High selectivity with low specificity: how SecB has solved the paradox of chaperone binding

Fundamental to the function of all molecular chaperones is their amazing ability to selectively and rapidly bind proteins in non-native states. Chaperones modulate a kinetic partitioning among the alternative pathways open to polypeptides within a cell, so that the proper pathway is taken. Here we review studies of SecB, a chaperone in Escherichia coli dedicated to facilitation of protein export, and emphasize the features that enable it to bind rapidly with high affinity and selectivity in the absence of consensus in sequence. The concepts discussed are likely to be generally applicable to chaperones.

Effect of exogenous fatty acids with greater than 22 carbon atoms (very long chain fatty acids) on superoxide production by human neutrophils

The effects of exogenous long chain fatty acids (LCFA) and very long chain fatty acids (VLCFA) on superoxide production by human neutrophils were compared. Superoxide production was greater and more rapid in response to arachidonic (20:4 (n-6)), eicosapentanoic (20:5 (n-3)), and docosahexanoic (22:6 (n-3)) acids than for triacontatetranoic (30:4 (n-6)), dotriacontatetranoic (32:4 (n-6)), and tetratriacontahexanoic (34:6 (n-3)) acids, although all of these fatty acids gave responses larger than FMLP. A similar decline in activity with increasing carbon chain length was observed for the monoenoic VLCFA (22:1 (n-9) to 34:1 (n-9)). 32:4 (n-6) did not affect responses to a maximally stimulatory concentration of 20:4 (n-6). However, the simultaneous addition of 20:4 (n-6) and 30:4 (n-6) gave additive responses if suboptimal dosages of 20:4 (n-6) were used. This suggests that the LCFA and VLCFA may use the same signal transduction systems. In addition, 30:4 (n-6) was only 10% as effective as was 20:4 (n-6) at gaining access to the organic solvent extractable cellular fraction. This figure correlated with the relative biologic potency of 20:4 (n-6) and 30:4 (n-6), suggesting that the extent of association with the cell may regulate the biologic activity of the fatty acids. The saturates, arachidic (20:0) and cerotic (26:0) acids, were either inactive or poor activators in all assay systems examined. The failure of 20:0 to induce superoxide production and the lower responses to 30:4 (n-6) and 34:6 (n-3) were not because of extracellular Ca2+, because the biologic potency of these fatty acids was not greatly enhanced by removing Ca2+ from the extracellular medium. In contrast, 20:4 (n-6)- and 22:6 (n-3)-induced superoxide production was markedly increased under Ca(2+)-free conditions.

Effect of exogenous fatty acids with greater than 22 carbon atoms (very long chain fatty acids) on superoxide production by human neutrophils

The effects of exogenous long chain fatty acids (LCFA) and very long chain fatty acids (VLCFA) on superoxide production by human neutrophils were compared. Superoxide production was greater and more rapid in response to arachidonic (20:4 (n-6)), eicosapentanoic (20:5 (n-3)), and docosahexanoic (22:6 (n-3)) acids than for triacontatetranoic (30:4 (n-6)), dotriacontatetranoic (32:4 (n-6)), and tetratriacontahexanoic (34:6 (n-3)) acids, although all of these fatty acids gave responses larger than FMLP. A similar decline in activity with increasing carbon chain length was observed for the monoenoic VLCFA (22:1 (n-9) to 34:1 (n-9)). 32:4 (n-6) did not affect responses to a maximally stimulatory concentration of 20:4 (n-6). However, the simultaneous addition of 20:4 (n-6) and 30:4 (n-6) gave additive responses if suboptimal dosages of 20:4 (n-6) were used. This suggests that the LCFA and VLCFA may use the same signal transduction systems. In addition, 30:4 (n-6) was only 10% as effective as was 20:4 (n-6) at gaining access to the organic solvent extractable cellular fraction. This figure correlated with the relative biologic potency of 20:4 (n-6) and 30:4 (n-6), suggesting that the extent of association with the cell may regulate the biologic activity of the fatty acids. The saturates, arachidic (20:0) and cerotic (26:0) acids, were either inactive or poor activators in all assay systems examined. The failure of 20:0 to induce superoxide production and the lower responses to 30:4 (n-6) and 34:6 (n-3) were not because of extracellular Ca2+, because the biologic potency of these fatty acids was not greatly enhanced by removing Ca2+ from the extracellular medium. In contrast, 20:4 (n-6)- and 22:6 (n-3)-induced superoxide production was markedly increased under Ca(2+)-free conditions.

