Subunit synthesis is toxic, resulting in a severe growth defect. (A) PapA synthesis is toxic in KS474 (pHJ2, filled squares) but tolerated in the parent strain KS272 (pHJ2, filled circles). (B) PapG synthesis is toxic in both KS272 and KS474 (pHJ8, filled symbols). (C) PapK synthesis is not toxic in either KS272 or KS474. (D) PapE synthesis is toxic (pHJ13, open circles) and toxicity is suppressed by co-synthesis of PapD (pHJ13+pHJ9203, open squares). (E) Non-aggregative PapA derivatives PapA-G150T (pHJ33, open circles), PapA-Y162L (pHJ34, filled triangles) and PapA-C22S (pHJ35, dashed line) are non-toxic in KS474. (F) Removal of the signal sequence (pRS4A, filled squares) renders PapA non-toxic in KS474.
View full figure (89 KB)Article
- The EMBO Journal (1997) 16, 6394 - 6406
- doi:10.1093/emboj/16.21.6394
The chaperone-assisted membrane release and folding pathway is sensed by two signal transduction systems
C. Hal Jones1, Paul N. Danese2, Jerome S. Pinkner1, Thomas J. Silhavy2 and Scott J. Hultgren1
- Department of Molecular Microbiology, Washington University Medical School, Box 8230, 660 South Euclid Avenue, St Louis, MO 63110, USA
- Department of Molecular Biology, Lewis Thomas Laboratory, Princeton University, Princeton, NJ 08544, USA
Correspondence to:
Scott J. Hultgren,
Received 2 June 1997; Revised 28 July 1997
Abstract
The assembly of interactive protein subunits into extracellular structures, such as pilus fibers in the Enterobacteriaceae, is dependent on the activity of PapD-like periplasmic chaperones. The ability of PapD to undergo a 
zippering interaction with the hydrophobic C-terminus of pilus subunits facilitates their folding and release from the cytoplasmic membrane into the periplasm. In the absence of the chaperone, subunits remained tethered to the membrane and were driven off-pathway via non-productive interactions. These off-pathway reactions were detrimental to cell growth; wild-type growth was restored by co-expression of PapD. Subunit misfolding in the absence of PapD was sensed by two parallel pathways: the Cpx two-component signaling system and the 
E modulatory pathway.
Keywords:
- chaperone,
- periplasmic stress,
- periplasmic targeting,
- protein folding,
- signal transduction
Introduction
Introduction
Periplasmic chaperones are an essential component of the machinery required for biogenesis of surface-localized structures such as bacterial pili (Hultgren et al., 1991, 1993, 1996, and references therein). These structures are typically associated with virulence, since they contain adhesins which mediate binding to complementary surfaces of host receptors (Hultgren et al., 1993, 1996; Roberts et al., 1994). P pili are produced specifically by uropathogenic strains of Escherichia coli and have been studied extensively as a model system to understand macromolecular assembly (Hultgren et al., 1991, 1993, 1996). Subunits destined for assembly into pili pass through a periplasmic intermediate state (Hultgren et al., 1991, 1996). Stability of pilus subunits in the periplasm is dependent on the activity of the PapD chaperone, which binds to each of the P pilus subunits to form bimolecular complexes (Hultgren et al., 1991, 1996). Periplasmic complexes between PapD and PapA, PapE, PapF, PapG and PapK have been identified and characterized (Hultgren et al., 1991, 1993, 1996; unpublished data). Once formed, chaperone-subunit complexes are targeted to outer membrane assembly sites, composed of proteins known as ushers, where pilus biogenesis takes place (Dodson et al., 1993). In the absence of the chaperone the subunits collapse into off-pathway aggregates that are proteolytically degraded (Bullitt et al., 1996; Hultgren et al., 1996). The studies described in this report provide new insights into the mechanism of how the chaperone guides translocation of subunits across membranes and facilitates their folding and assembly into pili.
PapD is the prototype member of a family of 26 periplasmic chaperones present in most members of the Enterobacteriaceae as well as in other bacteria (Hultgren et al., 1996; Hung et al., 1996). The 3-dimensional structure of PapD revealed that it possesses two globular immunoglobulin-like domains oriented such that a deep cleft is formed between the domains (Kuehn et al., 1993). The structural basis for part of the chaperone-subunit interaction was solved by co-crystallizing PapD with a peptide corresponding to the C-terminus of PapG (Kuehn et al., 1993).
Utilizing an in vitro assay we demonstrate that PapD facilitates the targeting of pilus subunits into the periplasm and drives the subunits down a productive pathway such that the subunits are kinetically partitioned away from an aggregative fate. We present data suggesting that PapD facilitates the final tertiary fold of the PapG adhesin. We also discovered that misfolded subunits activate one or both signal transduction pathways (Mecsas et al., 1993; Danese et al., 1995) that induce synthesis of multiple factors, such as the DegP protease, for recruitment to the periplasm. One pathway operates via a classic two component regulatory system, Cpx (Danese et al., 1995; Danese and Silhavy, 1997; Pogliano et al., 1997), while the second pathway results in direct modulation of a factor, E (Mecsas et al., 1993; Raina et al., 1995; Rouviere et al., 1995; De Las Penas et al., 1997; Missiakis and Raina, 1997; Missiakis et al., 1997).
Results
An assay to study the role of PapD in pilus biogenesis
Synthesis of P pilin subunits in the absence of PapD results in their proteolytic degradation (Hultgren et al., 1989; Slonim et al., 1992). We employed an isogenic pair of strains KS272 (MC1000) and KS474 (KS272 degP::kan) to test the role of the periplasmic protease DegP (Strauch et al., 1989; Lipinska et al., 1990) in degradation of subunits. It was anticipated that in this genetic background new insights into the mechanism of action of PapD could be gained, since subunits would accumulate in the absence of the chaperone and be more amenable to study.
In KS272 high level synthesis of PapG and PapE in the absence of PapD caused a profound growth defect, whereas PapA and PapK were tolerated when produced at similar or higher levels than PapE or PapG (Figure 1A–C and unpublished data). In KS474 PapA synthesis was found to be highly toxic, while PapK was tolerated when synthesized at similar or higher levels (Figure 1A and C). As expected from the results in KS272, PapG and PapE were highly toxic in KS474 (Figure 1B and D). Subunit synthesis (PapG, PapE and PapA) resulted in growth arrest; affected cultures showed no change in viable counts (
1
107 c.f.u./ml throughout the 6 h assay) when plated on LB agar lacking IPTG (unpublished data). We hypothesize that interactive surfaces on subunits promote non-productive interactions in the absence of the chaperone, leading to growth arrest. The inability of PapK synthesis to inhibit growth may be related to its inability to self-associate (Bullitt et al., 1996), which makes it less aggregative. The reason for PapG toxicity is not readily apparent. However, this protein is larger than the other pilins and previous studies have revealed a potential for it to aggregate in vitro (Kuehn et al., 1991; see also below).
Toxicity is a result of exposure of the subunit 
zipper motif and other interactive surfaces
The conserved C-terminus of subunits has been implicated in forming an interactive surface that drives subunit–subunit interactions (Bullitt et al., 1996). A Leu substitution at Tyr162 (Y162L) and a Thr substitution at Gly150 (G150T) in the C-terminal zipper motif of PapA have been shown to abolish the subunit–subunit interactions necessary for formation of pilus rods (Bullitt et al., 1996). However, both the mutant and wild-type proteins accumulated in the periplasmic space of both strains, KS272 and KS474, to the same extent when co-synthesized with the PapD chaperone (Bullitt et al., 1996). A Ser substitution for Cys22 (C22S) in PapA, which forms a conserved disulfide with Cys61, also abolished the subunit–subunit interactions necessary for pilus formation (R.Striker and S.J.Hultgren, personal communication), but was stable when expressed in strain KS474. Since these mutations disrupt interactive surfaces on the subunit, we tested whether these mutations would also abolish the toxic effect associated with expression of the subunit in the absence of the chaperone. Figure 1E shows that the non-aggregative PapA mutants (Y162L, G150T and C22S) are non-toxic when expressed in KS474, whereas a conserved substitution, Y162F, which assembles into pili, although to a lesser degree, remains toxic. These data suggest that exposure of the C-terminal zipper in the periplasm causes the observed toxic phenotype in strain KS474.
