HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text (PDF) All Versions of this Article: 16/14/1964 02-0395fjev1    most recent Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by PARK, S. Articles by NIXON, B. T. Search for Related Content PubMed PubMed Citation Articles by PARK, S. Articles by NIXON, B. T.

Two-component signaling in the AAA + ATPase DctD: binding Mg²⁺ and BeF₃^- selects between alternate dimeric states of the receiver domain ¹

Department of Biochemistry and Molecular Biology, and
^* Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania, USA

²Correspondence: Department of Biochemistry and Molecular Biology, 6 Althouse Laboratory, University Park, PA 16802, USA. E-mail: btn1{at}psu.edu

SPECIFIC AIM

The DctD protein of Sinorhizobium meliloti DctD is a ⁵⁴- or ^N-dependent enhancer binding protein (EBP) essential for symbiotic biological nitrogen fixation. Its central AAA+ ATPase domain is regulated by phosphorylation of an amino-terminal, two-component receiver domain. A novel dimeric receiver domain structure has been described. Its subunit interface consists of contacts along the 4–ß5–5 surface and includes contributions from a pseudo coiled coil made by extending helix 5 well into the adjacent linker that joins the receiver and ATPase domains. The integrity of this dimer interface was seen to be essential for inhibiting oligomeric assembly of the AAA+ ATPase into an active form, thus defining the "off-state" of the protein. Moreover, analytical ultracentrifugation data showed that phosphorylation altered the dimeric state but did not destabilize it. The aim of the present study was to compare the "activated" structure of the S. meliloti DctD receiver domain (N) and linker (L; hereafter this protein is called "DctDNL") with its "inactive" structure to reveal the molecular basis of two-component signal transduction in this EBP.

PRINCIPAL FINDINGS

Attempts to phosphorylate pregrown crystals of the unmodified DctDNL protein revealed that the dimeric structure of the inactive state is refractory to phosphorylation. Attempts to crystallize phosphorylated protein were consistent with this observation. When Mg²⁺ was bound to EDTA and removed by gel filtration chromatography, the half-time for dephosphorylation was increased from 1.6 to 4 h for the wild-type protein. This was not long enough to grow diffracting crystals. However, the half-time for dephosphorylation of DctDNL bearing the amino acid substitution E121K was 72 h, and a condition was identified that gave high quality crystals in 18 h. The structure of the protein in those crystals was essentially identical to that of the unphosphorylated protein; in fact, mass spectrometry indicated there was no phosphorylated protein in the crystals. Soaking pregrown crystals of wild-type or E121K DctDNL in Mg²⁺ and BeF_x also failed to reveal any changes and no electron density was present for metal or BeF₃^-.

More revealing results were obtained when crystals were grown in the presence of Mg²⁺ and BeF_x. Although no crystals for the wild-type protein were obtained in the presence of BeF_x, some were obtained for the E121K substitution. Since the crystallization conditions were very different from those previously used to determine the structure of the inactive state, control crystals were grown in the absence of the ligands but otherwise in the same mother liquor. To achieve this, the pH had to be raised from 7.0 to 7.5 and PEG 8K concentration had to be dropped from 19% to 5%. Most of the crystals grown in the presence of Mg²⁺ and BeF_x diffracted to only 6 Å resolution, but one that had grown in the P2₁2₁2₁ space group at a rarely formed interface between two phases that had separated in the hanging drop diffracted to 2.1 Å resolution. There are two molecules of the Mg²⁺-BeF₃^--bound DctD(E121K) protein in the asymmetric unit. A simulated annealing omit map calculated after removing a 5 Å sphere surrounding Mg²⁺ and BeF₃^- in the active sites of the models clearly shows density for the ligands in both structures. The trajectories of loops ß4–4 differ somewhat between the two structures, and from residue 136 to the carboxyl terminus the two structures are quite different; otherwise the phi and psi angles are similar, even though noncrystallographic symmetry was not used during refinement.

Compared with the inactive state structure, the activated one displays changes similar to those seen for the pairs Spo0AN and Ca²⁺-Spo0AN-P, FixJN and FixJN-P, CheY and Mn²⁺-BeF₃^--CheY, or Mg²⁺-BeF₃^--CheY-FliM(N). However, dramatic differences were observed for the packing of monomers in the activated DctDNL(E121K) crystal lattice vs. the one of inactive DctDNL. Four distinct protein–protein interfaces were present. One involves the 4–ß5 signaling surface and can be superimposed with the dimer interface between FixJN-P molecules (0.56 Å RMSD); the other three are tentatively believed to result from crystal packing forces.

