JOURNAL OF BACTERIOLOGY, Jan. 2002, p. 525–530 0021-9193/02/$04.00⫹0 DOI: 10.1128/JB.184.2.525–530.2002 Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Vol. 184, No. 2
Luteolin and GroESL Modulate In Vitro Activity of NodD Kuo-Chen Yeh,1† Melicent C. Peck,2 and Sharon R. Long1,2* Howard Hughes Medical Institute1 and Department of Biological Sciences,2 Stanford University, Stanford, California 94305-5020 Received 30 May 2001/Accepted 9 October 2001
In the early stages of symbiosis between the soil bacterium Sinorhizobium meliloti and its leguminous host plant, alfalfa, bacterial nodulation (nod) genes are controlled by NodD1, NodD2, and NodD3, members of the LysR family of transcriptional regulators, in response to flavonoid and other inducers released by alfalfa. To gain an understanding of the biochemical aspects of this action, epitope-tagged recombinant NodD1 and NodD3 were overexpressed in Escherichia coli. The DNA binding properties of the purified recombinant NodD proteins were indistinguishable from those of NodD isolated from S. meliloti. In addition, the E. coli GroEL chaperonin copurified with the recombinant NodD proteins. In this study, we showed that NodD proteins are in vitro substrates of the GroESL chaperonin system and that their DNA binding activity is modulated by GroESL. This confirmed the earlier genetic implication that the GroESL chaperonin system is essential for the function of these regulators. Increased DNA binding activity by NodD1 in the presence of luteolin confirmed that NodD1 is involved in recognizing the plant signal during the early stages of symbiosis. (11, 12). While in vivo studies using reporter fusions have shown that NodD1-dependent activation of nod gene expression requires flavonoid inducers, this has not been demonstrated in vitro, nor has direct binding of flavonoids to NodD protein been shown. Thus, the mechanism of NodD-mediated nod gene expression remains unclear. The GroESL chaperonin system, functioning as a tetradecamer of GroEL and a heptamer of GroES, forms a protein complex that associates with approximately 10% of Escherichia coli cytoplasmic proteins and facilitates folding and unfolding of proteins to their native forms in the presence of Mg2⫹ and ATP (7, 8, 21). In previous studies, we found that GroESL copurified with NodD1 and NodD3 from S. meliloti (10, 29). Genetic screening for other components involved in NodDmediated nod gene expression in S. meliloti revealed that a chromosomal copy of groESL is required (29). In addition, the DNA binding activity of NodD in a cell extract decreases in a groESL mutant background (29). Both genetic and biochemical results strongly suggest that the GroESL chaperonin system is involved in the folding and/or assembly of active NodD proteins to facilitate the regulation of nod gene expression. In this study, we investigated the biochemical properties of recombinant NodD1 and NodD3, including the effect of luteolin on the DNA binding activity of NodD1, and the involvement of GroESL in NodD protein folding.
Rhizobia establish nitrogen-fixing root nodules with their symbiotic-partner legumes. The symbiosis is initiated via chemical communication between bacteria and plants (24). In general, bacterial nod genes are induced by flavonoids and other compounds secreted by their host plants. The nod gene products participate in the production of Nod factor, which is responsible for the induction of early responses in host plants (5, 24). Through the exchange and recognition of molecular signals, host specificity is determined. Previous genetic studies revealed that expression of these nod genes requires NodD (28), a member of the LysR family of transcriptional activators (35). In Sinorhizobium meliloti, the symbiotic partner of alfalfa (Medicago sativa), there