Structure of the N-terminal Domain of the Circadian Clock-associated Histidine Kinase SasA

doi:10.1016/j.jmb.2004.07.010

J. Mol. Biol. (2004) 342, 9–17

Structure of the N-terminal Domain of the Circadian Clock-associated Histidine Kinase SasA Ioannis Vakonakis1, Douglas A. Klewer1, Stanly B. Williams2 Susan S. Golden2 and Andy C. LiWang1* 1

Department of Biochemistry and Biophysics, Texas A&M University, College Station TX 77843, USA 2 Department of Biology, Texas A&M University, College Station, TX 77843, USA

Circadian oscillators are endogenous biological systems that generate the w24 hour temporal pattern of biological processes and confer a reproductive fitness advantage to their hosts. The cyanobacterial clock is the simplest known and the only clock system for which structural information for core component proteins, in this case KaiA, KaiB and KaiC, is available. SasA, a clock-associated histidine kinase, is necessary for robustness of the circadian rhythm of gene expression and implicated in clock output. The N-terminal domain of SasA (N-SasA) interacts directly with KaiC and likely functions as the sensory domain controlling the SasA histidine kinase activity. N-SasA and KaiB share significant sequence similarity and, thus, it has been proposed that they would be structurally similar and may even compete for KaiC binding. Here, we report the NMR structure of N-SasA and show it to be different from that of KaiB. The structural comparisons provide no clear details to suggest competition of SasA and KaiB for KaiC binding. N-SasA adopts a canonical thioredoxin fold but lacks the catalytic cysteine residues. A patch of conserved, solventexposed residues is found near the canonical thioredoxin active site. We suggest that this surface is used by N-SasA for protein–protein interactions. Our analysis suggests that the structural differences between N-SasA and KaiB are the result of only a few critical amino acid substitutions. q 2004 Elsevier Ltd. All rights reserved.

*Corresponding author

Keywords: circadian; SasA; thioredoxin; cyanobacteria; evolution

Introduction Circadian clock systems generate the w24 hour temporal pattern of biological processes1,2 and can be found in evolutionarily diverse eukaryotes, including insects, plants, fungi and mammals,3 and have been demonstrated in a single prokaryotic group, the cyanobacteria.4 In general, proteins from different circadian clock systems have little sequence similarity. Despite the lack of homology with those of eukaryotes, the cyanobacterial clock has recently become the focus of structural research,5–8 as it is the simplest and possibly the oldest identified thus far.9 The core of the Present addresses: D. A. Klewer, Department of Chemistry and Biochemistry, Arizona State University, Tempe, AZ 85287, USA; S. B. Williams, Department of Biology, University of Utah, Salt Lake City, UT 84112, USA. Abbreviation used: NOE, nuclear Overhauser effect. E-mail address of the corresponding author: [email protected]

Synechococcus elongatus PCC 7942 clock comprises at least three interacting proteins, KaiA, KaiB and KaiC,4,10 that form complexes that oscillate in size and composition with w24 hour periodicity.11 Clock output controls the rhythmic expression patterns from virtually all S. elongatus promoter elements.4,12,13 SasA is an EnvZ-type, two-component histidine kinase protein14 previously shown to associate physically with KaiC.11,15 Disruption of the sasA gene or SasA kinase activity decreases expression from the kaiBC promoter and affects robustness of the circadian oscillator adversely, to the point that rhythmicity of expression is lost from many clockcontrolled genes.15 However, circadian rhythmicity is not abolished completely, as the kai genes are still expressed rhythmically. This places SasA function outside of the core clock oscillator and likely on the immediate clock output pathway.15 SasA is hypothesized to phosphorylate an as yet unidentified cognate response regulator similar to other histidine kinases of two-component signal transduction systems.15,16 KaiC is the largest of the three core clock proteins 4 and KaiC overexpression

0022-2836/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.

10 attenuates global transcription levels.4,13 Several studies show that KaiA and KaiB control the KaiC autophosphorylation state in vitro and in vivo by enhancing or attenuating the KaiC autokinase activity, respectively.6,17–19 The N-terminal domain of SasA (N-SasA) is necessary and sufficient for the SasA–KaiC interaction15 and, if SasA functions similarly to other histidine kinases, likely modulates the level of the SasA autokinase activity.15 N-SasA has significant sequence similarity to KaiB (36% identity and 58% similarity over 77 residues),15,20 which led to the hypothesis that the sasA gene arose from duplication and functional coupling to a histidine kinase of a kaiB-like ancestral gene.20 In addition, the alignment of several non-polar residues between N-SasA and KaiB implied a conserved hydrophobic core and, thus, similar protein structures.20 Because both proteins bind KaiC, proposed models suggested that N-SasA and KaiB possibly compete for KaiC binding.11,15,20 The recently-determined structure of KaiB from the cyanobacterium Anabaena sp. PCC 7120 revealed an a-b meander fold.8 KaiB crystallized as a dimer, which is consistent with previous reports of dimeric KaiB in vivo11 and formation of KaiB oligomers in vivo and in vitro.10,21,22 Here, we report the NMR structure of the S. elongatus N-SasA. Contrary to the sequencebased expectations, N-SasA is not structurally similar to KaiB, as it adopts a canonical thioredoxin fold23,24 and the isolated domain is monomeric in solution. Thus, a comparison of the N-SasA and KaiB structures does not suggest competition of SasA and KaiB for binding to KaiC. Despite the high

Structure of the N-terminal domain of SasA

level of structural similarity to other thioredoxinlike proteins, N-SasA lacks the cysteine residues necessary for redox activity. Instead, N-SasA acts as a KaiC-interaction module, possibly linking KaiC and SasA kinase activity. Our analysis of the N-SasA structure suggests that protein–protein interactions are likely mediated through a patch of conserved solvent-exposed residues near the canonical thioredoxin active site.

