The strains and the plasmids used for this study and DNA primers used for strain construction are listed in Tables 1-3, respectively. The marker-free deletion of rsiV (SBS352) was obtained as previously described [14] [21] and confirmed by polymerase chain reaction (PCR). Luria-Bertani (LB) medium was used. For plates, media were solidified with 1.5% agar. Tryptose blood agar (Difco) was used for preparation of B. subtilis strains for transformation. Synthetic media CI and CII were used for competence development for transformation [22] . When appropriate, antibiotics were included at the following concentrations (in µg/mL): ampicillin (100), chloramphenicol (5), erythromycin (0.5), neomycin (10), spectinomycin (100) and tetracycline (10). Lysozyme used was Wako Pure Chemical’s lysozyme hydrochloride from egg white. Cell growth was monitored with a mini photo 518R photometer (TAITEC) equipped with a 530- nm interference filter.

RsiV and RsiW were divided into N-terminal regions (RsiV_1-36 and RsiW_1-82), transmembrane regions (RsiV_37-59 and RsiW_83-105) and C-terminal regions (RsiV_60-285 and RsiW_106-208), respectively, based on the prediction of the transmembrane regions by TMHMM (URL:

http://www.cbs.dtu.dk/services/TMHMM/). Chimeric anti-σ factors, a chimera of the N-terminal region of RsiV and the transmembrane and C-terminal regions of RsiW (RsiV_N-RsiW_TMC) and a chimera of the N-terminal and transmembrane regions of RsiV and the C-terminal region of RsiW (RsiV_NTM-RsiW_C), were designed. These chimeras were constructed by PCR and Gibson Assembly Cloning Kit (New England Biolabs) on a vector pSG1729 [23] digested with KpnI and BamHI. The chimeric genes were transferred together with ribosome-binding site (RBS) on the vector pSG1729 by PCR and Gibson Assembly Cloning Kit to a vector pAPNC213 [24] digested with SalI. The resultant chimeric genes under

a ®, transformation with chromosomal or plasmid DNA.

P_spac and RBS were integrated into the chromosomal aprE locus. Constructs were verified by sequencing.

These were based on standard methods [25] [26] . The cycle number and the annealing temperatures for PCR were 40 cycles and 40˚C - 65˚C according to the T_m values of primers, respectively.

The β-galactosidase assay method using o-nitrophenyl-β-d-galactoside as sub-

strate and the unit definition were as described by Wang and Doi [27] . Proteins were separated by SDS gel electrophoresis with 15% polyacrylamide gel. Luminata Forte Western HRP Substrate (Millipore) was used for immunodetection according to the manufacturer’s instructions. For immunoblotting, anti-DYKDDDK tag (FLAG tag) monoclonal antibody (Wako Pure Chemicals) and anti-σ^A antiserum (a gift from K. Asai) were used. The secondary antibodies were peroxidase-conjugated anti-mouse IgG (H+L) and anti-rabbit IgG (H+L) (Jackson ImmunoResearch Laboratories). Immuno-stained bands were detected by C-Digit scanner (LI-COR), and the band intensities were measured by Image Studio software (LI-COR).

The activity of σ^V was monitored by β-galactosidase activity expressed from P_sigV-lacZ integrated into the chromosomal amyE locus. A ugtP deletion mutant showed an approximately two-fold activation of σ^V compared to wild type (Figure 1(a)). In ∆sigV and ∆sigV ∆ugtP mutants the activity was almost abolished, indicating that P_sigV-lacZ expression was dependent on σ^V activation. Deletion of the rsiV gene encoding anti-σ^V factor led to high σ^V activity, but additional introduction of a ugtP deletion mutation did not lead to additional activation of σ^V (Figure 1(a)). The rsiV deletion mutant was insensitive to the lack of glucolipids, which is consistent with the idea that the lack of glucolipids affects σ^V activity through RsiV.

The zinc chelator 1,10-phenanthroline (PT) inhibits the site-2 protease RasP [5] , as it does many other site-2 metalloproteases [28] . When ugtP⁺ cells were challenged by hen egg white lysozyme (HEWL) at low concentration (100 ng/mL), σ^V was activated several-fold. Addition of 10 mM PT returned the σ^V activity to the basal level (Figure 1(b)), most probably via inhibition of RasP. Approximately two-fold activation of σ^V in the ∆ugtP mutant compared to wild type was not changed by addition of 10 mM PT (Figure 1(b)), suggesting that the inhibition of RasP by PT did not affect activation. It seems that the activation of σ^V by the lack of glucolipids does not depend on the function of the site-2 protease RasP. Addition of HEWL to the PT-treated ∆ugtP cells did not increase the σ^V activity (Figure 1(b)). The σ^V activation by HEWL seems to depend on RasP in the ∆ugtP cells as well as in the ugtP⁺ cells.