In vitro activation of rat brain protein kinase C by polyenoic very-long-chain fatty acids

A variety of fatty acids including the cis-polyunsaturated very-long-chain fatty acids (VLCFA) (> 22 carbon atoms) common in retina, spermatozoa, and brain were examined for their ability to activate protein kinase C (PKC) purified from rat brain. Arachidonic [20:4(n-6)], eicosapentaenoic [20:5(n-3)], and docosahexaenoic [22:6(n-3)] acids as well as the VLCFA dotriacontatetraenoic [32:4(n-6)] and tetratriacontahexaenoic [34:6(n-3)] were equally capable of activating PKC in vitro with maximal activity being between 25 and 50 microM. The phorbol ester 12-O-tetradecanoylphorbol 13-acetate further enhanced the in vitro activation of PKC when added to the protein kinase assay system with the fatty acids. The fully saturated arachidic acid (20:0) was inactive in both assay systems. The potential significance of the in vitro activation of PKC by the VLCFA is discussed.

Neutrophil migration inhibitory properties of polyunsaturated fatty acids. The role of fatty acid structure, metabolism, and possible second messenger systems

The n-3 polyunsaturated fatty acids (PUFA) appear to have antiinflammatory properties that can be partly explained by their biological activity on leukocytes. Since leukocyte emigration is an essential component of the inflammatory response, we have examined the effects of the n-3 PUFA (eicosapentaenoic and docosahexaenoic acids) on neutrophil random and chemotactic movement. Preexposure of neutrophils for 15-30 min to 1-10 micrograms/ml PUFA reduced the random and chemotactic migration to both FMLP- and fungi-activated complement. The inhibitory effect diminished with increasing saturation and carbon chain length, and methylation abolished this activity. Arachidonic and docosahexaenoic acids were the most active fatty acids. The PUFA concentration required to inhibit migration was dependent on cell number, suggesting that the fatty acid effects on leukocyte migration in vivo may be governed by the stage of the inflammatory response. It was concluded that the PUFA rather than their metabolites were responsible for the inhibition since: (a) antioxidants did not prevent the PUFA-induced migration inhibition and the hydroxylated intermediates were less active, and (b) inhibitors of the cyclooxygenase and lipoxygenase pathways were without effect. Inhibitors of protein kinases and calmodulin-dependent enzyme system did not prevent the PUFA-induced migration inhibition, which was also independent of phospholipase D-catalyzed hydrolysis of phospholipids. It is also shown that PUFA decrease the FMLP-induced Ca2+ mobilization.

Recognition of ligands by SecB, a molecular chaperone involved in bacterial protein export

SecB is a molecular chaperone involved in protein export from Escherichia coli. It is a highly negatively charged, soluble, tetrameric protein with a monomer molecular mass of 16,400 kDa. It has two functions: it maintains precursors of some exported proteins in a conformation compatible with export, by preventing them from aggregating or from folding into their thermodynamically stable state in the cytoplasm, and it delivers both nascent and completed precursors to SecA, one of the components of the export apparatus that are on and in the plasma membrane. SecB recognizes completed precursors of soluble proteins, not by direct interaction with leader sequences but by virtue of the property, imposed by their leader sequences, that they fold slowly: i.e. there is a kinetic partitioning between folding and interaction with SecB. Only those polypeptides that fold slowly interact significantly with this molecular chaperone even though it is able to bind a wide variety of non-native proteins. Binding studies with purified peptides indicate that each SecB monomer has a binding site that can interact with flexible peptides having a net positive charge and a length of about ten residues, which may depend on the charge density. Binding of the hydrophobic fluorescent probe 1-anilino-naphthalene-8-sulphonate (ANS) indicates that simultaneous interaction of multiple peptides causes a conformational change that exposes a hydrophobic site on SecB. This hydrophobic region is thought to contribute an extra binding site for physiological ligands of SecB. A model of SecB binding to nonnative precursors is presented.