are three NodD proteins, designated NodD1, NodD2, and NodD3 (14, 19, 27, 32). NodD1 and NodD2 activate a nodC-lacZ reporter fusion in the presence of plant inducers, while NodD3 activates the same fusion in the absence of any inducer (27, 30, 31). A number of genetic and biochemical studies suggest that NodD functions as a sensor and/or receptor for plant signals. Using bacterial nod-lacZ fusions as reporters to monitor activity, NodD from different species exhibits sensitivity to specific flavonoid inducers (15, 17, 20, 38). Chimeric protein constructs and mutational studies also suggest the existence of at least two domains: a common one involved in DNA binding and a second one that specifies the spectrum of flavonoid inducers (2, 3, 26, 37). In vitro biochemical studies showed that extracts from bacterial strains expressing NodD bound to a consensus sequence in the promoter region of nod genes that was designated as the nod box (10, 11, 18). Immunoaffinity-purified NodD1 and ion-exchange-purified NodD3 were shown to protect the nod box from nuclease cleavage in footprinting assays
MATERIALS AND METHODS Plasmid construction for expression of NodD1 and NodD3. Two primers, NodD1:S1-XbaI (5⬘-GCTCTAGATAACGAGGGCAAAAAATGCGTTAGGG GCCTAG-3⬘) and NodD1:AS1-SalI (5⬘-CGGTCGACCACAGGTGTCCGACT AGG-3⬘), were designed to contain unique XbaI and SalI sites, respectively, for subsequent cloning of PCR products. nodD1 was amplified by PCR using cloned Pfu DNA polymerase (Stratagene) and pRmE39, which contains the entire nodD1 coding sequence (6). Primer NodD1:AS1-SalI was designed to remove the C-terminal stop codon to create an in-frame fusion of NodD1 to the Streptag (Biometra) (36), termed NodD1-ST. pNodD1-ST was constructed by cloning this XbaI-SalI PCR product into pASK75B after digestion with XbaI and SalI (39). Similarly, the coding region of nodD3 was amplified by PCR using pBGR2 (33) and primers NodD3:S1-XbaI (5⬘-GCTCTAGATAACGAGGGCAAAAAA TGCGTCAAAGGTCTTG-3⬘) and NodD3:AS1-SalI (5⬘-CGGTCGACCACGC CCAAGACACCCTGCG-3⬘). After digestion with XbaI and SalI, the PCR
* Corresponding author. Mailing address: Department of Biological Sciences, Gilbert Lab, 371 Serra Mall, Stanford University, Stanford, CA 94305-5020. Phone: (650) 723-3232. Fax: (650) 725-8309. E-mail: [email protected]
† Present address: Institute of BioAgricultural Sciences, Academia Sinica, Nankang, Taipei, Taiwan 11529, Republic of China. 525
YEH ET AL.
product was cloned into pASK75B to create pNodD3-ST, which contains an in-frame fusion of NodD3 to the Strep-tag (NodD3-ST). Expression of recombinant NodD1 and NodD3 in E. coli. For coexpression with GroESL, 30-ml overnight cultures of E. coli DH5␣ harboring either pNodD1-ST and pGroESL or pNodD3-ST and pGroESL (13) were used to inoculate 1 liter of L-broth medium supplemented with 100 g of ampicillin/ml and 50 g of chloramphenicol/ml to select for both plasmids. Cultures were incubated at 37°C with shaking until they reached an optical density at 550 nm of 0.5. Isopropyl-␤-D-thiogalactopyranoside (IPTG) was added (final concentration, 0.1 mM) to induce GroESL expression. To induce recombinant NodD expression, anhydrotetracycline (final concentration, 0.2 g/ml) was added 15 min after IPTG and the cultures were shaken for 3 h at 37°C. Cell pellets were collected by centrifugation, washed with wash buffer (20 mM Tris-HCl [pH 8.0], 20 mM NaCl, 0.05% NP-40), and stored at ⫺80°C. Cells were thawed, resuspended in 15 ml of lysis buffer (50 mM Tris-HCl [pH 8.0], 100 mM NaCl, 0.05% NP-40, 2.0 g of leupeptin/ml, 2 mM benzamidine, 1 mM dithiothreitol [DTT], 2 mM phenylmethylsulfonyl fluoride, 3 g of pepstatin A/ml), and lysed by passage through a BioNeb cell disruptor (Glas-Col) three times and followed by sonication (three times, 12 s each, with intervening 