Toxicity arises from the periplasm
The growth defect resulting from subunit production in the absence of PapD suggests that the toxic effect originates from the periplasm. Further support for this hypothesis was garnered by demonstrating that signal sequence processing is required for the observed toxicity. A PapA mutant lacking a signal sequence is not toxic when expressed in KS474 (Figure 1F). These results suggest that engagement of the Sec machinery and transport across the inner membrane is required for toxicity.
Toxicity suppression and correlation with complex formation
If the role of the periplasmic chaperone is to protect interactive subunit surfaces and block their aggregation, then production of the chaperone along with the toxic subunit should be sufficient to suppress toxicity. In both strains, KS272 and KS474, co-expression of the periplasmic chaperone PapD suppressed the toxicity of pilin subunit synthesis (Figures 1D, E, 2A and B and unpublished data).
Figure 2.
Subunit toxicity is relieved by co-synthesis of the PapD periplasmic chaperone. (A) PapG toxicity (pHJ31, open circles) is suppressed by co-synthesis of PapD (pL5101, filled circles). (B) PapA toxicity (pHJ2, open circles) is suppressed by co-synthesis of PapD (pHJ9203, filled circles) but not by co-synthesis of R8G PapD mutant protein (pHJ9204, open squares).
View full figure (44 KB)The invariant residues Arg8 and Lys112 in the crevice of the PapD cleft form a critical part of the subunit binding site (Slonim et al., 1992; Kuehn et al., 1993). Substitutions at Arg8 and Lys112 had no effect on PapD accumulation in the periplasm (Slonim et al., 1992; Kuehn et al., 1993). We tested the effect of mutating the invariant Arg8 residue to Gly (R8G) on the ability of PapD to suppress the toxic effects of PapA. The R8G mutation in PapD blocked its ability to suppress PapA toxicity (Figure 2B). These data confirm that the interactions, via the cleft residues, known to be essential for chaperone–subunit complex formation are also required for suppression of toxicity.
Periplasmic localization is dependent on PapD
In the light of precedents established with other systems, several explanations can be offered for the toxicity caused by overproduction of the Pap subunits. For example, it is possible that high level production of these subunits causes a deleterious 'jamming' of the general secretion machinery. This would inhibit envelope biogenesis and cause accumulation of the precursor forms of secreted proteins in the cytoplasm (Snyder and Silhavy, 1992). We ruled out the 'jamming' hypothesis by showing that the kinetics of signal sequence processing for periplasmic maltose binding protein (MBP) were unaffected by induction of either PapA or PapG (unpublished data). We also ruled out the related hypothesis that overproduced Pap subunits might interfere with discharge of proteins from the secretion machinery by showing that MBP release into the periplasm was also unaffected (unpublished data). These results demonstrate that overproduction of Pap subunits does not interfere with the function of the general protein secretion machinery.
The toxicity of overproduced Pap subunits could also result from improper targeting of these proteins to the periplasm (Matsuyama et al., 1995). Accordingly, we tested the periplasmic localization of pilin subunits when expressed in KS474 in the absence or presence of PapD (Figure 3A). Production of PapA, PapE, PapG and PapK in the absence of PapD resulted in poor accumulation of the subunits in the periplasm. These results suggest that efficient release of pilin subunits into the periplasm is dependent on PapD–subunit interactions.
Figure 3.
Targeting of pilin subunits to the periplasmic space is dependent on PapD. (A) Periplasmic fractions prepared from KS474 strains producing pilin subunits alone (odd numbered lanes) or producing subunits and PapD (even numbered lanes). Coomassie blue staining of SDS-PAGE clearly demonstrates that accumulation of stable pilin subunits PapA, PapE, PapG and PapK is dependent on the chaperone. (B) Western blot with anti-PapD–PapG antisera demonstrating PapD-mediated targeting of PapG to the periplasmic space. Fractionation of KS474/pHJ8/pHJ9203 reveals that PapG resides in the inner membrane fraction when synthesized in the absence (lane 4) of PapD, but is efficiently targeted to the periplasmic space by PapD (lane 2). In lanes 3 and 5 the samples were not exposed to -mercaptoethanol prior to SDS-PAGE, demonstrating that cysteine disulfides are formed in PapG prior to interaction with PapD.
The above results suggest that subunits synthesized alone are either not produced efficiently or are associated with another compartment of the cell, such as the inner membrane. The fate of subunits expressed in the absence of the chaperone was investigated by fractionating KS474/pHJ8 (papG)/pHJ9203 (papD) into periplasmic and inner membrane fractions (Figure 3B). The fractions were immunoblotted with polyclonal antisera directed against the PapD–PapG complex. As shown in Figure 3B, PapG is only detectable in the periplasm following co-induction of PapD synthesis; PapG is associated with the inner membrane fraction when produced in the absence of PapD. Similar results were obtained with PapA and PapE (unpublished data). We conclude that the chaperone facilitates release of the subunit from an inner membrane location to the soluble periplasmic compartment.
Spheroplast assay for testing the targeting function of PapD
To further define the role of the chaperone in mediating subunit compartmentalization we established a semi-in vitro assay utilizing spheroplasts. A related spheroplast system was employed by Matsuyama et al. (1995) to define the activity of the periplasmic lipoprotein carrier protein P20. In our system spheroplasts were prepared from KS474/pHJ8; the spheroplasts were induced and pulse labeled in the presence or absence of purified PapD added to the supernatant. As shown in Figure 4A, addition of purified PapD to the spheroplast suspension increased the amount of immunoprecipitable PapG in the supernatant by >30-fold (Figure 4A, compare lanes 1 and 2).
Figure 4.
Spheroplast release assay. PapG synthesis was induced in spheroplasts produced from KS474/pHJ8 by addition of 1 mM IPTG. (A) Prior to pulse labeling with 35S Trans label (cysteine + methionine), 7 
g purified PapD (lane 2), the R8A PapD derivative (lane 3) or purified histidine-tagged PapD (lane 6) were added and allowed to incubate for 10 min. Following a 10 min chase with unlabeled cysteine + methionine the spheroplast suspension was subjected to centrifugation to remove the spheroplasts and the supernatant immunoprecipitated with antibody containing reactivity to PapG. In lanes 4 and 5 purified PapD was added 10 (lane 4) or 30 (lane 5) min following the cold chase. (B) The F314S substitution in PapG reduces the efficiency of targeting relative to wild-type PapG (compare lanes 3 and 5). Both the double mutant F314S+G302V and deletion of the C-terminal 100 amino acids relieve PapG dependence on PapD for efficient partitioning to the supernatant fraction.
In a separate experiment PapD was added 10 or 30 min following the chase to test if targeting required that PapD interact with PapG during or shortly after translocation. In this experiment PapD-facilitated targeting would require that PapD act on post-translocated PapG that had accumulated in the inner membrane. PapD was fully capable of releasing PapG into the soluble fraction 10 and 30 min post-chase (Figure 4A, lanes 4 and 5). Furthermore, trypsin treatment of spheroplasts producing PapG released an 25 kDa fragment of PapG; demonstrating that a portion of PapG is exposed on the surface of the spheroplast (unpublished data). These findings suggest that membrane-associated PapG is stable and a substrate for PapD and we conclude that membrane-associated PapG is an intermediate in the biogenesis pathway.
