Structure of the Bacillus Cell Fate Determinant SpoIIAA in Phosphorylated and Unphosphorylated Forms

Structure, Vol. 9, 605–614, July, 2001, 2001 Elsevier Science Ltd. All rights reserved.

PII S0969-2126(01)00623-2

Structure of the Bacillus Cell Fate Determinant SpoIIAA in Phosphorylated and Unphosphorylated Forms Philippa R. Seavers,1 Richard J. Lewis,1,3 James A. Brannigan,1 Koen H. G. Verschueren,1 Garib N. Murshudov,1 and Anthony J. Wilkinson1,2 1 Structural Biology Laboratory Department of Chemistry University of York York YO10 5DD United Kingdom

Summary Background: The asymmetric cell division during sporulation in Bacillus subtilis gives rise to two compartments: the mother cell and the forespore. Each follow different programs of gene expression coordinated by a succession of alternate RNA polymerase ␴ factors. The activity of the first of these ␴ factors, ␴F, is restricted to the forespore although ␴F is present in the predivisional cell and partitions into both compartments following the asymmetric septation. For ␴F to become active, it must escape from a complex with its cognate anti-␴ factor, SpoIIAB. This relief from SpoIIAB inhibition requires the dephosphorylation of the anti-␴ factor antagonist, SpoIIAA. The phosphorylation state of SpoIIAA is thus a key determinant of ␴F activity and cell fate. Results: We have solved the crystal structures of SpoIIAA from Bacillus sphaericus in its phosphorylated and unphosphorylated forms. The overall structure consists of a central ␤-pleated sheet, one face of which is buried by a pair of ␣ helices, while the other is largely exposed to solvent. The site of phosphorylation, Ser57, is located at the N terminus of helix ␣2. The phosphoserine is exceptionally well defined in the 1.2 A˚ electron density maps, revealing that the structural changes accompanying phosphorylation are slight. Conclusions: Comparison of unphosphorylated and phosphorylated SpoIIAA shows that covalent modification has no significant effect on the global structure of the protein. The phosphoryl group has a passive role as a negatively charged flag rather than the active role it plays as a nucleus of structural reorganization in many eukaryotic signaling systems. Introduction The activation of alternate RNA polymerase ␴ factors is a widespread mechanism for changing the profile of gene expression in bacteria, for which spore formation by B. subtilis presents a textbook example [1, 2]. One of the simplest examples of cellular differentiation, sporulation begins with the formation of a polar septum lead2

Correspondence: [email protected] Present address: Laboratory of Molecular Biophysics, Rex Richards Building, University of Oxford, South Parks Road, Oxford, OX1 3QU, UK.

ing to an asymmetric cell division. This produces a larger mother cell and a smaller forespore. The mother cell proceeds to engulf the forespore and the two cells collaborate in constructing a complex proteinaceous coat around the developing spore. In the final stages the mother cell lyses, releasing the mature spore, which can lie dormant in the soil and germinate when favorable growth conditions are restored. Although the mother cell and the forespore contain identical chromosomes, they follow different programs of gene expression and have different fates. This is brought about by the sequential activation of alternate ␴ factors in the respective compartments, ␴F and ␴G in the forespore, ␴E and ␴K in the mother cell [3, 4]. ␴F is the first compartment-specific transcription factor to become active during sporulation. Although it is produced in the predivisional cell and partitions into both compartments upon formation of the asymmetric septum, ␴F is activated only in the forespore [5, 6]. ␴F (SpoIIAC) activity, and hence cell fate, is governed by a complex interplay with two other proteins encoded on the three cistron spoIIA operon, SpoIIAA and SpoIIAB [7]. SpoIIAB has two functions: it is an anti-␴ factor, which forms a tight complex with ␴F in which the latter is unavailable to bind core RNA polymerase (Figure 1), and it is also a protein kinase that specifically phosphorylates SpoIIAA [8, 9]. The phosphorylation state of SpoIIAA, which functions as an anti-␴ factor antagonist, is, in some sense, the ultimate arbiter of cell fate. In the predivisional cell, and in the mother cell following the asymmetric septation, SpoIIAA is present in its phosphorylated form, which is unable to bind to the anti-␴ factor, SpoIIAB [10]. The latter is therefore able to sequester the ␴ factor and prevent ␴F-directed gene expression. In the forespore compartment, SpoIIAA-P is selectively dephosphorylated and thereby activated by SpoIIE (Figure 1), which localizes to the polar septum [11–13]. The precise mechanism by which the phosphatase activity of SpoIIE is restricted to the forespore has not been unambiguously elucidated [14, 15]. Unphosphorylated SpoIIAA diverts the anti-␴ factor from its complex with ␴F, allowing the ␴ factor to combine with core RNA polymerase and initiate compartment-specific gene expression [16, 17]. Structural insights into the fascinating but complex interactions among these proteins are beginning to emerge with the determination of the solution structure of SpoIIAA from B. subtilis by NMR methods [18]. Sequence comparisons suggest that the architecture observed in SpoIIAA is shared by domains found in the sulfate transporter family of proteins, whose malfunction is associated with human disease [19]. The functional significance of the shared homology in STAS (sulphate transporters and anti-sigma factor antagonists) domain proteins is not yet clear. Our purpose in the present work is, first, to complement the NMR work with the

3

Key words: cell differentiation; crystallography; phosphorylation; sigma factor; sporulation; SpoIIAA

Structure 606

Figure 1. The Interactions among SpoIIA Proteins and SpoIIE in the Regulation of ␴F Activity (a) SpoIIAB (AB) is both an anti-␴ factor that can sequester ␴F (1) and a protein kinase that can phosphorylate SpoIIAA (AA) (2). SpoIIE (IIE) hydrolyses SpoIIAA-P (AA-P) (3). In the presence of unphosphorylated SpoIIAA, SpoIIAB can swap its ␴F partner for SpoIIAA (4). (b) In the larger mother cell compartment, the prevailing species are SpoIIAA-P and the SpoIIAB:␴F complex. SpoIIE (shaded circles) localizes to the sporulation septum but is active only in the forespore where it dephosphorylates SpoIIAA-P leading to ␴F activation.

determination of the crystal structure of SpoIIAA, and, second, to extend the structural insights to the phosphorylated species. Although we were able to grow crystals of SpoIIAA from B. subtilis, these crystals were unsuitable for X-ray analysis [20]. We therefore explored homologous proteins from other spore-forming Bacilli, and here we report the crystal structure of SpoIIAA from B. sphaericus in both its native and its phosphorylated forms.

