Activate NodD2 Protein in Rhizobium meliloti - NCBI

Plant Physiol. (1992) 99, 1526-1531 0032-0889/92/99/1 526/06/$01 .00/0

Received for publication February 11, 1992 Accepted March 4, 1992

Trigonelline and Stachydrine Released from Alfalfa Seeds Activate NodD2 Protein in Rhizobium meliloti' Donald A. Phillips*, Cecillia M. Joseph, and Carl A. Maxwell2 Department of Agronomy and Range Science, University of California, Davis, California 95616 ABSTRACT

nod gene inducer that is 10- to 20-fold more active than luteolin (19). A potentially important functional difference between these two natural inducers is that MCh activates both NodDl and NodD2 proteins (11), whereas luteolin activates only NodD1 (11, 12). The different nodD regulatory genes (nodD1, nodD2, and nodD3) in R. meliloti affect the rate at which this bacterium nodulates different host plants. At least two of the three nodD genes are required for rapid nodule formation on alfalfa (6), and only when all three genes are mutated is nodulation essentially eliminated on this host (13). Experiments with bacteria containing extra copies of various nodD genes showed that the three NodD proteins interact differently with seed rinses from various legume hosts (7). Alfalfa seed rinse, for example, strongly activated both NodDl and NodD2, but comparable fractions from four other species of legumes failed to activate NodD2 significantly (7). Subsequent studies of seed rinses separated on TLC did not locate a NodD2 activator in alfalfa seed rinse (8), and tests of flavonoids identified in alfalfa seed rinse showed only traces of NodD2 activation (11). This project was initiated to identify the unknown NodD2 activator reported previously (7) in crude alfalfa seed rinses. Understanding the chemical structure and traits of that compound may be important for defining ecochemical interactions between alfalfa and soil microbes around the germinating seed.

Spectroscopic data (nuclear magnetic resonance, mass spectrometry, ultraviolet-visible) in this study identify trigonelline and stachydrine as major components of alfalfa (Medicago sativa L.) seed rinse. Moreover, biological assays show that these natural products induce nodulation (nod) gene transcription in Rhizobium meliloti by activating the regulatory protein NodD2, but not the homologous NodDl protein. These findings contrast with the fact that the only previously identified NodD2 activator, 4,4'-dihydroxy-2'-methoxychalcone (MCh), also activates NodDl protein. Trigonelline and stachydrine induce nod genes only at much higher concentrations than MCh, but they are released from seeds in correspondingly greater amounts. The existence of these amphoteric, nonflavonoid nod gene inducers broadens our understanding of the biochemical processes and ecological mechanisms that a legume host uses to regulate its microbial symbiont.

Molecular data now support the concept that plants regulate microbial processes with chemical signals (24, 29). Root nodule formation by Rhizobium meliloti on alfalfa (Medicago sativa) requires transcription of bacterial nodulation (nod) genes, which are under the control of a family of regulatory nodD genes (18). Activation of NodD protein by a plant signal, such as a flavonoid, is one of the first steps in the nodulation process. Understanding how chemical signals from alfalfa activate different NodD proteins will clarify the process of rhizobial nodulation and is a crucial part of assessing which signal transduction pathways are important in the rhizosphere (25). Distinct zones of plant chemicals probably exist in the rhizosphere around a young alfalfa seedling. For example, luteolin, the first nod gene inducer identified from alfalfa (24), is discharged from imbibing seeds (10) but not from roots (19). In contrast, roots but not seeds release MCh3, a

MATERIALS AND METHODS

Biological Materials

Aqueous seed rinses, also termed exudates, were prepared from surface-sterilized alfalfa (Medicago sativa L., cv Moapa 69) seeds by soaking for 4 h (10). Rhizobium meliloti strains, which were generously supplied by Dr. S.R. Long, Stanford University, were derived from strain 1021 (20). The parent strain JM57 (22), which has a nodABC-lacZ translational fusion on the megaplasmid pSym, was modified to contain additional cloned copies of either nodDi (M57pRmJ3O; 14) or nodD2 (JM57pRmM137; 3). Transcription of nod genes was measured as ,3-galactosidase activity (21) from nodABC-lacZ with modifications described previously (19). Activities of uninduced bacterial controls and colorimetric controls of test compounds were subtracted.