When wild-type cells were treated with increasing concentrations (25, 50 and

100 ng/mL) of HEWL, the activity of σ^V increased accordingly (Figure 1(c)). A rasP::tet mutant showed a basal level σ^V activity in the absence of HEWL (Figure 1(c)). Whereas the addition of 100 ng/mL HEWL activated σ^V several-fold in rasP⁺ cells, the σ^V activity level in the rasP::tet mutant did not change in the presence of 100 ng/mL HEWL (Figure 1(c)), indicating that the activation of σ^V by HEWL requires RasP [5] . On the other hand, the approximately two-fold activation of σ^V in the ∆ugtP mutant did not change when, in addition, rasP was disrupted (Figure 1(c)), indicating that RasP has nothing to do with the σ^V activation by the lack of glucolipids. Addition of 100 ng/mL HEWL to the ∆ugtP mutant resulted in a remarkably high activation of σ^V (Figure 1(c)). The ∆ugtP mutant seems to become sensitized to the stress from HEWL addition. Addition of HEWL to ∆ugtP rasP::tet mutant cells did not change the σ^V activity (Figure 1(c)). The σ^V activation by HEWL required RasP in the ∆ugtP cells as well as in the ugtP⁺ cells.

RsiV is degraded by RIP in response to lysozyme. N-terminally FLAG tag-fused RsiV was expressed and analyzed by immunoblotting (Figure 2). When ugtP⁺ cells were treated with 25, 50 and 100 ng/mL HEWL, the amount of intact FLAG-RsiV protein was reduced to about 90%, about 70% and about 40%, respectively, compared with cells that were not treated with HEWL. This is consistent with the increasing activity of σ^V with increasing concentrations of HEWL (Figure 1(c)). In contrast, ∆ugtP cells showed no decrease in the amount of intact FLAG-tagged RsiV compared with the wild type not treated with HEWL. This result indicates that site-1 cleavage did not occur in ∆ugtP cells. Site-1 cleavage is a prerequisite for site-2 cleavage, and RasP cleavage is constitutive once site-1 cleavage has occurred [5] . The idea that RasP has nothing to do with the σ^V activation by lack of glucolipids fits into this scenario. When 100 ng/mL HEWL was added to ∆ugtP cells, the amount of intact FLAG-RsiV protein was reduced to about 7% (Figure 2). This corresponds well with the remarkably high activation of σ^V in the ∆ugtP mutant when 100 ng/mL HEWL was added (Figure 1(c)). The FLAG-tagged RsiV was functional: when the expression of P_spac-flag-rsiV was induced with IPTG, σ^V activity was repressed in both ugtP⁺ and ∆ugtP cells (Figure 1(d)).

Chimeric anti-σ factors, RsiV_N-RsiW_TMC and RsiV_NTM-RsiW_C, were constructed (see Materials and Methods). RsiW is an anti-σ factor for σ^W [1] [2] . Whereas σ^W is activated by RIP of RsiW [29] [30] in response to a variety of signals, including an alkaline shock [31] , it is not activated by lack of glucolipids [13] [14] .

Both chimeric anti-σ factors are functional as anti-σ^V: when their expression was induced with IPTG in a ∆rsiV background, σ^V activity was repressed (Figure 1(e)). The N-terminal region of RsiV seems to be sufficient to sequester σ^V.

In cells expressing either chimeric anti-σ factor, σ^V was not activated by the addition of 100 ng/mL HEWL (Figure 1(e)). This is consistent with reports that the C-terminal region of RsiV binds HEWL [6] [7] .

We found that σ^V was activated when the cells expressing RsiV_N-RsiW_TMC suffered alkaline stress (pH 8.9) (Figure 1(e)). Most probably the transmembrane and C-terminal extracytoplasmic regions of σ^W underwent RIP in response to the alkaline stress, which led to degradation of the chimeric anti-σ factor and to release of σ^V. In contrast, in the cells expressing RsiV_NTM-RsiW_C σ^V was not activated in response to alkaline stress (Figure 1(e)). The C-terminal extracytoplasmic region of RsiW may undergo site-1 cleavage, but it does not seem to lead to the site-2 cleavage of the fused transmembrane region of RsiV.

The expression of the chimeric anti-σ factors in ∆ugtP cells resulted in σ^V activity at the basal level of the ugtP⁺ cells (Figure 1(e)). This indicates that the C-terminal extracytoplasmic region is necessary for σ^V to respond to the lack of glucolipids.

We also constructed chimeric anti-σ factors, RsiW_N-RsiV_TMC and RsiW_NTM- RsiV_C, but they were not functional as anti-σ^W: IPTG induction of their expression in a ∆rsiW background even from multicopy plasmids did not lead to repression of σ^W activity (data not shown); the reason is unknown.