phorylated SpoIIAA crystals belong to space group P6522, with a single molecule in the asymmetric unit. These crystals diffract weakly and the data are limited to 2.7 A˚ spacing. The refined model comprises residues 2–92 and 96–114; the chain termini and residues 93–95 are not defined by the electron density maps and they are assumed to be disordered. The phosphorylated SpoIIAA crystals are in space group P212121, with two molecules in the asymmetric unit. This structure was solved and refined against diffraction data extending to 1.2 A˚ spacing, making this the highest resolution structure of a phosphophorylated protein yet determined. The final, refined model extends from residue 2–112 in both chains, Met1 and the C-terminal residues 113–117 being disordered. The details of data collection and refinement for the three structures are presented in Tables 1 and 2. The overall fold in the two B. sphaericus SpoIIAA crystal structures is similar to the averaged NMR structure of SpoIIAA from B. subtilis (1buz) [18]. The proteins from the two species share 39 identities over their 117 residue polypeptide chains. SpoIIAA is a single domain globular protein with a largely compact structure and overall dimensions of 25 ⫻ 30 ⫻ 35 A˚3. The molecule contains five ␤ strands (␤1–␤5) and four ␣ helices (␣1–␣4) in the order ␤1⫺␤2⫺␣1⫺␤3⫺␣2⫺␤4⫺␣3⫺␤5⫺␣4 (Figures 2a and 2b). The central element of the SpoIIAA structure is a ␤-pleated sheet formed by four prominent ␤ strands, the first of which is antiparallel to the other three. A shorter strand comprising residues 98–100 runs parallel to part of strand ␤4 (residues 79–81), extending the sheet to five strands. The local interactions of these segments are continued at Ala102, whose main chain amide and carbonyl groups each form hydrogen bonds to the side chain of Asn81. Two prominent and parallel ␣ helices, ␣1 and ␣2, pack against each other and onto one face of the ␤ sheet, giving rise to an extensive hydrophobic core. Two further ␣ helices at the C terminus pack onto the edge of this sheet in a less regular arrangement which has been termed a panhandle [19]. The third ␣ helix runs in an almost perpendicular direction to helices ␣1 and ␣2. The loops ␤2-␣1 and ␤3-␣2 exchange main chain hydrogen bonds as they run parallel to one another and to helix ␣3, across what we will arbitrarily refer to as the “top” of the molecule in Figure 2. The fourth ␣ helix packs onto the opposite face of the ␤ sheet largely through apolar interactions but also through a bidentate salt bridge between Glu104 and Arg16.

Results Structure Determination and Overall Topology The structure of B. sphaericus SpoIIAA was initially determined by multi-wavelength anomalous dispersion (MAD) methods applied to a SeMet derivative. The protein used to prepare the first unphosphorylated SpoIIAA crystals [20] was found on structure determination to contain an adventitious Met86 to Val point mutation, presumably introduced during polymerase chain reaction subcloning steps. The SpoIIAA(M86V) structure has been refined against data extending to 1.6 A˚ spacing. The structures of the authentic (wild-type) forms of SpoIIAA were subsequently solved by molecular replacement (see Experimental Procedures). The unphos-

The Phosphorylation Site The phosphorylatable Ser57 (Ser58 in B. subtilis [21]) in SpoIIAA is situated at the N terminus of helix ␣2 (Figures 2 and 3a). Its side chain is oriented away from the main body of the molecule and into the solvent. It forms no polar interactions with other protein atoms. The side chains of the flanking Asp56 and Ser58 residues form a charge-dipole interaction. The phosphoryl group in SpoIIAA-P is exceptionally well defined by the 1.2 A˚ resolution electron density maps (Figures 3b and 3c). The phosphoserine side chain has two conformations in molecule A (Figure 3c). In neither conformation do the phosphate atoms partici-

Crystal Structure of Phosphorylated SpoIIAA 607

Table 1. Data Collection Statistics Crystal—Data Set

SpoIIAA (M86V) ␭1

SpoIIAA (M86V) ␭2

SpoIIAA (M86V) ␭3

SpoIIAA (M86V)

Space group Cell dimension (A˚) a b c Beamline

P212121 36.4 53.3 100.7 ESRF BM14

P212121 36.5 53.3 100.7 ESRF BM14

P212121 36.2 53.0 99.9 ESRF BM14

Wavelength (A˚) Resolution range (A˚) (outer shell) Rmergea (%) All data Outer shell ⬍I/␴(I)⬎ All data Outer shell Measurements Unique measurements Completeness (%) All data Outer shell

0.9778 24.7–2.5 (2.56–2.50) 3.5 5.8 19.2 8.6 23,961 12,741 97.0 76.7

0.9787 24.7–2.5 (2.56–2.50) 3.2 5.3 19.2 6.9 23,875 12,713 96.7 75.6

0.9184 25.0–2.5 (2.56–2.50) 2.9 5.2 21.8 12.9 25,039 12,838 99.4 97.9

P212121 36.7 53.5 101.3 ESRF ID14 EH2 0.93 19.9–1.6 (1.67–1.61) 7.6 24 19.8 5.4 118,510 25,927 96.4 87.1

a

SpoIIAA Phosphate P212121 51.1 61.5 65.8 SRS 14.1

P212121 51.1 61.9 65.9 ESRF ID14 EH1 0.934 20.0–1.16 (1.20–1.16) 6.8 30.0 18.9 3.7 277,544 72,227 97.9 85.4

1.488 20–1.67 (1.70–1.67) 4.5 6.8 31.6 28.1 163,469 24,230 98.9 100.0

Ser57, Asp56, and Ser58 form what appears to be a strong charge-dipole interaction, with the Asp56 side chain also N-capping the helix ␣2. Although the situation of the phosphorylated Ser at the N terminus of an ␣ helix suggests the possibility of favorable electrostatic interactions between the phosphate group and the helix dipole, the distances between the phosphate oxygen atoms and the main chain nitrogens are generally ⬎4 A˚ suggesting that these interactions are weak. Comparison of SpoIIAA with SpoIIAA-P The two independent molecules in the asymmetric unit of the SpoIIAA-P crystals, which were refined without applying non-crystallographic restraints, can be superimposed with an rms⌬ of 0.7 A˚ for all equivalent C␣ atoms. When either of the SpoIIAA-P molecules is super-

Table 2. Refinement Statistics

Resolution range (A˚) Measurements Rcrysta (%) Free Rcrystb (%) Number of residues Missing residues A mol B mol Number of atoms Protein Solvent Tris Rms⌬ bonds (A˚) angles (⬚) Average B factor (A˚2) Main chain Side chain Solvent Tris Ramachandran plotc Most favored (%) Additional allowed (%)

