Effects of Carboxy-Terminal Modifications and pH on Binding of a Bacillus subtilis Small, Acid-Soluble Spore Protein to DNA

Abstract

Variants of the wild-type Bacillus subtilis α/β-type small, acid-soluble spore protein (SASP) SspC^wt were designed to evaluate the contribution of C-terminal residues to these proteins' affinity for DNA. SspC variants lacking one to three C-terminal residues were similar to SspC^wt in DNA binding, but removal of six C-terminal residues greatly decreased DNA binding. In contrast, a C-terminal extension of three residues increased SspC's affinity for DNA 5- to 10-fold. C-terminal and N-terminal changes that independently caused large increases in SspC-DNA binding affinity were combined and produced an additive effect on DNA binding; the affinity of the resulting variant, SspC^{ΔN11-D13K-C3}, for DNA was increased ≥20-fold over that of SspC^wt. For most of the SspC variants tested, lowering the pH from 7 to 6 improved DNA binding two- to sixfold, although the opposite effect was observed with variants having additional C-terminal basic residues. In vitro, the binding of SspC^{ΔN11-D13K-C3} to DNA suppressed the formation of cyclobutane-type thymine dimers and promoted the formation of the spore photoproduct upon UV irradiation to the same degree as the binding of SspC^wt. However, B. subtilis spores lacking major α/β-type SASP and overexpressing SspC^{ΔN11-D13K-C3} had a 10-fold-lower viability and far less UV and heat resistance than spores overexpressing SspC^wt. This apparent lack of DNA protection by SspC^{ΔN11-D13K-C3} in vivo is likely due to the twofold-lower level of this protein in spores compared to the level of SspC^wt, perhaps because of effects of SspC^{ΔN11-D13K-C3} on gene expression in the forespore during sporulation. The latter results indicate that only moderately strong binding of α/β-type SASP to DNA is important to balance the potentially conflicting requirements for these proteins in DNA transcription and DNA protection during spore formation, spore dormancy, and spore germination and outgrowth.

Bacterial spores of Bacillus and Clostridium species are among the most durable forms of life known and can survive environmental challenges, including dosages of UV radiation and elevated temperatures, that would be lethal to growing cells (32). For spores to survive such harsh treatment, their DNA must survive undamaged and thus must be protected in some fashion. This DNA protection is conferred in large part by one type of a family of small, acid-soluble spore proteins (SASP). The SASP are synthesized in the developing spore late in sporulation and are of two types, the α/β-type and the γ-type (14, 33). These proteins serve two major functions in spores. First, both types of SASP are rapidly degraded during spore germination, thus supplying an amino acid reservoir for the outgrowing spore (33). Second, the α/β-type SASP bind to and saturate the spore's chromosome, and this interaction protects the spore DNA against environmental insults, although the details of the molecular mechanism of this protection are not clear (28, 33). Interestingly, the binding of α/β-type SASP to DNA also protects the proteins themselves against damage such as deamidation and methionine oxidation (5, 7).

Spores of various Bacillus and Clostridium species contain multiple α/β-type SASP. In spores of Bacillus species, two major proteins, termed SASP-α and -β in Bacillus subtilis, generally comprise 75 to 80% of the α/β-type SASP pool, with from two to five minor proteins, depending on the species, constituting the remainder (32). All α/β-type SASP appear largely functionally redundant, since when either major or minor α/β-type SASP are overexpressed in spores, the overexpressed proteins provide spore DNA protection similar to that provided by the normal complement of these DNA binding proteins (16). The interaction of one minor B. subtilis α/β-type SASP, termed SspC^wt, with DNA has been extensively studied. This protein accounts for ∼10% of the total α/β-type SASP in B. subtilis spores but binds DNA with greater affinity than the two major α/β-type SASP of this species (C. S. Hayes, B. Setlow, and P. Setlow, unpublished results). It has been suggested (4) that SspC^wt and other minor α/β-type SASP are required to bind particular stretches of DNA in vivo. Although α/β-type SASP are not sequence-specific DNA binding proteins, all α/β-type SASP, including SspC^wt, bind with higher affinity to G-C-rich regions of DNA than to A-T-rich regions (6, 31). Presumably, this preference is a reflection of specific structural features of the α/β-type SASP-DNA complex, as DNA saturated with these proteins appears to assume an A-like helical conformation (3, 18, 19), a conformation that is more easily adopted by G-C-rich DNA. Since the protection of DNA in spores by α/β-type SASP binding depends on saturation of the DNA, one or more α/β-type SASP with a higher affinity for DNA may be required to protect particular DNA sequences or regions that only with difficulty adopt the protected structure in the α/β-type SASP-DNA complex.

Although α/β-type SASP with sufficiently high affinity for DNA are essential for protecting all spore DNA, the affinity of these proteins for DNA cannot be too high, as this may hinder the degradation of the proteins upon spore germination. In the first minutes of spore germination, the relatively dehydrated spore core rehydrates, and this event coincides with significant dissociation of α/β-type SASP from DNA (22). The DNA-free α/β-type SASP are then rapidly cleaved by the germination protease (GPR) and are eventually degraded to amino acids. At least one variant of SspC^wt with significantly enhanced affinity for DNA appears not to dissociate readily from DNA upon spore germination, and thus this protein is not degraded rapidly by GPR (8). As a consequence, normal DNA function is not restored early in the germination of these spores and many of them die (8). The potentially deleterious effect of α/β-type SASP binding to DNA has also been seen in Escherichia coli, as shortly after induction of SspC^wt synthesis in E. coli and binding of the protein to the cell's chromosome, cell growth and DNA replication cease, and in some genetic backgrounds, the cells die (25, 37).

The differences in the affinities of different α/β-type SASP for DNA are most likely due to variations in the amino acid sequences in the N- and C-terminal regions of these proteins. The amino acid sequences of α/β-type SASP are generally well conserved, both within and across Bacillus species, with ∼65% of residues being highly conserved among Bacillus species that diverged ≥500 million years ago (11, 33). However, this sequence homology is restricted to the central 75% of these proteins, with the sequences in both the N- and C-terminal regions being more divergent. For example, α/β-type SASP vary in length between 60 and 73 amino acids, with all of the variation in length in the N and C termini (11, 33). Previous studies focusing on the N-terminal region of SspC^wt identified some key residues affecting DNA binding (11, 33). The culmination of these studies was the production of the SspC^ΔN11-D13K variant, in which 11 residues at the N terminus were deleted while the positive charge in this region of the protein was retained; these changes resulted in a greater-than-10-fold improvement in the protein's affinity for pUC19 DNA over that of SspC^wt (4, 8). The SspC^ΔN11-D13K variant also provided far greater protection to DNA against deamination and restriction enzyme digestion in vitro than did SspC^wt (34).