Results N-SasA oligomeric state Full-length SasA has been shown to participate in homotypic interactions in vitro.15 Analytical ultracentrifugation equilibrium experiments of 15 N-enriched samples of the N-SasA domain dissolved in the NMR buffer (Figure 1(A)) show a particle mass consistent with the approximately 11.6 kDa molecular mass of our N-SasA construct. In contrast, a normalized simulated absorbance plot for a dimeric 23.2 kDa particle (shown in Figure 1(A) as a broken line) clearly does not fit the experimental data. Additionally, we were able to calculate the rotational correlation time (tc) for the N-SasA particle as described25 from backbone 15 N dynamics data (Figure 1(B) and (C)). The tc for N-SasA is approximately 6.8 ns at 25 8C, which is consistent with a monomeric particle. For comparison the 15 kDa KaiA135N has a tc of w8.2 ns at 25 8C,6 the 7.7 kDa vMIP-II has a tc of w4.7 ns at 25 8C25 and the 25 kDa ThKaiA180C has a tc of

Figure 1. Oligomeric state of the N-SasA domain. (A) Analytical ultracentrifugation equilibrium absorbance versus rotor radius. Data were acquired on a 20 mM sample of 15N-enriched N-SasA in the NMR buffer and fit to an ideal monodisperse model shown as a continuous line. Residuals of the fit are plotted against rotor radius at the top of the graph. The molecular mass from the fit is consistent with an N-SasA monomer (11.6 kDa). For comparison, a normalized simulated plot for a 23.2 kDa particle is shown as a broken line. (B) 15N{–1H} NOE and C, 15N T1/T2 ratio data are plotted here versus residue number. Data were collected on a 0.6 mM 15N-enriched N-SasA sample in the NMR buffer at 25 8C and 14.1 T (600 MHz 1H frequency). The rotational correlation time (tc) of N-SasA was calculated25 to be w6.8 ns, which is consistent with a monomer in solution.

11

Structure of the N-terminal domain of SasA

Table 1. NMR structure statistics and restraints Experimental restraints NOE Intraresidue (iKjZ0) Sequential (iKjZ1) Short range (1!iKj!5) Long range (iKjR5) Ambiguous Hydrogen bondsa Dihedral angles (deg.) f j c1 c2 3 HNHA J couplings 13 a 13 b C , C chemical shifts Total number of restraints Structure quality RMSDs from experimental restraints ˚) Distance restraints (A Dihedral (deg.) Chemical shifts (ppm) 13 a C 13 b C 3 HNHA J couplings (Hz) ˚ Distance violations O0.3 A Dihedral angle violations O58 RMSDs from idealized geometry ˚) Bonds (A Angles (deg.) Impropers (deg.) Ramachandran statisticsb (%) Most-favored regions Additionally allowed regions Generously allowed regions Disallowed regions Structure precisionb,d ˚) Backbone atoms (A ˚) All heavy-atoms (A

1767 447 463 357 477 23 34 63 63 67 4 71 198 2267 25 Structure ensemble

Averaged minimized structure

0.018G0.001 0.32G0.04

0.012 0.26

1.06G0.03 1.01G0.03 0.62G0.03 0 0

1.01 0.96 0.61 0 0

0.0022G0.0001 0.526G0.008 0.37G0.01

0.0022 0.52 0.42

92.1 5.4 2.5 0.1c

91.8 4.1 2.7 1.4c

0.34G0.06 0.7G0.1

a Distance restraints were applied between amide proton and oxygen atoms, and between amide nitrogen and oxygen atoms for each hydrogen bond restraint. b Mobile residues (15N{–1H} NOE!0.6) were excluded (Figure 1(B)). Included residues are 13–40, 46–65, 68–76 and 78–101. c Q101 falls in the disallowed region in one out of the 25 structures of the ensemble and in the average minimized structure. d RMSD from the average structure.