SpoIIAA (M86V)

SpoIIAA-Phosphate

SpoIIAA

19.9–1.6 25,927 15.9 21.4 230 1, 117 1, 117

20.0–1.16 72,227 13.2 16.4 222 1, 113–117 1, 113–117

20–2.74 4,022 25.0 30.2 110 1, 93–95, 115–117

1,815 206 — 0.021 1.801

1,829 423 16 0.025 1.921

872 9 — 0.020 2.021

17.4 22.0 30.0 —

9.8 13.0 27.2 30.6

35.8 37.0 54.3 —

94 6

90.8 9.2

85 15

Rcryst ⫽ 100 ⫻ ⌺hkl||Fobs| ⫺ k|Fcalc||/⌺hkl|Fobs| Free Rcryst ⫽ 100 ⫻ ⌺hkl傺T||Fobs| ⫺ k|Fcalc||/⌺hkl傺T|Fobs|, where hkl傺T represents the test data set of 5% of the diffraction data. c Calculated in PROCHECK [50]. b

P6522 81.6 81.6 71.6 ESRF ID14 EH3 0.930 20–2.74 (2.84–2.74) 7.2 34.2 23.5 5.3 26,991 4,022 99.8 99.7

Rmerge ⫽ 100 ⫻ ⌺hkl⌺i| I ⫺ ⬍I⬎|/⌺hkl⌺iI

pate in direct intramolecular interactions. Instead, the phosphoryl group points away from the protein and into the solvent, forming hydrogen bonds to water molecules and polar groups on surrounding protein molecules in the lattice. These include the hydroxyl of Thr9 and the guanidino group of Arg112 from one neighboring molecule in the crystal, and the imidazole of His24 from another. In the B molecule, a single side chain conformation has been modeled, with the phosphate oxygen atoms making interactions with a series of water molecules and the imidazole of His24 in a neighboring subunit. The face of the ring of the His24 side chain packs onto that of the adjacent His23, and this ring stacking extends in the crystal to include the equivalent residues from a neighboring molecule (Figure 3c). As for the native protein, in SpoIIAA-P, the side chains of the residues flanking

a

SpoIIAA (Wild-Type)

Structure 608

Figure 2. The Overall Fold of SpoIIAA (a) Ribbon tracing with the C␣ and side chain atoms of Ser57 shown in ball-and-stick. (b) Stereo C␣ trace with the N- and C-termini and every 10th residue labeled. This and subsequent figures were drawn with the program MOLSCRIPT [49].

imposed onto the structure of native SpoIIAA, the rms⌬ on equivalent C␣ atoms (residues 2–92 and 96–112) is 1.1 A˚ (Figure 4a). Similar comparison of these crystal structures with the averaged NMR structure [18] produces rms⌬s in the range 3.0–3.4 A˚ for superposition of 100 equivalent C␣ atoms (residues 12–111). However, a more careful analysis shows that the principal difference between the NMR and the X-ray structures is a significant displacement of the helices ␣1, ␣2, and ␣3 relative to the ␤ sheet (Figure 4b). These differences may reflect intrinsic flexibility within SpoIIAA. Alternatively, they may reflect sequence differences between species. In this regard the bulkier Trp47 residue on strand ␤3 in B. sphaericus SpoIIAA, which replaces a Leu in the B. subtilis protein, may alter the nature of the interactions between the ␤ sheet and helices ␣1 and ␣2. A third cause of the structural differences may be a lack of distance constraints between the ␤ sheet and the ␣ helices in the NMR data. The crystal structures of the native and phosphorylated wild-type proteins are also more similar to one another than either is to the mutant SpoIIAA in which Met86 is substituted by Val. Comparing SpoIIAA and SpoIIAA(M86V), the mean rms⌬ on C␣ atom positions of residues 2–92 and 96–112 is 1.8 A˚, while for SpoIIAA-P and SpoIIAA(M86V), a similar comparison based on residues 2–112 gives a mean pairwise rms⌬ of 1.7 A˚. In

the latter there are noticeable differences in the orientation of the ␤3-␣2 loop and helix ␣3 (Figure 4c). This is likely to be attributable to the mutation since Met86, which is located on helix ␣3, packs closely against Met55 which is situated on the ␤3-␣2 loop (Figure 4c). Similar changes take place in both molecules of the asymmetric unit of the crystal, arguing against an effect of crystal packing. Even so, it is interesting to note that Val occurs at position 86 in the majority of SpoIIAA homologs, including that from B. subtilis. Regardless, in discussing the structural changes arising from phosphorylation of SpoIIAA, we will confine our comparison to the crystal structures of the wild-type protein. The main structural differences between SpoIIAA and SpoIIAA-P are in residues 83–98, which encompass helix ␣3 and its flanking loops (Figure 4a). This segment contains residues 93–95, which are disordered in the SpoIIAA structure. The region of greatest variability between the independent molecules in the asymmetric unit of the SpoIIAA-P crystals spans these same residues, implying that they are intrinsically rather flexible. There are no hydrogen bonding or electrostatic interactions tethering residues 82–98 to the rest of the protein. Instead, the packing of this segment onto the protein core involves apolar contacts. Overall, therefore, we conclude that there are no concerted movements of the protein accompanying the arrival of the phosphoryl group. The only structural changes attributable to phosphorylation are the extra bulk of the phosphoryl group and two extra negative charges.

Discussion Interactions with Partner Proteins The genes of the spoIIA operon are switched on in the predivisional cell in response to deteriorating environmental conditions, which bring about phosphorylation of the response regulator Spo0A and activation of its transcription regulatory functions [22]. SpoIIAA is phosphorylated by SpoIIAB and ATP. SpoIIAB then forms a tight ternary complex with ␴F and an adenine nucleotide, presumably ATP [23]. Phosphorylated Spo0A also stimulates the expression of spoIIE, which encodes SpoIIE, an integral membrane protein with ten putative membrane-spanning segments at its N terminus and a PP2Clike phosphatase domain at its C terminus. SpoIIAA-P persists, however, until SpoIIE is assembled into polar rings at the potential site of septum formation in a process dependent on FtsZ [24–26]. In a reaction that is confined to the forespore by a mechanism that is not yet fully understood, SpoIIE hydrolyses the serine-phosphate ester bond in SpoIIAA-P, thus activating the anti-␴ factor antagonist. SpoIIAA diverts SpoIIAB from its complex with ␴F such that the ␴ factor can combine with RNA polymerase core enzyme and initiate compartment-specific gene expression (Figure 1). The simultaneous actions of SpoIIE and SpoIIAB in the forespore do not lead to an extensive futile cycle of SpoIIAA phosphorylation and dephosphorylation because SpoIIAB is much slower to turnover than SpoIIE [23]. After catalyzing phosphate transfer from ATP to SpoIIAA, SpoIIAB appears to be detained in a long-