The present study sought to modify the C-terminal region of SspC^wt to elucidate the contribution of residues in this region of the protein to DNA binding. Modifications to the C terminus of SspC^wt that resulted in improved DNA binding were performed on the SspC^ΔN11-D13K variant, and the effects of these multiple changes in SspC^wt on DNA binding were determined in vitro and in spores. Finally, since the pH in developing spores of Bacillus species decreases from ∼8 to as low as 6.3 during sporulation and then rises again to near 8 early in spore germination (29, 35), the binding of SspC^wt and its variants to DNA in vitro was characterized at pH values spanning those observed in sporulating cells and in developing, dormant, and germinated spores.

MATERIALS AND METHODS

Construction of sspC mutants.

The bacterial strains and plasmids used in these studies are described in Table 1. Mutations in the B. subtilis sspC^wt gene were generated by PCR from sspC^wt in plasmid pPS708 (19). The alterations to the C terminus-encoding region of sspC^wt used the upstream primer SSPC-JK (5′ CGCGCCATGGCTCAACAAAGTAG 3′), which had an NcoI restriction site (underlined) at the translation-initiating methionine codon (boldface), as well as flanking sequence to aid PCR and subsequent restriction enzyme digestion. Downstream primers altered the appropriate codons to produce the following mutant genes encoding SspC variants (Fig. 1): sspC^ΔC9 (5′ CCGGATCCTAAGCTAAGCGAACTAAACGTTTAG 3′), sspC^ΔC6 (5′ CCGGATCCTAGTTTTGTTGACCTAAGCGAAC 3′), sspC^ΔC3 (5′ CCGGATCCTAACCGCCCATGTTTTGTTGAGC 3′), sspC^ΔC1 (5′ CATAGGATCCTAAAATTGACCGCCC 3′), sspC^C1+ (5′ CCGGATCCTTCATTTATGAAATTGACCGCCC 3′), sspC^C3+ (5′ CCGGATCCCTCACTTCTGTCCATGAAATTGACCGC 3′), and sspC^C3 (5′ CCGGATCCCTCACTGCTGTCCATGAAATTGACCGC 3′). All downstream primers contained a BamHI restriction site (underlined) after the translation stop codon (boldface), as well as flanking sequence to aid PCR and subsequent restriction enzyme digestion. The sspC^{ΔN11-D13K-C3} gene was generated with the sspC^ΔN11-D13K upstream primer (5′ CCATGGCTAAATTACTAATTCCTCAAGCAG 3′) (8) containing an NcoI site (underlined) at the translation initiation codon (boldface; note that this mutant sspC gene was previously called sspC^Δ11-D13K, but we have renamed it sspC^ΔN11-D13K to distinguish between the N- and C-terminal mutations) and the sspC^C3 downstream primer. Hot-start PCR was performed using Taq DNA polymerase and 30 cycles of the following program: 94°C for 1 min, 47°C for 1 min, and 72°C for 2 min, finishing with a 10-min extension step at 72°C.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Characteristics or genotypea	Source or reference
B. subtilis
PS356	α−β−	15
PS578	α−β−, pUB110, Kmr	14
PS832	Wild type	Laboratory stock
PS2972	α−β−, pPS1393 with sspCwt, Kmr	4
PS3019	α−β−, pPS1393 with sspCΔN11-D13K, Kmr	4
PS3574	α−β−, pPS1393 with sspCΔC6, Kmr	This work
PS3575	α−β−, pPS1393 with sspCC3, Kmr	This work
PS3576	α−β−, pPS1393 with sspCΔN11-D13K-C3, Kmr	This work
E. coli
BL21(DE3)	T7 RNA polymerase under control of the lac promoter	Stratagene
PS708	JM107, pPS708	25
Plasmids
pCR2.1	Overhanging 3′ T ends for Taq cloning; Ampr Kmr	Invitrogen
pET11d	T7 promoter; Ampr	Stratagene
pPS708	pDG148 with a 0.6-kb fragment containing sspCwt; Ampr Kmr	25
pPS1393	pUB110 derivative with pUC19, sspB, and PsspB; Ampr Kmr	37
pPS2798	pUC19 containing sspB and PsspB; Ampr	4

^aAbbreviations: Km^r, kanamycin (10 μg/ml) resistance; Amp^r, ampicillin (100 μg/ml) resistance.

FIG. 1.

Alignment of amino acid sequences of SspC variants. Deletion variants are denoted SspC^ΔNX or SspC^ΔCX, where N and C indicate deletions at the N and C termini, respectively, and X is the number of residues deleted. Variants with additional residues are denoted SspC^CX+, where C is an addition at the C terminus, X is the number of residues added, and + means that one of the additional residues is positively charged. The N-terminal methionine, which is removed posttranslationally, is shown in all variants, but the numbering of residues begins at the adjacent alanine. Residues conserved in α/β-type SASP from different Bacillus species are underlined. The site where GPR cleaves these proteins is enlarged and in boldface.

Subcloning of mutant sspC genes.

All mutant sspC PCR products were ligated into the pCR2.1 cloning vector (Invitrogen, Carlsbad, Calif.) and cloned in E. coli TG1. After verification sequencing, the genes were excised from recombinant plasmids by digestion with NcoI and BamHI and inserted between NcoI and BamHI sites in the E. coli expression vector pET11d (Stratagene, La Jolla, Calif.). For expression of the SspC variants in B. subtilis, sspC genes were excised from pET11d by digestion with NcoI and BamHI and inserted between the same sites in plasmid pPS2798 (4), placing the sspC genes under the control of the strong forespore-specific promoter of the sspB gene (P_sspB) of B. subtilis. After the recombinant plasmid was cloned in E. coli, the pPS2798 derivatives were digested with EcoRI and BamHI, and the 0.45-kb P_sspB-sspC fragments were purified by agarose gel electrophoresis. The E. coli/B. subtilis shuttle vector pPS1393 (37) was digested with EcoRI and BamHI, and the 3.7-kb fragment containing the pUB110 origin of replication and a kanamycin resistance gene was purified by agarose gel electrophoresis. These two purified fragments were then ligated, forming a B. subtilis expression vector with the various sspC genes under the control of P_sspB. These plasmids were transformed into B. subtilis, with selection for resistance to kanamycin (10 μg/ml) (1).

Expression and purification of SspC variants from E. coli.