CONCLUSIONS AND SIGNIFICANCE

Although BeF₃^- might not activate all two component receiver domains, this study supports the growing consensus that it mimics phosphorylation and triggers a molecular switch common among two-component receiver domains. Two novel insights revealed by the DctD studies are that 1) a specific dimeric structure of a receiver domain can maintain the highly conserved active site in a state that is refractory to phosphorylation; and 2) maintaining the active site in its activated conformation can stabilize a dramatically different dimeric state of the receiver domain, providing a molecular mechanism for regulating the function of an associated output domain, in this case a AAA+ ATPase. Comparing these results with those reported for NtrC, another EBP that shares 38% sequence identity with DctD, indicates that diverse signaling mechanisms can be present even among two-component response regulators that belong to a single subfamily.

For DctD, we hypothesize that the common two component signaling switch toggles between distinct dimeric states of the receiver domain (Fig. 1 ) to regulate the assembly of a AAA+ ATPase domain used to stimulate ⁵⁴-dependent transcription at the dctA promoter. In this hypothesis, the inactive state is a stable dimer in which a strong dimerization determinant resides in the receiver domain. This dimeric conformation of the receiver domain is refractory to phosphorylation and able to inhibit further oligomerization of the ATPase. A rarer dimeric conformation of the receiver domain or an intermediate that occurs in monomers is then stabilized by phosphorylation or BeF₃^- binding in the presence of Mg²⁺. The stabilized, active conformation of the receiver domain is predicted to fail to inhibit assembly of the ATPase (it may also stimulate assembly of the oligomer, but no evidence supports this notion). Interaction with the cognate kinase DctB may overcome the refractory state of the off-state receiver domain before phosphoryl-transfer.

View larger version (44K): [in this window] [in a new window]   Figure 1. Overlay of "off" and "on" state dimers of the DctDNL(E121K) receiver domain. One monomer of the off-state dimer (white) has been superimposed with one monomer of the Mg2+-BeF3--bound dimer (red). C atoms of the active site residue Asp55 are displaced by 21.7 Å in the unaligned partner molecules. To superimpose these monomers, one would have to apply a translation vector of 25.1 Å, 12.8 Å, and -18.5 Å and accommodate an angle between the rotation axis and centroid vector of 47.2°.

It is not known how the receiver domain, linker, and AAA+ ATPase domain physically interact in either the unphosphorylated or phosphorylated states of DctD, but the dramatic differences in the two dimeric conformations of the receiver domain and linker must somehow be exploited to suppress ATPase oligomer formation on the one hand, then permit or enable it on the other. It is therefore interesting to contrast known features of two-component signaling in DctD and NtrC (Fig. 2 ). Both proteins are members of the AAA+ subfamily of ATPases, both use two-component receiver domains to control the ATPase activity, and both use helix-turn-helix DNA binding domains to target their respective genes. DctD and NtrC proteins are dimers in their unphosphorylated forms and probably form hexamers or dodecamers on phosphorylation. [Proteins in the AAA+ ATPase family tend to form hexamers or dodecamers; earlier reports clearly show that oligomerization follows activation of NtrC. Recent cryo EM data are consistent with a hexamer or dodecamers for the EBP PspF (M. Buck, X. Zhang, and J. Schumacher, personal communication).] Given these similarities, one might expect similar signal transduction mechanisms in DctD and NtrC. However, important differences appear to reflect divergent strategies for using two-component signal transduction in these proteins to modulate the equilibrium between dimeric and hexameric AAA+ ATPase domains. NtrC depends on an FIS-like dimerization determinant in its carboxyl-terminal DNA binding domain to maintain the dimeric, unphosphorylated state. A similar element is not present in DctD, which instead has a dimeric receiver domain that starkly contrasts with the monomeric one of NtrC. Instead of an intrinsically active ATPase domain like that present in DctD, the central domain of NtrC is inert and unable to form an ATPase activity unless assisted by its receiver domain or altered by amino acid substitutions. In NtrC, the receiver domain thus appears to act passively in the unphosphorylated state, but on phosphorylation it must facilitate activation of the ATPase. In DctD, the receiver domain actively represses the ATPase domain until the former is phosphorylated.

View larger version (35K): [in this window] [in a new window]   Figure 2. Contrasting models for two-component regulation of AAA+ ATPase assembly in DctD and NtrC. A dimeric off-state is maintained by an amino-terminal receiver domain and linker element in DctD and a carboxyl-terminal FIS-like element in NtrC. The dimeric structure of the DctD receiver domain and linker actively represses assembly of a competent ATPase domain. The monomeric receiver domain of NtrC is passive, not needing to inhibit an incompetent ATPase domain. Phosphorylation of the receiver domain remodels loop ß4–4 through strand ß5 differently, creating an open signaling surface in DctD and a closed one in NtrC (indicated surfaces are colored by charge and are drawn with and without embedded backbone and selected side chains for residues 83–104 of DctD and residues 82–103 of NtrC; helix 4, red, is to the left of strand ß5). Stabilizing an alternate dimer of the receiver domain of DctD fails to inhibit assembly of an ATPase hexamer (option 1). On the other hand, activation facilitates a positive interaction between monomeric receiver and ATPase domains in NtrC, stabilizing hexamer assembly (option 2). An alternative possibility that a positive interaction exists between the activated receiver and ATPase domains of DctD is also depicted (option 3).