'This work was supported by U.S. Department of Agriculture National Research Initiative Competitive Grants Program award 9137305-6513 and grant US-1884-90 from the U.S.-Israel Binational Agricultural Research and Development Fund. 2Present address: Plant Biology Division, The Samuel Roberts Noble Foundation, Inc., P.O. Box 2180, Ardmore, OK 73402. 3Abbreviations: MCh, 4,4'-dihydroxy-2'-methoxychalcone; COSY, two-dimensional homonuclear shift correlation spectroscopy; 5H, chemical shift of proton; 2H, two protons, etc.; C-1, carbon one, etc.; FAB, fast atom bombardment; H4,5, interaction between protons on C-4 and C-5; 150, concentration for half-maximal induction; MeOH, methanol; s, singlet; m, multiplet; t, triplet; m/z, mass/charge ratio.

Purification of NodD2 Activators Initial tests with 4 g of alfalfa seed produced an aqueous seed rinse that was diluted to 50% MeOH, injected into a 1526

RHIZOBIUM NodD2 ACTIVATORS FROM ALFALFA SEED

Waters HPLC system (Millipore Corp., Milford, MA) equipped with a reverse-phase column (Lichrosorb 5RP18; Phenomenex, Rancho Palos Verdes, CA), and eluted with a 0- to 90-min linear gradient from 20:79:1 (v/v) water:MeOH:acetic acid to 99:1 (v/v) MeOH:acetic acid. Eluant was monitored with a photodiode array detector, which recorded the maximum absorbance between 260 and 400 nm every second. Fractions were collected at 1-min intervals and assayed for NodD2-dependent nod gene-inducing activity. Large-scale purification studies used rinses that were collected after 4 h from 50 g of seeds and freeze dried. The sample was redissolved in a small volume of water and passed through C18 Maxi-Clean cartridges (Alltech Associates, Inc., Deerfield, IL) to remove flavonoids. Water rinses from cartridges were freeze dried, and then the sample was redissolved in water and applied to a cation-exchange HPLC column (Partisil 10 scx; Whatman Chemical Separation, Inc., Clifton, NJ). The column was rinsed with water for 10 min at a flow rate of 2.0 mL min-', and then a 30-min linear gradient from water to 40:60 (v/v) MeOH:water containing 3 mM HNO3 eluted the NodD2-activating fraction.

Analyses of NodD2 Activators NMR data were collected by the authors at the University of California at Davis NMR facility. One-dimensional 'HNMR, two-dimensional 'H-NMR (COSY), and '3C-NMR measurements were recorded in [U-'H]water on a GN-300 Omega NMR spectrometer (General Electric Co., Fremont, CA). FAB and high-resolution MS data were collected at the University of California at Davis Facility for Advanced Instrumentation on a recharge basis with a ZAB-HS-2F MS (VG Analytical, Wythenshawe, UK) using positive ionization (xenon, 8 keV, 1 mA) with a VG Dynamic FAB probe. FAB MS measurements were made on samples dissolved in MeOH and injected with 95:5 (v/v) water:glycerol at a flow rate of 5 gL min-'. For high-resolution mass determinations, the sample was injected with PEG 300. UV-visible spectra were recorded on a Lambda 6 dual-beam spectrophotometer (Perkin Elmer, Norwalk, CT).

Chemicals Authentic trigonelline was purchased from Sigma Chemical Co. (St. Louis, MO). Stachydrine was obtained from American Tokyo Kasel, Inc. (Portland, OR) and later synthesized from O-methyl-N,N'-diisopropylisourea (Aldrich Chemical Co., Milwaukee, WI) and L-proline (Sigma) (23). MCh was synthesized by published procedures (2). RESULTS

Purification of Factors Three regions of nod gene-inducing activity were detected with the NodD2-containing R. meliloti strain JM57pRmM137 in HPLC fractions from alfalfa seed rinse (Fig. 1). Most of the NodD2-activating capacity eluted in the void volume of the reverse-phase C18 column. The fractions that formed peak x could be retained on a normal-phase column in pure MeOH,

1 527

5 0.4 .0

0.2 b.