Due to the toxicity of even small amounts of SspC^wt in E. coli (25), transformed stocks of E. coli strains carrying sspC genes are difficult to maintain; consequently, use of fresh transformants is recommended. E. coli BL21(DE3) (Stratagene) was transformed with the pET11d expression vector containing the various sspC genes, and transformants were selected at 37°C on Luria-Bertani (LB) medium agar plates (10 g of tryptone, 5 g of yeast extract, 10 g of NaCl, 15 g of agar, and 1 ml of 1 N NaOH per liter) with ampicillin (100 μg/ml). Starting with a fresh single colony, each strain was grown at 37°C in 3 ml of LB medium plus ampicillin to an optical density at 600 nm (OD₆₀₀) of ∼1, diluted 1:500 into 1.2 liters of 2× YT medium (16 g of tryptone, 10 g of yeast extract, and 5 g NaCl per liter) plus ampicillin (100 μg/ml), and grown at 37°C with shaking at 250 rpm. When the OD₆₀₀ reached 1.5, IPTG (isopropyl-β-d-thiogalactopyranoside) was added to 0.5 mM to induce expression of SspC^wt and its variants, and cultures were incubated a further 2 h. The cultures were harvested by centrifugation at 8,000 × g for 10 min at 4°C, the resulting pellets were washed by centrifugation with 0.15 M NaCl at 4°C, and the final cell pellets were frozen and lyophilized. The yield was ∼1 g of lyophilized cells per 1.2 liters of culture.

For extraction and purification of SspC^wt and its variants, ∼2 g of lyophilized cells was disrupted for 2 min in a dental amalgamator (Wig-L-Bug) with 100 mg of dry cells, 150 mg of glass beads, and a steel ball bearing in each Wig-L-Bug capsule. Proteins were extracted with 80 ml of ice-cold 3% acetic acid-30 mM HCl per g of dry cells, with stirring of the suspension for 30 min at 4°C. The suspension was centrifuged (8,000 × g, 10 min, 4°C), and the pellet was reextracted with 56 ml of the cold acid mixture per g of dry cells. Both supernatant fluids were pooled and dialyzed overnight at 4°C in Spectra/Por no. 3 tubing (molecular weight cutoff, 3,500) against two changes of 2 liters of 1% acetic acid. The dialyzed protein was then passed through a 15-ml DEAE-cellulose column equilibrated in 1% acetic acid at 4°C. SspC^wt and its variants did not adhere to this column, but a number of negatively charged contaminants were removed, and the column flowthrough was collected and lyophilized. The lyophilized DEAE-cellulose column flowthrough was dissolved in 3 ml of 8 M urea-10 mM Tris-maleate (pH 6.1) at room temperature and loaded on a 35-ml carboxymethyl (CM)-cellulose column equilibrated in 10 mM Tris-maleate (pH 6.1). At pH 6.1, SspC^wt reversibly bound to this column and was eluted at room temperature with a gradient of 0 to 0.4 M NaCl (100 ml of each solution) in the Tris-maleate buffer. SspC^wt did not adhere to the column in the same buffer at pH 7.5, whereas, at pH ≤5.5, the protein would not elute in the salt gradient. SspC variants with a greater net negative charge than SspC^wt (SspC^ΔC1, SspC^ΔC3, and SspC^ΔC6) did not bind strongly to the CM-cellulose column at pH 6.1, but their elution was sufficiently retarded for them to be separated from contaminants. The remaining SspC variants behaved similarly to SspC^wt on the CM-cellulose column. Column fractions were analyzed for protein by the Bradford assay (Bio-Rad) and by polyacrylamide gel electrophoresis (PAGE) at low pH (21). Appropriate fractions were pooled, dialyzed in Spectra/Por no. 3 tubing against four changes of 1 liter of H₂O at 4°C, lyophilized, and dissolved in 10 mM sodium phosphate, pH 7.0. Purified proteins were stable for less than 1 month at 4°C but were stable for at least 1 year when frozen at −80°C. Protein concentrations were determined by amino acid analysis, and all proteins used in this work were ≥95% pure as judged by PAGE at low pH (21).

Analysis of protein-DNA binding affinity by circular dichroism spectroscopy (CD).

Synthetic polynucleotides (Sigma) for analysis of protein-DNA binding were dialyzed extensively in Spectra/Por no. 3 tubing against 10 mM sodium phosphate, pH 7, at 4°C and stored at −20°C. One synthetic polynucleotide, poly(dG):poly(dC) (pdG:pdC), was also sonicated briefly to reduce the viscosity of the solution. Plasmid pUC19 was purified from E. coli with the Qiagen miniprep kit and linearized by digestion with PstI. For experiments analyzing the effect of pH on protein-DNA complex formation, protein and DNA were dialyzed separately overnight in Spectra/Por no. 3 tubing at 4°C in 10 mM sodium phosphate at pH 6, 7, or 8.

All CD measurements were obtained at 20°C on a Jasco 715 spectropolarimeter, with a quartz cuvette of 1 cm path length. The DNA binding of SspC^wt and its variants was reflected in an increase in the molar ellipticity of the protein at 222 nm (Θ₂₂₂) (22), indicating the appearance of α-helical structure (6). Spectra were corrected for the contributions of buffer, DNA, and free protein. Standard CD operating parameters were as follows: 250 to 190 nm scanned, 0.5-nm step resolution, 50-nm/min scan speed, six accumulations averaged, 8-s response time, 1-nm bandwidth, and 50-millidegree sensitivity. Protein-DNA complexes were allowed to come to equilibrium before each scan, as judged by the stability of the Θ₂₂₂ signal and previously measured kinetics of binding (6). Reverse titrations were performed with ∼2 μM protein and increasing amounts of pdG:pdC. As described previously (10), these reverse titrations established the maximum and minimum Θ₂₂₂ values for free protein and protein bound to DNA, as well as the binding site size. Forward titrations were performed by using these parameters with ∼15 μM DNA and increasing amounts of protein. The DNA used in forward titrations was linear pUC19, poly(dAdT):poly(dAdT) (pdAdT:pdAdT), or pdG:pdC, with the selection of the DNA dictated by the binding strength of the protein in order to keep the protein-DNA binding affinity in a range measurable by CD. The data were used to construct titration curves according to the model of von Hippel et al. (10, 17), which allows calculation of an apparent binding constant, Kω, which is composed of both an intrinsic binding constant, K, and a cooperativity factor, ω. Hill plots (36) were also constructed to compare the relative cooperativities of the binding of the SspC variants to DNA. In this method, a cooperativity factor of 1 implies noncooperative binding, a cooperativity factor >1 implies positive cooperativity, and a factor <1 implies negative cooperativity.

Expression of SspC^wt and its variants in B. subtilis.

All strains of B. subtilis used in these studies are derivatives of strain 168. Strain PS356, which lacks the genes encoding SASP-α and -β, was transformed (1) with derivatives of plasmid pPS1393 carrying various sspC genes under the control of the strong, forespore-specific P_sspB. Consequently, in spores of these transformed strains, SspC^wt or a variant is the only major α/β-type SASP. Spores of B. subtilis strains were prepared by nutrient exhaustion on 2× SG medium agar plates (21) that were incubated in a plastic bag at 37°C for ∼3 days until sporulation was complete. The plates were removed from the bags and held at room temperature for ∼4 days to allow lysis of any growing cells. Spores were scraped from the plates, washed as described previously (21), and stored in water at 4°C protected from light. All spores used in this work were free (>98%) from growing and sporulating cells. SspC^wt and its variants were extracted from 2.5 mg (dry weight) of lyophilized spores as described above for extraction from E. coli, except that 8 min of shaking in the Wig-L-Bug was required to disrupt the spores. The acid extracts were dialyzed as described above, lyophilized, and dissolved in 15 μl of 8 M urea, and aliquots were subjected to PAGE at low pH (21).