FimC, HifB and SfaE, the periplasmic chaperones required for assembly of type 1, Haemophilus and S pili respectively, and the newly discovered EcpD (E.coli PapD) (Raina et al., 1993) were purified and tested in the spheroplast assay. None of these chaperones were able to stimulate release of PapG into the supernatant fraction (unpublished data).
The structural basis of the targeting reaction was investigated by testing the effect of site-directed mutations in both PapD and PapG of residues that are known to be critical for PapD–PapG complex formation based on the co-crystal structure of PapD with the PapG C-terminal peptide (Kuehn et al., 1993). Arg8 forms a critical part of the invariant subunit binding site in the chaperone cleft. The targeting activity of PapD was severely compromised by an R8A substitution, arguing that the invariant cleft is critical for this function (Figure 4A, lane 3). Phe314 and Gly302 are two conserved C-terminal PapG residues that are part of the conserved zipper motif that is recognized by PapD. The Phe was changed to Ser (F314S PapG) as a single mutation and it was also combined with a Gly mutation to Val at position 302 (F314S+G302V PapG) as part of a double mutation. The C-terminal 100 residues of PapG were also deleted, yielding the PapG
100 truncate. The F314S mutation in PapG reduced the efficiency with which PapD released this derivative into the supernatant (Figure 4B, lane 5). F314S+G302V PapG was efficiently released to the supernatant in the absence of PapD, in contrast to wild-type PapG (Figure 4B, lanes 6 and 7). Moreover, the presence of PapD had no effect on release of this double mutant protein. Similarly, deletion of the C-terminal 100 amino acids resulted in a stable 24 kDa PapG derivative that was released into the supernatant independent of PapD (Figure 4B, lanes 8 and 9). These results argue that the zipper region of a subunit facilitates its retention in the inner membrane in the absence of PapD and that binding of PapD to this surface is required for release from the membrane.
Overproduction of PapG activates a two component signal transduction system
Since the Cpx and E modulatory systems both respond to overproduction of extracytoplasmic proteins (Mecsas et al., 1993; Danese et al., 1995; Danese and Silhavy, 1997; De Las Penas et al., 1997; Missiakis and Raina, 1997; Missiakis et al., 1997), we wished to determine if these systems were also affected by overproduction of the PapE and PapG subunits. Accordingly, we transformed strains PND2000 (MC4100, degP–lacZ), SP558 (PND2000, cpxA::cam) and SP559 (PND2000, cpxR::
) with pHJ8 (overproducing PapG) and pMMB66 (vector control for pHJ8) and then determined the amount of degP–lacZ transcription generated from these transformants. Lanes 1 and 2 of Figure 5A show that overproduction of PapG stimulates degP–lacZ transcription 6.7-fold. Interestingly, overproduction of PapG only stimulates degP–lacZ transcription 2.4-fold in the absence of CpxA (compare lanes 3 and 4 of Figure 5A). Finally, in a cpxR-, cpxA- background (SP559) the stimulatory effect of PapG overproduction on degP–lacZ transcription is only 2.2-fold (compare lanes 5 and 6 of Figure 5A). Thus overproduction of PapG stimulates degP transcription, in part, by activation of the Cpx signal transduction system.
Figure 5.
Pilus subunit synthesis in the absence of PapD results in promoter activation via the Cpx and E signal transduction pathways. (A) degP–lacZ promoter activation as a result of PapG production utilizes both the Cpx and E modulation pathways. degP–lacZ activity was monitored in strains PND2000, SP558 and SP559 following induction of synthesis of PapG from pHJ8. (B) Co-synthesis of PapD abrogates degP–lacZ promoter activation. degP–lacZ activity was monitored in PND2000 following induction of PapG synthesis (pHJ9208) or PapG and PapD synthesis (pHJ9208 + pHJ6). (C) Synthesis of PapG activates the rpoHP3–lacZ promoter, indicating that PapG misfolding is sensed by the E modulatory system. rpoHP3–lacZ activity was monitored in SP616 following induction of PapG synthesis (pHJ8). (D) PapG activation of the cpxP–LacZ promoter is Cpx dependent. cpxP–lacZ activity was monitored in SP594 and SP620 following induction of PapG synthesis (pHJ8). (E) Synthesis of PapE activates the degP–lacZ promoter in a Cpx-dependent manner. degP–lacZ activity was monitored in PND2000, SP558 and SP559 following induction of PapE synthesis (pHJ13). (F) The cpxP–lacZ promoter is activated by PapE synthesis and is Cpx dependent. cpxP–lacZ activity was monitored in SP594, SP619 and SP620 following induction of PapE synthesis (pHJ13). (G) Pilus biogenesis is monitored by the Cpx and E signal transduction pathways. degP–lacZ activity was monitored in PND2000, SP558 and SP559 following induction of pilus synthesis from pFJ29 (pap) and pFJ29-71 (papD-). degP–lacZ promoter activation is seen, especially following induction of the pilus operon lacking a functional papD gene (pFJ29-71, papD-). (H) degP–lacZ promoter activation resulting from initiation of pilus biogenesis is not entirely Cpx dependent, as seen in triggering of the E modulatory pathway. rpoHP3–lacZ activity was monitored following induction of the pilus operon from pFJ29 and pFJ29-71 in SP616.
The results presented in Figures 1 and 2 indicated that co-synthesis of the PapD chaperone blocks the growth defect associated with overproduction of P pilin subunits. Thus we hypothesized that co-synthesis of PapG and PapD would suppress PapG-mediated induction of degP transcription. As shown in Figure 5B, production of PapG with simultaneous synthesis of PapD significantly reduced the amount of degP transcription generated from PND2000. This result implies that E.coli monitors the assembly status of PapG and uses this information to modulate degP transcription.
PapG synthesis stimulates 
E activity
Figure 5A shows that synthesis of PapG causes a residual induction of degP transcription even in the absence of the Cpx pathway (compare lanes 1 and 2 with lanes 5 and 6). Therefore, a second signal transduction system must also sense the unfolded state of PapG and stimulate degP transcription. The E modulatory system has been shown to activate both the degP and the rpoHP3 promoters (Mecsas et al., 1993). Because the rpoHP3 promoter is solely controlled by E and not by the Cpx system (Danese et al., 1995), we used an rpoHP3–lacZ operon fusion to determine if PapG synthesis also stimulated E activity. Figure 5C shows that PapG synthesis (from pHJ8) stimulates rpoHP3–lacZ [SP616 (MC4100, rpoHP3–lacZ)] transcription 2.4-fold, which closely agrees with the 2.2-fold stimulation of degP transcription in the absence of the Cpx proteins (Figure 5A, lanes 5 and 6).
Production of PapG stimulates transcription from a second Cpx-regulated locus
In an effort to corroborate the conclusion that synthesis of unchaperoned PapG stimulates the Cpx pathway, we assayed a second Cpx-regulated locus, cpxP, whose transcription is wholly dependent on CpxR and is independent of E (P.N.Danese, doctoral thesis, Princeton University). To this end, we determined the -galactosidase activities of strains SP594 (MC4100, cpxP–lacZ) and SP620 (SP594, cpxR::) following induction of PapG synthesis. As shown in Figure 5D, PapG synthesis in the parental strain (SP594) stimulates cpxP transcription 10.3-fold. In a cpxR-, cpxA- background (SP620) cpxP transcription is not induced following PapG synthesis.