w9.4 ns at 50 8C.7 Therefore, the isolated N-SasA domain is monomeric in solution under the experimental conditions described here and we expect that the reported SasA homotypic interactions15 are mediated through a domain homologous to the core four-helix bundle domain of EnvZ.26 N-SasA structure The solution structure of N-SasA was calculated from a total of 2267 experimental restraints (Table 1), with an average of 24.6 restraints per residue for residues 11–101. The final ensemble (Figure 2(B)) consists of the 25 lowest-energy structures, and has an RMSD against the average ˚ for the backbone and 0.7(G0.1) A ˚ of 0.34(G0.06) A for all heavy-atoms of low-mobility residues (Table 1). As seen in Figure 2(B), the first ten residues of N-SasA are unstructured in solution, which is consistent with the lack of long-range nuclear Overhauser effects (NOEs) from these

residues and the small deviations from random coil observed in the 13 Ca and 13 Cb chemical shifts.27 In addition, 15N{–1H} NOE values in this region are small or negative (Figure 1(B)), which indicates rapid, large-amplitude local motions. N-SasA consists of a four-stranded b-sheet in 2p1a3a4 arrangement flanked by two parallel a-helices (a1 and a3) on one side and partially covered on the other side by a small a-helix (a2) connecting b2 and b3. The overall architecture is that of the canonical thioredoxin fold.23 Indeed, a structure-based search by Dali28 using the average minimized structure of N-SasA as query returned statistically significant superpositions of N-SasA with thioredoxin, thioredoxin-like and glutaredoxin proteins (see the Supplementary data); other structure-based search methods29,30 returned similar results (not shown). Figure 2(C) compares thioredoxin and glutaredoxin structures to that of N-SasA. N-SasA shows a mixture of thioredoxin and glutaredoxin features. Similar to glutaredoxins,31

12

Structure of the N-terminal domain of SasA

Figure 2. Solution structure of the N-SasA domain and comparison to other thioredoxin-fold proteins. Shown here are (A) the average minimized structure of N-SasA and (B) the 25-structure ensemble in two mutually orthogonal views. The N- and C-termini of the protein are indicated. The ensemble of structures was superimposed using the backbone of lowmobility residues (Table 1). (C) Comparison of bacteriophage T4 glutaredoxin (PDB-ID 1AAZ), the N-SasA domain and Escherichia coli thioredoxin (2TRX). The unstructured first ten residues of N-SasA, and the first 20 residues of thioredoxin that have no structural homologues in N-SasA are not shown. The structures were superimposed using the Ca atoms of the b-sheet.

N-SasA lacks the N-terminal b-strand and short a-helix typically found in thioredoxins.23,24,32 However, N-SasA has longer secondary structure elements than glutaredoxins23 and the most significant superpositions found by all structure-based search methods were thioredoxin proteins. In addition, N-SasA appears to retain some features of thioredoxin proteins, such as a highly conserved, positively charged residue (usually lysine) at the end of thioredoxin b4 (equivalent to K73 at the end of b3 in N-SasA). In thioredoxins, that positively charged residue interacts with and stabilizes the C-terminal helix through electrostatic and hydrogen bond interactions.24,32 The C-terminal helix of N-SasA is elongated compared to that of canonical thioredoxins24,32 and, thus, K73 does not stabilize the C-terminal helix in the same fashion. Instead, K73 packs against W98 of a3 and is likely involved in charge–dipole interactions. Nevertheless, the similarities of N-SasA and thioredoxin-fold proteins are strictly structural, as N-SasA lacks the cysteine residues necessary for redox activity. Although thioredoxin and thioredoxin-like proteins have been known to play regulatory roles in protein kinase C,33 SasA has an unusual arrangement in which a thioredoxin-like domain serves as the

sensory domain for a two-component histidine kinase. Putative N-SasA interaction surface Sequence conservation among six cyanobacterial N-SasA domains is relatively low and covers primarily residues important for structure, such as the two cis-proline residues (P69 and P76) and hydrophobic residues of the b-sheet (Figure 3(A)). Interestingly, the a2 area of N-SasA has a higher proportion of highly conserved residues compared to the other two helices, with at least two of these residues completely exposed to solvent (E61 and V66). The side-chains of these two residues are near the active site in thioredoxins and are flanked by three other conserved or conservatively substituted residues (P57, R64, Q80) as shown in Figure 3(B). The thioredoxin active site area is known to mediate non-catalytic protein–protein interactions such as those seen in the thioredoxin-T7 DNA polymerase complex.34 Similarly, a mutation near the same area in the non-catalytic thioredoxin-like N-terminal domain of Drosophila Wind is known to affect the interaction of Wind with Pipe, which results in subcellular mislocalization of Pipe.35 Thus, we

Structure of the N-terminal domain of SasA

13

Figure 3. Sequence alignment of cyanobacterial N-SasA proteins and putative protein interaction surface. (A) A CLUSTAL-W multiple sequence alignment of different cyanobacterial N-SasA domains. Sequences correspond to the following strains: PCC 7942, S. elongatus PCC 7942; PCC 6803, Synechocystis sp. PCC 6803; PCC 7120, Anabaena sp. PCC 7120; BP-1, Thermosynechococcus elongatus BP-1; CCMP 1378, Prochlorococcus marinus subsp. pastoris CCMP 1378/MED 4; and WH8120, Synechococcus sp. WH8120. Highly conserved residues likely involved in protein–protein interactions are indicated by yellow bars. (B) Accessible surface area of residues highlighted in (A). The unstructured first ten residues of N-SasA are not shown.