1-min pauses). The soluble-protein fraction was obtained by centrifugation at 20,000 ⫻ g for 40 min. Soluble proteins were further fractionated by adding 0.3 g of ultrapure ammonium sulfate (ICN) per ml of extract, incubating on ice at least 1 h, and centrifuging at 17,000 ⫻ g and 4°C for 30 min. The resulting protein precipitate was redissolved in 8 ml of buffer W (100 mM Tris-HCl [pH 8.0], 1 mM EDTA, 1 mM DTT) supplemented with 1 mM phenymethylsulfonyl fluoride and clarified by centrifugation for 15 min at 38,000 ⫻ g. Affinity purification of NodD1-ST and NodD3-ST. The recombinant proteins were purified from 1-liter volumes of E. coli cultures by using 2.5- to 3-ml (bed volume) streptavidin-agarose (Sigma) affinity columns as described previously (36, 39). The columns were washed with 5 to 10 column volumes of buffer W prior to use. Avidin (40 g/ml) was added to ammonium sulfate-fractionated protein for 30 min on ice to block biotinylated proteins. This protein solution was applied to the streptavidin-agarose column, and the column was washed with 12 to 15 column volumes of buffer W to remove nonspecifically bound proteins. Affinity-purified protein was eluted with 5 bed volumes of buffer E (100 mM Tris-HCl [pH 8.0], 1 mM EDTA, 3 mM diaminobiotin). Most of the protein was eluted at 3 to 9 ml, and it was concentrated and exchanged with storage buffer (25 mM Tris-HCl [pH 7.5], 10% glycerol, 1 mM EDTA, 1 mM DTT, 100 mM KCl) by using an Ultrafree-15 filter unit (30K NMWL membrane, Millipore). Samples were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 12.5% gels, using the Laemmli buffer system (23). Protein sequencing was carried out by Edman degradation at the Protein and Nucleic Acid facility at Stanford University. SEC-HPLC analysis and sample collection.Size exclusion chromatography (SEC)–high-performance liquid chromatography (HPLC) separations were performed using a Dionex DX500 system equipped with a gradient pump and a variable-wavelength UV-visible light detector interfaced with the Dionex PeakNet program. The affinity-purified NodD1 preparation was injected into a 4.6- by 250-mm Zorbax Bio Series GF-250 column (Hewlett-Packard) equilibrated in the mobile-phase buffer (50 mM Tris-HCl [pH 7.5], 200 mM NaCl, 0.5 mM EDTA, 0.5 mM DTT). Molecular size calibration was performed using Bio-Rad gel filtration standards (catalog no. 151-1901). The desired peaks were collected and concentrated using a Millipore Ultrafree-MC (polysulfone, 30K NMWL membrane) centrifugal filter concentrator. GroESL incubation and luteolin treatment. In our standard assay, 5 l of affinity- or SEC-HPLC-purified NodD protein (1 M purified recombinant NodD protein and 5 to 10 M associated GroEL) was incubated with the GroESL cycling system (17 M GroES, 10 mM KCl, 5 mM MgCl2, 2 mM ATP) in a 10-l reaction volume for various periods of time at room temperature. Luteolin was dissolved in N,N⬘-dimethyl formamide to make a concentrated stock solution and stored at ⫺20°C. This concentrated stock solution was diluted in TED100 (50 mM Tris-HCl [pH 8.0], 0.5 mM EDTA, 0.5 mM DTT, 100 mM NaCl) to make a working solution with a luteolin concentration that was 10-fold higher than the final concentration used in a given experiment. An appropriate volume of this working solution was added to samples for 2 to 5 min at room temperature. DNA-protein complex electrophoretic mobility shift assay (EMSA). The ability of purified recombinant NodD proteins to bind to the nodF nod box was monitored by measuring the specific shift in mobility during gel electrophoresis. Treated NodD samples (2 l) were incubated at room temperature for 10 min with the end-labeled nodF nod box DNA fragment in binding buffer as described previously (10). The gels were dried, and the migration of DNA fragments was