Heat and UV resistance of spores.

Analysis of spore killing by heat and UV radiation was carried out essentially as described previously (8, 15, 16). For heat killing, spores at an OD₆₀₀ of 1 were incubated at 85°C, samples were withdrawn at appropriate intervals, and 10-μl aliquots of serial dilutions were spotted on dry LB agar plates containing no antibiotics. The plates were incubated for 16 to 20 h at 30°C, and colonies were counted. For UV resistance studies, 2 ml of spores at an OD₆₀₀ of 1 was placed in a 35-mm-diameter petri dish on a rotating surface underneath a short-wave UV lamp with maximum output at 254 nm. The UV fluence on the spores was measured with a BLAK-RAY UV meter (UVP, San Gabriel, Calif.). The irradiated spores were diluted, spotted, and counted as in the heat resistance studies.

Analysis of the UV photochemistry of DNA in vitro.

Analysis of the effect of SspC variants on the UV photochemistry of DNA was carried out with PstI-linearized [³H]thymidine-labeled pUC19 DNA as described previously (20, 30). Briefly, the DNA was exposed to UV radiation at 254 nm in 100 μl of buffer containing 6.7 mM BIS-Tris propane (1,3-bis[tris{hydroxymethyl}methylamino]propane)-HCl, 6.7 mM MgCl₂, 0.6 mM dithiothreitol, and 3.3 mM NaP0₄, pH 7, and, in the presence or absence of SspC variants, the DNA was acid hydrolyzed, photoproducts were separated by descending-paper chromatography, and the paper was cut into strips and counted in a scintillation counter. Individual photoproducts were identified by their migration positions relative to that of thymine (20, 30). Previous work has shown that the same spectrum of UV photoproducts is formed in SspC^wt-DNA complexes irradiated at pH 7 and 6 (B. Setlow and P. Setlow, unpublished data).

RESULTS

Effects of deletion of C-terminal residues from SspC^wt.

SspC^wt, a minor α/β-type SASP from B. subtilis, was chosen for study for a number of reasons. First, although SspC^wt is a minor α/β-type SASP, its effects on DNA in vitro and in vivo are very similar to those of the major α/β-type SASP (19). Second, SspC^wt can be overexpressed to high levels in E. coli and is easy to purify (19). Finally, there have been many previous studies examining the interaction of SspC^wt with DNA, including one examining the effects of a number of alterations in the N-terminal region of this protein (4).

The C-terminal regions of α/β-type SASP differ in length by up to eight residues and display little of the sequence conservation that characterizes the majority of the protein (2) (Fig. 1). The C-terminal region of SspC^wt is of average length and is unusual only in that it terminates in a histidine rather than the more usual lysine or arginine. Truncations at the C terminus of SspC^wt were designed to successively delete three amino acids, generating SspC^ΔC3, SspC^ΔC6, and SspC^ΔC9, as well as one variant (SspC^ΔC1) in which only the C-terminal histidine was removed (Fig. 1). Three of these four proteins were expressed at high levels in E. coli, although SspC^ΔC9 was not (data not shown). Consequently, no further C-terminal truncations were prepared.

Although SspC^wt by itself possesses very little secondary structure, it assumes a highly α-helical structure upon binding to DNA (6). This feature can be used to monitor and quantitate the extent to which SspC^wt and its variants bind to DNA, since CD is exquisitely sensitive to the presence of α-helices in proteins. Indeed, CD has been used to analyze both forward titrations (adding protein to a set amount of DNA) and reverse titrations (adding DNA to a set amount of protein) to determine apparent equilibrium binding constants between different DNAs and a number of SspC variants (4). Therefore, this method was used to quantitate the interaction between DNA and the SspC variants with C-terminal truncations. The increase in helicity as SspC variants interacted with DNA was assumed to correspond directly to binding, and binding was assumed to be a single-step procedure, as was consistent with previous results (6). Linearized pUC19 was used to simulate an infinitely long, random-sequence, unstructured, double-stranded DNA lattice.

There was little difference in the apparent binding constants characterizing the binding of SspC^wt and the C-terminal truncation variants SspC^ΔC1 and SspC^ΔC3 with linearized pUC19 (Table 2). This is perhaps not surprising, since the deleted C-terminal residues are not conserved in α/β-type SASP. In contrast, the binding affinity of SspC^ΔC6 for pUC19 DNA was significantly lower than that of SspC^wt, and only an upper limit for the binding constant characterizing the binding of this protein with DNA could be estimated (Table 2). However, the interaction between SspC^ΔC6 and pdG:pdC, the DNA to which α/β-type SASP bind the most tightly (binding constant at pH 7 is ≥10⁷ M⁻¹ [6]), was measurable by CD (Table 2), proving that this variant is not devoid of DNA binding capability.

TABLE 2.

Binding constants for SspC variants with DNA at three pH values^a

SspC variant	Apparent binding constantb (M−1) at pH:			DNA
	8	7	6
SspCwt	3.9 × 105	5.2 × 105	2.7 × 106	Linear pUC19f
C-terminal deletions
SspCΔC1	NDc	6.1 × 105	2.6 × 106	Linear pUC19
SspCΔC3	ND	4.6 × 105	2.9 × 106	Linear pUC19
SspCΔC6	ND	≤105	≤105	Linear pUC19
SspCΔC6	ND	ND	1.6 × 105	pdG:pdCd
C-terminal additions
SspCC1+	ND	9.5 × 105	≤105	Linear pUC19
SspCC3	ND	4.8 × 106	1.4 × 107	Linear pUC19
SspCC3+	ND	3.8 × 106	1.5 × 106	Linear pUC19
Combination variants
SspCΔN11-D13K	ND	6.4 × 105	1.3 × 106	pdAdT:pdAdTe
SspCΔN11-D13K-C3	ND	1.5 × 106	9.9 × 106	pdAdT:pdAdT

^aConstants were determined by forward titrations as described in Materials and Methods.

^bErrors are +/− 15% based on duplicate titrations and previous observations (4, 6).

^cND, not determined.

^dBinding of SspC^wt to pdG:pdC is 30- to 50-fold stronger than binding to linear pUC19 at pH 7 (6).

^eBinding of SspC^ΔN11-D13K to linear pUC19 is ≥10-fold stronger than that of SspC^wt at pH 7 (4).

^fBinding of SspC^wt to linear pUC19 is approximately twofold stronger than binding to pdAdT:pdAdT at pH 7 (6).