propose that the surface shown in Figure 3(B) is used by N-SasA for interactions with KaiC or interdomain interactions with the SasA histidine kinase domain. Comparison of N-SasA and KaiB A CLUSTAL-W sequence alignment between S. elongatus N-SasA and KaiB (Figure 4(A)) shows

28.6% identical and 55.2% similar residues distributed equally through the sequences, while BLAST searches with the S. elongatus KaiB or N-SasA sequences return statistically significant alignments of the two proteins with 36% identity and 58% similarity over 77 residues and an expectation value (E) of 2.2!10K2 (not shown). The sasA gene has been proposed to have evolved from fusion of a kaiB-like ancestral gene to a histidine kinase gene.20

Figure 4. Structure comparison of the N-SasA domain and KaiB. (A) Shown here is a CLUSTAL-W alignment of S. elongatus N-SasA and KaiB. The secondary structure elements from the solution structure of N-SasA (top line), and the crystal structure of the Anabaena sp. PCC 7120 KaiB8 (highly homologous to S. elongatus KaiB; bottom line) are shown. Residues likely to be critical for the differences in the two structures are indicated by yellow bars. (B) Secondary structure prediction for S. elongatus N-SasA and KaiB. Prediction was performed by PSIPRED36 and the regular secondary structure elements identified are shown. N-SasA residues 1–8 are not shown. (C) Structures and (D) protein architectures of S. elongatus N-SasA and Anabaena sp. KaiB.8 The unstructured first ten residues of N-SasA are not shown. The protein architectures are highly similar for the first half of the sequences but diverge in the second half.

14 Both the SasA and KaiB proteins are known to interact with KaiC,10,15 and the SasA–KaiC interaction is mediated by the N-SasA domain.15 Thus, it was proposed that N-SasA and KaiB might also share similar tertiary structures15,20 and likely compete for the same binding site in KaiC.11,15,20 However, comparison of our N-SasA domain structure to that of the recently determined Anabaena sp. PCC 7120 KaiB crystal structure8 shows that KaiB and N-SasA have different folds. As KaiB proteins from S. elongatus and Anabaena are conserved (87% identity and 94% similarity over 97 residues, EZ5!10K42) to a much higher degree than N-SasA and KaiB, we expect that the S. elongatus and Anabaena KaiB proteins will have highly similar structures. Comparison of the KaiB and N-SasA structures reveals differences at the secondary, tertiary and quaternary structural levels. As shown in Figure 4(A), the secondary structure elements of the two proteins align well for the first half of the sequence but differ after b2. As a result, the protein architectures and structures diverge after the initial b-a-b motif (Figure 4(C) and (D)). Whereas N-SasA follows a canonical thioredoxin fold, KaiB does not show significant similarity to any other protein of known structure. Indeed, structure-based search methods28–30 were unable to find alignments of continuous residues when using KaiB as query (not shown). The only region of KaiB that superimposed well in these methods was the initial b-a-b motif shared between N-SasA and KaiB. In addition, KaiB forms a homodimer in the crystal, using primarily b3 and the loop that connects b2 and b3 as the dimer interface. The KaiB homodimer is physiological, as previous studies revealed homotypic KaiB associations in vivo.11 In contrast, the isolated N-SasA domain is monomeric in solution. The identical amino acid residues between N-SasA and KaiB indicated by our CLUSTAL-W alignment do not appear to form similar accessible surfaces in the two structures.

Discussion The differences observed between the N-SasA and KaiB structures were surprising, given the sequence similarity of the two proteins. Nevertheless, these differences can explain previous experimental results for SasA and KaiB. KaiB is involved in homotypic interactions and it would be reasonable to expect a KaiB-like N-SasA to interact with KaiB. However, no SasA–KaiB interactions have been detected in vitro or in yeast two-hybrid assays.15 Also, KaiB is known to act as an attenuator of KaiC autophosphorylation,6,18,19 which is an activity not exhibited by SasA.19 Both KaiB and N-SasA bind to KaiC; thus, they could potentially compete for the same KaiC interaction surface. However, in vivo data suggest that both KaiB and SasA can associate with KaiC simultaneously.11 Note that although most of the residues forming the