J. BACTERIOL. visualized and/or quantitated by autoradiography and/or phosphorimaging (BioRad GS-363 system).
RESULTS Recombinant NodD proteins bind the nodF nod box. For purposes of clarity, we chose to compare and contrast two proteins: NodD1, which is the best-studied example of a flavonoid-dependent transcriptional activator, and NodD3, the best-studied inducer-independent form of this protein family. Previous biochemical and genetic studies showed that a functional GroESL chaperonin system is required for NodD to activate transcription at nod gene promoters (29). We asked whether GroEL and GroES are required for in vitro activity of purified NodD proteins. To do this, we purified NodD from an E. coli strain that was also overexpressing E. coli GroES and GroEL. We hypothesized that the elevated levels of GroESL would increase the solubility of the recombinant NodD, resulting in the production of active NodD proteins. Both recombinant NodD1 and NodD3 were expressed at their expected molecular sizes (33 and 35 kDa, respectively) (Fig. 1A). The purification yield was approximately twofold higher for both NodD proteins when coexpressed with GroESL (data not shown). A 60-kDa E. coli protein copurified with both NodD1 and NodD3 (Fig. 1A). We determined the amino-terminal sequence of this protein and confirmed that it was the molecular chaperonin GroEL (data not shown). Using amino-terminal sequencing, we also identified a 70-kDa protein in the affinity-purified NodD3 preparation as the heat shock protein DnaK (data not shown). We failed to detect significant amounts of GroES by Coomassie blue staining (Fig. 1A). Trace amounts of DnaK were detected in the NodD1 preparation. Taken together with results of our previous genetic studies (29), these data suggest that GroES, GroEL, and DnaK directly interact with both NodD1 and NodD3. To examine whether the affinity-purified recombinant NodD proteins were active, we measured DNA binding by EMSA. Affinity-purified recombinant NodD1 and NodD3 could bind the nodF nod box (Fig. 1B). Formation of the NodD-nod box complex was previously shown only with NodD isolated from S. meliloti cells (10, 11, 29). GroESL modulates the DNA binding activity of NodD1. A direct interaction between GroESL and NodD implicates the chaperonin system in the potential regulation of NodD. In other systems, addition of GroESL, K⫹, Mg2⫹, and ATP to unfolded proteins has shown that the protein can refold through a series of cycles of dissociations of polypeptide ligands from preformed complexes (24a). To address this possibility here, a GroESL cycling system, as described in Materials and Methods, was incubated with purified NodD prior to performance of the DNA binding assay. We predicted that by activating the chaperonin system in vitro, some of the NodD that was in a form that was not competent to bind DNA would be correctly folded into an active form and thus made available for DNA binding. As predicted, preincubation of the chaperonin cycling system with NodD3 resulted in a 60% increase in its ability to bind the regulatory nod box sequence (Fig. 2B). To our surprise, however, incubation with the chaperonin cycling system decreased the DNA binding activity of NodD1 (Fig. 2A). This decrease in the binding activity of NodD1 was not due to a lowering of NodD1 protein levels, as determined by examination of Coomassie-stained SDS gels (data not shown).
VOL. 184, 2002
MODULATION OF NodD ACTIVITY BY LUTEOLIN AND GroESL
FIG. 1. Affinity purification and DNA binding activity of recombinant NodD1 and NodD3. (A) Strep-tag-purified protein preparations were resolved by SDS-PAGE and stained with Coomassie blue R250. Molecular size standards (in kilodaltons) are indicated (MW). (B) EMSA was performed by mixing purified recombinant NodD1 (100 ng) or NodD3 (200 ng) with an end-labeled nodF nod box DNA fragment as described in Materials and Methods. The negative (⫺) and positive (⫹) controls show the migration of the DNA fragment in the absence or presence of excess S. meliloti NodD3 (12), respectively. The arrow indicates the single electrophoretically retarded complex.
Luteolin stimulates the DNA binding activity of NodD1. We were curious about whether the different effects of GroESL on NodD1 versus NodD3 reflected intrinsic differences between the two proteins. Because NodD1, but not NodD3, requires plant inducers to activate nod gene expression, it is plausible that the differential effects of GroESL on the two transcriptional activators are a reflection of this requirement. Indeed, we found that the flavone luteolin, a natural ligand of NodD1, actually stimulates the DNA binding activity of NodD1, and this stimulation depends on GroESL (Fig. 3A, middle panel). These results are in contrast to those shown in Fig. 2A, which demonstrated that Strep-tag-purified NodD1 that had not been treated with GroESL could still bind to the nod box; Fig. 3A (left panel) shows that this DNA binding activity is not affected by luteolin. We hypothesize that the addition of GroES and ATP to complete the chaperonin cycling system facilitated correct folding of NodD1 with its ligand, luteolin, resulting in an increase in DNA binding activity (Fig. 3A, middle panel). In contrast, luteolin had no effect on the DNA binding activity of NodD3 that had been treated with GroESL (Fig. 3A, right