Examination of the role of the proteolytically-activated form of protein kinase C in the differentiation of human haemopoietic cells

In neutrophils, the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) induced the translocation of the Ca(++)- and phospholipid-dependent protein kinase, protein kinase C (PK-C) from the soluble to the particulate fraction. At the same time there was a corresponding increase in the amount of Ca(++)- and phospholipid-independent protein kinase activity recovered in the soluble fraction. This soluble Ca(++)- and phospholipid-independent protein kinase presumably reflects proteolytic activation of the particulate associated PK-C. Bone marrow and undifferentiated HL-60 cells also translocated PK-C to the particulate fraction in response to TPA but did not accumulate the soluble Ca(++)- and phospholipid-independent form of the enzyme. Similar results were obtained using HL-60 cells induced to differentiate with dimethyl sulphoxide (DMSO), recombinant human granulocyte-macrophage colony-stimulating factor (rh GM-CSF) or 1 alpha,25-dihydroxyvitamin D3. There was also no significant change in either the number or time of expression of differentiation-specific cell surface antigens observed on HL-60 cells induced to differentiate with either DMSO, 1 alpha,25-dihydroxyvitamin D3 or TPA in the presence of cyclosporin A, an agent reported to inhibit the proteolytic breakdown of PK-C to the Ca(++)- and phospholipid-independent form. Likewise, cyclosporin A did not affect the rate of extent of differentiation of primary bone marrow cell cultures. These results suggest that the proteolytically activated and phospholipid-independent form of PK-C is probably not involved in haemopoietic cell differentiation.

The neutrophil respiratory burst. Responses to fatty acids, N-formylmethionylleucylphenylalanine and phorbol ester suggest divergent signalling mechanisms

The oxygen-dependent respiratory burst is a key neutrophil function required for the killing of bacteria. However, despite intensive investigation, the molecular events which initiate the respiratory burst remain unclear. Recent reports have suggested the agonist-induced hydrolysis of cellular phosphatidylcholine (PtdCho) by phospholipase D may be an essential requirement for initiating or mediating the respiratory burst. We have investigated the effects of the chemotactic peptide N-formylmethionylleucylphenylalanine (fMLF), the phorbol ester 12-O-tetradecanoyl-phorbol 13-acetate (TPA) and the polyunsaturated fatty acids arachidonic [20:4 (n-6)] and docosahexaenoic [22:6 (n-3)] acids in light of this hypothesis. Ethanol-inhibited superoxide production in response to 20:4, 22:6 and fMLF, in a dose-dependent fashion, suggesting an involvement of phospholipase D. The phosphatidic-acid phosphohydrolase inhibitor DL-propranolol completely inhibited superoxide production induced by both 20:4 and 22:6, and partially inhibited the response to TPA. In contrast, superoxide production in response to fMLF was increased by propranolol. fMLF and TPA, but not the fatty acids, stimulated phospholipase D as indicated by the accumulation of phosphatidic acid and, in the presence of ethanol, phosphatidylethanol derived from PtdCho. Extracellular Ca2+ was found to be an essential requirement for fMLF-induced superoxide production. However, responses to the fatty acids were dramatically enhanced under Ca(2+)-free conditions. Responses to TPA were independent of the extracellular Ca2+ concentration. Both fatty acids and fMLF, but not TPA, mobilised Ca2+ from intracellular stores, a response insensitive to the effects of both ethanol and propranolol. These results show that, unlike fMLF and TPA, the fatty acids do not cause hydrolysis of PtdCho by phospholipase D. However, the data indirectly suggests that the fatty acids may initiate the phospholipase-D-catalysed hydrolysis of phospholipids other than PtdCho.

A kinetic partitioning model of selective binding of nonnative proteins by the bacterial chaperone SecB

An in vitro assay for the interaction of SecB, a molecular chaperone from Escherichia coli, with polypeptide ligands was established based on the ability of SecB to block the refolding of denatured maltose-binding protein. Competition experiments show that SecB binds selectively to nonnative proteins with high affinity and without specificity for a particular sequence of amino acids. It is proposed that selectivity in binding is due to a kinetic partitioning of polypeptides between folding and association with SecB.