Structure of the N-terminal domain of SasA

putative N-SasA protein interaction surface are conserved or substituted conservatively in KaiB, the different folds of the two proteins result in different spatial positions for these residues. Thus, the structural data presented here provide no evidence to suggest that SasA and KaiB compete for KaiC binding. Indeed, based on the structural differences between N-SasA and KaiB, and their different effect on KaiC, we expect that they bind at different sites on KaiC. It is important to point out that the proposal for the evolutionary relationship between kaiB and sasA, based on the statistically significant sequence similarity,20 is not invalidated by our work. The fact that N-SasA belongs to the well-known and often encountered thioredoxin fold, while KaiB is structurally novel, suggests a high probability of common lineage to a thioredoxin-like gene. A copy of that ancestral gene may have fused to a histidine kinase gene, thereby creating sasA; while a second copy accumulated mutations giving rise to kaiB. An interesting question is whether the structural divergence between the N-SasA domain and KaiB is the result of changes in secondary structure likelihood in the two protein sequences or if a few key amino acid changes can trigger this transformation. Prediction of secondary structure elements by PSIPRED36 returns effectively the same results for S. elongatus KaiB and N-SasA (Figure 4(B)). Interestingly, in the second half of the sequences where the two structures diverge, the predicted secondary structure does not match well the secondary structure from either the S. elongatus N-SasA or Anabaena KaiB structures but, instead, features elements from both. Thus, we propose that the differences observed are more likely due to specific critical substitutions in the two proteins. A comparison of the two sequences suggests such possibly critical substitutions on N-SasA in the a2 helix and the b3-b4 hairpin, and along the KaiB dimerization interface. Specifically, N-SasA V60 and Y63 contribute stabilizing hydrophobic interactions for N-SasA a2, which are lost upon substitutions to A53 and D56, respectively, and may be responsible for the extended coil configuration of the equivalent region in KaiB. Formation of the N-SasA b3-b4 hairpin is likely hindered in KaiB by substitution of a small flexible residue, G77, for L70. In addition, KaiB residues involved in stabilizing hydrophobic interactions across the KaiB dimer interface (L47 and I87) are missing from N-SasA (A54 and N93), thereby making dimer formation less favorable. Despite the high level of structural similarity to thioredoxin and glutaredoxin proteins, N-SasA lacks the catalytic cysteine residues of those enzymes. Instead, the thioredoxin-fold is used here as a KaiC interaction module to couple KaiC and SasA histidine kinase activity and, thus, control clock output. Residues in a relatively small accessible surface of N-SasA near the thioredoxin active site are conserved significantly among cyanobacterial species (Figure 3(B)). It is possible that N-SasA uses this area as an interface for interactions with

15

Structure of the N-terminal domain of SasA

KaiC or with the SasA histidine kinase domain. Both KaiC–SasA and SasA interdomain interactions would facilitate signal transduction and likely serve as clock output from the core oscillator, which results ultimately in rhythmic transcription of virtually all S. elongatus genes.

Materials and Methods Protein preparation The gene fragment encoding S. elongatus SasA residues 1–105 (N-SasA) was cloned in a pGEX-2T expression vector (Pharmacia), thereby creating a glutathione-Stransferase (GST)/N-SasA fusion. The N and C termini of N-SasA were altered slightly during cloning and are GSSLS instead of MGESLS, and QEGIF instead of QEAF, respectively. This construct was transformed in Escherichia coli BL21(DE3) and the cells were grown at 37 8C in minimal medium containing 15N-enriched NH4Cl (Spectra Stable Isotopes) and unenriched or 13C-enriched glucose (Spectra Stable Isotopes) as sole nitrogen and carbon sources, respectively. Protein expression was induced by adding isopropyl-b-D-thiogalactopyranoside (IPTG, Calbiochem) to a final concentration of 1 mM. The cells were harvested after five hours, resuspended in 200 mM NaCl, 20 mM Tris–HCl (pH 7.5), 1% (v/v) Triton-X, and lysed by French press. Cell lysates were centrifuged at 20,000g for 30 minutes and the supernatant was loaded onto a GSTrap column (Pharmacia) equilibrated with 20 mM Tris–HCl (pH 7.5), 100 mM NaCl. The bound protein was released by addition of 10 mM reduced glutathione (Calbiochem) to the same buffer. The protein was buffer-exchanged to 20 mM Tris–HCl (pH 8.0), 100 mM NaCl, and cleaved by thrombin (Sigma) for four hours. After cleavage, the pH was adjusted to 7.5 and thrombin was inactivated by addition of excess pAPMSF (Calbiochem). Finally, cleaved N-SasA was separated from GST through a GSTrap column and exchanged to the final NMR buffer (20 mM Na2HPO4 (pH 7.0), 100 mM NaCl). Protein purity was established by SDS-PAGE and N-SasA was concentrated to approximately 0.6–0.8 mM by stirred ultrafiltration. Protein concentration was determined by measurement of absorbance at 280 nm, using a corrected extinction coefficient as suggested by Pace et al.37 Approximately 6 mg of purified N-SasA were obtained per liter of doubly labeled culture. Analytical ultracentrifugation Analytical ultracentrifugation equilibrium experiments were performed on 15N-enriched samples of N-SasA (calculated mass of 11,615 Da for the fully enriched protein) in the NMR buffer using a Beckman Optima XL-A analytical ultracentrifuge. UV absorbance was monitored at 235 nm. The centrifugation was for 48 hours at 25,000 rpm in a Beckman An-60Ti rotor, at 4 8C. The data were fit to an ideal monodisperse model using the program Origin (Microcal Software Inc., Northampton, MA). NMR experiments All experiments were performed at 25 8C using Varian Inova 600 and 500 MHz spectrometers equipped with triple-axis gradient probes. Sequential 13C, 15N and 1H,