panel). We were also curious as to whether the effect of luteolin on NodD1 was dependent on the concentration of luteolin. The DNA binding activity increased up to 10-fold with increasing luteolin concentrations, peaking at 10 to 100 M luteolin (Fig. 3B). At luteolin concentrations higher than 100 M, the DNA binding activity of NodD1 decreased. This dose-response effect strongly suggested that the formation of a complex between the nodF nod box and NodD1 in the presence of GroESL is modulated by luteolin. NodD1 activity requires the GroESL chaperonin system. As seen in Fig. 2A, Strep-tag-purified NodD1 that had not been treated with the in vitro chaperonin cycling system could still bind the nod box regulatory sequence. Furthermore, the DNA binding activity was not affected by luteolin (Fig. 3A, left pan-
el), but this activity differed from one preparation to another (data not shown). It is possible that this variability reflected differences in proportions of active and inactive NodD1 in the various preparations. Treatment of purified NodD1 with the chaperonin cycling system reduced nod box binding activity, but binding activity was modulated by luteolin in a GroESLdependent manner (Fig. 3A, middle panel). To further probe the role of the chaperonin GroESL in producing functional NodD1, we used SEC-HPLC to separate NodD1 complexed with GroEL from other forms of NodD1 in our affinity-purified preparation. As shown in Fig. 4A and B, SEC-HPLC allowed us to isolate a high-molecular-mass (⬎850-kDa) GroEL-NodD1 complex. This fraction failed to bind the nod box regulatory sequence unless luteolin, GroES, and ATP were added (Fig. 4C). This result clearly demonstrated the direct involvement of the GroESL chaperonin system in the formation of active NodD1. Apparently, NodD1 requires a fully functional GroESL system to develop its ligand-mediated DNA recognition specificity. We also used SEC-HPLC to isolate a similar GroEL-NodD3 complex. In contrast to the results we saw with NodD1, restoration of nod box binding activity to NodD3 did not require luteolin (Fig. 4D). These in vitro data are fully consistent with the previously observed differences between NodD1 and NodD3 with regard to inducer dependence (27). DISCUSSION Previous studies have shown that nodD is required for expression of nod genes (28). NodD1 activates nod gene expression in response to luteolin and 4,4⬘-hydroxy-6-methoxychalcone, while NodD3 is inducer independent (25, 27). In addition, genetic and biochemical studies have shown that NodD activity is also dependent on the GroESL chaperonin system (29). In this study, we established that recombinant tagged forms of NodD1 and NodD3 purified from E. coli bind
YEH ET AL.
FIG. 2. In vitro GroESL chaperonin treatment affects the DNA binding activity of NodD proteins. Purified recombinant NodD1 (A) and NodD3 (B) were preincubated with the GroESL chaperonin cycling system for the indicated time periods prior to EMSA. The samples were then mixed with labeled nod box DNA fragment and incubated an additional 10 min at room temperature before being loaded onto native polyacrylamide gels. Note that chaperonin-dependent cycling is ongoing throughout this second incubation period before electrophoresis commences. As a control, DNA binding activity in the absence of the chaperonin cycling system was measured (⫺). Autoradiographs were scanned, and the relative amount of shifted complex normalized to the control is indicated at the bottom of each lane.
to the nod box. Coexpression of GroESL with NodD1 or NodD3 in E. coli enhances protein yield. Here we demonstrated in a purified system that GroESL is required for NodD1 to respond to its inducer, luteolin (Fig. 3A and 4C); this interaction results in an increase in DNA binding activity. In contrast, the DNA binding activity of a NodD3-GroESL complex is not stimulated by luteolin (Fig. 3A and 4D). Our in vitro results are consistent with previous in vivo assays of reporter gene expression and mRNA abundance (27), which showed a dependence of NodD1 on flavonoids for induction, in contrast to the constitutive activity of NodD3. Since the discovery of the NodD and flavonoid regulatory circuit, the exact role of the flavonoid has been uncertain, and direct specific binding of NodD to flavonoids has not been demonstrated in vitro. It is interesting that other LysR regulatory proteins have been similarly elusive in demonstrating direct binding to coinducers. Here, in a purified system, we showed that luteolin has a specific effect on NodD1. This provides the first in vitro evidence for a luteolin-mediated effect that relates to NodD1 function. Another dimension of flavonoid signaling, as yet unex-
FIG. 3. Luteolin-stimulated DNA binding activity of NodD1 is dependent on GroESL treatment. (A) EMSA was performed with preparations of NodD1 (left and middle panels) and NodD3 (right panel) in the presence (⫹) or absence (⫺) of the GroESL cycling system (GroE) and/or luteolin. The arrow indicates the single electrophoretically retarded complex. (B) Dosage response of luteolin-stimulated DNA binding activity of NodD1. The indicated concentrations of luteolin were added to NodD1 that had been pretreated with the GroESL cycling system prior to EMSA. The bar graph represents the relative amounts of retarded complex based on phosphorimager quantification.
