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: 15/7/1326 00-0516fjev1    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 MEYER, M. G. Articles by NIXON, B. T. Search for Related Content PubMed PubMed Citation Articles by MEYER, M. G. Articles by NIXON, B. T.

A dimeric two-component receiver domain inhibits the 54-dependent ATPase in DctD¹

Department of Biochemistry and Molecular Biology, The Pennsylvania State University, University Park, Pennsylvania 16802, USA; and
^* Department of Plant and Microbial Biology, University of California, Berkeley, California 94720-3102, USA

³Correspondence: Biochemistry and Molecular Biology, The Pennsylvania State University, 6 Althouse Lab, University Park, PA 16802, USA. E-mail: btn1{at}psu.edu

SPECIFIC AIM

Two-component signal transduction systems control a variety of behaviors such as adaptation to altered osmolarity, chemotaxis, nutrient acquisition, beneficial and harmful symbioses, and complex development pathways. Here we explore the structural basis of signal transduction in DctD, a two-component response regulator and bacterial enhancer binding protein that is essential for biological nitrogen fixation by Sinorhizobium meliloti.

PRINCIPAL FINDINGS

The DctD protein has three domains: an amino-terminal regulatory domain; a central 54-dependent ATPase domain; and a carboxyl-terminal DNA binding domain. Earlier studies showed that the carboxyl-terminal third of the DctD receiver domain exerts a negative affect on its central ATPase domain. Phosphorylation of the receiver domain overcomes this inhibition via an unknown mechanism.

1. DctD (2–143) fragment is phosphorylatable and shows catalytic autodephosphorylation
Functional studies and sequence alignments define the end of the receiver domain at residue 125 and the beginning of the central domain between residues 145 and 149. Fragment 2–128 was ill-behaved; in contrast, fragment 2–143 Dc + DNL was a highly soluble, stable protein. Like other two-component receiver domains, that of DctD could be phosphorylated by a cognate kinase and by small organic phosphates. At 25°C, the half-life of phosphoryl-DctDNL was 4.4 min in the presence of the kinase fragment and Mg²⁺, 1.6 h in the absence of the kinase but presence of Mg²⁺, and 4.3 h in the absence of both.

2. X-ray structure of DctDNL: a novel receiver domain dimer
The X-ray structure for DctDNL fragment was first solved using MAD. Data for native crystals were collected at room temperature (295 K) using Cu anode radiation and refined using the prior model. Each monomer has five parallel ß-strands forming a single sheet that is surrounded by five -helices, typical of receiver domains. The eight monomers in this I222 unit cell form three different dimers. An alternative crystal form for DctDNL substitution variant D55C (M. G. Meyer, D. Kolbe, and B. T. Nixon, unpublished) and genetic evidence (presented below) identified the biologically relevant dimer. DctD’s phosphorylation site is similar to those present in other receiver structures like FixJN and CheY. Compared with these structures, loops 2-ß3 (residues 45–49) and 3-ß4 (residues 74–77) swing toward an extended helix 5 in DctDNL. Comprised of residues 108–143, helix 5 is > 50 Å long and joins the receiver domain to the output domains.

Dimerization buries 850 Ų of solvent accessible surface. The residues becoming inaccessible have among the lowest B factors of the structure. They come from helix 4, loop 4-ß5, strand 5 and helix 5, including the linker region, which forms a pseudocoiled coil. Sequence analysis suggests that 4.5% of 1148 known two-component receiver domains are flanked by such coiled coils. Interacting in a tails-to-tails orientation, two dimers form a tetramer in the crystal.

3. DctDNL fragment forms stable homodimers
Sedimentation equilibrium was reached after 9 to 12 h of centrifugation at 6000 to 32,000 rpm and 298 K. At 35 µM or higher loading concentrations, distribution of the protein was well modeled as a single dimeric species. Similar results were obtained with protein loaded at concentrations of 1 to 30 µM. However, these data could be modeled as a monomer-dimer system, with a dissociation constant of 0.4 (0.2, 0.7) µM.

4. Genetic and biochemical data for amino acid substitution variants suggest that the dimer is necessary to inhibit the basal activity of the ATPase domain
The entire dctD gene was randomly mutagenized and screened for constitutive alleles whose encoded proteins showed no evidence of proteolysis in Western blots of cell extracts. All 15 of the independently isolated substitutions map to the receiver domain or linker region. Eight substitutions (Y100D, Y100H, D101G, E121K, K122E, N129S, N129H, and L132Q) occur in the dimer interface; five substitutions (P78S, I80N, I103T, L112P; L112P was isolated twice) occur directly beneath it; and two substitutions (V7M and L44P) are far removed. Single substitutions D101K, K122D, and T83I, or the double substitution D101K/K122D, were made and found to give constitutive activity, but substitutions T83S, S54C, D55C, and D55E showed no constitutive activation.