Crystal Structure of Phosphorylated SpoIIAA 609

Figure 3. Electron Density Surrounding the Phosphorylatable Ser (a) Ser57 in unphosphorylated SpoIIAA(M86V), (b) the phosphorylated serine in the B molecule of SpoIIAA-P, and (c) the Ser-phosphates and the nearby His23 and His24 residues in adjacent molecules of the lattice. The Ser57-phosphate (SEP57) side chain has two conformations in molecule A. In the high-resolution maps, dual conformations of Met55 of molecule B and Ser57-phosphate of molecule A are evident. The atoms are colored according to their type: carbon is light blue, oxygen is red, nitrogen is blue, sulfur is yellow, and phosphorus is green. The electron density of the neighboring molecule in (c) is colored orange. In each case a 2Fo-Fc map is displayed and contoured at one standard deviation above the mean value of the electron density.

lived complex with ADP [27] by a second molecule of unphosphorylated SpoIIAA [28]. A long-lived SpoIIAASpoIIE complex has also been invoked based on experiments with SpoIIE fusion proteins and SpoIIE mutants, which suggest that ␴F may be regulated at a step subsequent to dephosphorylation of SpoIIAA-P [14]. These data suggest that following dephosphorylation of SpoIIAA-P by SpoIIE, SpoIIAA may be retained in complex with SpoIIE pending the completion of the sporulation septum. It may be significant in regard to the proposed longlived complexes with SpoIIAB and SpoIIE that the phos-

phorylatable Ser57 in B. sphaericus SpoIIAA is contiguous with an extended exposed apolar surface formed by (i) Val60, Leu64, Met67, and the aliphatic stem of the side chain of Arg68 on helix ␣2, (ii) Val89 and Phe92 on helix ␣3, and (iii) Leu95 and Trp98 on the ␣3-␤5 loop (Figure 5a). Extended hydrophobic surfaces frequently mediate stable protein-protein interactions [29]. Furthermore, the flexibility inherent in helix ␣3 and the loop that follows it adds further support for a role for these residues in protein-protein interactions, since binding surfaces on proteins often undergo a disorder-to-order transition on complex formation [30]. A corresponding hydrophobic

Structure 610

Figure 4. SpoIIAA Structural Comparisons Comparison of the overall structure of unphosphorylated SpoIIAA from B. sphaericus (green) with (a) the B molecule from the SpoIIAAP crystal structure (cyan), (b) the average NMR structure of SpoIIAA from B. subtilis (red) [18], and (c) the A molecule of the SpoIIAA(M86V) crystal structure (magenta). The figure is a stereo plot of the course of the protein backbone. In (a) and (c) the structures were superimposed by least squares minimization of the rms⌬ of the C␣ atom positions of residues 2–111. In (b) the superposition was based on the C␣ positions of 21 equivalent atoms from strands 1–4 of the ␤ sheet (rms⌬ ⫽ 0.67 A˚). The side chain and C␣ atoms of Ser57 in unphosphorylated SpoIIAA are displayed in (a) while those of Met55 and Met86 are shown in (c). For clarity, residues 103 to the C terminus have been omitted. Note that the green traces are broken between residues 92 and 96 because the intervening residues are disordered in the unphosphorylated SpoIIAA structure.

patch has been noted in SpoIIAA from B. subtilis where site-directed mutagenesis experiments have implicated residues contributing to this surface in binding either SpoIIAB or SpoIIE [28]. That this surface may be used to mediate proteinprotein interactions is illustrated by an extensive lattice contact made between the two molecules of the asymmetric unit of the SpoIIAA(M86V) crystals (Figure 6). The

molecules pack tightly about a noncrystallographic twofold axis of symmetry, burying 1250 A˚2 of largely apolar surface area on helix ␣2 and the ␣3-␤5 loop. Despite the presence of these dimers in the SpoIIAA(M86V) crystals, we have not observed dimers of either mutant or wild-type SpoIIAA in solution (unpublished data). A homolog of SpoIIAA, the TM1442 gene product from Thermotoga maritima has recently been purified and shown to exist as a monomer-dimer equilibrium in solution. The dimer form has been crystallized [31]. It will be interesting to compare the TM1442 protein dimer with the SpoIIAA(M86V) dimer shown in Figure 6. The importance of surface hydrophobic residues on the front face of the molecule (Figure 5a) in determining SpoIIAA function is also implied by sequence comparisons. In Figure 7, we have extended an alignment of anti-␴ factor antagonist sequences presented by Kovacs et al. [18] to include recently determined sequences of SpoIIAA proteins from evolutionarily distant endosporeforming bacteria. The hydrophobic nature of residues highlighted in Figure 5a is conserved. The sequence comparison emphasizes the clustering of strongly conserved residues (Gly19-Glu20-Leu21-Asp22 and Phe54Met55-Asp56) on the loops preceding the long ␣ helices. Glu20 and Phe54 protrude prominently on the upper surface of the molecule as shown in Figure 5b. These loops are followed by a pair of conserved surface histidines (His23 and His24) and the phosphorylatable Ser respectively. The sequence conservation extends into helix ␣2 and is apparent in the ␣3-␤5 loop, which includes the invariant Ser93 (Figures 5 and 7). Four of the six invariant Gly residues can be assigned obvious structural roles either in facilitating turns or, in the case of Gly59, in close packing of the protein core. However, this cannot be the case for Gly61 or Gly65, which are situated on the outer surface of helix ␣2. It is possible that their conservation is important in facilitating the close approach of partner proteins. In this regard, substitution of Gly62 in SpoIIAA from B. subtilis (corresponding to Gly61 in the B. sphaericus protein) by Asp leads to inefficient phosphorylation of SpoIIAA by SpoIIAB and dephosphorylation of SpoIIAA-P by SpoIIE in vitro and a failure to sporulate in vivo [32]. We can speculate that SpoIIAB and/or SpoIIE bind to SpoIIAA forming (1) predominantly polar interactions with protruding side chains on the top surface of the molecule that confer specificity and (2) predominantly apolar interactions with the front face of the molecule that confer higher affinity and, importantly, low dissociation rate constants (Kd ⵑ50 nM and koff ⵑ0.015 s⫺1 for the SpoIIAA:SpoIIAB complex in the presence of ADP [23]). Figure 7 also includes the sequences of RsbV-like anti-␴ factor antagonists which regulate stress responses in a range of bacteria. Many of the residues conserved among the SpoIIAA-like proteins are also conserved in the RsbV-like proteins, suggesting a similar mechanism of action. These two sets of proteins can be distinguished, based on primary sequence, by “signature” residues surrounding the phosphorylatable Ser. The invariant Ser58, Ser93, and Phe54 in the SpoIIAA sequences appear as Thr, Thr, and Tyr, respectively, in the RsbV proteins. The conserved His23, His24, and Arg66 in SpoIIAA are not present in RsbV (Figure 7). These