plained, relates to inducer structure and specificity. Each plant produces a cocktail of inducers, most of which are flavonoids; inhibitors are also present in seed and root exudates. Distinct flavonoids are found to be optimal for each plant-Rhizobium pair (9, 22, 30–32, 40). Examples of inducing compounds include luteolin and methoxychalcone for alfalfa-S. meliloti, 4⬘,7dihydroxyflavone for clover-Rhizobium leguminosarum bv. trifoli, daidzein and genistein for soybean-Bradyrhizobium japonicum, and eriodictyol and hesperitin for pea and Vicia sativa-R. leguminosarum bv. viciae. It has been shown that switching the nodD allele in a bacterium is sufficient to change the bacterial flavonoid response specificity (17, 38; M. C. Peck and S. R. Long, unpublished data). How flavonoid structural
VOL. 184, 2002
MODULATION OF NodD ACTIVITY BY LUTEOLIN AND GroESL
FIG. 4. The GroESL chaperonin system and luteolin are responsible for formation of active NodD1. (A) Affinity-purified NodD1 was further fractionated by SEC-HPLC as described in Materials and Methods. The indicated fraction was collected and concentrated for SDS-PAGE analysis (B) and EMSA experiments (C). The secondary peak at 7.3 min is diaminobiotin (data not shown). Molecular size standards (in kilodaltons) are indicated in panel B (MW). (C) EMSA was performed following treatment with (⫹) or without (⫺) GroES, ATP, and/or 10 M luteolin as indicated. (D) EMSA was similarly conducted with SEC-HPLC-purified NodD3 after treatment with (⫹) or without (⫺) GroES, ATP, and/or luteolin. The arrow indicates the single electrophoretically retarded complex.
specificity relates to NodD activity is an important subject for future study. The mechanism by which chaperonins modulate the activity of the NodD transcriptional activators is not clear. It has been shown that the GroESL chaperonin system is also required for the function of Vibrio fischeri LuxR expressed in E. coli (4). In Bacillus subtilis, GroESL modulates the activity of the repressor HrcA and is thus involved in control of the heat shock response in that organism (26a). An interaction between GroESL and a LysR family regulator other than NodD has not
been reported. Moreover, a specific molecular mechanism for the role of GroESL has not been described for any prokaryotic transcription activator. In eukaryotes, however, at least some examples are known, the most obvious one being the role of the heat shock protein complex in the activity of the steroid receptor proteins (reviewed in reference 1). Specifically, the association of steroid aporeceptors with heat shock protein 90 (HSP90) and other heat shock proteins results in an aporeceptor complex. The presence of steroids results in dissociation of the aporeceptor from the HSP90 complex (34). Then, the
YEH ET AL.
activated steroid receptors dimerize and regulate nuclear gene expression. It is possible that in the present case a variation on this behavior occurs in which chaperonin systems are important for the ability of flavones to interact productively with the bacterial regulatory molecule NodD1. We showed in vitro that the presence of luteolin stimulates the DNA binding activity of NodD1 in the presence of GroESL. Intriguingly, peptide homology between NodD and steroid receptors has been suggested (16). Although the role of flavonoids as ligands is still not clear in this system, the simplest model is that flavonoids bind to the folded and released NodD1 and regulate its activity. Alternatively, flavonoid inducers might dissociate the NodD1-GroESL complex and stimulate the dimerization or oligomerization of NodD1, just as steroids do in eukaryotic systems. The NodD system will provide a genetically and biochemically tractable system for the study of chaperonin action in a ligand-dependent prokaryotic gene transcription system. ACKNOWLEDGMENTS We thank Manish Raizada for preparing GroES and GroEL proteins and Judith Frydman for advice on their use. We also thank Robert Fisher and Shu-Hsing Wu for critical reading of the manuscript and members of the Long lab for helpful discussions. This work was supported by NIH grant GM 30962 to S.R.L. and by the Howard Hughes Medical Institute. REFERENCES 1. Bohen, S. P., and K. R. Yamamoto. 