Different Mechanisms for Thermal Inactivation of Bacillus subtilis Signal Peptidase Mutants*

The type I signal peptidase SipS ofBacillus subtilis is of major importance for the processing of secretory precursor proteins. In the present studies, we have investigated possible mechanisms of thermal inactivation of five temperature-sensitive SipS mutants. The results demonstrate that two of these mutants, L74A and Y81A, are structurally stable but strongly impaired in catalytic activity at 48 °C, showing the (unprecedented) involvement of the conserved leucine 74 and tyrosine 81 residues in the catalytic reaction of type I signal peptidases. This conclusion is supported by the crystal structure of the homologous signal peptidase of Escherichia coli (Paetzel, M., Dalbey, R. E., and Strynadka, N. C. J. (1998) Nature 396, 186–190). In contrast, the SipS mutant proteins R84A, R84H, and D146A were inactivated by proteolytic degradation, indicating that the conserved arginine 84 and aspartic acid 146 residues are required to obtain a protease-resistant conformation. The cell wall-bound protease WprA was shown to be involved in the degradation of SipS D146A, which is in accord with the fact that SipS has a large extracytoplasmic domain. As WprA was not involved in the degradation of the SipS mutant proteins R84A and R84H, we conclude that multiple proteases are responsible for the thermal inactivation of temperature-sensitive SipS mutants.

Bacterial proteins that are exported from the cytoplasm via the general secretion pathway are synthesized with an aminoterminal signal peptide. The signal peptide is required for the targeting of pre-proteins to the membrane and the initiation of protein translocation across this membrane (for review, see Ref. 1). During, or shortly after, the translocation of secretory pre-proteins, signal peptides are removed by type I signal peptidases (SPases), 1 which is a prerequisite for the release of the mature protein from the trans side of the membrane (for review, see Ref. 2).
Type I SPases belong to a special class of serine peptidases (peptidase classification: clan SF, family S26; Ref. 3) with con-served serine and lysine residues, which are essential for catalysis, most likely by forming a serine-lysine catalytic dyad (see Refs. 2,4,and 5). As demonstrated by the crystallographic analysis of the type I SPase of Escherichia coli (also known as leader peptidase), the active site serine residue acts as a nucleophile attacking the carbonyl carbon of the scissile peptide bond at the SPase recognition site, whereas the unprotonated form of the lysine ⑀-amino group probably serves to activate the hydroxyl group of the serine residue. A similar catalytic mechanism has been proposed for the structurally related LexA-like proteases (4 -8), which employ a serine-lysine catalytic dyad for self-cleavage (7,9,10).
In the Gram-positive bacterium Bacillus subtilis five chromosomally encoded type I SPases have been identified, which are denoted SipS, SipT, SipU, SipV, and SipW (11)(12)(13). SipS and SipT are of major importance for protein secretion, whereas SipU, SipV, and SipW contribute only to a minor extent to the processing of secretory pre-proteins (12)(13)(14). Cells depleted of functional SipS and SipT stop growing and lyse, showing that the presence of either SipS or SipT is essential for growth and viability. This was demonstrated with a B. subtilis strain in which the chromosomal copies of the sipS and sipT genes were disrupted and functionally replaced by one of five different plasmid-borne genes for temperature-sensitive SipS mutant proteins (i.e. L74A, Y81A, R84A, R84H, and D146A; Ref. 13). In the present study, we have investigated the mechanism of thermal inactivation of these five SipS mutant proteins. The results show that SipS L74A and Y81A are structurally stable at high temperature, indicating that the substituted residues are, in addition to the putative active site serine (Ser-43) and lysine residues (Lys-83), involved in catalysis. By contrast, SipS R84A, R84H, and D146A are prone to proteolytic degradation, in particular at high temperature, showing that the residues at these positions are required to maintain resistance to proteases. Interestingly, a cell wallbound protease, WprA, was shown to be involved in the degradation of SipS D146A, but not in the degradation of SipS R84A and R84H. Table I lists the plasmids and bacterial strains used. TY medium was prepared as described in Ref. 15. GCHE medium was prepared as described in Ref. 16. Antibiotics were used in the following concentrations: chloramphenicol, 5 g/ml; erythromycin, 1 g/ml; kanamycin, 10 g/ml.

Plasmids, Bacterial Strains, and Media-
Transformation of B. subtilis-B. subtilis was transformed by growth in GCHE medium until an optical density at 600 nm (OD 600 ) of approximately 1.0, subsequent addition of DNA to the culture, and continued growth for 4 h.
Western Blot Analysis-Western blotting was performed using a semi-dry system as described by Kyhse-Andersen (17). After separation by SDS-polyacrylamide gel electrophoresis, proteins were transferred to polyvinylidene difluoride membranes (Roche Molecular Biochemi-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by Biotechnology Grant Bio2-CT93-0254 from the European Union.