backbone and side-chain assignments were described earlier.27 Stereospecific assignments of valine Hg methyl groups and c1 angles for valine, isoleucine and threonine residues were obtained by determining long-range 15N and 13C 0 J-couplings to side-chain methyl groups.38,39 Stereospecific assignments of methyline Hb protons and additional c1 angle determinations were performed by using HNHB40 and HACAHB41 experiments. Stereospecific assignments of leucine Hd methyl groups were inferred from the NOE spectroscopy (NOESY) spectra. Isoleucine c2 angle determination was performed by using a long-range 13C-13C J-coupling experiment.42 Interproton NOE distance restraints were obtained by 13 C/13C-edited 4D NOESY, 13C/15N-edited 4D NOESY and 15N-edited 3D NOESY spectra. Hydrogen bond restraints were applied based on hydrogen-exchange protection data collected by NMR, as well as the existence of expected regular secondary structure NOEs.43 The isomerization states of five out of seven proline residues were determined as trans; P69 and P76 are in the cis isomerization state.43 Backbone dynamics measurements (15N T1, 15N T2 and 15N{–1H} NOE) were performed and analyzed as described.25 NMR structure calculations The f and j dihedral angle values were derived from TALOS.44 Dihedral angle restraint boundaries were set to two standard deviations for residues with all ten dihedral angle predictions in the same region of the Ramachandran plot, and to two standard deviations plus 108 for residues with nine out of ten predictions in the same region of the Ramachandran plot. The XPLOR-NIH software package45 was used for all stages of NMR structure calculations. Only NOE, dihedral angle and 3JHNHA coupling potential energy terms were used as restraints during simulated annealing. Additional potential energy terms were applied during the structure refinement process. These include a radius of gyration restraint, with a calculated value of 11.63 nm applied to residues 14–40, 46–64, and 68–101;46 a conformational database potential term47 and direct refinement against 13 Ca and 13 Cb chemical shifts.48 The final ensemble consists of the 25 lowest-energy structures (Table 1). Figure and sequence notes Structure Figures were prepared by SPOCK.49 The first two and last three residues of the N-SasA construct (cloning artifacts) are not shown. All sequences shown were drawn from the Swiss-Prot/TrEMBL databases and have the following accession numbers: S. elongatus PCC 7942 SasA, Q06904; Synechocystis sp. PCC 6803 SasA, Q55630; Anabaena sp. PCC 7120 SasA, Q8YR50; Thermosynechococcus elongatus BP-1 SasA, Q8DMT2; Prochlorococcus marinus subspecies pastoris CCMP 1378/MED4 SasA, Q7V113; Synechococcus sp. WH8120 SasA, Q7U871; S. elongatus PCC 7942 KaiB, Q9Z3H3. Data Bank accession codes Sequential 13C, 15N and 1H, backbone and side-chain assignments have been deposited in the BioMagResBank under accession number 5141. The structures and structure calculation restraints have been deposited in the Protein Data Bank under accession numbers 1T4Y and 1T4Z for the average minimized structure and the 25 structure ensemble, respectively.

16

Structure of the N-terminal domain of SasA

Acknowledgements We thank Dr Emil F. Pai for providing us with the Anabaena KaiB atomic coordinates prior to public release, and Dr Karl Koshlap for technical assistance with the NMR instrumentation. Funding was provided to A.C.L. by the National Institutes of Health (GM064576) and to S.S.G. by the National Science Foundation (MCB-9982852). S.B.W. was supported by an NRSA fellowship (NIH GM19644). The NMR instrumentation in the Biomolecular NMR Laboratory at Texas A&M University was supported by the National Science Foundation grant DBI-9970232.

13.

14. 15.

16.

Supplementary data Supplementary data associated with this article can be found, in the online version, at doi: 10.1016/j.jmb.2004.07.010

17.

18.

References 1. Bu¨nning, E. (1973). The Physiological Clock (3rd edit.). Springer, New York. 2. Dunlap, J. C., Loros, J. J. & DeCoursey, P. J. (2004). Editors of Chronobiology: Biological Timekeeping. Sinauer Associates, Sunderland, MA. 3. Dunlap, J. C. (1999). Molecular bases for circadian clocks. Cell, 96, 271–290. 4. Ishiura, M., Kutsuna, S., Aoki, S., Iwasaki, H., Andersson, C. R. Tanabe, A. et al. (1998). Expression of a gene cluster kaiABC as a circadian feedback process in cyanobacteria. Science, 281, 1519–1523. 5. Ye, S., Vakonakis, I., Ioerger, T. R., LiWang, A. C. & Sacchettini, J. C. (2004). Crystal structure of circadian clock protein KaiA from Synechococcus elongatus. J. Biol. Chem. 279, 20511–20518. 6. Williams, S. B., Vakonakis, I., Golden, S. S. & LiWang, A. C. (2002). Structure and function from the circadian clock protein KaiA of Synechococcus elongatus: a potential clock input mechanism. Proc. Natl Acad. Sci. USA, 99, 15357–15362. 7. Vakonakis, I., Sun, J., Wu, T., Holzenburg, A., Golden, S. S. & LiWang, A. C. (2004). NMR structure of the KaiC-interacting C-terminal domain of KaiA, a circadian clock protein: implications for the KaiA– KaiC interaction. Proc. Natl Acad. Sci. USA, 101, 1479– 1484. 8. Garces, R. G., Wu, N., Gillon, W. & Pai, E. F. (2004). Anabaena circadian clock proteins KaiA and KaiB reveal a potential common binding site to their partner KaiC. EMBO J. 23, 1688–1698. 9. Dvornyk, V., Vinogradova, O. & Nevo, E. (2003). Origin and evolution of circadian clock genes in prokaryotes. Proc. Natl Acad. Sci. USA, 100, 2495–2500. 10. Iwasaki, H., Taniguchi, Y., Ishiura, M. & Kondo, T. (1999). Physical interactions among circadian clock proteins KaiA, KaiB, and KaiC in cyanobacteria. EMBO J. 18, 1137–1145. 11. Kageyama, H., Kondo, T. & Iwasaki, H. (2003). Circadian formation of clock protein complexes by KaiA, KaiB, KaiC, and SasA in cyanobacteria. J. Biol. Chem. 278, 2388–2395. 12. Liu, Y., Tsinoremas, N. F., Johnson, C. H., Lebedeva,