1994. Modulation of steroid receptor signal transduction by heat shock proteins, p. 313–334. In R. I. Morimoto, A. Tissie´res, and C. Georgopoulos (ed.), The biology of heat shock proteins and molecular chaperonins. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 2. Burn, J., L. Rossen, and A. W. B. Johnston. 1987. Four classes of mutations in the nodD gene of Rhizobium leguminosarum biovar viciae that affect its ability to autoregulate and/or activate other nod genes in the presence of flavonoid inducers. Genes Dev. 1:456–464. 3. Burn, J. E., W. D. Hamilton, J. C. Wootton, and A. W. B. Johnston. 1989. Single and multiple mutations affecting properties of the regulatory gene nodD of Rhizobium. Mol. Microbiol. 3:1567–1577. 4. Dolan, K. M., and E. P. Greenberg. 1992. Evidence that GroEL, not 32, is involved in transcriptional regulation of the Vibrio fischeri luminescence genes in Escherichia coli. J. Bacteriol. 174:5132–5135. 5. Downie, J. A., and S. A. Walker. 1999. Plant responses to nodulation factors. Curr. Opin. Plant Biol. 2:483–489. 6. Egelhoff, T. T., R. F. Fisher, T. W. Jacobs, J. T. Mulligan, and S. R. Long. 1985. Nucleotide sequence of Rhizobium meliloti 1021 nodulation genes: nodD is read divergently from nodABC. DNA 4:241–248. 7. Ewalt, K. L., J. P. Hendrick, W. A. Houry, and F. U. Hartl. 1997. In vivo observation of polypeptide flux through the bacterial chaperonin system. Cell 90:491–500. 8. Fink, A. L. 1999. Chaperonin-mediated protein folding. Physiol. Rev. 79: 425–449. 9. Firmin, J. L., K. E. Wilson, L. Rossen, and A. W. B. Johnston. 1986. Flavonoid activation of nodulation genes in Rhizobium reversed by other compounds present in plants. Nature 324:90–92. 10. Fisher, R. F., T. T. Egelhoff, J. T. Mulligan, and S. R. Long. 1988. Specific binding of proteins from Rhizobium meliloti cell-free extracts containing NodD to DNA sequences upstream of inducible nodulation genes. Genes Dev. 2:282–293. 11. Fisher, R. F., and S. R. Long. 1989. DNA footprint analysis of the transcriptional activator proteins NodD1 and NodD3 on inducible nod gene promoters. J. Bacteriol. 171:5492–5502. 12. Fisher, R. F., and S. R. Long. 1993. Interactions of NodD at the nod box: NodD binds to two distinct sites on the same face of the helix and induces a bend in the DNA. J. Mol. Biol. 233:336–348. 13. Goloubinoff, P., A. A. Gatenby, and G. H. Lorimer. 1989. GroE heat shock proteins promote assembly of foreign prokaryotic ribulose bisphosphate carboxylase oligomers in Escherichia coli. Nature 337:44–47. 14. Gottfert, M., B. Horvath, E. Kondorosi, P. Putnoky, F. Rodriguez-Quinones, and A. Kondorosi. 1986. At least two nodD genes are necessary for efficient nodulation on alfalfa by Rhizobium meliloti. J. Mol. Biol. 191:411–420. 15. Gyorgypal, Z., N. Iyer, and A. Kondorosi. 1988. Three regulatory nodD
alleles of diverged flavonoid specificity are involved in host-dependent nodulation by Rhizobium meliloti. Mol. Gen. Genet. 212:85–92. 16. Gyorgypal, Z., and A. Kondorosi. 1991. Homology of the ligand-binding regions of Rhizobium symbiotic regulatory protein NodD and vertebrate nuclear receptors. Mol. Gen. Genet. 226:337–340. 17. Gyorgypal, Z., E. Kondorosi, and A. Kondorosi. 1991. Diverse signal sensitivity of NodD protein homologs from narrow and broad host range rhizobia. Mol. Plant-Microbe Interact. 4:356–364. 18. Hong, G. F., J. E. Burn, and A. W. B. Johnston. 1987. Evidence that DNA involved in the expression of nodulation (nod) genes in Rhizobium binds to the product of the regulatory gene nodD. Nucleic Acids Res. 15:9677–9690. 19. Honma, M. A., and F. M. Ausubel. 1987. Rhizobium meliloti has three functional copies of the nodD symbiotic regulatory gene. Proc. Natl. Acad. Sci. USA 84:8558–8562. 20. Horvath, B., C. W. B. Bachem, J. Schell, and A. Kondorosi. 