Crystal Structure of Phosphorylated SpoIIAA 611

Figure 5. Space Filling Views of the SpoIIAA Surface (a) View similar to that in Figure 2a illustrating the hydrophobic surface contiguous with the phosphorylatable Ser (SEP57, green). Other residues are colored according to type; positively charged residues (Arg, Lys, and His) are blue, negatively charged residues (Asp and Glu) are red, polar residues (Asn, Gln, Ser, and Thr) are yellow, and apolar residues are in gray. Selected residues referred to in the text are labeled. (b) View looking down onto SEP57 illustrating residues conserved among SpoIIAA homologues. SEP57 is again in green; residues invariant in 11 of the 12 SpoIIAA-like sequences are in yellow with the exceptions of Gly61 and Gly65, which are in red.

residues are candidates for conferring specificity on the binding of SpoIIAA and RsbV to their cognate kinases SpoIIAB and RsbW, respectively, since there appears to be no crosstalk between these systems [33]. Ser/Thr Phosphorylation in Other Regulatory Systems In the phosphorylation control of ␴F activity, SpoIIAB is the kinase which phosphorylates SpoIIAA as well as the sensor of the phosphorylation signal. The structural data presented here demonstrate that phosphorylation does not perturb the gross structure of SpoIIAA and that the phosphate acts as a flag as opposed to a trigger of conformational change. SpoIIAB needs only to sense whether SpoIIAA is phosphorylated or not, and this can presumably be achieved by surveying a surface centered on, or overlapping with, the active Ser. Isocitrate dehydrogenase (IDH) and the His-containing phosphocarrier protein (HPr) constitute two other regulatory systems where negligible conformational changes result from Ser/Thr phosphorylation [37, 38]. Phosphorylation of Ser113 inhibits IDH so that the CO2-evolving steps of the tricarboxylic acid cycle can be bypassed as isocitrate is diverted into the glyoxylate pathway. The Ser-phosphate partially occupies the isocitrate binding site and prevents substrate binding by an electrostatic blocking mechanism. Phosphorylation of

Ser46 prevents HPr from acting as a phosphocarrier substrate in the phosphoenol-pyruvate:sugar phosphotransferase system associated with carbohydrate entry into bacterial cells. Modeling studies suggest that phosphorylated HPr is prevented from binding to the phosphodonor enzyme-I (EI) by electrostatic repulsion between the Ser-phosphate on HPr and a glutamate residue on EI [38]. We would predict that phosphorylation control in SpoIIAA is also based on electrostatic repulsion between the phosphorylated Ser and the phosphate groups of adenine nucleotides bound to SpoIIAB. Interestingly, in SpoIIAA, as in IDH and HPr, the effects of phosphorylation of the wild-type protein can be mimicked by replacement of the active Ser by acidic residues [17, 36, 38]. In glycogen phosphorylase and in the MAP/cyclin dependent kinases which regulate the cell cycle and cell development in eukaryotes, Arg residues from different domains/subunits are recruited to counter the additional negative charges arising from phosphorylation [34, 35]. Comparison of the phospho- and dephospho-forms of the enzymes reveals long-range conformational changes accompanying the formation of a phosphate recognition site. The Arg movements extend the “sphere of influence” of the phosphoryl group to structurally remote sites where new or altered binding surfaces may be created or where active residues may be reconfigured Figure 6. A SpoIIAA Dimer in the Crystal Stereo ribbon diagram illustrating the packing of the two molecules in the asymmetric unit of the SpoIIAA(M86V) crystals. The N- and C-termini and helices ␣2 and ␣3 are labeled. The close association of the ␣3-␤5 loops and the ␣2 helices about a pseudo two-fold noncrystallographic symmetry axis is apparent.

Structure 612

Figure 7. Alignment of Anti-␴ Factor Antagonist Sequences (1–12) SpoIIAA-like proteins from 1. B. sphaericus (AJ278291), 2. B. subtilis (P10727), 3. B. licheniformis (P26777), 4. B. anthracis (http:// www.tigr.org/tdb/mdb/mdbinprogress.html), 5. B. coagulans (P70877), 6. B. stearothermophilus (O32726), 7. B. megaterium (P35147), 8. B. halodurans (Q9KCN3), 9. Paenibacillus polymyxa (O32720), 10. Carboxydothermus hydrogenoformans (http://www.tigr.org/tdb/mdb/ mdbinprogress.html), 11. C. acetobutylicum (http://www.genomecorp.com/genesequences/clostridium/clospage.html), and 12. C. difficile (http://sanger.ac.uk/Projects/C_difficile/). Residues invariant in at least 11 of the 12 sequences are picked out in the consensus, Spo. (13–19) RsbV-like sequences: 13. B. anthracis (Q9K5J8), 14. Staphylococcus epidermidis (AAG23811), 15. Staphylococcus aureus (P95842), 16. B. halodurans (Q9Kff2), 17. B. subtilis (P17903), 18. B. licheniformis (O50230), and 19. Listeria monocytogenes (O85016). Residues invariant in at least 6 of the 7 sequences are picked out in the consensus Rsb. Con is the consensus sequence for anti-␴ factor antagonists, given in bold if strictly conserved over the family. Other residues highlighted in bold represent the “signature” by which SpoIIA/RsbV families can be identified from their primary sequence. Semicolons denote conservation of amino acid residue based on physiochemical property, e.g., hydrophobicity or charge.

[36]. In contrast to SpoIIAA, HPr and IDH, acidic residue substitutions of the active Ser in glycogen phosphorylase and the MAP/cyclin dependent kinases, do not mimic the effects of phosphorylation.