Effects of addition of C-terminal residues to SspC.

Since C-terminal truncations either had no effect or only weakened the affinity of SspC for DNA, we added C-terminal residues. Three variants were designed to examine the effect of adding a positive charge, additional residues, or both to the C terminus of SspC^wt (Fig. 1). A C-terminal lysine was added to one variant, SspC^C1+, because most α/β-type SASP terminate in lysine or arginine (2), residues that would likely be positively charged at physiological pH. In contrast, SspC^wt terminates in histidine, which may or may not be charged at physiological pH. Three additional residues were added to SspC^C3, which terminates in glycine-glutamine-glutamine. Many, although not all, α/β-type SASP contain a glycine three residues from the C terminus (2), so that pattern was repeated for SspC^C3. Glutamine is the most common amino acid (except for positively charged residues) in the C-terminal region of α/β-type SASP (2), is relatively inert, and is amenable to α-helix formation, so two glutamines were added after the glycine. The third variant, SspC^C3+, is a combination of the two modifications of charge and length, with lysine as the C-terminal residue.

All additions to the C terminus increased SspC binding to DNA at pH 7 (Table 2). SspC^C1+ had a 2-fold-higher affinity for DNA than SspC^wt, while the affinity of SspC^C3 for DNA was almost 10-fold higher. Interestingly, the combination of charge and extra residues did not produce an additive effect, for SspC^C3+ had a slightly lower affinity for DNA than did SspC^C3, although its affinity was still higher than that of SspC^C1+. The trend, therefore, is that additional C-terminal residues are beneficial for SspC-DNA binding, although an additional positive charge in the C-terminal region is not.

Effects of pH on the binding of SspC^wt and its variants to DNA.

The effect of pH on the binding of SspC^wt and its variants was also tested (Table 2) for two reasons. First, as noted above, the charge in the C-terminal regions of SspC variants affects the binding of these proteins to DNA and, although the C-terminal histidine of SspC^wt seems not to contribute to DNA binding at pH 7, this residue may be protonated and have a greater effect on DNA binding at a lower pH. Second, and more importantly, the internal pH of spores of Bacillus species has been measured at close to 6, although the pH in both sporulating cells and germinated spores is 7.5 to 8 (12, 13, 29, 35). The affinity of SspC^wt for pUC19 decreased only very slightly as the pH was raised from 7 to 8; consequently, no other SspC variants were tested for their binding to DNA at pH 8. In contrast to the slight decrease in the affinity of SspC^wt for DNA in going from pH 7 to 8, there was a twofold increase in the affinity in going to pH 6. While we had hoped to examine the binding of SspC^wt to DNA at even lower pH values, at a pH of 5.5 or less, the protein-DNA complex began to precipitate. The truncation variants SspC^ΔC1 and SspC^ΔC3 had affinities for DNA that were very similar to that of SspC^wt at both pH 7 and 6 (Table 2). Since the C-terminal histidine is deleted in both SspC^ΔC1 and SspC^ΔC3, these data indicate that this residue is not particularly important in DNA binding.

Concerning variants with extra C-terminal residues, the variant with three extra residues, SspC^C3, had a twofold-higher affinity for pUC19 DNA at pH 6 than at pH 7, similar to what was seen with SspC^wt. Surprisingly, for both SspC variants with an additional C-terminal positive charge, SspC^C1+ and SspC^C3+, the affinity for DNA decreased with a decrease in pH (Table 2). For SspC^C3+ the decrease in the affinity for DNA was minimal in going from pH 7 to 6, but any decrease in affinity for DNA going from 7 to 6 is in contrast to the behavior of the other variants. For SspC^C1+, the drop in the affinity of this protein for pUC19 DNA at pH 6 was so drastic that the binding constant for the protein-DNA complex was too low to be measured by CD.

SspC variants with changes in both the N- and C-terminal regions.

On the basis of the results given above, a potentially very strong DNA-binding SspC variant was designed by combining alterations at both the N and C termini that increased protein-DNA affinity. In this variant, the change in SspC^C3 was combined with the N-terminal changes in SspC^ΔN11-D13K previously shown to greatly increase protein-DNA affinity (4), giving SspC^{ΔN11-D13K-C3}. This variant could be overexpressed in and purified from E. coli, although the growth of this strain was noticeably slower than that of E. coli strains overexpressing SspC^wt or other C-terminal variants (data not shown). The slower growth of the E. coli strain carrying sspC^{ΔN11-D13K-C3} was seen even when expression of the protein was not being induced, showing that even a small amount of leaky expression of this protein is enough to slow cell growth. Similar results have been noted with another tight-binding α/β-type SASP (8, 25).

The apparent binding constant characterizing the binding of SspC^{ΔN11-D13K-C3} to pUC19 DNA at pH 7 was greater than 10⁷ M⁻¹ (Table 2), which is close to the upper limit of measurement by CD. The stoichiometry for the binding of this protein to pdG:pdC was 1 protein per 4 bp, similar to what was found for SspC^wt (8, 16). Because of the very tight binding of SspC^{ΔN11-D13K-C3} and SspC^ΔN11-D13K to pUC19, the DNA used routinely to quantitate protein-DNA binding was pdAdT:pdAdT, a polynucleotide for which α/β-type SASP have lower affinity than for pUC19 (4, 6). At pH 7, SspC^{ΔN11-D13K-C3} had a twofold-higher affinity for pdAdT:pdAdT than did SspC^ΔN11-D13K (Table 2), suggesting that the effects of C-terminal and N-terminal modifications on DNA binding are additive. When the pH was lowered to 6, the affinity of SspC^{ΔN11-D13K-C3} for DNA improved even further, to almost 10⁷ M⁻¹, a value sevenfold higher than that at pH 7. It has been shown previously that at pH 7 SspC^ΔN11-D13K binds ≥10-fold better to linear pUC19 DNA than does SspC^wt (8), and, at pH 6, SspC^{ΔN11-D13K-C3} bound 15-fold better to pdAdT:pdAdT than did SspC^ΔN11-D13K at pH 7 (Table 2). While these comparisons are with different DNAs, these data suggest that the total improvement in DNA binding going from SspC^wt at pH 7 to SspC^{ΔN11-D13K-C3} at pH 6 is ≥100-fold.

Cooperativity in the binding of SspC and its variants to DNA.