Full-length DctD and its E121K substitution variant were purified. The substitution clearly resulted in elevated ATPase activity, which could nonetheless be activated further upon phosphorylation or binding of BeF_x, which has been shown to mimic phosphorylation of several receiver domains.

Ultracentrifugation data for DctDNL fragments bearing substitutions K122E and E121K at loading concentrations between 1 and 30 µM were fit best by a monomer-dimer model with estimated K_D and 95% confidence intervals of 7.2 (8.2, 6.5) µM and 11.0 (10.4, 11.4) µM, respectively. Fluorescence anisotropy experiments for wild-type and E121K fragments were consistent with these results, which suggest that these constitutive substitutions reduce dimer stability 10- to 20-fold.

5. The structure of substitution variant K122E is essentially identical to wild-type
Substitution K122E gave the highest observed transcription activation activities. The K122E fragment crystallized in the same unit cell and space group as did wild-type fragment. The K122E substitution causes a charge reversal resulting in the loss of two ion pairs, a disruption of the hydrogen bonding network, and different water structure at the dimer interface but no large structural rearrangements. In wild-type, K122 participates in a salt bridge interaction with E121, as well as in an intramolecular interaction with D76, which is an important packing interaction for helix 5 against the rest of the protein.

6. Phosphorylation of DctDNL fragment does not destabilize the dimer
Equilibrium ultracentrifugation experiments were performed at 22,700 rpm, 20°C, on phosphorylated DctDNL fragment. Data were collected at 4 h and every 2 h thereafter for 18 h, with loading concentrations of 13 to 130 µM, which were brought to 67 mM MgCl₂ and 100 mM carbamyl phosphate at time zero. Equilibrium was achieved in 9 h, and the data clearly indicated that the phosphorylated protein is dimeric in the 5 to 200 µM concentration range. These data could also be fit to a monomer-dimer model with essentially the same variance of fit, yielding a K_D of 0.2 (0.05, 0.8) µM.

DISCUSSION AND CONCLUSIONS

These results show that the DctDNL fragment is a good substrate for interaction with its cognate kinase DctB and that it has catalytic autodephosphorylation activity. Furthermore, the unphosphorylated DctD receiver domain and linker form a dimer that is novel among the known two-component receiver domain structures. The 4-ß5-5 face of the receiver domain forms the subunit interface, with an extension of helix 5 through the adjoining linker forming a coiled coil-like structure that appears necessary to support the dimer. We propose that this dimeric structure is needed to maintain the ‘off’ state of the central ATPase domain in DctD (Fig. 1) . This could explain why random substitutions that cause the central domain ATPase to partially escape repression by the receiver domain cluster in the dimer interface and why at least two of these, K122E and E121K, reduce dimer stability in vitro.

View larger version (17K): [in this window] [in a new window]   Figure 1. Diagram of regulation in DctD. The receiver domain/linker dimer is hypothesized to maintain repression of the 54-dependent ATPase domain in the unphosphorylated state. Two mechanisms are illustrated, differing in the presence or absence of a dimeric receiver domain in the activated state. The reported data favor option 2.

A simple, consistent model for activation is that phosphorylation destabilizes the receiver domain dimer. It is expected from studies of other two-component receiver domains (see below) that phosphorylation alters the structure of the dimer interface of the DctD receiver domain. Preliminary ¹H-¹⁵N HSQC spectra confirm that phosphorylation or binding BeF_x alters the backbone environment of DctD residue Y100, which is clearly involved in the dimer interface. However, phosphorylation of the DctD fragment did not reduce dimer stability in vitro. The simple model may thus be incorrect; rather, activation may involve changing an equilibrium between inactive and active, dimeric conformations (option 2 in Fig. 1 ).

The emerging view of two-component signal transduction is that phosphorylation alters the 3-ß4-4-ß5 face of the receiver domain to a greater or lesser extent, depending upon the actual receiver domain being studied. This view arises from work on CheY, NtrC, FixJ, and Spo0A, for which structures of the phosphorylated or BeF_x-bound forms of these proteins or fragments bearing their receiver domains were reported within the last year. For CheY, phosphonomethylation of D57C was found to mimic phosphorylation of D57 and supports some of the structural changes seen upon binding BeF_x. Comparison of these ‘activated’ structures with the ‘inactive’ ones gives us our first views of the structural bases of two-component signal transduction.

