed to require SRP as an alternative chaperone. Of the five proteins tested, only f3-lactamase export was adversely affected by loss of 4.5s RNA (Poritz et al., 1990; Ribes et al.;1990). However, even in this case, the evidence indicated to the investigators themselves (Poritz et al., 1990) that precursor p-lactamase accumulation may have resulted indirectly from induction of the heat shock response. Third, in contrast to loss of 4.5s RNA, mutations in each of the known secgenesresult in aclear unambiguouspleiotropic defect in protein export. Fourth, obligate cotranslational export has not been demonstrated for a single E. coli protein. Furthermore, both in vivo and in vitro studies have failed to establish a difference between the cotranslational and posttranslational export modes; mechanistically, both appear identical and both involve each of the Set proteins. Finally, although there is no evidence for a role of bacterial SRP in protein export, genetic and physiological analyses of 4.5s RNA function have indicated a role for the molecule in protein synthesis (Brown, 1989). The sequence similarities noted between the mammalian SRP subunits and docking protein and certain E. coli molecules are intriguing but, in themselves, do not constitute evidence for conserved functional roles. It certainly would not be surprising to find that molecules with a common evolutionary origin have diverged with respect to function. Although it may turn out eventually that this bacterial SRP does have a key role in initiating the obligate, cotranslational translocation of some subset of E. coli proteins that have yet to be identified, as has been suggested (Poritz et al., 1990; Rapoport, 1991) there is currently no data supporting this postulate. On the contrary, the available evidence clearly establishes that the known Set proteins comprise the primary protein translocation system in E. coli. Phil Bassford,’ Jon Beckwith, Koreaki lto,3 Carol Kumamoto,-’ Shoji Mirushima,5 Don Oliver,‘j Linda Randall,’ Tom Silhavy,* P. C. Tai,g and Bill Wickner’O ‘Department of Microbiology and Immunology, University of North Carolina, Chapel Hill, North Carolina 27514; *Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, Massachusetts 02115; 31nstitute for Virus Research, Kyoto University, Kyoto 606, Japan; 4Physiology Department, Tufts University School of Medicine, Boston, Massachusetts 02111; 51nstitute of Applied Microbiology, University of Tokyo, Tokyo 113, Ja-
pan; 6Department of Microbiology, SUNY at Stony Brook, Stony Brook, New York 11794; 7Biochemistry/Biophysics Program, Washington State University, Pullman, Washington 99164-4660; 8Department of Molecular Biology, Princeton University, Princeton, New Jersey 08544; gDepartment of Metabolic Regulation, Boston Biomedical Research Institute, Boston, Massachusetts 02114; 10Molecular Biology Institute, UCLA School of Medicine, Los Angeles, California 90024-l 570. References Bernstein, H. D., Poritz, M. A., Strub, K., Hoben, and Walter, P. (1989). Nature 340, 482-486. Bieker,
K. L., and Silhavy,
Bieker, membr.
K. L., Phillips, 22,291-310.
Brown,
S. (1989).
T. J. (1990).
T. (1990).
J. Bioenerg.
Emr, S. D., Hanley-Way,
E., Driessen,
Proc. Natl. Acad.
S., and Silhavy,
Fikes, J. D., and Bassford,
Kumamoto,
C. A. (1989).
Kumamoto,
C. A., and Beckwith,
Proc.
Kumamoto, C. A., and 11554-11558.
Sci. USA 82, 4384Cell 23,79-88.
J. Bacterial.
7 77.402-409.
E., Hendrick, Natl. Acad.
Gannon,
A. J. M., and
T. J. (1981).
P. J., Jr. (1989).
Hard, F.-U., Lecker, S., Schiebel, (1990). Cell 63, 269-279.
J. (1985).
J. P., and Wickner,
S., Akimaru,
P. M. (1988).
J., and Mizushima,
Muller, M., and Blobel, G. (1984). 7425.
763,267-274.
J. Biol. Chem.
M. A., Strub,
M. A., Bernstein, H. D., Strub, K., Zopf, P. (1990). Science 250, 1111-1117.
Rhoads, 70.
T. (1991).
FEBS
K., and Walter,
P. (1988).
Cell 55, 4-6. D., Wilhelm,
H., and
Cell 33, 231-240. Nature
349, 107-108.
D. B., Tai, P. C., and Davis,
Ribes, V., Rijmisch, K., Giner, (1990). Cell 63, 591-600.
8. (1984).