19.

20.

21.

22.

23. 24. 25.

26.

27.

28.

N. V., Golden, S. S., Ishiura, M. & Kondo, T. (1995). Circadian orchestration of gene expression in cyanobacteria. Genes Dev. 9, 1469–1478. Nakahira, Y., Katayama, M., Miyashita, H., Kutsuna, S., Iwasaki, H., Oyama, T. & Kondo, T. (2004). Global gene repression by KaiC as a master process of prokaryotic circadian system. Proc. Natl Acad. Sci. USA, 101, 881–885. Dutta, R., Qin, L. & Inouye, M. (1999). Histidine kinases: diversity of domain organization. Mol. Microbiol. 34, 633–640. Iwasaki, H., Williams, S. B., Kitayama, Y., Ishiura, M., Golden, S. S. & Kondo, T. (2000). A KaiC-interacting sensory histidine kinase, SasA, necessary to sustain robust circadian oscillation in cyanobacteria. Cell, 101, 223–233. Bourret, R. B., Hess, J. F. & Simon, M. I. (1990). Conserved aspartate residues and phosphorylation in signal transduction by the chemotaxis protein CheY. Proc. Natl Acad. Sci. USA, 87, 41–45. Iwasaki, H., Nishiwaki, T., Kitayama, Y., Nakajima, M. & Kondo, T. (2002). KaiA-stimulated KaiC phosphorylation in circadian timing loops in cyanobacteria. Proc. Natl Acad. Sci. USA, 99, 15788–15793. Kitayama, Y., Iwasaki, H., Nishiwaki, T. & Kondo, T. (2003). KaiB functions as an attenuator of KaiC phosphorylation in the cyanobacterial circadian clock system. EMBO J. 22, 2127–2134. Xu, Y., Mori, T. & Johnson, C. H. (2003). Cyanobacterial circadian clockwork: roles of KaiA, KaiB and the kaiBC promoter in regulating KaiC. EMBO J. 22, 2117– 2126. Dvornyk, V., Deng, H.-W. & Nevo, E. (2004). Structure and molecular phylogeny of sasA genes in cyanobacteria: insights into evolution of the prokaryotic circadian system. Mol. Biol. Evol. 21, 1468–1476. Iwase, R., Imada, K., Hayashi, F., Uzumaki, T., Namba, K. & Ishiura, M. (2004). Crystallization and preliminary crystallographic analysis of the circadian clock protein KaiB from the thermophilic cyanobacterium Thermosynechococcus elongatus BP-1. Acta Crystallog. sect. D, 60, 727–729. Xu, Y., Piston, D. W. & Johnson, C. H. (1999). A bioluminescence resonance energy transfer (BRET) system: application to interacting circadian clock proteins. Proc. Natl Acad. Sci. USA, 96, 151–156. Martin, J. L. (1995). Thioredoxin—a fold for all reasons. Structure, 3, 245–250. Katti, S. K., LeMaster, D. M. & Eklund, H. (1990). Crystal structure of thioredoxin from Escherichia coli at ˚ resolution. J. Mol. Biol. 212, 167–184. 1.68 A LiWang, A. C., Cao, J. J., Zheng, H., Lu, Z., Peiper, S. C. & LiWang, P. J. (1999). Dynamics study on the antihuman immunodeficiency virus chemokine viral macrophage-inflammatory protein-II (vmip-II) reveals a fully monomeric protein. Biochemistry, 38, 442–453. Tomomori, C., Tanaka, T., Dutta, R., Park, H., Saha, S. K. Zhu, Y. et al. (1999). Solution structure of the homodimeric core domain of Escherichia coli histidine kinase EnvZ. Nature Struct. Biol. 6, 729–734. Klewer, D. A., Williams, S. B., Golden, S. S. & LiWang, A. C. (2002). Sequence-specific resonance assignments of the N-terminal, 105-residue KaiC-interacting domain of SasA, a protein necessary for a robust circadian rhythm in Synechococcus elongatus. J. Biomol. NMR, 24, 77–78. Holm, L. & Sander, C. (1998). Touring protein fold space with Dali/FSSP. Nucl. Acids Res. 26, 316–319.