PapE synthesis stimulates Cpx but not E activity
We also performed the assays described above with the P pilus fibrillar component PapE (pHJ13). Surprisingly, unlike PapG, PapE production stimulates only the Cpx signal transduction pathway. As shown in Figure 5E, induction of PapE synthesis in PND2000 caused a 4-fold stimulation of degP transcription (compare lanes 1 and 2), whereas in SP558 and SP559 the amount of degP–lacZ transcription is not elevated above the background level (compare lanes 3 and 4 and lanes 5 and 6). High level synthesis of PapE also stimulated cpxP–lacZ transcription in a Cpx-dependent fashion (Figure 5F) and failed to stimulate transcription from the rpoHP3–lacZ fusion (unpublished data). Thus while the Cpx pathway can monitor the level of both PapE and PapG, the E modulatory system can only monitor PapG levels.
The Cpx pathway as a monitor of pilus assembly
We also observed a significant activation of degP–lacZ transcription when the entire P pilus operon was expressed in the absence of the PapD chaperone (Figure 5G, compare lanes 1, 2 and 3). This P pilus-mediated induction of degP transcription is significantly reduced in the absence of the Cpx pathway (compare lanes 2 and 3 with lanes 8 and 9); however, a residual induction in degP transcription was still detected (compare lanes 7–9). This residual induction (2.1-fold) appears to be due to an increase in E activity because high level expression of the P pilus operon also stimulated rpoHP3–lacZ transcription 2.1-fold (Figure 5H). The implication from these results is that the Cpx and E modulatory systems monitor the assembly status of the P pilus subunits and use this information to modulate synthesis of the periplasmic protease DegP in response to this input.
Chaperone folding assay
The presence of misfolded or partially denatured proteins in the periplasm is thought to be a signal that leads to activation of degP transcription. PapD is known to be required to stabilize pilus subunits, so we decided to investigate further the folded state of pilin subunits synthesized in the absence of PapD. PapG activity can be monitored in vitro by binding to the Gal
(1-4)Gal digalactoside (galabiose) receptor (Hultgren et al., 1991, 1996). Lectin binding activity has been shown to be a function of tertiary structure (Weis et al., 1988). We tested the ability of periplasmically localized full-length PapG to bind to galabiose when produced alone or together with wild-type PapD. Nearly 100% of the immunoprecipitable PapG in the periplasm bound to digalactoside when PapD was present. However, when PapD was absent <10% of the immunoprecipitable periplasmic PapG bound to receptor (Figure 6). These results indicate that in the absence of PapD, PapG is unable to achieve or be maintained in a native-like receptor binding conformation.
Figure 6.
PapD-mediated folding of PapG as assayed by galabiose chromatography. (A) Labeled periplasms containing PapG only (lanes 1 and 7) or PapG along with PapD (lanes 2 and 8), PapD R8A (lanes 3 and 9), PapD K112A (lanes 4 and 10), FimC (lanes 5 and 11) or the PapD–FimC chimera (lanes 6 and 12) were subjected to immunoprecipitation (lanes 1–6) or galabiose–Sepharose chromatography (lanes 7–12). (B) Phosphorimager quantification of lanes 1–6 of the gel in (A). (C) Quantification of binding efficiency of PapG when expressed alone or in the context of chaperones and chaperone mutants, expressed as a percentage of immunoprecipitable material (A, lanes 1–6) that bound to galabiose–Sepharose (A, lanes 7–12).
View full figure (50 KB)The ability of PapD to maintain PapG in a receptor binding conformation depends on the invariant chaperone anchor
The R8A and K112A mutations in PapD, which reduced the ability of the chaperone to facilitate PapG targeting, also failed to assist PapG to fold or be maintained in a native-like receptor binding conformation, since periplasmic PapG was unable to bind receptor (Figure 6A, lanes 3 and 4, and C). Similarly, FimC was unable to stimulate PapG targeting to the periplasm and apparently unable to guide the protein down a productive folding pathway (Figure 6A, lane 5, and C). A PapD–FimC chimeric chaperone was constructed by genetically fusing domain 1 (encoding residues 1–112) of papD to domain 2 (encoding residues 112–205) of fimC; this protein is stable in the periplasm (unpublished data). The chimera was unable to mediate the assembly of either type 1 or P pili (unpublished data), however, in contrast to FimC, it was able to suppress the toxicity associated with PapA expression (unpublished data). The chimera also partially stimulated PapG targeting to the periplasm; however, the resulting periplasmic PapG was unable to bind receptor (Figure 6).
Discussion
The assembly of a family of extracellular structures, including pili, afimbrial adhesins and proteinaceous capsules, in bacteria is a conserved process that utilizes a family of periplasmic chaperones (Hultgren et al., 1991, 1993, 1996; Hung et al., 1996). Bacteria face a multitude of problems when assembling extracellular structures, including: (i) translocation of monomeric proteins across two membranes and their transport through the highly concentrated periplasm; (ii) prevention of spontaneous assembly in membranes and/or the periplasm; (iii) blocking of off-pathway processes that lead to misfolding, aggregation and proteolysis. Periplasmic chaperones, along with cognate outer membrane ushers, constitute a novel molecular machinery (Hultgren et al., 1991, 1993, 1996) necessary for guiding biogenesis of adhesive organelles in Gram-negative bacteria. Pilus subunit proteins cross the inner membrane via the Sec pathway (Dodd et al., 1984; Pugsley and Possot, 1994); as all subunits have a classic signal sequence that is processed to form the mature pilin in the periplasm (Dodd et al., 1984; Hultgren et al., 1991, 1993, 1996; unpublished data). However, we discovered that the Sec machinery does not efficiently release pilus subunits into the periplasm. The appearance in the periplasm of soluble pilin subunits PapA, PapE, PapG and PapK is greatly facilitated by interaction with the periplasmic PapD chaperone. In the absence of PapD, PapG was found to be associated with the inner membrane. Mutagenesis studies combined with in vitro and in vivo studies demonstrated that targeting of subunits into the periplasm was dependent on an anchoring interaction of the terminal carboxylate group of the subunits to the invariant Arg8 and Lys112 residues in the crevice of the chaperone cleft. In addition to this molecular anchoring interaction, zippering of the conserved C-terminus of the subunits along the exposed edge of the G1 -strand of PapD also facilitated release of the subunits from the cytoplasmic membrane into the periplasm. Interestingly, it is the presence of the hydrophobic C-terminal motif that makes subunit targeting dependent on PapD: this motif tethers the subunits to the cytoplasmic membrane. Deletion of the C-terminal 100 residues of PapG renders periplasmic localization independent of PapD. Mutations in many of the conserved alternating hydrophobic residues in the C-terminal zipper motif of PapG were found to behave similarly to the PapG100 protein, suggesting that altering a feature of the zipper (hydrophobicity) or the conformation of the C-terminal domain results in a protein that behaves as if the entire C-terminal domain is missing (unpublished data). This same C-terminal motif has been shown, in the case of PapA, to be part of an interactive surface that participates in subunit–subunit interactions required for assembly of pilus rods (Bullitt et al., 1996). Premature subunit–subunit interactions are prevented in the periplasm by PapD, which binds to and caps this surface (Kuehn et al., 1993; Bullitt et al., 1996). Our current studies have shown that the ability of PapD to release subunits from the membrane also depends upon this interaction.