0 .0 (U 0

0 (a

to

0.0 60 40 20 0 0

60 30 Retention Time (min)

90

Figure 1. HPLC characteristics and nod induction assays of alfalfa seed rinses collected during the first 4 h of imbibition. A, Maximum absorbance (260-400 nm) of exudate from the equivalent of 5.3 mg of seeds (i.e. two seeds) fractionated on a reverse-phase C18 column. Peaks indicated with letters have been identified previously as follows: a, luteolin-7-O-glucoside; b, quercetin-3-O-galactoside; c, a quercetin conjugate: d, 5-methoxyluteolin; e, 3',5dimethoxyluteolin; f, luteolin; g, chrysoeriol (9, 10). Peak x is the subject of this investigation. B, f3-Galactosidase activity induced from nodABC-lacZ under the control of NodD2 protein in R. meliloti strain JM57pRmM137 by 50% of each fraction from chromatogram A. Peak x was subsequently shown to contain trigonelline and stachydrine. U, Units.

but they eluted with traces of water in the buffer. All activity from peak x was retained on a cation-exchange column when introduced in water, and a single peak with NodD2-activating capacity eluted in a linear gradient to 40% MeOH containing 3 mm HNO3. The UV-visible spectrum of the fraction showed an absorbance peak at 264 nm with a shoulder at 272 nm. NMR Experiments

One-dimensional 'H-NMR spectra (Fig. 2) of the NodD2activating fraction purified from peak x showed the following resonances: 6H ppm, 9.57 (1H, s), 9.23 to 9.17 (2H, m), 8.39 (1H, t, J = 6.7 Hz), 4.69 (3H, s), 4.56 (1H, t, J = 9.8 Hz), 4.00 to 3.96 (1H, m), 3.84 (1H, dd, J = 9.8, 19.6 Hz), 3.58 (3H, s), 3.38 (3H, s), 2.84 to 2.72 (1H, m), 2.70 to 2.55 (1H, m), 2.50 to 2.38 (2H, m). Resonances at 4.69 ppm and greater were consistent with trigonelline (4) and allowed proton assignments from 4.69 to 9.57 ppm as N-CH3, C-5, C-4/C-6, and C-2, respectively. Resonances below 4.69 were consistent with stachydrine (1, 28), and COSY experiments (Fig. 3) allowed the following assignments to resonances from 2.50 to 4.56 ppm: C-4, C-3, N-CH3, N-CH3, C-5, and C-2. Onedimensional 'H-NMR spectra (Fig. 2) did not eliminate the possibility that split resonances were caused by interactions between low-field and high-field protons. The COSY spectrum (Fig. 3), however, showed that interactions were separated into two regions: those >8 ppm and those < 4.69 ppm. Thus, a necessary, but not sufficient, requirement was ful-

1 528

PHILLIPS ET AL.

Figure 2. 'H-NMR spectrum of the NodD2activating fraction purified from peak x in Figure 1. The signal associated with the solvent [U2H]water has been removed. Values above resonances indicate the number of protons measured by integration of the area under each peak.

Plant Physiol. Vol. 99, 1992

3.50 1 .9'9

1.02

3.38

3.39

1.00 0.98 0.97

f

I

10

0.96j

8

6 Proton Shift (ppm)

2.07

1.07 0.99 \

f/I 2

4

filled for concluding that two separate molecules were present in the sample. Resonances in '3C-NMR experiments with the NodD2-activating fraction were consistent with those reported for trigonelline (27) and stachydrine (1).

and C7H1402N (stachydrine) respectively.