Structure of the N-terminal domain of SasA

29. Krissinel, E. & Henrick, K. (2003). Protein structure comparison in 3D based on secondary structure matching (SSM) followed by Ca alignment, scored by a new structural similarity function. In Proceedings of the 5th International Conference on Molecular Structural Biology (Kungle, A. J. & Kungl, P. J., eds), p. 88, Vienna. 30. Madej, T., Gibrat, J. F. & Bryant, S. H. (1995). Threading a database of protein cores. Proteins: Struct. Funct. Genet. 23, 356–369. 31. Eklund, H., Ingelman, M., Soderberg, B. O., Uhlin, T., Nordlund, P. Nikkola, M. et al. (1992). Structure of oxidized bacteriophage T4 glutaredoxin (thioredoxin). Refinement of native and mutant proteins. J. Mol. Biol. 228, 596–618. 32. Weichsel, A., Gasdaska, J. R., Powis, G. & Montfort, W. R. (1996). Crystal structures of reduced, oxidized, and mutated human thioredoxins: evidence for a regulatory homodimer. Structure, 4, 735–751. 33. Witte, S., Villalba, M., Bi, K., Liu, Y., Isakov, N. & Altman, A. (2000). Inhibition of the c-Jun N-terminal kinase/AP-1 and NF-kappaB pathways by PICOT, a novel protein kinase C-interacting protein with a thioredoxin homology domain. J. Biol. Chem. 275, 1902–1909. 34. Doublie, S., Tabor, S., Long, A. M., Richardson, C. C. & Ellenberger, T. (1998). Crystal structure of a bacterio˚ resoluphage T7 DNA replication complex at 2.2 A tion. Nature, 391, 251–258. 35. Ma, Q., Guo, C., Barnewitz, K., Sheldrick, G. M., Soling, H. D., Uson, I. & Ferrari, D. M. (2003). Crystal structure and functional analysis of Drosophila Wind, a protein-disulfide isomerase-related protein. J. Biol. Chem. 278, 44600–44607. 36. Jones, D. T. (1999). Protein secondary structure prediction based on position-specific scoring matrices. J. Mol. Biol. 292, 195–202. 37. Pace, C. N., Vajdos, F., Fee, L., Grimsley, G. & Gray, T. (1995). How to measure and predict the molar absorption coefficient of a protein. Protein Sci. 4, 2411–2423. 38. Grzesiek, S. & Bax, A. (1993). Amino acid type determination in the sequential assignment procedure of uniformly 13C/15N-enriched proteins. J. Biomol. NMR, 3, 185–204.

17 39. Vuister, G. W., LiWang, A. C. & Bax, A. (1993). Measurement of three-bond nitrogen–carbon J couplings in proteins uniformly enriched in 15N and 13C. J. Am. Chem. Soc. 115, 5334–5335. 40. Madsen, J. C., Sorensen, O. W., Sorensen, P. & Poulsen, F. M. (1993). Improved pulse sequences for measuring coupling constants in 13C, 15N-labeled proteins. J. Biomol. NMR, 3, 239–244. 41. Grzesiek, S., Kuboniwa, H., Hinck, A. P. & Bax, A. (1995). Multiple-quantum line narrowing for measurement of Ha K Hb J couplings in isotopically enriched proteins. J. Am. Chem. Soc. 117, 5312–5315. 42. Bax, A., Max, D. & Zax, D. (1992). Measurement of long-range 13C–13C J couplings in a 20-kDa protein– peptide complex. J. Am. Chem. Soc. 114, 6924–6925. 43. Wu¨thrich, K. (1986). NMR of Proteins and Nucleic Acids. Wiley, New York. 44. Cornilescu, G., Delaglio, F. & Bax, A. (1999). Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR, 13, 289–302. 45. Schwieters, C. D., Kuszewski, J. J., Tjandra, N. & Marius Clore, G. (2003). The Xplor-NIH NMR molecular structure determination package. J. Magn. Reson. 160, 65–73. 46. Kuszewski, J., Gronenborn, A. M. & Clore, G. M. (1999). Improving the packing and accuracy of NMR structures with a pseudopotential for the radius of gyration. J. Am. Chem. Soc. 121, 2337–2338. 47. Kuszewski, J., Gronenborn, A. M. & Clore, G. M. (1996). Improving the quality of NMR and crystallographic protein structures by means of a conformational database potential derived from structure databases. Protein Sci. 5, 1067–1080. 48. Kuszewski, J., Qin, J., Gronenborn, A. M. & Clore, G. M. (1995). The impact of direct refinement against 13 C alpha and 13C beta chemical shifts on protein structure determination by NMR. J. Magn. Reson. ser. B, 106, 92–96. 49. Christopher, J. A. (1999). SPOCK: The Structural Properties Observation and Calculation Kit. College Station, TX: Center for Macromolecular Design, Texas A&M University.

Edited by M. F. Summers (Received 25 May 2004; received in revised form 8 July 2004; accepted 9 July 2004)