A., Dobberstein,
J. Bacterial.
159, 63-
B., and Tollervey,
Rdmisch, K., Webb, J., Herz, J., Prehn, S., Frank, R., Vingron, Dobberstein, B. (1989). Nature 340, 478-482. Schatz, 240.
P. J., and Beckwith,
Schiebel, E., Driessen, Cell 64, 927-939.
J. (1990).
Annu.
A. J. M., Hartl, F.-U.,
Tani, K., Shiozuka, K., Tokuda, Chem. 264, 18582-l 8588. Wickner, W., Driessen, Biochem., in press.
Lett.
Proc. Natl. Acad. Sci. USA 81,7421-
Poritz, Walter,
L. L. (1983).
263,
D., and Wick-
S. (1990).
Poritz,
Randall,
W.
Sci. USA 86, 5320-5324. J. Bacterial.
Lill, FL, Cunningham, K., Brundage, L. A., Ito, K., Oliver, ner, W. (1989). EMBO J. 8, 961-966.
Rapoport,.
Bio-
J. Mol. Biol. 209, 79-90.
L., and Tai, P. C. (1985).
Matsuyama, 269, 96-100.
S.,
Cell 51, 833-842.
G. J., and Silhavy,
Brundage, L., Hendrick, J. P., Schiebel, Wickner, W. (1990). Cell 62, 649-657. Chen, 4388.
P. J., Brenner,
Rev. Genet. and Wickner,
H., and Mizushima,
A. J. M., and Hartl,
M., and 24, 215-
W. (1991).
S. (1989).
F.-U. (1991).
D.
Annu.
J. Biol. Rev.
May 3, 1991, Copyright
0 1991 by Cell Press
Matters Arising
The Primary Pathway of Protein Export in E. coli Escherichia coli proteins (Ffh and F&Y) and 4.5s RNA species (Ffs) have been identified that exhibit sequence similarities to subunits of mammalian signal recognition particle (SRP) and docking protein (Poritz et al., 1988; Riimisch et al., 1989; Bernstein et al., 1989). Since the latter are known to be involved in targeting proteins for insertion into, or transit across; the membranes of the rough endoplasmic reticulum of mammalian cells, this raises the question as to whether these E. coli homologs are components of the protein export machinery. Both groups of researchers that have recently reported on this question (Ribes et al., 1990; Poritz et al., 1990). have proposed that the E. coli ribonucleoprotein particle containing 4.5s RNA does have a role in protein export. Most recently, it has been suggested (Rapoport, 1991) that this bacterial SRP homolog is a key component of the primary protein export pathway of E. coli. Such a finding would indicate that the signals for membrane translocation in mammalian and bacterial cells are recognized by molecules that are highly evolutionarily conserved. Although this is no doubt an attractive hypothesis, the current evidence does not justify the enthusiasm with which it has been set forth. In E. coli, the unparalleled ease of combining genetic, biochemical, and physiological approaches has already greatly facilitated studies of fundamental protein export mechanisms. Prior to the excitement concerning the discovery of bacterial SRP, a great deal had been learned concerning protein export in E. coli. Genetic approaches (Bieker et al., 1990) had led to the identification of various prl alleles selected on the basis of their ability to restore export of proteins with defective leader/signal peptides and conditional set alleles that resulted in cells exhibiting severe protein export defects under nonpermissive conditions. At least six previously unrecognized genes-secA (pr/D), secB, se&, se& (pr/G), secf, and secY (prlA)whose mutant phenotypes strongly indicated a role in protein export were identified in this manner. A variety of genetic and in vivo physiologic studies have provided a wealth of information concerning the function and order of the six Set proteins in the export pathway. Biochemical approaches began with in vitro translocation reactions (Rhoads et al., 1984; Muller and Blobel, 1984), subsequently established a requirement for most of the Set proteins for in vitro protein translocation, and, using strains engineered to overproduce Set proteins, have progressed to the isolation and functional reconstitution of most of the proteins required for translocation across the E. coli cytoplasmic membrane (Brundage et al., 1990). Current reviews of genetic and biochemical studies of this bacterial export pathway have appeared (Bieker et al., 1990; Schatz and Beckwith, 1990; Wickner et al., 1991).