Stability of SipS Mutant
Proteins at 48 0 C-We have shown previously that the SipS mutant proteins R84A, R84H, and D146A are prone to proteolytic degradation at 37°C, showing very low levels of SPase activity (6). Nevertheless, at this temperature sufficient active molecules of SipS R84A, R84H, or D146A are produced in B. subtilis to replace the wild-type forms of SipS and SipT (13). To investigate whether the inactivity of these three mutant proteins at 48°C (13) was due to increased proteolytic degradation, Western blotting experiments were carried out with B. subtilis 8G5 sipS (⌬S; lacking the chromosomal sipS gene) transformed with plasmid pS-R84A, pS-R84H, or pS-D146A (Table I). As shown in Fig. 1 (A and B), SipS R84A, R84H, or D146A were detectable at 37°C, but not at 48°C. Furthermore, at 37°C all three mutant proteins were present in reduced amounts compared with wildtype SipS, the D146A mutant protein being most unstable. Examination of the accumulation of the hybrid precursor pre(A13i)-␤-lactamase (11) in B. subtilis ⌬ST (lacking the chromosomal sipS and sipT genes) transformed with pS-R84A, pS-R84H, or pS-D146A, showed that SPase activity was strongly reduced at 48°C, compared with B. subtilis ⌬ST transformed with pGDL41, specifying wild-type SipS (Fig. 1C). Thus, impaired SPase activity at 48°C was paralleled by the disappearance of the SipS mutant proteins, indicating that SipS R84A, R84H, and D146A were inactivated by proteolysis.
Similarly, the stability of the SipS mutant proteins L74A and Y81A, which compared with SipS R84A, R84H, and D146A, showed a relatively high activity at 37°C (6) and which were unable to replace SipS and SipT at 48°C (13), was investigated by Western blotting, using B. subtilis ⌬S transformed with plasmids pS-L74A or pS-Y81A. Unexpectedly, both at 37 and 48°C, the levels of SipS L74A and Y81A were comparable with those of wild-type SipS (Fig. 1, A and B). The impaired SPase activity of SipS L74A and Y81A at 48°C was reflected by the increased accumulation of pre(A13i)-␤-lactamase in B. subtilis ⌬ST transformed with plasmids pS-L74A or pS-Y81A, as compared with B. subtilis ⌬ST producing wild-type SipS (Fig. 1C). These results show that leucine 74 and tyrosine 81 are very important for catalysis at 48°C, but not for protease resistance of SipS.
SipS D146A Is a Substrate for the Cell Wall-bound Protease WprA-As SipS is a type II membrane protein, the largest (carboxyl-terminal) part of which is exposed to the extracytoplasmic side of the membrane (11), it seems likely that proteases residing in the membrane or cell wall are responsible for the degradation of the SipS R84A, R84H, and D146A mutant proteins. As a first approach to identify the proteases responsible for their degradation, we tested the stability of SipS D146A, the most unstable mutant, in B. subtilis strains lacking (putative) membrane-bound proteases such as FtsH, HtrA, protease IV or Tsp (for a review on the E. coli homologues of these proteases, see Ref. 18), or the cell wall-bound protease WprA (19,20). Strikingly, none of the membrane-bound proteases appeared to be involved in the degradation of SipS D146A (data not shown). In contrast, at 37°C, B. subtilis W⌬S (depleted of WprA, lacking wild-type SipS) accumulated approximately 10fold higher levels of SipS D146A than B. subtilis ⌬S (Fig. 2), showing that WprA is involved in the degradation of SipS D146A. Interestingly, the levels of SipS R84A and R84H were subtilis ⌬S (lacking the chromosomal sipS gene) containing pGDL41 (wild-type (wt)), pS-L74A, pS-Y81A, pS-R84A, pS-R84H, or pS-D146A were grown in TY medium until the mid-exponential growth phase (OD 600 Ϸ 0.7). Next, the cultures were divided into two aliquots, which were incubated for 3 h at either 37 (A) or 48°C (B). Cells were separated from the growth medium by centrifugation, and the cellular levels of the different SipS proteins were analyzed by SDS-polyacrylamide gel electrophoresis and Western blotting. The position of SipS is indicated. C, the accumulation of pre(A13i)-␤-lactamase in B. subtilis ⌬ST containing pGDL41 (wild-type (wt)), pS-L74A, pS-Y81A, pS-R84A, pS-R84H, or pS-D146A was determined by Western blotting as in B. p, precursor; m, mature. Suppression of the Temperature Sensitivity of B. subtilis ⌬ST (pS-D146A) by WprA Depletion-We have shown previously that the production of SipS D146A, to levels that are below the detection limit for Western blotting, can be sufficient for growth of strains lacking SipS and SipT, provided that the temperature is not raised above 42°C (13). To determine whether the depletion of WprA in B. subtilis ⌬ST (pS-D146A) could suppress the temperature sensitivity of this strain, even though the depletion of WprA did not result in the accumulation of detectable amounts of SipS D146A at 48°C (see above), a wprA mutation was introduced into this strain. As shown in Fig. 3, the depletion of WprA resulted in the suppression of the temperature sensitivity of B. subtilis ⌬ST (pS-D146A); while B. subtilis ⌬ST (pS-D146A) stopped growing and started to lyse upon a temperature shift from 37 to 48°C, the corresponding WprA-depleted strain continued to grow normally at 48°C. These observations show that depletion of WprA resulted in increased levels of active SipS D146A even at 48°C. DISCUSSION In the present studies, we show that temperature-sensitive mutants of SipS are inactivated via different mechanisms. Two major classes of temperature-sensitive SipS mutants were identified. The first class of mutant proteins, consisting of SipS L74A and Y81A, is structurally stable but has impaired catalytic activity at 48°C, while the second class of mutant proteins, consisting of SipS R84A, R84H, and D146A, is sensitive to proteolysis, in particular at 48°C.