No specific recognition of leader peptide by SecB, a chaperone involved in protein export

Most proteins destined for export from Escherichia coli are made as precursors containing amino-terminal leader sequences that are essential for export and that are removed during the process. The initial step in export of a subset of proteins, which includes maltose-binding protein, is binding of the precursor by the molecular chaperone SecB. This work shows directly that SecB binds with high affinity to unfolded maltose-binding protein but does not specifically recognize and bind the leader. Rather, the leader modulates folding to expose elements in the remainder of the polypeptide that are recognized by SecB.

Mouse autoantibodies bind to a phospholipase-C-sensitive structure on red blood cells

After culturing mouse peritoneal cells in vitro for 4 days, high numbers of cells can be detected that secrete autoantibodies against isologous red blood cells (RBC), modified with the proteolytic enzyme bromelain (Brom). Plaque-forming cell numbers against mouse Brom RBC were significantly reduced by pretreating mouse Brom RBC prior to haemolytic assay with phospholipase C, an enzyme that hydrolyzes phospholipids, notably phosphatidylcholine. In contrast, further treatment of mouse Brom RBC with Brom, neuraminidase, beta-chymotrypsin, trypsin, or papain had no effect on plaque-forming cell numbers. These results show that phosphatidylcholine is an integral part of the mouse RBC autoantigen exposed by Brom treatment.

Phorbol ester induces differential membrane-association of protein kinase C subspecies in human platelets

Human platelets contained proteins which cross-reacted with antisera specific for brain protein kinase C-alpha and -beta. When platelets were incubated with 12-O-tetradecanoylphorbol-13-acetate there was a rapid accumulation of protein kinase C-alpha in the particulate fraction associated with a loss of this subspecies from the soluble fraction. No particulate accumulation or soluble loss of protein kinase C-beta could be detected when platelets were incubated with the phorbol ester.

Unity in function in the absence of consensus in sequence: role of leader peptides in export

Passage of proteins across membranes during export from their site of synthesis to their final destination is mediated by leader peptides that paradoxically exhibit a unity of function in spite of a diversity of sequence. These leader peptides act in at least two stages of the export process: at entry into the pathway and subsequently during translocation across the membrane. How selectivity is imposed on the system in the absence of a consensus among the sequences of leader peptides is the main issue discussed here.

Biochemical investigation of protein export in Escherichia coli

Export of proteins from the bacterial cytoplasm to a final destination in the periplasm and outer membrane is one example of the fundamental process occurring in all cells whereby polypeptides are transferred across biological membranes. Investigations on a variety of different systems have indicated similarities in the mechanism of this process. In the cases of bacterial protein export and the transfer of polypeptides across the endoplasmic reticulum in eukaryotic cells the processes are so similar that understanding gleaned from studies of the one is usually directly applicable to the other. The study of protein export in E. coli has two advantages over that of eukaryotic secretion. Not only is there the possibility of doing sophisticated genetic experiments, but also one can carry out biochemical investigations in vivo, a facility not so readily available with eukaryotic organisms. Such studies have, for example, shown that membrane translocation can occur both cotranslationally and post-translationally, that export requires protonmotive force, that some component of the export apparatus prevents the exported protein from assuming its native structure in the cytosol, and that there are probably at least two functions for the leader sequence, one in targeting the protein to the export pathway and one in translocation across the membrane.

Coordinated assembly of multisubunit proteins: oligomerization of bacterial enterotoxins in vivo and in vitro

In this paper we study the assembly, in vivo and in vitro, of a family of hexameric, heat-labile enterotoxins produced by diarrheagenic bacteria. The toxins, which consist of an A subunit and five B subunits, are assembled by a highly coordinated process that ensures secretion of the holotoxin complex. We show that (i) oxidation of cysteine residues in the B subunits is a prerequisite step for in vivo formation of B-subunit pentamers, (ii) reduction of dissociated B subunits in vitro abolishes their ability to reassemble, (iii) the kinetics of B-pentamer assembly in vivo can be mimicked under defined conditions in vitro, (iv) A subunits cannot associate with fully assembled B pentamers in vitro, and (v) A subunits cause an approximately 3-fold acceleration in the rate of B-subunit pentamerization in vivo, implying that A subunits play a coordinating role in the pathway of holotoxin assembly. The last finding is likely to be of general significance, since it provides a mechanism for preferentially excluding or favoring certain intermediates in the assembly of multisubunit proteins.