In brief, translocation can be initiated either co- or posttranslationally; the former requires the nascent polypeptide to reach a certain critical length, but is not driven by translation (Randall, 1983). Many precursor proteins are “chaperoned” by forming a complex with the cytoplasmic SecB protein (Kumamoto, 1989). This prevents misfolding and binding of the precursor to nonproductive sites (Hart1 et al., 1990). The SecBlprecursor complex then binds to the membrane via an interaction with the peripheral cytoplasmic membrane protein SecA through the affinities of SecA for SecB (Hart1 et al., 1990) and for both the leader and mature domains of the precursor. Genetic analyses have strongly suggested direct interactions between the signal peptide and SecA (Fikes and Bassford, 1989), as well as with the integral membrane proteins SecE and SecY (Emr et al., 1981). SecY and SecE interact within the membrane (Brundage et al., 1990; Matsuyama et al., 1990; Bieker and Silhavy, 1990). The SecA protein is bound to the membrane by its affinities for SecY/E (Hart1 et al., 1990) and acidic phospholipids. Upon binding a precursor, SecA is activated as an ATPase, and this ATPase activity is required for translocation (Lill et al., 1989). Progress of the chain through the membrane (Tani et al., 1989) is driven by ATP hydrolysis (Chen and Tai, 1985) and the proton motive force (Schiebel et al., 1991). Integral cytoplasmic membrane proteins SecD and SecF may be involved in a late step in translocation, possibly in the release of newly processed proteins. There clearly is still much to be learned concerning this Set-mediated protein translocation pathway. What are the exact roles of each component in this process? How do ATP and the membrane potential drive translocation? Does the protein cross the membrane through the lipid phase or through a proteinaceous pore involving SecYlE? Nevertheless, the basic pathway is well established. Indeed, this is the only translocation reaction for which mutants are available in each relevant component, the genes have all been cloned and sequenced, and most of the functional proteins have been purified and reconstituted into translocation-competent proteoliposomes. How do the E. coli homologs of SRP (Ffh and Ffs) and docking protein (FtsY) fit into the Set-mediated protein export pathway? Rapoport (1991) has suggested that SRP-dependent, cotranslational translocation may constitute the primary export pathway of the bacterial cell, and that SecB chaperone function constitutes a posttranslational salvage pathway for those precursor proteins that fail to be cotranslationally targeted for export by SRP. This suggestion ignores in vivo studies (Randall, 1983) demonstrating that most precursor proteins normally display both cotranslational and posttranslational modes of export, as well as the documented secB null mutant defects (Kumamoto and Beckwith, 1985): the mutant accumulates a variety of precursor proteins, is defective in cotranslational export of proteins (Kumamoto and Gannon, 1988), the proposed function of the bacterial SRP, and is considerably delayed in posttranslational protein export.
Both Ribes et al. (1990) and Poritz et al. (1990) have argued strongly for a primary role of bacterial SRP in protein export. However, such a role would appear to be ruled out by the current data: First, bacterial SRP and docking protein homologs have not been identified in geneticselections for export mutants or in the biochemical dissection of in vitro translocation reactions. Second, direct tests of protein export in cells that have been depleted of functional 4.5s RNA showed normal export of maltose-binding protein, alkaline phosphatase, outer membrane protein A (OmpA), and ribose-binding protein. Secretion of ribosebinding protein and alkaline phosphatase does not require SecB, and therefore might have been proposed to require SRP as an alternative chaperone. Of the five proteins tested, only f3-lactamase export was adversely affected by loss of 4.5s RNA (Poritz et al., 1990; Ribes et al.;1990). However, even in this case, the evidence indicated to the investigators themselves (Poritz et al., 1990) that precursor p-lactamase accumulation may have resulted indirectly from induction of the heat shock response. Third, in contrast to loss of 4.5s RNA, mutations in each of the known secgenesresult in aclear unambiguouspleiotropic defect in protein export. Fourth, obligate cotranslational