Previous studies have shown that SspC binds to DNA cooperatively (3, 6, 19). The degree of cooperativity of binding varies with the DNA substrate, with binding to pdAdT:pdAdT being the most cooperative, binding to pdG:pdC being the least cooperative, and binding to pUC19 being intermediate in cooperativity (31). All of the variants tested displayed positive cooperativity in binding to DNA in Hill plots, with the exception of SspC^{ΔN11-D13K-C3} at pH 6 (Table 3). Truncation variants SspC^ΔC1 and SspC^ΔC3 had binding cooperativities similar to that of SspC^wt and exhibited similar cooperativities at pH 6 and 7. Truncation variant SspC^ΔC6 was unable to bind pUC19 but bound pdG:pdC with a cooperativity factor lower than that for SspC^wt binding to pUC19. However, the cooperativity for the binding of SspC^wt to pUC19 is significantly greater than for the binding of this protein to GC-rich DNA (6). The binding of the more positively charged variants, SspC^C1+ and SspC^C3+, to DNA was also reduced in cooperativity, more notably at pH 7 than at pH 6. Conversely, the C-terminal extension variant SspC^C3 exhibited substantially enhanced cooperativity in binding to DNA at both pH values. The two tight-binding SspC variants had to be tested with pdAdT:pdAdT instead of pUC19, which tends to increase the cooperativity of SspC binding (6). While SspC^ΔN11-D13K appeared slightly enhanced in cooperativity over SspC^wt at both pH levels, SspC^{ΔN11-D13K-C3} appeared similar in cooperativity to SspC^wt at pH 7 but was reduced in cooperativity at pH 6 to the point of being characterized as noncooperative. While the latter comparisons between SspC^wt and SspC^ΔN11-D13K and SspC^{ΔN11-D13K-C3} were made with different DNAs, pUC19 for SspC^wt and pdAdT:pdAdT for SspC^ΔN11-D13K and SspC^{ΔN11-D13K-C3}, previous work (6) has shown that the binding of SspC^wt to pdAdT:pdAdT is much more cooperative than is binding to pUC19. Consequently, it appears most likely that the binding of SspC^ΔN11-D13K and SspC^{ΔN11-D13K-C3} to DNA is less cooperative than is the binding of SspC^wt.

TABLE 3.

Cooperativity factors for binding of SspC variants to DNA at two pH values

SspC variant	Cooperativity factora at pH:		DNA
	7	6
SspCwt	1.8 ± 0.1	1.6 ± 0.1	Linear pUC19
SspCΔC1	1.8 ± 0.1	2.0 ± 0.5	Linear pUC19
SspCΔC3	1.8 ± 0.1	1.7 ± 0.1	pdAdT:pdAdT
SspCΔC6	—b	—	Linear pUC19
SspCΔC6	NDc	1.4 ± 0.1	pdG:pdC
SspCC1+	1.4 ± 0.1	—	Linear pUC19
SspCC3	2.9 ± 0.9	2.5 ± 0.4	Linear pUC19
SspCC3+	1.3 ± 0.1	1.5 ± 0.1	Linear pUC19
SspCΔN11-D13K	2.1 ± 0.1	2.1 ± 0.1	pdAdT:pdAdT
SspCΔN11-D13K-C3	1.8 ± 0.1	0.9 ± 0.1	pdAdT:pdAdT

^aCooperativity factors were determined from Hill plots as described in Materials and Methods. Cooperativity errors are the standard deviations of the slope of the Hill plot.

^b—, unable to be determined due to insufficient binding.

^cND, not determined.

Expression of SspC and its variants in spores and effects on spore survival and resistance.

To examine the effects of SspC variants on DNA in vivo, spores of B. subtilis lacking the two normally major α/β-type SASP (termed α⁻β⁻ spores) were engineered to overexpress SspC^wt or selected variants. SspC^wt and its variants are the predominant α/β-type SASP in these spores, due to the α⁻β⁻ genetic background. Previous work (8) has shown that α⁻β⁻ spores overexpressing SspC^ΔN11-D13K have ∼10-fold-lower viability than α⁻β⁻ spores overexpressing SspC^wt or wild-type spores, and this was also found in the present work (data not shown). The α⁻β⁻ spores overexpressing SspC^{ΔN11-D13K-C3} also had a 10-fold-lower viability than wild-type spores (data not shown). In contrast, α⁻β⁻ spores overexpressing other SspC variants exhibited viability identical (within 20%) to that of wild-type spores or α⁻β⁻ spores overexpressing SspC^wt (data not shown).

As a further measure of the effect of the SspC variants on DNA properties in vivo, the UV and heat resistance of α⁻β⁻ spores overexpressing these proteins was measured (Fig. 2), since both of these treatments can kill α⁻β⁻ spores by DNA damage. In agreement with previous results (15), α⁻β⁻ spores that lack the major α/β-type SASP are much less resistant than are wild-type spores to UV radiation and heat. However, overexpressing SspC^wt in α⁻β⁻ spores restored spore UV and heat resistance to that of wild-type spores, again as found previously (15). This result demonstrates the functional redundancy of different α/β-type SASP. Of the α⁻β⁻ spores overexpressing SspC variants with C-terminal changes, spores with high levels of SspC^ΔC6, a variant that binds DNA poorly, displayed little more UV and heat resistance than α⁻β⁻ spores. Surprisingly, SspC^ΔN11-D13K, which binds quite well to DNA in vitro, provided less protection to spore DNA against UV and heat than did SspC^wt, as was noted previously (8). Even more striking was the finding that α⁻β⁻ spores overexpressing SspC^C3 or SspC^{ΔN11-D13K-C3} exhibited UV and heat resistance similar to that of α⁻β⁻ spores, even though these two SspC variants bind more tightly to DNA in vitro than SspC^wt.

FIG. 2.

UV (A) and wet heat (B) resistance of B. subtilis spores expressing SspC variants. Purified spores of various strains were exposed to UV radiation (42 J/m² · min) or wet heat (85°C), and survival was determined as described in Materials and Methods. Values for survivors are the averages of duplicate determinations with less than twofold differences between them; error bars are omitted for clarity. The strains used are as follows: PS356, α⁻β⁻ (•); PS2972, α⁻β⁻ SspC^wt (▪); PS3019, α^−β− SspC^ΔN11-D13K (▴); PS3574, α⁻β⁻ SspC^ΔC6 (○); PS3575, α⁻β⁻ SspC^C3 (▪); PS3576, α⁻β⁻ SspC^{ΔN11-D13K-C3} (▵). The UV and heat resistance of spores of strains PS578 (α⁻β⁻ pUB110) and PS832 (wild-type) was virtually identical to that of spores of strains PS356 (α⁻β⁻) and PS2972 (α⁻β⁻ SspC^wt), respectively (data not shown).