It remains to be determined whether PapG is associated with Sec components or a novel site in the inner membrane prior to its interaction with PapD. However, since the function of the general secretion machinery is not impaired in any way in the absence of PapD, we favor membrane association. The C-terminal zipper motif is extremely hydrophobic (Kuehn et al., 1993; Hung et al., 1996) and may insert into the membrane in a manner similar to a classic stop transfer sequence (Kuroiwa et al., 1991). Alternatively, the C-terminus of the subunits may be part of an edge strand of a -sheet that 'lies' on top of the membrane, inserting its alternating hydrophobic sidechains into the membrane. In either case, we suggest that the membrane-associated state may represent an intermediate in pilus biogenesis, functioning to protect the C-terminal interactive surface and prevent it from participating in non-productive interactions. A spheroplast assay showed directly that purified PapD added to whole spheroplasts facilitated release of PapG from the surface into the supernatant in a process that can occur long after synthesis is complete. The ability of PapD to partition fully translocated PapG supports the hypothesis that the membrane-associated form of PapG represents a stable intermediate in biogenesis.
Our studies are reminiscent of those recently reported by Matsuyama et al. (1995) describing the function of the P20 protein, which acts as a carrier protein for assembly of LPP (major outer membrane lipoprotein) into the outer membrane. Using a spheroplast system these investigators demonstrated that P20 facilitated release of LPP from the inner membrane into the soluble fraction. LPP is thought to be tethered to the cytoplasmic membrane by the lipid portion of the lipoprotein. The interaction of P20 with LPP results in its release from the membrane and formation of a specific complex between LPP and P20.
A pilus subunit in a membrane-associated state most likely does not exist in a native conformation. In vivo in the absence of PapD, subunits are susceptible to proteolytic degradation by the DegP protease, which is thought to act on denatured or partially unfolded substrates (Kolmar et al., 1996). In a degP- strain and in the absence of PapD the majority of PapG protein fractionates with the cytoplasmic membrane, although some can be recovered in the periplasm. Less than 10% of the PapG protein recovered from periplasmic extracts in these cells is capable of binding to the Gal(1-4)Gal receptor. In contrast, when co-expressed with the PapD chaperone nearly 100% of the recoverable PapG protein binds to receptor. This result argues that the chaperone-assisted release reaction may drive folding of a subunit to completion and/or be necessary to maintain the subunit in a native conformation. The membrane-bound state may represent a molten globule or a highly folded intermediate with partially denatured local structures. Based on circular dichroism and Chou–Fasman predictions, there is evidence that pilin monomers are predominantly -sheet proteins (unpublished results). Thus the chaperone could provide a predefined context in the form of a template for initial -strand formation and thereby predispose subsequent formation of -sheet structure.
PapD may facilitate membrane release and folding of subunits via a two-step mechanism. Previous studies have demonstrated that PapD recognizes a 'second site' on PapG separate from the C-terminal zipper (Xu et al., 1995). Our current hypothesis is that the second site is exposed in the membrane-associated state and is recognized by PapD, possibly via domain 2 (Figure 7). Presumably, residues in domain 2 of PapD also participate in driving the folding process, since a PapD–FimC chimera containing domain 1 of PapD and domain 2 of FimC was unable to guide PapG into a receptor binding conformation. This binding event may trigger a conformational change in the subunit that results in exposure of its C-terminus and allows its partitioning from a membrane-associated state to an extended strand along the G1 -strand of PapD. This two-step pathway explains how membrane release is coupled to template-assisted folding (Figure 7).
Figure 7.
A model of signal transduction pathways that monitor pilus biogenesis. The chaperone-mediated protein translocation, folding and targeting pathway is illustrated (see text for details). Off-pathway products of this pathway result in activation of one or both of the Cpx and E modulatory signal transduction pathways. Triggering the Cpx pathway results in activation of the E-dependent promoter degP and four E-independent promoters (Danese et al., 1995; Danese and Silhavy, 1997; Pogliano et al., 1997). Triggering the E modulatory pathway via RseAB results in up-regulation of the degP promoter as well as three others (Mecsas et al., 1993; Raina et al., 1995; Rouviere et al., 1995; De Las Penas et al., 1997; Missiakis and Raina, 1997; Missiakis et al., 1997). Activation of the Cpx pathway by NlpE (Snyder et al., 1995) and the E modulatory pathway by OMP expression (Mecsas et al., 1993) are also depicted.
Interestingly, the PapK subunit is not toxic in KS474; however, this subunit is still dependent on the chaperone for periplasmic targeting. Previous studies (Hultgren et al., 1993, 1996) have shown that PapK is present in the pilus in single or low copy and that it does not aggregate into oligomeric sub-assemblies in vitro (Bullitt et al., 1996; unpublished data). This finding suggests that the aggregation potential of subunits is, in part, a cause of toxicity. The properties of the zipper C-terminal motif that increase the aggregation potential of a subunit may be the same properties that contribute to inner membrane retention, since inner membrane-localized PapK has not been demonstrated (unpublished data). This proposal is supported by the finding that PapA mutants that fail to multimerize and/or are not assembly competent (Bullitt et al., 1996) are also non-toxic. Although PapG is believed to exist in single or very low copy in the pilus (Kuehn et al., 1992) it remains a highly toxic protein. Earlier work showed that PapG diluted out of urea formed an aggregate that failed to enter an isoelectric focusing gel, suggesting that PapG aggregates can form under certain conditions (Kuehn et al., 1991). PapG is also larger (35 kDa) than the other pilins (13-17 kDa) due to the presence of the N-terminal Gal(1-4)Gal binding domain. We suggest that toxicity is related to specific properties in the C-terminal zipper that influence inner membrane retention, as well as other hydrophobic surfaces that may be exposed during folding or in the membrane-retained state, but the precise cause of toxicity is not clearly understood.
Our results show that off-pathway products of pilus biogenesis require new gene products to be synthesized and recruited to the periplasm to prevent toxicity. There are at least two pathways that monitor the assembly state of extracytoplasmic proteins (Mecsas et al., 1993; Cosma et al., 1995; Danese et al., 1995; Raina et al., 1995; Danese and Silhavy, 1997; De Las Penas et al., 1997; Missiakis and Raina, 1997; Missiakis et al., 1997; Figure 7). The Cpx two-component and the E modulatory pathways respond to different extracytoplasmic signals and each pathway uses these external stimuli to activate degP transcription in addition to several other genes (Danese and Silhavy, 1997; De Las Penas et al., 1997; Missiakis and Raina, 1997; Missiakis et al., 1997). In this study we have shown that PapG synthesis stimulates degP transcription by activating both the Cpx and E modulatory pathways. PapE synthesis only results in activation of degP via the Cpx pathway. Co-production of PapD suppresses transcriptional activation, suggesting that the signal sensed in the periplasm is misfolding or aggregation of the subunits. However, the signal could also result from membrane perturbation due to PapG association with the inner membrane. In either case, these non-productive associations appear to be triggered by the C-terminal zipper. If the subunit zipper is not correctly positioned along the exposed edge of the G1 -strand of PapD it apparently engages in the non-productive associations that stimulate degP transcription.
The precise mechanism of activation of the Cpx and E modulatory pathways is unclear (Danese and Silhavy, 1997; De Las Penas et al., 1997; Missiakis and Raina, 1997; Missiakis et al., 1997). It appears that these systems evolved to ameliorate perturbations in certain protein assembly processes by up-regulating the synthesis of protein folding agents, such as DsbA (Danese and Silhavy, 1997; Pogliano et al., 1997) and peptidyl-prolyl isomerases (Missiakis et al., 1996; Danese and Silhavy, 1997; Pogliano et al., 1997), and proteases, like DegP, that can destroy misfolded proteins. We suggest that, in terms of pilus biogenesis, the Cpx and E modulatory pathways may function to monitor pilus assembly and use this information to control synthesis of extracytoplasmic proteases and protein folding agents that can ultimately regulate the assembly process. Furthermore, these two signal transduction pathways could also conceivably be used to transmit signals relevant to pathogenesis, such as signaling attachment to the host epithelium.