Concluding Remarks Anti-␴ factors are utilized elsewhere in bacteria to regulate the transcription of genes of diverse function including those involved in flagella biosynthesis, phage development, virulence, and the response to oxidative stress (reviewed in [39, 40]). In these systems, a number of different mechanisms exist for regulating the anti-␴ factor. Inhibition by FlgM of ␴28, which controls flagella biosynthesis in Escherichia coli, is overcome by export of the anti-␴ factor from the cell. The activity of RsrA, which inhibits ␴R-directed expression of oxidative stress genes in Streptomyces coelicolor, is controlled by redox-induced disulphide bond formation [40]. In contrast, development and the stress response in B. subtilis are regulated through a phosphorylation-sensitive anti-␴ factor antagonist. The structures of SpoIIAA [18] and SpoIIAA-P described here provide the first structural insights into anti-␴ factor regulation. The key observation is that the structural changes caused by phosphorylation are slight, with the implication that the mere presence or absence of a ⫺PO32⫺ group on Ser57 determines cell fate. The essence of a signal is, of course, in its interpretation, and a full appreciation of phosphorylation in the control ␴F demands crystal structures of SpoIIAA in complex with the sensor, SpoIIAB.

Biological Implications Opportunities to study cellular differentiation and alternate developmental pathways are rare in tractable systems such as unicellular organisms. The process of sporulation in B. subtilis is the ultimate response of the bacterium to a hostile environment and has proved a fertile ground for research in cell development. Sporulation involves an asymmetric cell division giving rise to two cells of unequal size that follow different programs of gene expression but collaborate in the maturation of the smaller cell into a resistant spore. The coordinated and sequential activation of alternate RNA polymerase ␴ factors in the mother cell and forespore compartments lies at the heart of transcription regulation and differential gene expression. The activity of the first of these ␴ factors, ␴F, is restricted to the forespore, although ␴F is present in both cell compartments. This is because ␴F is normally held in an inactive complex with SpoIIAB. Similar anti-␴ factor based mechanisms control a number of cellular responses, including the production of flagella and virulence factors. The escape of ␴F from inhibition requires a third protein, SpoIIAA, whose activity is controlled by phosphorylation and dephosphorylation. SpoIIAA is normally maintained in the inactive state by the specific protein kinase activity of SpoIIAB. It is activated by SpoIIE, a phosphatase associated with the polar septum. Thus, the presence or absence of the phosphoryl group in SpoIIAA governs the fate of the cell. We have determined the crystal structure of SpoIIAA

Crystal Structure of Phosphorylated SpoIIAA 613

in its active form and in its inactive state in which the protein is phosphorylated on Ser57. The lack of conformational rearrangement caused by the arrival of the phosphoryl group is striking and contrasts with the effects of phosphorylation in other regulatory proteins involved in signaling. It appears, therefore, that the loss of the bulk of the phosphorus and three oxygen atoms together with the associated ⫺2 charge represents the switch from the inactive to the active form of SpoIIAA. Experimental Procedures Protein Preparation and Crystallization Recombinant SpoIIAA and SpoIIAA-P proteins were purified from E. coli BL21 strains harboring pET28a derivatives containing B. sphaericus DNA sequences coding for either spoIIAA alone or both spoIIAA and spoIIAB. Coexpression of the two proteins leads to intracellular phosphorylation of SpoIIAA by SpoIIAB. SpoIIAA and SpoIIAA-P were purified and crystallized as described previously [20]. When the structure of the unphosphorylated form was solved, it became clear that residue 86 was Val rather than Met, as a result of an A-to-G transition mutation introduced during a PCR subcloning step. The introduction of this mutation turns out to be serendipitous as crystals of the SpoIIAA(M86V) mutant are much more strongly diffracting than crystals of authentic unphosphorylated B. sphaericus SpoIIAA (Table 1). Selenomethionine substituted SpoIIAA(M86V) was prepared from the methionine auxotrophic E. coli strain B834 (DE3) by growth in suitable liquid media containing SeMet. Purification and crystallization of SeMet-SpoIIAA was as for the native protein. Structure Determination SeMet derivative data were collected from a single SpoIIAA(M86V) crystal on beamline BM14 at the ESRF (European Synchrotron Radiation Facility), Grenoble, with a MARCCD detector. Ninety degrees of data were collected at three wavelengths at a crystal to detector separation of 160 mm yielding data to a nominal 2.5 A˚ spacing. Data were acquired in 0.5⬚ oscillations. The data collection statistics are summarized in Table 1. The positions of the same nine selenium sites were determined using the programs Shake’n’Bake and SOLVE, the latter in its automatic Patterson search mode [41, 42]. During refinement of the Se parameters in MLPHARE [43], a tenth Se position was identified in difference maps. MAD phasing from these atoms produced an electron-density map with a clear boundary between protein and solvent, with a mean figure-of-merit (FOM) of 0.78–2.5 A˚. These phases were improved by solvent flattening and histogram matching in the program DM [44], the mean FOM rising to 0.88. The map was readily interpretable, allowing almost the entire structure (two molecules in the asymmetric unit) to be built using the X-AUTOFIT routines in the program QUANTA (MSI). Prior to refinement of the model, the phases were extended to 1.6 A˚ (Table 1). This produced a high quality electron-density map with a mean FOM of 0.77. After making appropriate adjustments to the model, it was subjected to refinement using the maximum likelihood methods implemented in REFMAC [45]. The Rfactor and the free Rfactor (Rfree) dropped immediately from ⵑ42.6% to 29.0% and 31.6%, respectively. Successive rounds of rebuilding and refinement were carried out until refinement converged, at which point water molecules were added using the X-SOLVATE routine of QUANTA. Water molecules were subsequently deleted if they failed to meet chemical and refinement criteria. Further rounds of refinement were carried out, including a final round of anisotropic temperature factor refinement in the program REFMAC [46]. The final Rfactor (Rfree) was 15.9% (21.4%). Details of refinement and the final model are given in Table 2. Before initiating the MAD phasing experiments, we failed to solve the structure of B. sphaericus SpoIIAA(M86V) by molecular replacement using the NMR coordinates of B. subtilis SpoIIAA as a search model. After the structure was solved, the MR calculations were repeated, this time successfully, as one of three test cases of newly developed routines implemented in the program MOLREP that