MS Studies

On the basis of the NMR and MS data, we concluded that the NodD2-activating fraction purified from peak x contained a mixture of trigonelline and stachydrine. That interpretation was supported by separate 1H-NMR experiments that showed that authentic trigonelline and stachydrine produced the same resonances as the nod gene-inducing fraction from peak x (data not shown). Confirmation that trigonelline and stach-

138.0555 and 144.1025,

Confirmation of Identification

MS experiments with the NodD2-activating fraction purified from HPLC peak x gave strong molecular ions, MH', at 138 and 144 m/z. High-resolution analysis of those ions measured 138.0557 and 144.1031 m/z, respectively. Calculated values for MH+ ions from C7H802N (trigonelline)

Figure 3. COSY 1H-NMR spectrum of the NodD2-activating fraction purified from peak x in Figure 1. Interactions between protons on

are

0

adjacent carbon atoms are indicated as H4,5 for C-4 and C-5, etc. Because all proton-proton

interactions can be explained by

resonances

2

in

separate molecules of trigonelline (4.69-9.57

ppm) and stachydrine (2.50-4.56 ppm), this COSY is consistent with the presence of two separate molecules in the NodD2-activating fraction purified from peak x.

-c

*0

4

0.

0n 6 a

0

8110 a

10

8

6

a

I

0

4

2

0

Proton Shift (ppm)


ydrine induce transcription of a nodC-lacZ fusion by activating NodD2 protein, but not NodDl protein, was obtained in direct biological assays with authentic compounds (Fig. 4). Activity of Inducers

Trigonelline and stachydrine are much less active than MCh and produced I50 values for activation of NodD2 protein that were approximately 4 orders of magnitude higher than MCh (Fig. 4). The presence of both trigonelline and stachydrine in seed rinses may enhance nod gene-inducing activity relative to the two individual compounds because mixtures of the two molecules gave significantly higher activity than the two separately (Table I). Trigonelline produces a strong UV absorbance, log E = 3.55 at 264 nm in water, but stachydrine has no distinctive absorbance band. Thus, direct quantitative comparisons could not be made between the biological activity of the mixture purified from peak x (Fig. 1) and nod-inducing activity of authentic compounds. Amounts of Compounds

Three samples of peak x (Fig. 1) purified separately contained ratios of trigonelline:stachydrine ranging from 2:1 to 1:2, as determined by 'H-NMR ratios. In one case with three replicates, rinses from 2000 seeds (5 g) yielded 5.4 ± 0.3 mg of NodD2-activating material from peak x. Ratios between resonances on 'H-NMR spectra indicated that those samples contained 1.8 mg of trigonelline and 3.6 mg of stachydrine. Correcting for losses in purification, which were estimated by supplementing with authentic trigonelline, those amounts represent 8.6 and 16.4 nmol seed-' of trigonelline and stachydrine, respectively.

125

100 5

cn

75 50 25

U) 0

la

125

CD 100 0

40

0

0o

75 50 25 0

0

10 9 10o

10 7 1010

104 10

Concentration (M) Figure 4. Induction of nodABC-lacZ in R. meliloti under the control of different NodD proteins. Authentic trigonelline, stachydrine, or MCh was supplied to R. meliloti JM57pRmM137 containing extra NodD2 protein (A) or to R. meliloti JM57pRmJ3O containing extra NodDl protein (B). SE bars are obscured by the data points. U, Units.

1 529

Table 1. Effects of Combining Trigonelline and Stachydrine on NodD2 Activation in R. meliloti JM57pRmM 137 Mean values ± SE of fl-galactosidase activity reflect transcription of a nodC-lacZ fusion under control of NodD2 protein. Compound Concentration Trigonelline Stachydrine units of :-galactosidase MM 1 0.4±0.2 9.5±1.0 10 35.0 ± 1.6 25.8 ± 0.4 20 43.3±2.0 28.2±1.1 1 trigonelline + 1 stachydrine 10 trigonelline + 10 stachydrine