The observation that SipS L74A and Y81A are structurally stable, showing a particularly reduced enzymatic activity at 48°C, indicates that leucine 74 and tyrosine 81 are involved in catalysis. This is a novel conclusion, which was not evident from previous site-directed mutagenesis experiments with SipS or other type I SPases (for review, see Ref. 2), but which is supported by the recently published crystal structure of the E. coli SPase I (5). As evidenced by the latter structure, the side chains of phenylalanine 133, tyrosine 143, and methionine 270, and the main chain atoms of methionine 270, methionine 271, glycine 272, and alanine 279 make van der Waals contacts with the side chain of the active site lysine residue 145. Thus, the ⑀-amino group of lysine 145 is located in a hydrophobic environment, which is probably essential to lower its pK a to such an extent that it can function as a general base (5). As leucine 74 and tyrosine 81 of SipS are the equivalents of phenylalanine 133 and tyrosine 143 of E. coli SPase I, it seems likely that their side chains are required to lower the pK a of the active site lysine residue 83 of SipS. Even though the latter hypothesis provides an explanation for the reduced activities of the SipS L74A and Y81A mutant proteins, it does not explain why these mutant proteins display some residual activity at 37°C (6, 13), but not at 48°C. It is, however, conceivable that the pK a of lysine 83 is further increased at elevated temperature due to local unfolding events. Alternatively, local unfolding in the L74A and Y81A mutant proteins at 48°C may affect the interaction between the active site serine 43 and lysine 83 residues.
In contrast to SipS L74A and Y81A, the thermal inactivation of the SipS mutants R84A, R84H, and D146A seems to be based on proteolysis, suggesting that these mutations strongly retard the folding of SipS to a protease-resistant conformation or that they make the folded SipS protein protease-sensitive. Based on the crystal structure of E. coli SPase I (5), we favor the latter hypothesis, because the equivalent residues of arginine 84 and aspartic acid 146 of SipS in the E. coli SPase I (i.e. arginine 146 and aspartic acid 273) form a salt bridge, 2 which may be required to rigidify the structure of type I SPases, such as SipS. Our present results show that the wall-bound protease WprA (19), which was shown recently to be involved in the degradation of folding intermediates of secreted proteins (20), is involved in the degradation of SipS D146A, but not of SipS R84A and R84H. Furthermore, depletion of WprA did not result in the accumulation of SipS D146A to wild-type levels. Thus, it seems that other, as yet unidentified, proteases at the membrane-cell wall interface are also involved in the degradation of structurally unstable SipS mutant proteins. These could even include other type I SPases, such as SipU, SipV, or SipW, as suggested by the observation that the homologous Sec11p subunit of the SPase complex in the yeast endoplasmatic reticular membrane is involved in protein degradation (21). Finally, we are currently unable to exclude the possibility that cytosolic proteases are involved the degradation of SipS mutant pro-2 R. E. Dalbey, personal communication. In the resulting strain W⌬S, the transcription of the wprA gene is IPTG-dependent (20). Cells of B. subtilis ⌬S or W⌬S, which were transformed with pGDL41 (wild-type (wt)), pS-R84A, pS-R84H, or pS-D146A, were grown at 37°C in TY medium without IPTG until the beginning of the post-exponential growth phase. Next, the cellular levels of SipS wildtype and mutant proteins were analyzed by SDS-polyacrylamide gel electrophoresis and Western blotting as in Fig. 1, A and B. To visualize the lack of effect of WprA depletion on wild-type SipS, which is produced in much higher amounts than SipS R84A, R84H, or D146A (see Fig. 1), the film was exposed for a shorter period of time. The position of SipS is indicated. subtilis W⌬ST (pS-D146A) (indicated with the f) were grown in TY medium without IPTG at 37°C. When the cells reached the mid-exponential growth phase (OD 600 Ϸ 0.7; indicated with an arrow), the temperature was shifted to 48°C and incubation continued. Samples were taken for optical density readings at half-hourly intervals. teins, either before, during, or after membrane insertion. The latter possibility is particularly intriguing since it would require retrograde transport to the cytoplasm, as documented recently for the degradation of ER lumenal proteins (22)(23)(24)(25).