The question remains: in the activated oligomer, are the interactions between receiver and ATPase domains similar or different in NtrC and DctD? The hydrophobic side chains of the upper portion of helix 4 that are believed to contact the AAA+ ATPase domain in NtrC are engaged in a dimer interface in Mg²⁺-BeF₃^--DctDNL, suggesting that different modes of contact must occur in the activated proteins. Indeed, contact may not occur between the domains in DctD, where derepression would be sufficient to activate the ATPase. If contact does occur, the twofold symmetry of the DctDNL dimer and sixfold symmetry of a hexameric AAA+ ATPase ring dictate that each monomer of the receiver domain dimer must have a different spatial relationship with the ring. This need not be the case for the monomeric receiver domain of NtrC, which could be packed against the ring using its sixfold symmetry.

Significant differences exist in the signaling surfaces of the DctD and NtrC receiver domains (Fig. 2) that may help explain how triggering the common two-component switch can variously regulate the assembly of a AAA+ ATPase domain. In DctDNL, signaling appears to create an open cleft and three hydrophobic projections (Ile88, Val92, and Ile95) used to support a homodimer of the activated receiver domain by rotating the monomers 180° and inserting the projections into the cleft, forming two parallel stacks of three hydrophobic side chains. Ile95 of DctD is replaced with Tyr94 in NtrC, and remodeling of its signaling surface arranges the side chains of residues Ala83, Leu87, and Tyr94 to fill the corresponding cleft in a way that would appear to prevent homodimerization but provide a surface for stabilizing the fully assembled ATPase. The different remodeling of the signaling surfaces appears to depend at least in part on substitution of Gly84 and Leu95 of DctD with Ala83 and Tyr94 in NtrC (Fig. 2) . Nearly 16% of the 1556 receiver domains present in the PFAM database (including that of FixJ) have glycine in the first of these two positions. Like DctD and FixJ, most of these receiver domains (80%) have a hydrophobic residue in the second position, with none having a tyrosine as found in NtrC. Glycine at the base of loop ß4–4 may thus serve to open the hydrophobic signaling surface in some receiver domains to facilitate homo-dimerization in the manner seen for the DctD and FixJ proteins of S. meliloti.

FOOTNOTES

¹ To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.02-0395fje; to cite this article, use FASEB J. (October 4, 2002) 10.1096/fj.02-0395fje

This article has been cited by other articles:

				J. G. Malone, R. Williams, M. Christen, U. Jenal, A. J. Spiers, and P. B. Rainey The structure-function relationship of WspR, a Pseudomonas fluorescens response regulator with a GGDEF output domain Microbiology, April 1, 2007; 153(4): 980 - 994. [Abstract] [Full Text] [PDF]

				R. Gao, A. Mukhopadhyay, F. Fang, and D. G. Lynn Constitutive Activation of Two-Component Response Regulators: Characterization of VirG Activation in Agrobacterium tumefaciens. J. Bacteriol., July 1, 2006; 188(14): 5204 - 5211. [Abstract] [Full Text] [PDF]

				S. De Carlo, B. Chen, T. R. Hoover, E. Kondrashkina, E. Nogales, and B. T. Nixon The structural basis for regulated assembly and function of the transcriptional activator NtrC Genes & Dev., June 1, 2006; 20(11): 1485 - 1495. [Abstract] [Full Text] [PDF]

				N. Sal-Man, D. Gerber, and Y. Shai The Identification of a Minimal Dimerization Motif QXXS That Enables Homo- and Hetero-association of Transmembrane Helices in Vivo J. Biol. Chem., July 22, 2005; 280(29): 27449 - 27457. [Abstract] [Full Text] [PDF]

				X. F. Yang, Y. Ji, B. L. Schneider, and L. Reitzer Phosphorylation-independent Dimer-Dimer Interactions by the Enhancer-binding Activator NtrC of Escherichia coli: A THIRD FUNCTION FOR THE C-TERMINAL DOMAIN J. Biol. Chem., August 27, 2004; 279(35): 36708 - 36714. [Abstract] [Full Text] [PDF]

				S.-Y. Lee, A. De La Torre, D. Yan, S. Kustu, B. T. Nixon, and D. E. Wemmer Regulation of the transcriptional activator NtrC1: structural studies of the regulatory and AAA+ ATPase domains Genes & Dev., October 15, 2003; 17(20): 2552 - 2563. [Abstract] [Full Text] [PDF]

This Article Full Text (PDF) All Versions of this Article: 16/14/1964 02-0395fjev1    most recent Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by PARK, S. Articles by NIXON, B. T. Search for Related Content PubMed PubMed Citation Articles by PARK, S. Articles by NIXON, B. T.