26.1

±

0.6

87.9 ± 2.9

DISCUSSION The direct spectroscopic techniques used in this study allow one to conclude that the major NodD2-activating fraction released from Moapa 69 alfalfa seeds (Fig. 1, peak x) contains both trigonelline and stachydrine. Although these compounds could not be separated by the purification methods used, 'H-NMR, "3C-NMR, and MS techniques revealed no other natural products in the final biologically active fraction. The possibility that a conjugate of the two compounds is the active factor was decreased by the fact that purifying the active fraction from different seed samples yielded varying ratios (2:1 to 1:2) of the two compounds. A biological requirement for such a hypothetical conjugate was removed by the demonstration that the two molecules independently activate NodD2 protein in R. meliloti (Fig. 4A). These data, therefore, identify trigonelline and stachydrine as the first reported compounds that regulate nod gene induction in R. meliloti through NodD2 protein with no apparent activation of NodDl (Fig. 4). More subtle regulatory interactions between NodDl and NodD2 in the presence of these compounds remain to be tested because strains used in this study contained single copies of each nodD gene on the symbiotic plasmid, pSym. The compounds are released naturally from alfalfa seeds and probably make up the previously unidentified NodD2-activating factor reported by others in studies with alfalfa seed rinse (7). The quantities of trigonelline and stachydrine released from Moapa 69 alfalfa seeds (8-16 nmol seed-') qualify these compounds as major components in seed exudate because only 1.8 nmol seed-' of the dominant flavonoid, quercetin-3-O-galactoside, is discharged from the same seeds (9). Both trigonelline and stachydrine have been reported previously in alfalfa. Crude extracts of water-stressed alfalfa root nodules contain 'H-NMR resonances that were interpreted as N-methyl protons of these compounds (16), and stachydrine was identified in 1918 from alfalfa hay extracts (30). Although trigonelline occurs in seeds of many other legumes (31), data reported here are ecologically important because they establish its presence in alfalfa seed rinse. Trigonelline was previously described as a very weak inducer of nodA transcription in R. meliloti (26), but there was no indication

1530

PHILLIPS ET AL.

of how the signal was transduced. An interesting aspect of that report is that 10 Mm trigonelline represented an Io value similar to the current results (Fig. 4A), even in the absence of extra nodD2 genes. This comparison suggests that measurements made with the reporter gene system in R. meliloti JM57pRmM137 are physiologically relevant. Trigonelline and stachydrine apparently are the major NodD2 activators in alfalfa seed rinses (Fig. 1). Two other minor factors eluting at 64 and 75 min (Fig. 1) may be 4',7dihydroxyflavone and chrysoeriol. Those compounds activate NodD2 slightly and have retention times on chromatograms of alfalfa seed exudate similar to those shown in Figure 1 (10, 11). MCh, which is released only from roots of Moapa 69 alfalfa, not seeds, apparently could fulfill any requirement for activation of NodD2 in RP meliloti cells that are not present in the seed zone. That function of MCh is accomplished at an I50 value 4 orders of magnitude lower than either trigonelline or stachydrine (Fig. 4A), but alfalfa seedlings release correspondingly less MCh (1 pmol plant h-1) (19) than trigonelline or stachydrine. A comparison of trigonelline and stachydrine structures with those of other nod gene inducers from alfalfa (Fig. 5) indicates that not all portions of the MCh molecule are required for activation of NodD2 protein. MCh probably has a 10- to 20-fold lower I5o than luteolin with NodDl protein (19) because its three carbon-carbon bonds with unhindered rotation create a more flexible molecule. One might extend the same reasoning to explain how MCh, but not luteolin, activates NodD2 protein. However, results in Figure 4A suggest that simple flexibility of the MCh molecule is only part of the answer. Because trigonelline and stachydrine