search simultaneously for multiple copies of the search molecule in the asymmetric unit [47]. Structure of SpoIIAA-Phosphate Data from the SpoIIAA-P crystals were initially collected on beamline 14.2 at the SRS to 1.67 A˚ spacing. The structure was solved by molecular replacement in MOLREP [48] using the refined SpoIIAA(M86V) structure as the search molecule. Two solutions were obtained giving an initial Rfactor of 54.1% (Rfree ⫽ 54.9%). Refinement was carried out in the program REFMAC interspersed with manual model building sessions in the program QUANTA. Water molecules were introduced and refined, and finally the atoms of phosphorylated serine (SEP) were added. At this point, a higher resolution data set was collected on beamline ID14.1 at the ESRF, Grenoble. These 1.16 A˚ data were combined with the 1.67 A˚ SRS data and refinement of the model was continued. The higher resolution data revealed alternate conformations for a number of residues. Unphosphorylated SpoIIAA For the wild-type SpoIIAA crystals, data were collected on beamline 14.3 at the ESRF, Grenoble. The data are limited to 2.7 A˚ spacing. The structure was solved by molecular replacement in MOLREP using molecule B of the refined SpoIIAA-P structure as the search molecule. A single solution was obtained with an Rfactor of 49.2% and a correlation coefficient of 0.46. It was apparent in the first maps that movements of residues 83–98 had taken place. These residues were deleted from the initial model and rebuilt manually. In the course of refinement in REFMAC, it became clear that residues 83–98 have high mobility and that residues 93–95 are disordered. Acknowledgments This work was funded by Wellcome Trust Grant 057339/Z/99/Z. P.R.S. was the recipient of an EPSRC (Engineering and Physical Sciences Research Council) studentship and K.H.G.V. and G.N.M. were supported by the BBSRC (Biotechnology and Biological Sciences Research Council). We are grateful to the BBSRC for its support of the York Structural Biology Laboratory and to the ESRF and the SRS ( Synchrotron Radiation Source) for access to and support on their beam-lines. We would like to thank Lorraine Hewitt, Nicholas Tabouriech, Johan Turkenburg, Anabelle Varrot, and Lisa Wright for assistance with data collection and Eleanor Dodson and Michael Yudkin for helpful discussions. R.J.L. is currently a Wellcome Trust Career Development Fellow. Received: March, 14 2001 Revised: May, 23 2001 Accepted: June, 6 2001 References 1. Losick, R., and Pero, J. (1981). Cascades of sigma factors. Cell 25, 583–584. 2. Stragier, P., and Losick, R. (1996). Molecular genetics of sporulation in Bacillus subtilis. Annu. Rev. Genet. 30, 297–341. 3. Losick, R., and Stragier, P. (1992). Criss-cross regulation of celltype-specific gene expression during development in B. subtilis. Nature 355, 601–604. 4. Kroos, L., and Yu, Y.-T.N. (2000). Regulation of ␴ factor activity during Bacillus subtilis development. Curr. Opin. Microbiol. 3, 553–560. 5. Gholamhoseinian, A., and Piggot, P.J. (1989). Timing of spoII gene expression relative to septum formation during sporulation of Bacillus subtilis. J. Bacteriol. 171, 5747–5749. 6. Margolis, P., Driks, A., and Losick, R. (1991). Establishment of cell type by compartmentalized activation of a transcription factor. Science 254, 562–565. 7. Fort, P., and Piggot, P.J. (1984). Nucelotide sequence of the sporulation locus spoIIAA in Bacillus subtilis. J. Gen. Microbiol. 130, 2147–2153. 8. Duncan, L., and Losick, R. (1993). SpoIIAB is an anti-␴ factor that binds to and inhibits transcription by regulatory protein ␴F from Bacillus subtilis. Proc. Natl. Acad. Sci. USA 90, 2325–2329.

Structure 614

9. Min, K.-T., Hilditch, C.M., Diederich, B., Errington, J., and Yudkin, M.D. (1993). ␴F, the first compartment-specific ␴ factor of Bacillus subtilis is regulated by an anti-␴ factor that is also a protein kinase. Cell 74, 735–742. 10. Magnin, T., Lord, M., Errington, J., and Yudkin, M.D. (1996). Establishing differential gene expression in sporulating Bacillus subtilis: phosphorylation of SpoIIAA (anti-anti-␴F) alters its conformation and prevents formation of a SpoIIAA/SpoIIAB complex. Mol. Microbiol. 19, 901–907. 11. Arigoni, F., Pogliano, K., Webb, C.D., Stragier, P., and Losick, R. (1995). Localisation of protein implicated in establishment of cell type to sites of asymmetric division. Science 270, 637–640. 12. Duncan, L., Alper, S., Arigoni, F., Losick, R., and Stragier, P. (1995). Activation of cell specific transcription by a serine phosphatase at the site of asymmetric division. Science 270, 641–644. 13. Bara´k, I., et al., and Youngman, P. (1996). Structure and function of the Bacillus SpoIIE protein and its localization to sites of sporulation septum assembly. Mol. Microbiol. 19, 1047–1060. 14. King, N., Dreesen, O., Stragier, P., Pogliano, K., and Losick, R. (1999). Septation, dephosphorylation and the activation of ␴ during sporulation in Bacillus subtilis. Genes Dev. 13, 1156– 1167. 15. Frandsen, N., Bara´k, I., Karnazyn-Campelli, C., and Stragier, P. (1999). Transient gene asymmetry during sporulation and establishment of cell specificity in Bacillus subtilis. Genes Dev. 13, 394–399. 16. Alper, S., Duncan, L., and Losick, R. (1994). An adenosine nucleotide switch controlling the activity of a cell type-specific transcription factor in B. subtilis. Cell 77, 195–205. 17. Diederich, B., Wilkinson, J.F., Magnin, T., Najafi, S.M.A., Errington, J., and Yudkin, M.D. (1994). Role of interactions between SpoIIAA and SpoIIAB in regulating cell-specific transcription factor ␴F in Bacillus subtilis. Genes Dev. 8, 2653–2663. 18. Kovacs, H., Comfort, D., Lord, M., Campbell, I.D., and Yudkin, M.D. (1998). Solution structure of SpoIIAA, a phosphorylatable component of the system that regulates transcription factor sigma(F) of Bacillus subtilis. Proc. Natl. Acad. Sci. USA 95, 5067–5071. 19. Aravind, L. and Koonin, E. (2000). The STAS domain — a link between anion transporters and antisigma-factor antagonists. Curr. Biol. 10, R53–R55. 20. Seavers, P.R., Lewis, R.J., Brannigan, J.A., and Wilkinson, A.J. (2001). Crystallisation and preliminary X-ray analysis of the sporulation factor SpoIIAA in its native and phosphorylated forms. Acta Crystallogr. D 57, 292–295. 21. Najafi, S.M.A., Willis, A.C., and Yudkin, M.D. (1995). Site of phosphorylation of SpoIIAA the anti-anti-sigma factor for sporulation-specific ␴F of Bacillus subtilis. J. Bacteriol. 177, 2912–2913. 22. Hoch, J.A. (1993). Regulation of the phosphorelay and the initiation of sporulation in Bacillus subtilis. Annu. Rev. Microbiol. 47, 441–465. 23. Magnin, T., Lord, M., and Yudkin, M.D. (1997). Contribution of partner switching and SpoIIAA cycling to regulation of ␴F activity in sporulating Bacillus subtilis. J. Bacteriol. 179, 3922–3927. 24. Levin, P.A., Losick, R., Stragier, P., and Arigoni, F. (1997). Localization of the sporulation protein SpoIIE in Bacillus subtilis is dependent upon the cell division protein FtsZ. Mol. Microbiol. 25, 839–846. 25. Khvorova, A., Zhang, L., Higgins, M.L., and Piggot, P.J. (1998). The spoIIE locus is involved in the Spo0A-dependent switch in the location of FtsZ rings in Bacillus subtilis. J. Bacteriol. 180, 1256–1260. 26. Lucet, I., Feucht, A., Yudkin, M.D., and Errington, J. (2000). Direct interaction between the cell division protein FtsZ and the cell differentiation protein SpoIIE. EMBO J. 19, 1467–1475. 27. Najafi, S.M.A., Harris, D.A., and Yudkin, M.D. (1997). Properties of the phosphorylation reaction catalysed by SpoIIAB that help to regulate sporulation of Bacillus subtilis. J. Bacteriol. 179, 5628–5631. 28. Lee, C.-S., Clarkson, J., Shu, J.-C., Campbell, I.D., and Yudkin, M.D. (2001). Bacillus subtilis mutations that alter the pathway of phosphorylation of the anti-anti-␴F factor SpoIIAA lead to a Spo- phenotype. Mol. Microbiol. 40, 9–19. 29. Lo Conte, L., Chothia, C., and Janin, J. (1999). The atomic struc-