Correlation of competence for export with lack of tertiary structure of the mature species: a study in vivo of maltose-binding protein in E. coli

Sensitivity to proteolytic degradation was used to monitor folding of polypeptides in vivo. A correlation between competence for export and lack of stable tertiary structure was established by comparing the kinetics of folding of mutated precursor maltose-binding protein that carries a defective leader peptide with the kinetics of folding of wild-type precursor that is competent for export. It is proposed that during export a kinetic competition exists between productive translocation and folding of precursor intracellularly into a stable conformation that is not compatible with transfer.

Autoantibodies against mouse bromelain-modified RBC are specifically inhibited by a common membrane phospholipid, phosphatidylcholine

Sera from mice injected 4 days earlier with lipopolysaccharide lysed mouse RBC treated with bromelain (brom). This lytic activity was totally inhibited by including phosphatidylcholine at final concentrations of about 2 micrograms/ml, or more, in the lytic mixtures. In contrast, the lytic activity of antibodies against rat RBC was not inhibited, even at concentrations of phosphatidylcholine up to 2.5 mg/ml. Various components of the phosphatidylcholine molecule, and other lipids including the closely-related molecule dipalmitoyl phosphatidyl-dimethyl-ethanolamine which is identical to dipalmitoyl phosphatidylcholine, except for the absence of a CH2 group on the polar head group, did not inhibit lysis by the autoantibodies. Autoantibodies against mouse brom RBC, but not antibodies against rat RBC, bound to, and could be eluted from, phosphatidylcholine molecules attached to an insoluble matrix. Liposomes of phosphatidylcholine prepared in the presence of phosphatidic acid or phosphatidylinositol did not inhibit the lysis of mouse brom RBC by autoantibodies to the same extent as liposomes of only phosphatidylcholine. This suggests that phosphatidylcholine is recognized by the autoantibodies only if presented in a certain configuration. We suggest that the function of these autoantibodies may be to facilitate the removal of membrane-damaged cells from the body. Such cells may arise by the process of ageing, or because of the effects of infectious agents such as viruses.

Export of protein in bacteria

Images

Mechanism of toxin secretion by Vibrio cholerae investigated in strains harboring plasmids that encode heat-labile enterotoxins of Escherichia coli

A genetically engineered Vibrio cholerae strain from which the cholera toxin genes had previously been deleted was used as a host in which to study the expression and secretion of related toxins and their subunits. Recombinant plasmids encoding heat-labile enterotoxins (LTs) from Escherichia coli of human and porcine origin were expressed in the V. cholerae host, and this resulted in the secretion of the LTs into the extracellular milieu. The secreted LTs were isolated and it was found that the A subunits of human and porcine LT were "unnicked" polypeptides, which indicates that nicking is not obligatory for toxin secretion. V. cholerae strains were also constructed that harbored plasmids encoding either the A or the B subunits of human LT (A+B-, or A-B+). Approximately 90% of the B subunits were secreted from the A-B+ strain, while all of the A subunits expressed by the A+B- strain remained cell associated. This implies that strains synthesizing both subunits assemble the A and B subunits prior to their secretion. We propose that the entry of the toxin into the secretory step of the export pathway is mediated by a secretory apparatus that recognizes structural domains within the B subunit of LT.

Cellular location of heat-labile enterotoxin in Escherichia coli

We demonstrated that both the A and B subunits of heat-labile enterotoxin from Escherichia coli are located in the periplasm. The toxin was shown to form aggregates in Tris-EDTA buffers which are routinely used for isolating membranes. The aggregates pellet upon centrifugation, and this may explain why several previous investigators have concluded that enterotoxin is associated with membranes.

Assembly in vivo of enterotoxin from Escherichia coli: formation of the B subunit oligomer

An oligomer of the B subunit of heat-labile enterotoxin of Escherichia coli has been observed in minicells and in whole cells. There is a delay after synthesis of the B subunit before it appears in the oligomer. The delay is not due to slow processing of the precursor. A similar delay in oligomerization of the major outer membrane protein OmpF is also described.