export has not been demonstrated for a single E. coli protein. Furthermore, both in vivo and in vitro studies have failed to establish a difference between the cotranslational and posttranslational export modes; mechanistically, both appear identical and both involve each of the Set proteins. Finally, although there is no evidence for a role of bacterial SRP in protein export, genetic and physiological analyses of 4.5s RNA function have indicated a role for the molecule in protein synthesis (Brown, 1989). The sequence similarities noted between the mammalian SRP subunits and docking protein and certain E. coli molecules are intriguing but, in themselves, do not constitute evidence for conserved functional roles. It certainly would not be surprising to find that molecules with a common evolutionary origin have diverged with respect to function. Although it may turn out eventually that this bacterial SRP does have a key role in initiating the obligate, cotranslational translocation of some subset of E. coli proteins that have yet to be identified, as has been suggested (Poritz et al., 1990; Rapoport, 1991) there is currently no data supporting this postulate. On the contrary, the available evidence clearly establishes that the known Set proteins comprise the primary protein translocation system in E. coli. Phil Bassford,’ Jon Beckwith, Koreaki lto,3 Carol Kumamoto,-’ Shoji Mirushima,5 Don Oliver,‘j Linda Randall,’ Tom Silhavy,* P. C. Tai,g and Bill Wickner’O ‘Department of Microbiology and Immunology, University of North Carolina, Chapel Hill, North Carolina 27514; *Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, Massachusetts 02115; 31nstitute for Virus Research, Kyoto University, Kyoto 606, Japan; 4Physiology Department, Tufts University School of Medicine, Boston, Massachusetts 02111; 51nstitute of Applied Microbiology, University of Tokyo, Tokyo 113, Ja-
pan; 6Department of Microbiology, SUNY at Stony Brook, Stony Brook, New York 11794; 7Biochemistry/Biophysics Program, Washington State University, Pullman, Washington 99164-4660; 8Department of Molecular Biology, Princeton University, Princeton, New Jersey 08544; gDepartment of Metabolic Regulation, Boston Biomedical Research Institute, Boston, Massachusetts 02114; 10Molecular Biology Institute, UCLA School of Medicine, Los Angeles, California 90024-l 570. References Bernstein, H. D., Poritz, M. A., Strub, K., Hoben, and Walter, P. (1989). Nature 340, 482-486. Bieker,
K. L., and Silhavy,
Bieker, membr.
K. L., Phillips, 22,291-310.
Brown,
S. (1989).
T. J. (1990).
T. (1990).
J. Bioenerg.
Emr, S. D., Hanley-Way,
E., Driessen,
Proc. Natl. Acad.
S., and Silhavy,
Fikes, J. D., and Bassford,
Kumamoto,
C. A. (1989).
Kumamoto,
C. A., and Beckwith,
Proc.
Kumamoto, C. A., and 11554-11558.
Sci. USA 82, 4384Cell 23,79-88.
J. Bacterial.
7 77.402-409.
E., Hendrick, Natl. Acad.
Gannon,
A. J. M., and
T. J. (1981).
P. J., Jr. (1989).
Hard, F.-U., Lecker, S., Schiebel, (1990). Cell 63, 269-279.
J. (1985).
J. P., and Wickner,
S., Akimaru,
P. M. (1988).
J., and Mizushima,
Muller, M., and Blobel, G. (1984). 7425.
763,267-274.
J. Biol. Chem.
M. A., Strub,
M. A., Bernstein, H. D., Strub, K., Zopf, P. (1990). Science 250, 1111-1117.
Rhoads, 70.
T. (1991).
FEBS
K., and Walter,
P. (1988).
Cell 55, 4-6. D., Wilhelm,
H., and
Cell 33, 231-240. Nature
349, 107-108.
D. B., Tai, P. C., and Davis,
Ribes, V., Rijmisch, K., Giner, (1990). Cell 63, 591-600.
8. (1984).
A., Dobberstein,
J. Bacterial.
159, 63-
B., and Tollervey,
Rdmisch, K., Webb, J., Herz, J., Prehn, S., Frank, R., Vingron, Dobberstein, B. (1989). Nature 340, 478-482. Schatz, 240.
P. J., and Beckwith,
Schiebel, E., Driessen, Cell 64, 927-939.
J. (1990).
Annu.
A. J. M., Hartl, F.-U.,
Tani, K., Shiozuka, K., Tokuda, Chem. 264, 18582-l 8588. Wickner, W., Driessen, Biochem., in press.
Lett.
Proc. Natl. Acad. Sci. USA 81,7421-
Poritz, Walter,
L. L. (1983).
263,
D., and Wick-
S. (1990).
Poritz,
Randall,
W.
Sci. USA 86, 5320-5324. J. Bacterial.
Lill, FL, Cunningham, K., Brundage, L. A., Ito, K., Oliver, ner, W. (1989). EMBO J. 8, 961-966.
Rapoport,.
Bio-
J. Mol. Biol. 209, 79-90.
L., and Tai, P. C. (1985).
Matsuyama, 269, 96-100.
S.,
Cell 51, 833-842.
G. J., and Silhavy,
Brundage, L., Hendrick, J. P., Schiebel, Wickner, W. (1990). Cell 62, 649-657. Chen, 4388.
P. J., Brenner,
Rev. Genet. and Wickner,
H., and Mizushima,
A. J. M., and Hartl,
M., and 24, 215-
W. (1991).
S. (1989).
F.-U. (1991).
D.
Annu.
J. Biol. Rev.