aC/ cS

0

Trigonelline

c

Stachydrine

Plant Physiol. Vol. 99, 1992

activate NodD2, the A ring of MCh (Fig. 5) probably is primarily responsible for the same function. Thus, one can suggest that luteolin fails to activate NodD2 protein because its C and B rings impair binding of the A ring to a NodD2 activation site. One significance of this conclusion is that it may facilitate the design of synthetic molecules that can activate NodD2 protein and be used to regulate transcription of novel genes delivered to crop plants by genetically altered bacteria. These observations offer a new perspective on the possibility that availability of flavonoids limits root nodule initiation in alfalfa. Although reports show that adding flavonoids to germinating alfalfa seeds increases nodulation significantly in some cultivars (15, 17), tests with other cultivars do not always produce similar increases (Y. Kapulnik, personal communication). One complication in such experiments could be related to unknown differences in the amount of compounds such as trigonelline and stachydrine. Thus, a more complete understanding of how amounts of these compounds and nod gene-inducing flavonoids vary in seeds of different alfalfa populations may help optimize nodulation on that crop plant. The biochemical and ecological implications of these observations suggest that several benefits accrue to alfalfa by using nonflavonoid molecules for inducing nod genes in R. meliloti. First, because trigonelline and stachydrine are formed by metabolic pathways separate from those responsible for flavonoids, the plant is not prevented from producing nod gene-inducing signals if flavonoid synthesis is decreased by unexpected genetic or biochemical events. Second, the excellent solubility of trigonelline and stachydrine in water may allow them to diffuse more easily than flavonoids in soil. This potentially larger ecochemical zone could increase the number of R. meliloti cells that respond to the presence of the alfalfa seedling. However, because positive charges on these molecules could impair movement in soil despite their solubility, direct tests for mobility in soil are being conducted. Third, the symbiotic plasmid, pSym, in R. meliloti carries genes for degrading millimolar concentrations of trigonelline and stachydrine (5). Thus, it is possible that rhizobia remove these signaling compounds from the soil to permit communication with the next generation of alfalfa plants. In light of the interesting chemical traits of these NodD2 activators, it seems desirable to obtain more information concerning rhizobial genes that are regulated by NodD2. ACKNOWLEDGMENTS

4,4'-Dihydroxy-2'-methoxychalcone of

Luteolin Figure 5. Structures of trigonelline and stachydrine identified as NodD2 activators in peak x (Fig. 1) compared with MCh and luteolin.

We thank Dr. A.D. Jones of the University of California at Davis Facility for Advanced Instrumentation for skillful and timely MS analyses. LITERATURE CITED 1. Ahman VU, Basha A, Rahman A (1975) Identification and C13 N.M.R. spectrum of stachydrine from Cadaba fruticosa. Phytochemistry 14: 292-293 2. Carlson RE, Dolphin DH (1982) Pisum sativum stress metabolites: two cinnamylphenols and a 2'-methoxychalcone. Phytochemistry 21: 1733-1736 3. Fisher RF, Egelhoff TT, Mulligan JT, Long SR (1988) Specific binding of protein from Rhizobium meliloti cell-free extracts containing NodD to DNA sequences upstream of inducible nodulation genes. Genes Dev 2: 282-293


4. Ghosal S, Dutta SK (1971) Alkaloids of Abrus precatorius. Phytochemistry 10: 195-198 5. Goldmann A, Boivin C, Fleury V, Message B, Lecoeur L, Maille M, Tepfer D (1991) Betaine use by rhizosphere bacteria: genes essential for trigonelline, stachydrine, and camitine catabolism in Rhizobium meliloti are located on pSym in the symbiotic region. Mol Plant Microbe Interact 4: 571-578 6. Gottfert M, Horvath B, Kondorosi E, Putnoky P, RodriguezQuinones F, Kondorosi A (1986) At least two different nodD genes are necessary for efficient nodulation on alfalfa by Rhizobium meliloti. J Mol Biol 191: 411-420 7. Gyorgypal Z, Iyer N, Kondorosi A (1988) Three regulatory nodD alleles of diverged flavonoid-specificity are involved in host-dependent nodulation by Rhizobium meliloti. Mol Gen Genet 212: 85-92 8. Gyorgypal Z, Kondorosi E, Kondorosi A (1991) Diverse signal sensitivity of NodD protein homologs from narrow and broad host range rhizobia. Mol Plant Microbe Interact 4: 356-364 9. Hartwig UA, Joseph CM, Phillips DA (1991) Flavonoids released naturally from alfalfa seeds enhance growth rate of Rhizobium meliloti. Plant Physiol 95: 797-803 10. Hartwig UA, Maxwell CA, Joseph CM, Phillips DA (1990) Chrysoeriol and luteolin released from alfalfa seeds induce nod genes in Rhizobium meliloti. Plant Physiol 92: 116-122 11. Hartwig UA, Maxwell CA, Joseph CM, Phillips DA (1990) Effects of alfalfa nod gene inducing flavonoids on nodABC transcription in Rhizobium meliloti strains containing different nodD genes. J Bacteriol 172: 2769-2773 12. Honma MA, Asomaning M, Ausubel FM (1990) Rhizobium meliloti nodD genes mediate host-specific activation of nodABC. J Bacteriol 172: 901-911 13. Honma MA, Ausubel FM (1987) Rhizobium meliloti has three functional copies of the nodD symbiotic regulatory gene. Proc Natl Acad Sci USA 84: 8558-8562 14. Jacobs TW, Egelhoff TT, Long SR (1985) Physical and genetic map of a Rhizobium meliloti nodulation gene region and nucleotide sequence of nodC. J Bacteriol 162: 469-476 15. Jain V, Garg N, Nainawatee HS (1990) Naringenin enhanced efficiency of Rhizobium meliloti-alfalfa symbiosis. World J Microbiol Biotech 6: 434-436 16. Jones GP, Naidu BP, Starr RK, Paleg LG (1986) Estimates of solutes accumulating in plants by 'H nuclear magnetic resonance spectroscopy. Aust J Plant Physiol 13: 649-658 17. Kapulnik Y, Joseph CM, Phillips DA (1987) Flavone limitations