Materials and methods
Strains and genetic constructs
KS272 {MC1000 [F- (ara-leu. 7697 galE galK lacX74 rpsL (strr)]} and the isogenic protease-deficient strain KS474 (KS272 degP::kan) were kindly provided by J.Beckwith and have been previously described (Strauch and Beckwith, 1988; Strauch et al., 1989). DH5[supE44 lacU169 (
80 lacZM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1] (Hanahan, 1983) was used in all cloning steps and in expression studies with papG. Other constructs used were PND2000 (MC4100, 
RS88[degP–lacZ]) (Danese et al., 1995), SP558 (PND2000, cpxA::cam), SP559 (PND2000, cpxR::), SP616 (MC4100, RS45 [rpoHP3–lacZ]) (Mecsas et al., 1993), SP594 (MC4100, RS88 [cpxP–lacZ]) (P.N.Danese, doctoral thesis, Princeton University), SP619 (SP594, cpxA::cam) and SP620 (SP594, cpxR::). pMMB66 was used as an expression plasmid to place subunits under control of the Ptac promoter. pHJ2 was constructed by moving the papA-containing BamHI–EcoRI fragment from pPAPAY162 (a gift of M.-J.Wick, Washington University) into pMMB66 (Furste et al., 1986). pHJ8 was constructed by moving the BamHI–EcoRI fragment, containing papG, from pPAP55 (Hultgren et al., 1989) into pMMB66. pHJ13 was constructed by moving the BamHI–EcoRI fragment, containing papE, from pPAP63 (Lindberg et al., 1989) into pMMB66. pTK5 was constructed by cloning the BamHI–SalI fragment, containing papK, from pFJ11 (Jacob-Dubuisson et al., 1993) into pMMB66. The arabinose promoter plasmid pMON6235 was kindly provided by M.Obukowitz (Monsanto Corporate Research, St Louis, MO). pMON6235cat was created by digesting pMON6235 with NcoI and HindIII (removing the cat gene) and religating the large fragment. pHJ9 was constructed by cloning a spectinomycin cassette (kindly provided by M.Caparon, Washington University) into the EcoRI and PvuII sites of pMON6235cat, truncating the resident amp gene. papD was cloned under the control of the Para promoter in pHJ9 in two steps. First, papD was moved as a BamHI–EcoRI fragment from pLS101 (Slonim et al., 1992) into pIC20H, creating pHJ3. papD was then moved from pHJ3 as a BamHI–BglII fragment into pHJ9, creating pHJ9203. Point mutant derivatives in PapD (pHJ9204 carries the R8A derivative of PapD) were moved into pHJ9 following a similar protocol. pHJ23, containing a C-terminal truncated version of papG, was previously described (Xu et al., 1995). pHJ7 (F314S) and pHJ30 (G302V+F314S) were constructed using the dut- ung- protocol essentially as described previously (Slonim et al., 1992). pLSR8A and pK112A have been previously described (Slonim et al., 1992; Kuehn et al., 1993). Mutants were constructed using the dut- ung- protocol essentially as described previously (Zoller and Smith, 1983). pHJ31 was constructed by restricting pHJ21 with BamHI and EcoRI and filling in the ends with Klenow fragment. The blunt fragment was then cloned into the SmaI site of pTRC99A, placing the papG gene under control of the trc promoter. pHJ9208 was constructed in two steps. First, the papG-containing BamHI–EcoRI fragment was moved from pHJ8 into pUC19, creating pHJ21. The papG-containing BamHI–BglII fragment was then moved from pHJ21 into pHJ9, creating pHJ9208. The PapD–FimC chimera was constructed by cloning a PCR product containing domain 2 of FimC into HindIII- and BamHI-restricted pLS101, replacing domain 2 of PapD. pRS3A and pRS4A carry the wild-type and leaderless versions of papA respectively in pTRC99A (R.T.Striker, unpublished data). pHJ33, pHJ34 and pHJ35, carrying the G150T, Y162L and C22S mutations in papA respectively, have been previously described (Bullitt et al., 1996). pFJ29 and pFJ29-71, containing the whole pap operon and the papD null allele respectively, have been previously described (Jacob-Dubuisson et al., 1994).
Toxicity assay
Overnight cultures, grown at 30°C in the presence of appropriate antibiotics, were diluted 100-fold into Luria broth containing appropriate antibiotics and inducers. Toxicity was achieved with three of the subunits studied using different levels of isopropyl--D-thiogalactoside (IPTG). PapG toxicity required 80 M IPTG, while toxicity of PapA required 0.3 mM IPTG. PapE was toxic when induced at 35 M IPTG, while PapK was non-toxic at 0.5 mM IPTG. Arabinose was added to the culture at 4% at the onset of growth to induce PapD and PapD mutants in the pHJ9 vector for suppression of PapA toxicity. pHJ31 (papG) and pHJ31/pLS10 (papG, papD) were used to demonstrate PapG toxicity and suppression by PapD, due to an apparent suppressant effect of arabinose precluding the use of pHJ9203 (papD) for induction of PapD by arabinose addition. This suppressant effect of arabinose was observed to a lesser degree with PapE toxicity, but was not observed with PapA toxicity and suppression with PapD induced from pHJ9203. Following a 2 h lag, the OD600 of the cultures was monitored at hourly intervals. Growth was followed for at least 6 h. Viable counts (c.f.u.) were plated at each time point, revealing that 'toxicity' was a manifestation of a failure of the culture to grow following subunit induction, i.e. 1107 c.f.u. at time 0 and 6 h later, whereas in the presence of the chaperone the c.f.u. reached >1109 over the growth period.
Chaperone and subunit induction and bacterial fractionation
Bacterial cultures were grown in Luria broth in the presence of appropriate antibiotics at 30°C to an OD600 of 0.6, at which time IPTG was added to 0.5 mM, -D-arabinose to 0.2% and growth allowed to proceed for 60 min. Fractionation of bacteria was performed as previously described (Dodson et al., 1993). Briefly, periplasm was isolated from equivalent gram quantities of bacteria by the sucrose/lysozyme method (Jones et al., 1993). Following periplasm isolation, cytoplasm and membrane fractions were obtained by sonication followed by centrifugation. The membrane pellet was fractionated into inner and outer membrane fractions by differential solubility in sarkosyl (Dodson et al., 1993).
Spheroplast release assay
Spheroplasts were produced following the procedure of Witholt et al. (1976) as follows. KS474/pHJ8 was grown in minimal medium A with 0.2% glycerol overnight and diluted 1/10 into fresh minimal medium A. At an A600 of 0.8 the bacteria were harvested by centrifugation and resuspended in 1 ml 0.2 M Tris–HCl, pH 8. The cell suspension was then diluted by 1/2 into 0.2 M Tris–HCl, pH 8, 1 M sucrose, 0.5 mM EDTA. Lysozyme was added to 60 g/ml final concentration. The suspension was again diluted 2-fold with dH2O and incubated at 25°C for 20 min. Spheroplast formation was monitored by light microscopy, with >95% spheroplast formation occurring within 15 min. Spheroplasts were stabilized by addition of 0.02 M MgCl2. The spheroplasts were collected by centrifugation and resuspended in minimal medium A plus 0.02 M MgCl2. To assay the partitioning of PapG from the spheroplast suspension, 1 mM IPTG was added and incubation allowed to proceed for 5 min. The suspension was then pulsed with 35S Trans label (cysteine + methionine) (ICN, Costa Mesa, CA) for 1 min, followed by a 10 min unlabeled chase with 0.1 M cysteine + methionine. Purified PapD (2–10 g) was added at either the IPTG induction point or following a 10–30 min cold chase. The spheroplast pellet was removed by centrifugation and the supernatant subjected to immunoprecipitation with anti-Pap tip antisera (Medimunne Inc., Gaithersburg, MD), which has strong reactivity against PapG. The immunoprecipitations were quantitated with a Phosphorimager (BioRad, Hecules, CA) and Molecular Analyst software (BioRad).