These results suggest that there may be no direct correlation between the affinity of a particular SspC variant for DNA and the ability of that overexpressed variant to provide protection to DNA in spores. Levels of total α/β-type SASP in spores are close to those needed to saturate the spores' DNA, and several studies have shown that any spore DNA not saturated with these proteins is susceptible to lethal damage from heat and UV radiation and that, for UV radiation, the susceptibility of DNA free of α/β-type SASP to lethal damage is actually greater than that of DNA in growing cells (15, 16, 24). Consequently, one obvious explanation for the low heat and UV resistance of spores overexpressing SspC^C3 or SspC^{ΔN11-D13K-C3}is that these tight-binding SspC variants may be present at lower levels than SspC^wt in α⁻β⁻ spores, levels that are too low to saturate the DNA in the spore. Indeed, while α⁻β⁻ spores overexpressing all SspC variants appeared to have similar levels of SASP-γ (which is not bound to spore DNA), these spores contained less of the tighter-binding SspC variants, SspC^C3, SspC^ΔN11-D13K, and SspC^{ΔN11-D13K-C3}, than of SspC^wt (Fig. 3). Analyses of different amounts of extracts from several different independent spore preparations by PAGE indicated in particular that the level of SspC^{ΔN11-D13K-C3} was only one-half that of SspC^wt, with levels of SspC^ΔN11-D13K and SspC^C3 being intermediate between those of SspC^{ΔN11-D13K-C3} and SspC^wt (Fig. 3B and data not shown).

FIG. 3.

SASP levels in spores expressing SspC variants. SASP were extracted, aliquots were run on PAGE gel at low pH, and the gels were stained as described in Materials and Methods. (A) Samples are from equal amounts of spores. Lane 1, PS356 (α⁻β⁻); lane 2, PS578 (α⁻β⁻ pUB110); lane 3, PS832 (wild type); lane 4, PS2972 (α⁻β⁻ SspC^wt); lane 5, PS3019 (α⁻β⁻ SspC^ΔN11-D13K); lane 6, PS3574 (α⁻β⁻ SspC^ΔC6); lane 7, PS3575 (α⁻β⁻ SspC^C3); lane 8, PS3576 (α⁻β⁻ SspC^{ΔN11-D13K-C3}); lane 9, purified SspC^wt standard. (B) The samples run are different amounts of extracts from equal amounts of spores of strains PS2972 (SspC^wt) (lanes 1 to 4) and PS3576 (SspC^{ΔN11-D13K-C3}) (lanes 5 to 8). The amounts of sample run on PAGE gels were as follows: lanes 1 and 5, 2 μl; lanes 2 and 6, 4 μl; lanes 3 and 7, 8 μl; lanes 4 and 8, 12 μl. Lanes are from the same gel, with spacer lanes cut out. Results are representative of several extractions. For both panels, the migration positions of SASP-α, -β, -γ and SspC variants are indicated to the right.

Effects of SspC variants on the UV photochemistry of DNA in vitro.

While the previous results strongly suggested that the lack of protection of DNA in spores by tight-binding SspC variants was due to the low level of these proteins in spores, an alternative explanation is that the binding of these particular SspC variants does not alter DNA structure in the same way as does that of SspC^wt. To examine this possibility in detail, we analyzed the UV photochemistry of SspC-pUC19 complexes in vitro (Table 4). The major photoproducts formed by UV irradiation of DNA in vitro are cyclobutane-type thymine dimers (T◊T) formed between adjacent thymine residues on the same DNA strand (20) (Table 4). As expected, SspC^ΔC6, a variant that binds pUC19 DNA poorly if at all, had no significant effect on the UV photochemistry of this DNA (Table 4). However, as shown previously (20) and in this work (Table 4), UV irradiation of an SspC^wt-pUC19 complex generates very little T◊T, but rather a thyminyl-thymine adduct termed the spore photoproduct, which is also formed between adjacent thymine residues on the same DNA strand. Strikingly, UV irradiation of the complex between SspC^{ΔN11-D13K-C3} and pUC19 generated a spectrum of photoproducts that was the same as that of photoproducts that formed with the SspC^wt-pUC19 complex (Table 4). Thus, the tightest-binding SspC variant, SspC^{ΔN11-D13K-C3}, appears to have the same effects on DNA in vitro as does SspC^wt and thus likely has the same effect on DNA structure as SspC^wt.

TABLE 4.

Effect of SspC variants on the UV photochemistry of DNA^a

Sampleb	% of total thymine as:
	T◊T	SPd
DNA only	6.3	≤0.5c
DNA + SspCwt	≤1.3c	3.5
DNA + SspCΔ6	6.0	≤0.5c
DNA + SspCΔN11-D13K-C3	≤1.3c	3.9

^a[³H]thymidine-labeled pUC19 DNA with or without SspC variants was exposed to UV radiation, and photoproducts were identified as described in Materials and Methods. The ratios of SspC variants, when present, to DNA were 6/1 (wt/wt). This ratio of protein to DNA is significantly above that needed to saturate the DNA; if an SspC variant binds DNA with high affinity, the saturation ratio is ∼3/1 (wt/wt).

^bThe UV dose for DNA only was 10 kJ/m²; that for DNA plus SspC^wt and DNA plus SspC^Δ6 was 30 kJ/m²; that for DNA plus SspC^{ΔN11-D13K-C3} was 10 kJ/m², but similar results were obtained with a dose of 30 kJ/m².

^cThis represents a variety of nonspecific photoproducts formed at the high UV doses used in these experiments.

^dSP, spore photoproduct.

DISCUSSION

Among the goals of these studies were to (i) determine the role of the C-terminal region of SspC in DNA binding, (ii) make alterations that might improve this binding, and (c) combine productive C-terminal changes with any other factors (N-terminal changes and pH) that enhance DNA binding. The results obtained indicate that all three factors, the C-terminal sequence, the N-terminal sequence, and pH, play a role in determining the affinity of SspC binding to DNA and further that these three factors can be manipulated independently. The overall amino acid sequence similarities between α/β-type SASP suggest that these results can be generalized to all α/β-type SASP.

It is particularly interesting that the N- and C-terminal regions of SspC appear to function independently. Among the known α/β-type SASP, there is no obvious correlation between the lengths of the N-terminal and C-terminal regions (2). Such variation may contribute significantly to the differences in the affinities of the various α/β-type SASP for DNA, while allowing an overlap in the function of the complement of these proteins in a spore. For example, high-affinity α/β-type SASP may be necessary to bind to A-T-rich regions of DNA to ensure the saturation of the chromosome, whereas lower-affinity α/β-type SASP may be important to initiate the process of dissociation from DNA in the first minutes of spore germination. This variation in DNA binding affinity may be an example of how α/β-type SASP as a group must be flexible in their binding to DNA in order to fulfill their many roles in a spore.