Phosphorylation of NtrCN displaces ß -strands 4 and 5, and -helices 3 and 4 away from the active site, and causes a register shift and an axial rotation in helix 4. These changes expose a hydrophobic patch at the amino-terminal end of helix 4, which is hypothesized to activate the central ATPase domain by acting as a contact surface for another part of the protein. Phosphorylation dependent cleavage from Fe-BABE conjugated to the amino-terminal portion of helix 4 occurred in the partner subunit of NtrC in the beginning of its central domain. This evidence supports the hypothesis that helix 4 interacts with the beginning of the central domain for signal propagation in NtrC. Although activation in DctD could occur by a similar mechanism, there are differences between it and NtrC. Unlike DctDN, NtrCN is a stable monomer in the absence of the linker region between it and the ATPase domain, and the linker in NtrC is believed to be a random coil. Removing some or all of the amino terminal domain of NtrC makes proteins that bind to DNA, but these proteins do not hydrolyze ATP or activate transcription unless they also bear substitutions in the central domain. This contrasts starkly with the constitutive activity found in similarly truncated DctD proteins. Substitutions that give rise to constitutively active NtrC proteins (D54E, D86N and or A89T) map in the site of phosphorylation or in the amino-terminal portion of helix 4; none of the substitutions in DctD map to the corresponding residues. Analogous to D54E in NtrC, the D55E substitution in DctD failed to activate DctD in vivo whereas it did activate NtrC. Contrasting results were also seen for substitution T83I: in DctD it activated the ATPase, whereas the corresponding substitution inactivates NtrC (and CheY). The major dimerization element of NtrC is provided by an extra helix preceding those of the helix-turn-helix DNA binding domain; this extra helix is absent in DctD. These observations suggest that the subunit–subunit and interdomain interactions in DctD are different from those found in NtrC.

NMR structure of beryllofluoride-CheY shows T87 has moved to permit hydrogen bonding between its hydroxyl group and the phosphate analog, creating a hydrophobic cavity that is filled by inward rotation of the Y106 side chain. This motion was anticipated by structural and phenotypic studies of amino acid substitution variants of CheY. The two residues are highly conserved in two-component receiver domains, and the same reorganization of them occurs upon phosphorylation of FixJN and Spo0AN. Only the Y106 sidechain moves in phosphonomethylated D57C CheY, which nonetheless shows evidence of being activated by the chemical modification. Similar movements permit a dimerization interface to develop between -helix 4 and ß-strand 5 in FixJN, and dimerization of FixJ is correlated with increased DNA binding and transcription activation by its associated output function.

The orientation of these conserved residues—D55, T83, and F102 in DctDNL—is very similar to that found in CheY and FixJN and less so to that of NtrCN. Substitution T83I was made and found to strongly activate the DctD protein, but substitution T83S did not. It thus appears that structural rearrangements arising from changes in the hydrophobic pocket occupied by T83 could be involved in activating DctD. The dimer interface in FixJN-P is quite distinct from that seen in DctDNL: partner subunits have to be moved 18 Å from one structure to position them as in the other. It remains to be seen whether the FixJN-P dimer or those of Spo0AN or ETR1N are relevant to the dimeric structure(s) of phosphorylated DctDNL.

We hypothesize that an extended, coiled-coil helix 5 is involved in signal transduction in 54 of the 1148 known two-component receiver domains. The FixJ protein of S. meliloti is among those strongly predicted to have a coiled-coil linker following the receiver domain. This receiver domain has been studied as an isolated fragment. The reported dimer for its phosphorylated form is incompatible with the dimer predicted to contain the coiled-coil linker region. Either the prediction is wrong or structural studies with FixJ that include the linker region might reveal new information. Also, the predicted coiled-coil linkers do not appear in all members of a given output class of two-component signal transduction proteins, or even in genetic homologs. For example, NtrC is not predicted to be like DctD and the coiled-coil linker predicted for FixJ of S. meliloti is not predicted for FixJ proteins of other species. If these predictions prove to be correct, they will dramatically underscore the diversity among structural mechanisms that link common changes in phosphorylated receiver domains with changes in the functional state of response regulator output domains.

FOOTNOTES

¹ To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.00-0516fje ; to cite this article, use FASEB J. (March 20, 2001) 10.1096/fj.00-0516fje

² Present address: National Institutes of Health, National Center for Research Resources, Bethesda, MD, USA.

This article has been cited by other articles:

				S. N. Joslin and D. R. Hendrixson Analysis of the Campylobacter jejuni FlgR Response Regulator Suggests Integration of Diverse Mechanisms To Activate an NtrC-Like Protein J. Bacteriol., April 1, 2008; 190(7): 2422 - 2433. [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]

				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: 15/7/1326 00-0516fjev1    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 MEYER, M. G. Articles by NIXON, B. T. Search for Related Content PubMed PubMed Citation Articles by MEYER, M. G. Articles by NIXON, B. T.