-galactosidase assays
Bacteria were grown overnight in Luria broth, subcultured (1:50) and grown for 60 min, at which point IPTG was added to 0.5 mM to induce pilus subunit proteins. The induced strains were allowed to grow for 90 min and then harvested for -galactosidase assay. -galactosidase activities were determined utilizing a microtiter plate assay (Slauch and Silhavy, 1991) and expressed as (U. A600)103, where U = mol product formed/min. A minimum of four independent assays were performed on each strain and averaged. pHJ8 (papG) and pHJ13 (papE) were used to demonstrate promoter activation following induction of synthesis by IPTG addition. pHJ9208 (papG) and pHJ6 (papD) were used to demonstrate suppression of promoter activation by PapG synthesis.
Gal
(1-4)Gal chromatography
The PapD–PapG complex was purified from periplasmic fractions on Gal(1-4)Gal–Sepharose beads as previously described (Hultgren et al., 1989). For the folding studies 35S-labeled periplasm (see below) was applied to Gal(1-4)Gal–Sepharose beads in 1 M NaCl, 1 PBS (0.12 M NaCl, 2.7 mM KCl, 0.067 M Na2HPO4, 0.014 M KH2PO4, pH 7.4) for 30 min at 25°C. Following extensive washing with 1 M NaCl, 1 PBS the bound material was eluted with soluble Gal(1-4)Gal (17 mg/ml) in 1 M NaCl, 1 PBS by gentle rocking for 15 min at 25°C.
Other methods
SDS-PAGE, isoelectric focusing and Western blot analysis were performed as previously described (Slonim et al., 1992; Dodson et al., 1993; Jones et al., 1993; Xu et al., 1995). Pulse/chase of periplasmic contents was performed essentially as described previously (Stader et al., 1986; Snyder and Silhavy, 1992). Bacteria were diluted, from an overnight culture in minimal medium A, 1/10 into fresh minimal medium A and allowed to grow to an A600 of 0.8, at which time the culture was induced with 1 mM IPTG for 15 min. An aliquot of 1 ml of the induced culture was then pulsed with 35S Trans label (cysteine + methionine) for 1 min and chased for 10 min with 0.1 M unlabeled cysteine + methionine. Labeling was stopped by dilution of the label mixture into ice-cold 50 mM Tris-HCl, pH 8, 100 g/ml chloramphenicol. Following periplasm preparation, immunoprecipitation was carried out with anti-Pap tip antisera (Medimmune Inc., Gaithersburg, MD), which contains strong reactivity against PapG.
Acknowledgements
Top of pageThe authors would like to thank K.Dodson and S.Knight for critical reading of the manuscript. C.H.J. is supported by NIH Postdoctoral Fellowship F32AIO8665. P.N.D. gratefully acknowledges NIH Training Grant GM07388. This work was supported by NIH grant RO1A129549 to S.J.H. and NIGMS grant GM34821 to T.J.S.
References
Top of pageBullitt E, Jones CH, Striker R, Soto G, Jacob-Dubuisson F, Pinkner J, Wick M-J, Makowski L and Hultgren SJ (1996) Development of pilus organelle subassemblies in vitro depends on chaperone uncapping of a beta zipper. Proc Natl Acad Sci USA, 93, 12890–12895. | Article | PubMed | ChemPort |
Cosma CL, Danese PN, Carlson JH, Silhavy TJ and Snyder WB (1995) Mutational activation of the Cpx signal transduction pathway of Escherichia coli suppresses the toxicity conferred by certain envelope-associated stresses. Mol Microbiol, 18, 491–505. | Article | PubMed | ISI | ChemPort |
Danese PN and Silhavy TJ (1997) The E and the Cpx signal transduction systems control the synthesis of periplasmic protein-folding enzymes in Escherichia coli. Genes Dev, 11, 1183–1193. | PubMed | ISI | ChemPort |
Danese PN, Snyder WB, Cosma CL, Davis LJB and Silhavy TJ (1995) The Cpx two-component signal transduction pathway of Escherichia coli regulates degP transcription. Genes Dev, 9, 387–398. | Article | PubMed | ISI | ChemPort |
De Las Penas A, Connolly L and Gross CA (1997) The E-mediated response to extracytoplasmic stress in Escherichia coli is transduced by RseA and RseB, two negative regulators of E. Mol Microbiol, 24, 373–385. | Article | PubMed | ISI | ChemPort |
Dodd DC, Bassford PJJ and Eisenstein BI (1984) Dependence of secretion and assembly of type 1 fimbrial subunits of Escherichia coli on normal protein transport. J Bacteriol, 159, 1077–1079. | PubMed | ISI | ChemPort |
Dodson KW, Jacob-Dubuisson F, Striker RT and Hultgren SJ (1993) Outer membrane PapC usher discriminately recognizes periplasmic chaperone–pilus subunit complexes. Proc Natl Acad Sci USA, 90, 3670–3674. | Article | PubMed | ChemPort |
Furste JP, Pansegrau W, Frank R, Blocker H, Scholz P, Bagdasarian M and Lanka E (1986) Molecular cloning of the plasmid RP4 primase region in a multi-host-range tacP expression vector. Gene, 48, 119–131. | Article | PubMed | ISI | ChemPort |
Hanahan D (1983) Studies in transformation of Escherichia coli with plasmids. J Mol Biol, 166, 557–580. | Article | PubMed | ISI | ChemPort |
Hultgren SJ, Lindberg F, Magnusson G, Kihlberg J, Tennent JM and Normark S (1989) The PapG adhesin of uropathogenic Escherichia coli contains separate regions for receptor binding and for the incorporation into the pilus. Proc Natl Acad Sci USA, 86, 4357–4361. | Article | PubMed | ChemPort |
Hultgren SJ, Normark S and Abraham SN (1991) Chaperone-assisted assembly and molecular architecture of adhesive pili. Annu Rev Microbiol, 45, 383–415. | Article | PubMed | ISI | ChemPort |
Hultgren SJ, Abraham SN, Caparon MG, Falk P, St Geme JW and Normark S (1993) Pilus and non-pilus bacterial adhesins: assembly and function in cell recognition. Cell, 73, 887–901. | Article | PubMed | ISI | ChemPort |
Hultgren SJ, Jones CH and Normark SN (1996) Bacterial adhesins and their assembly. In Neidhardt,F.C. (ed.), Escherichia coli and Salmonella; Cellular and Molecular Biology. ASM Press, Washington, DC, pp. 2730–2756.