C-terminal truncations of SspC^wt appear generally inconsequential up to glycine 68, so it is somewhat surprising that the length of the C-terminal region in α/β-type SASP is generally more conserved than the length of the N-terminal region (33). The region prior to glycine 68 is thought to participate in DNA-protein contacts (23), so the lack of DNA binding by SspC^ΔC6, in which glycine 68 is also removed, is not surprising. There is also no α/β-type SASP in Bacillus species with as short a C-terminal region as SspC^ΔC6, and perhaps α/β-type SASP require a certain length for proper function (32). SspC^ΔC9, in which five residues beyond glycine 68 are deleted, is not accumulated in E. coli. We presume, as found for SspC variants with very large N-terminal deletions (4), that removal of too much of the C-terminal region of SspC^wt, as in SspC^ΔC9, results in a protein that is rapidly degraded, since the protein cannot bind DNA and is thus not protected against proteolysis (27). C-terminal extensions of SspC^wt, however, appear to increase DNA binding. The far-C-terminal region is not thought to participate in DNA contacts but could affect binding in other ways. The extra residues may favor protein-protein contacts on DNA, as reflected in the increased cooperativity of SspC^C3 binding. However, no variants with more than three additional residues were constructed, so the extent to which further C-terminal extension or different residues could improve the binding of SspC^wt to DNA is not known.

The cooperativity of the SspC^{ΔN11-D13K-C3} variant's binding to DNA is intriguing. At pH 7, although the cooperativity of the binding of SspC^{ΔN11-D13K-C3} to pdA:pdT is less than that of SspC^ΔN11-D13K, it is still positive, but, at pH 6, the Hill factor falls to ∼1, indicating a transition to noncooperative binding. Furthermore, at pH 6, this variant was able to bind a 5-bp dG:dC oligonucleotide (data not shown) that can likely accommodate only a single α/β-type SASP. In contrast, SspC^wt cannot bind to this oligonucleotide (6). The fact that SspC^{ΔN11-D13K-C3} appears to exhibit negative cooperativity in its binding to DNA (even though this negative value may not be significant) is likely an artifact of the Hill plot, which underestimates cooperativity because it does not take into account the obstruction of contiguous binding sites along a DNA lattice (17). A preferable way to quantitate the cooperativity of protein-DNA binding is the McGhee-von Hippel equation (10). However, this equation is intended for modeling protein binding to homogeneous polynucleotides, and its use results in poor fitting with pUC19. Moreover, the modeling is only valid for proteins with high cooperativity. Therefore, to better compare all SspC variants, the Hill cooperativity factor is a more broadly applicable parameter.

Characteristics of the C-terminal region aside from length alone also determine the affinity with which SspC variants bind DNA. In variants terminating in lysine, the charged additional residue counteracts the benefits of extra C-terminal length. Possibly the confluence of charges—negative at the C-terminal carboxyl, partially positive at the histidine, and positive at the terminal residue—obstructs an important protein contact or creates a repulsion between the C-terminal residues and some other region of the protein, perhaps the amino terminus. Certainly the cooperativity of DNA binding by these variants is less than that of SspC^wt. In addition, these variants exhibit a decrease in affinity for DNA in going from pH 7 to 6, in contrast to other SspC variants. Most α/β-type SASP terminate in a lysine or arginine (2), making it likely that the additional positive charge in the C-terminal regions of SspC^C1+ and SspC^C3+, rather than the presence of a basic C-terminal residue alone, is the cause of the reduction in DNA binding affinity. However, the only natural α/β-type SASP with two basic residues in the C-terminal region, the Bacillus cereus protein BceI, has one of the highest affinities for binding to DNA (9, 31). Further clarification of the role of specific C-terminal residues of SspC^wt in DNA binding may need to await a high-resolution structure of an α/β-type SASP bound to DNA.

For most SspC variants, pH is one of the most easily manipulated and most significant determinants of the strength of their binding to DNA, as in vitro a decrease in pH from 7 to 6 resulted in a two- to sixfold increase in the affinity for DNA. However, the molecular explanation for the improved binding of SspC^wt to DNA at low pH is unclear. The C-terminal histidine, a residue that usually has a pK_a of ∼6, might become protonated at lower pH values, and the resulting positive charge might facilitate ionic interactions with DNA. However, this seems unlikely, since the affinities of the SspC^ΔC1 and SspC^ΔC3 variants for DNA are similar to that of SspC^wt at both pH 6 and 7. Therefore, protonation of another residue between pH 6 and 7 may be responsible for the increased SspC-DNA binding at pH 6. However, the identity of this residue is not clear, and we were unable to decrease the pH of binding reaction mixtures further because of the precipitation of the samples below pH 6.

α/β-type SASP are well-designed to take advantage of conditions in the developing forespore. A few hours after the start of sporulation, the forespore undergoes acidification, as the forespore pH falls to between 6 and 7 while the pH of the mother cell remains at 7.5 to 8 (12, 13, 29, 35). This drop in pH might be just enough to enhance the binding of α/β-type SASP to DNA at this critical juncture. Other factors in the forespore also encourage α/β-type SASP binding. For example, the binding of α/β-type SASP alters the conformation of DNA from a B-helix to an A-like helix, a DNA conformation favored under conditions of low humidity, and a large decrease in the forespore core water content is also a landmark event in sporulation (3, 18, 19). Therefore, the DNA in the developing spore is already primed to accept its protective coat of α/β-type SASP, and the decreased forespore pH may facilitate this binding. Conversely, during spore germination, the loss of the stored spore cations, including H⁺, is one of the earliest events, along with core hydration, and is followed by degradation of the α/β-type SASP (22). The rise in spore core pH early in germination almost certainly aids in the dissociation of α/β-type SASP from DNA, and this dissociation allows rapid degradation of these proteins by GPR, enabling the resumption of gene expression.

As well as α/β-type SASP are designed to bind spore DNA, they are not intended to bind permanently, and their intrinsic binding affinity for DNA is rather low. Most likely, this moderate binding reflects the conflicting purposes of α/β-type SASP: to bind and protect DNA during sporulation and spore dormancy but to dissociate from DNA and be degraded during spore germination. Both previous and present work has shown that α/β-type SASP that are not degraded rapidly are detrimental to spore outgrowth (8, 24). Perhaps it is not so surprising, then, that SspC variants with significantly enhanced affinity for DNA do not fulfill their role in the protection of the spore. If these variants bind DNA too tightly, they will inevitably become self-repressing as they nonspecifically bind to their own promoter sequences and other proximal regions. Indeed, α/β-type SASP have been shown to have global effects on forespore gene expression during sporulation (26) and in particular can regulate the expression of other ssp genes (33). If α/β-type SASP synthesis is repressed to the point where there is insufficient protein in spores to saturate the spore chromosome, some damage to the exposed DNA is probable. As demonstrated above, α⁻β⁻ spores that have only ∼15% of the α/β-type SASP content of wild-type spores have DNA that is much more susceptible to both UV and heat damage.

DNA is not the component of spores most susceptible to damage, but it is the most vital. The germinating spore can repair damage to its membranes and resynthesize much of its protein complement, but it cannot regenerate a new chromosome. One of the dilemmas a spore faces is between leaving its DNA vulnerable and having it accessible during the potentially long time from late in sporulation, through dormancy, and on into germination. With their moderate and flexible binding to DNA, α/β-type SASP appear designed to accommodate these conflicting needs and facilitate spore survival.