18. 19.

20.

21. 22. 23.

24. 25. 26.

27. 28. 29.

30. 31.

1531

to root nodulation and symbiotic nitrogen fixation in alfalfa. Plant Physiol 84: 1193-1196 Long SR (1992) Genetic analysis of Rhizobium nodulation. In G Stacey, RH Burris, HJ Evans, eds, Biological Nitrogen Fixation. Chapman and Hall, New York, pp 560-597 Maxwell CA, Hartwig UA, Joseph CM, Phillips DA (1989) A chalcone and two related flavonoids released from alfalfa roots induce nod genes of Rhizobium meliloti. Plant Physiol 91: 842-847 Meade HM, Long SR, Ruvkun GB, Brown SE, Ausubel FM (1982) Physical and genetic characterization of symbiotic and auxotrophic mutants of Rhizobium meliloti induced by transposon Tn5 mutagenesis. J Bacteriol 149: 114-122 Miller JH (1972) Experiments in Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, pp 352-355 Mulligan JT, Long SR (1985) Induction of Rhizobium meliloti nodC expression by plant exudate requires nodD. Proc Natl Acad Sci USA 82: 6609-6613 Musich JA, Rapoport H (1977) Reaction of O-methyl-N,N'diisopropylisourea with amino acids and amines. J Org Chem 42: 139-141 Peters NK, Frost JW, Long SR (1986) A plant flavone, luteolin, induces expression of Rhizobium meliloti nodulation genes. Science 233: 977-980 Phillips DA (1992) Flavonoids: plant signals to soil microbes. Rec Adv Phytochem 26: 201-231 Schmidt J, John M, Wieneke U, Krussmann H-D, Schell J (1986) Expression of the nodulation gene nodA in Rhizobium meliloti and localization of the gene product in the cytosol. Proc Natl Acad Sci USA 83: 9581-9585 Sciuto S, Chillemi R, Piattelli M (1988) Onium compounds from the red alga Pterocladia capillacea. J Nat Prod 51: 322-325 Smith GM, Pettigrew GW (1980) Identification of N,N-dimethylproline as the N-terminal blocking group of Crithidia oncopelti cytochrome C557. Eur J Biochem 110: 123-130 Stachel SE, Messens E, Van Montagu M, Zambryski P (1985) Identification of the signal molecules produced by wounded plant cells that activate T-DNA transfer in Agrobacterium tumefaciens. Nature 318: 624-629 Steenbock H (1918) Isolation and identification of stachydrine from alfalfa hay. J Biol Chem 35: 1-13 Tramontano WA, McGinley PA, Ciancaglini EF, Evans LS (1986) A survey of trigonelline concentrations in dry seeds of the Dicotyledoneae. Environ Exp Bot 26: 197-205