Hung DL, Knight SD, Woods RM, Pinkner JS and Hultgren SJ (1996) Molecular basis of two subfamilies of immunoglobulin-like chaperones. EMBO J, 15, 3792–3805. | PubMed | ISI | ChemPort |
Jacob-Dubuisson F, Heuser J, Dodson K, Normark S and Hultgren SJ (1993) Initiation of assembly and association of the structural elements of a bacterial pilus depend on two specialized tip proteins. EMBO J, 12, 837–847. | PubMed | ChemPort |
Jacob-Dubuisson F, Pinkner J, Xu Z, Striker R, Padmanhaban A and Hultgren SJ (1994) PapD chaperone function in pilus biogenesis depends on oxidant and chaperone-like activities of DsbA. Proc Natl Acad Sci USA, 91, 11552–11556. | Article | PubMed | ChemPort |
Jones CH, Pinkner JS, Nicholes AV, Slonim LN, Abraham SN and Hultgren SJ (1993) FimC is a periplasmic PapD-like chaperone that directs assembly of type 1 pili in bacteria. Proc Natl Acad Sci USA, 90, 8397–8401. | PubMed | ChemPort |
Kolmar H, Waller PRH and Sauer RT (1996) The DegP and DegQ periplasmic endoproteases of Escherichia coli: specificity for cleavage sites and substrate conformation. J Bacteriol, 178, 5925–5929. | PubMed | ISI | ChemPort |
Kuehn MJ, Normark S and Hultgren SJ (1991) Immunoglobulin-like PapD chaperone caps and uncaps interactive surfaces of nascently translocated pilus subunits. Proc Natl Acad Sci USA, 88, 10586–10590. | PubMed | ChemPort |
Kuehn MJ, Heuser J, Normark S and Hultgren SJ (1992) P pili in uropathogenic E. coli are composite fibres with distinct fibrillar adhesive tips. Nature, 356, 252–255. | Article | PubMed | ISI | ChemPort |
Kuehn MJ, Ogg DJ, Kihlberg J, Slonim LN, Flemmer K, Bergfors T and Hultgren SJ (1993) Structural basis of pilus subunit recognition by the PapD chaperone. Science, 262, 1234–1241. | Article | PubMed | ISI | ChemPort |
Kuroiwa T, Sakaguchi M, Mihara K and Omura T (1991) Systematic analysis of stop transfer sequence for microsomal membrane. J Biol Chem, 266, 9251–9255. | PubMed | ISI | ChemPort |
Lindberg F, Tennent JM, Hultgren SJ, Lund B and Normark S (1989) PapD, a periplasmic transport protein in P-pilus biogenesis. J Bacteriol, 171, 6052–6058. | PubMed | ISI | ChemPort |
Lipinska B, Zylicz M and Georgopoulos C (1990) The HtrA (DegP) protein, essential for Escherichia coli survival at high temperatures, is an endopeptidase. J Bacteriol, 172, 1791–1797. | PubMed | ISI | ChemPort |
Matsuyama S-I, Tajima T and Tokuda H (1995) A novel periplasmic carrier protein involved in the sorting and transport of Escherichia coli lipoproteins destined for the outer membrane. EMBO J, 14, 3365–3372. | PubMed | ISI | ChemPort |
Mecsas J, Rouviere PE, Erickson JW, Donohue TJ and Gross CA (1993) The activity of E, an Escherichia coli heat-inducible sigma factor, is modulated by outer membrane proteins. Genes Dev, 7, 2618–2628. | PubMed | ISI | ChemPort |
Missiakis D and Raina S (1997) Signal transduction pathways in response to protein misfolding in the extracytoplasmic compartments of E.coli: role of two new phophoprotein phophatases PrpA and PrpB. EMBO J, 16, 1670–1685. | Article | PubMed | ISI | ChemPort |
Missiakis D, Betton J-M and Raina S (1996) New components of protein folding in extracytoplasmic compartments of Escherichia coli SurA, FkpA and Skp/OmpH. Mol Microbiol, 21, 871–884. | Article | PubMed | ChemPort |
Missiakis D, Mayer MP, Lemaire M, Georgopoulos C and Raina S (1997) Modulation of the Escherichia coli E (RpoE) heat-shock transcription-factor activity by the RseA, RseB and RseC proteins. Mol Microbiol, 24, 355–371. | Article | PubMed | ChemPort |
PoglianoJ, Lynch AS, Belin D, Lin ECC and Beckwith J (1997) Regulation of Escherichia coli cell envelope proteins involved in protein folding and degradation by the Cpx two-component system. Genes Dev. 11, 1169–1182.
Pugsley A and Possot G (1994) The general secretory pathway of Klebsiella oxytaca: no evidence for relocalization or assembly of pilin-like PulG protein into a multiprotein complex. Mol Microbiol, 10, 665–674.
Raina S, Missiakas D, Baird L, Kumar S and Georgopoulos C (1993) Identification and transcriptional analysis of the Escherichia coli htrE operon which is homologous to pap and related pilin operons. J Bacteriol, 175, 5009–5021. | PubMed | ChemPort |
Raina S, Missiakis D and Georgopoulos C (1995) The rpoE gene encoding the sigma E (sigma 24) heat shock sigma factor of Escherichia coli. EMBO J, 14, 1043–1055. | PubMed | ISI | ChemPort |
Roberts JA et al. (1994) The Gal (1-4) Gal-specific tip adhesin of Escherichia coli P-fimbriae is needed for pyelonephritis to occur in the normal urinary tract. Proc Natl Acad Sci USA, 91, 11889–11893. | Article | PubMed | ChemPort |
Rouviere PE, de las Penas A, Mecsas J, Lu CZ, Rudd KE and Gross CA (1995) rpoE, the gene encoding the second heat-shock sigma factor, sigma E, in Escherichia coliEMBO J., 14, 1032–1042.
Slauch JM and Silhavy TJ (1991) cis-acting ompF mutations that result in OmpR-dependent constitutive expression. J Bacteriol, 173, 4039–4048. | PubMed | ISI | ChemPort |
Slonim LN, Pinkner JS, Branden CI and Hultgren SJ (1992) Interactive surface in the PapD chaperone cleft is conserved in pilus chaperone superfamily and essential in subunit recognition and assembly. EMBO J, 11, 4747–4756. | PubMed | ISI | ChemPort |
Snyder WB and Silhavy TJ (1992) Enhanced export of -galactosidase fusion proteins on prlF mutants is lon dependent. J Bacteriol, 174, 5661–5668. | PubMed | ChemPort |
Snyder WB, Davis LJ, Danese PN, Cosma CL and Silhavy TJ (1995) Overproduction of NlpE, a new outer membrane lipoprotein, suppresses the toxicity of periplasmic LacZ by activation of the Cpx signal transduction pathway. J Bacteriol, 177, 4216–4223. | PubMed | ISI | ChemPort |
Stader J, Benson SA and Silhavy TJ (1986) Kinetic analysis of lamB mutant suggests the signal sequence plays multiple roles in protein export. J Biol Chem, 261, 15075–15080. | PubMed | ChemPort |
Strauch KL and Beckwith J (1988) An Escherichia coli mutation preventing degradation of abnormal periplasmic proteins. Proc Natl Acad Sci USA, 85, 1576–1580. | PubMed | ChemPort |
Strauch KL, Johnson K and Beckwith J (1989) Characterization of degP, a gene required for proteolysis in the cell envelope of Escherichia coli at high temperature. J Bacteriol, 171, 2689–2696. | PubMed | ISI | ChemPort |
Weis W, Brown JH, Cusack S, Paulson JC, Skehel JJ and Wiley DC (1988) Structure of the influenza virus haemagglutinin complexed with its receptor, sialic acid. Nature, 333, 426–431. | Article | PubMed | ISI | ChemPort |
Witholt B, Boekhout M, Brock M, Kingman J, van Heerikhuizen H and de Leig L (1976) An efficient and reproducible procedure for the formation of spheroplasts from variously grown Escherichia coli. Anal Biochem, 74, 160–170. | Article | PubMed | ISI | ChemPort |
Xu Z, Jones CH, Haslam D, Dodson K, Kihlberg J and Hultgren SJ (1995) Molecular dissection of PapD interaction with PapG reveals two chaperone binding sites. Mol Microbiol, 16, 1011–1020. | Article | PubMed | ISI | ChemPort |
Zoller MJ and Smith M (1983) Oligonucleotide-directed mutagenesis of DNA fragments cloned into M13 vectors. Methods Enzymol, 100, 468–500. | PubMed | ChemPort |