30.

31.

32.

33.

34.

35.

36. 37.

38.

39. 40. 41. 42. 43.

44.

45.

46.

47.

48.

49.

50.

ture of protein-protein recognition sites. J. Mol. Biol. 285, 2177– 2198. Daughdrill, G.W., Chadsey, M.S., Karlinsey, J.E., Hughes, K.T., and Dahlquist, F.W. (1997). The C-terminal half of the anti-sigma factor, FlgM, becomes structured when bound to its target, sigma28. Nat. Struct. Biol. 4, 285–291. Ha, K.S. et al., and Suh, S.W. (2001). Crystallization and preliminary X-ray crystallographic analysis of the TM1442 gene product from Thermotoga maritima, a homologue of Bacillus subtilis anti-anti-sigma factors. Acta Crystallogr. D 57, 276–278. Barilla, D., Lucet, I., Kuhlmann, A., and Yudkin, M.D. (1999). Genotype, phenotype, and protein structure in a regulator of sporulation: effects of mutations in the spoIIAA gene of Bacillus subtilis. J. Bacteriol. 181, 3860–3863. Alper, S., Dufour, A., Garsin, D.A., Duncan, L., and Losick, R. (1996). Role of adenosine nucleotides in the regulation of a stress-response transcription factor in Bacillus subtilis. J. Mol. Biol. 260, 165–177. Russo, A.A., Jeffrey, P.D., and Pavletich, N.P. (1996). Structural basis of cyclin-dependent kinase activation by phosphorylation. Nat. Struct. Biol. 3, 696–700. Canagarajah, B.J., Khokhlatchev, A., Cobb, M.H., and Goldsmith, E.J. (1997). Activation mechanism of the MAP kinase ERK2 by dual phosphorylation. Cell 90, 859–869. Johnson, L.N., and O’Reilly, M. (1996). Control by phosphorylation. Curr. Opin. Struct. Biol. 6, 762–769. Hurley, J.H., Dean, A.M., Thorsness, P.E., Koshland, D.E., Jr., and Stroud, R.M. (1990). Regulation of isocitrate dehydrogenase by phosphorylation involves no long-range conformational change in the free enzyme. J. Biol. Chem. 265, 3599–3602. Audette, G.F., et al., and Delbaere, L.T.J. (2000). The 1.9 A˚ resolution structure of phospho-serine 46 HPr from Enterococcus faecalis. J. Mol. Biol. 303, 545–553. Hughes, K.T., and Mathee, K. (1998). The anti-sigma factors. Annu. Rev. Microbiol. 52, 231–286. Kang, J.-G., et al., and Roe, J.-H. (1999). RsrA, an anti-sigma factor regulated by redox change. EMBO J. 18, 4292–4298. Weeks, C.M., and Miller, R. (1999). The design and implementation of SnB version 2.0. J. Appl. Crystallogr. 32, 120–124. Terwilliger, T.C., and Berendzen, J. (1999). Automated MAD and MIR structure solution. Acta Crystallogr. D 55, 849–861. CCP4 (Collaborative Computational Project 4) (1994). The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D 50, 760–763. Cowtan, K. (1994). An automated procedure for phase improvement by density modification. Joint CCP4 ESF-EACBM Newslett. Protein Crystallogr. 31, 34–38. Murshudov, G.N., Vagin, A.A., and Dodson, E.J. (1997). Refinement of macromolecular structures by the Maximum-Likelihood Method. Acta Crystallogr. D 53, 240–255. Murshudov, G.N., Lebedev, A., Vagin, A.A., and Dodson, E.J. (1999). Efficient anisotropic refinement of macromolecular structures using FFT. Acta Crystallogr. D 55, 247–255. Vagin, A.A., and Teplyakov, A. (2000). An approach to multicopy search in molecular replacement. Acta Crystallogr. D 56 1622– 1624. Vagin, A.A. and Teplyakov, A. (1997). MOLREP: an automated program for molecular replacement. J. Appl. Crystallogr. 30, 1022–1025. Kraulis, P.J. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallogr. 24, 946–950. Laskowski, R.A., MacArthur, M.W., Moss, D.S., and Thornton, J.M. (1993). PROCHECK: a program to check the stereochemical quality of protein structure coordinates. J. Appl. Crystallogr. 26, 283–291.

Accession Numbers Coordinates and structure factors for SpoIIAA (1h4z), SpoIIAA-P (1h4x), and SpoIIAA(M86V) (1h4y) have been deposited with the EBI.