1987. Hostspecific regulation of nodulation genes in Rhizobium is mediated by a plant signal interacting with the nodD gene product. EMBO J. 6:841–848. 21. Houry, W. A., D. Frishman, C. Eckerskorn, F. Lottspeich, and F. U. Hartl. 1999. Identification of in vivo substrates of the chaperonin GroEL. Nature 402:147–154. 22. Kosslak, R. M., R. Bookland, J. Barkei, H. E. Paaren, and E. R. Appelbaum. 1987. Induction of Bradyrhizobium japonicum common nod genes by isoflavones isolated from Glycine max. Proc. Natl. Acad. Sci. USA 84:7428–7432. 23. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685. 24. Long, S. R. 1996. Rhizobium symbiosis: nod factors in perspective. Plant Cell 8:1885–1898. 24a.Martin, J., and F. U. Hartl. 1997. Chaperonin-assisted protein folding. Curr. Opin. Struct. Biol. 7:41–52. 25. Maxwell, C. A., U. A. Hartwig, C. M. Joseph, and D. A. Phillips. 1989. A chalcone and two related flavonoids released from alfalfa roots induce nod genes of Rhizobium meliloti. Plant Physiol. 91:842–847. 26. McIver, J., M. A. Djordjevic, J. J. Weinman, G. L. Bender, and B. G. Rolfe. 1989. Extension of host range of Rhizobium leguminosarum bv. trifolii caused by point mutations in nodD that result in alterations in regulatory function and recognition of inducer molecules. Mol. Plant-Microbe Interact. 2:97– 106. 26a.Mogk, A., G. Homuth, C. Scholz, L. Kim, F. X. Schmid, and W. Schumann. 1997. The GroE chaperonin machine is a major modulator of the CIRCE heat shock regulon of Bacillus subtilis. EMBO J. 16:4579–4590. 27. Mulligan, J. T., and S. R. Long. 1989. A family of activator genes regulates expression of Rhizobium meliloti nodulation genes. Genetics 122:7–18. 28. Mulligan, J. T., and S. R. Long. 1985. Induction of Rhizobium meliloti nodC expression by plant exudate requires nodD. Proc. Natl. Acad. Sci. USA 82:6609–6613. 29. Ogawa, J., and S. R. Long. 1995. The Rhizobium meliloti groELc locus is required for regulation of early nod genes by the transcription activator nodD. Genes Dev. 9:714–729. 30. Peters, N. K., J. W. Frost, and S. R. Long. 1986. A plant flavone, luteolin, induces expression of Rhizobium meliloti nodulation genes. Science 233:977– 980. 31. Phillips, D. A., F. D. Dakora, E. Sande, C. M. Joseph, and J. Zon. 1994. Synthesis, release and transmission of alfalfa signals to rhizobial symbionts. Plant Soil 161:69–80. 32. Putnoky, P., and A. Kondorosi. 1986. Two gene clusters of Rhizobium meliloti code for early essential nodulation functions and a third influences nodulation efficiency. J. Bacteriol. 167:881–887. 33. Rushing, B. G., M. M. Yelton, and S. R. Long. 1991. Genetic and physical analysis of the nodD3 region of Rhizobium meliloti. Nucleic Acids Res. 19:921–927. 34. Sanchez, E. R., S. Meshinchi, W. Tienrungroj, M. J. Schlesinger, D. O. Toft, and W. B. Pratt. 1987. Relationship of the 90-kDa murine heat shock protein to the untransformed and transformed states of the L-cell glucocorticoid receptor. J. Biol. Chem. 262:6986–6991. 35. Schell, M. A. 1993. Molecular biology of the LysR family of transcriptional regulators. Annu. Rev. Microbiol. 47:597–626. 36. Schmidt, T. G. M., and A. Skerra. 1994. One step affinity purification of bacterially produced proteins by means of the strep tag and immobilized recombinant core streptavidin. J. Chromatogr. 676:337–345. 37. Spaink, H. P., C. A. Wijffelman, R. J. H. Okker, and B. E. J. Lugtenberg. 1989. Localization of functional regions of the Rhizobium nodD product using hybrid nodD genes. Plant Mol. Biol. 12:59–73. 38. Spaink, H. P., C. A. Wijffelman, E. Pees, R. J. H. Okker, and B. J. J. Lugtenberg. 1987. Rhizobium nodulation gene nodD as a determinant of host specificity. Nature 328:337–340. 39. Yeh, K. C., S. H. Wu, J. T. Murphy, and J. C. Lagarias. 1997. A cyanobacterial phytochrome two-component light sensory system. Science 277:1505– 1508. 40. Zaat, S. A. J., C. A. Wijffelman, H. P. Spaink, A. A. N. Van Brussel, R. J. H. Okker, and B. J. J. Lugtenberg. 1987. Induction of the nodA promoter of Rhizobium leguminosarum Sym plasmid pRL1JI by plant flavanones and flavones. J. Bacteriol. 169:198–204.