|
|
|
|||||||
|
|||||||||
(Received for publication, December 8, 1994) From the
Phaseolus vulgaris (common bean) can be nodulated by
different Rhizobium species. A new species has been recently
proposed: Rhizobium etli. Following transcriptional activation
of the bacterial nodulation genes using naringenin or bean seed
exudate, we have isolated, purified, and characterized R. etli extracellular nodulation factors. They are chitopentameric
compounds that are N-methyl-N-vaccenoylated at their
non-reducing end. At position 6 of the reducing N-acetyl-D-glucosamine, they are
4-O-acetyl-L-fucosylated. Minor compounds bear a
carbamate group on the terminal non-reducing saccharidic residue.
Azorhizobium, Bradyrhizobium, and Rhizobium, collectively referred to as Rhizobia, specifically
trigger stem or root nodule organogenesis on leguminous plants. In
these symbiotically elicited organs, the bacteria reduce atmospheric
nitrogen to ammonia(1) . During the infection process,
signal exchange between the two symbionts controls the specificity of
this interaction(2) . Legume roots secrete flavonoids that
regulate the expression of the nodulation genes of the bacteria (the nod and nol genes)(3) . In turn, these genes
are involved in the synthesis and excretion of the so-called nodulation
factors (Nod factors) ( Nod factors are
lipooligosaccharides that share the same common backbone: an
oligochitin core bearing a fatty acid that N-acylates the
non-reducing end. These factors may also carry a variety of
substitutions that complete the backbone and are involved in the
specificity of the interaction between the plant and the
bacteria(6) . At picomolar to nanomolar concentrations, these
molecules provoke the root hair deformation effect in
nodulation-competent root hairs(2, 7, 8) .
They also induce rapid membrane depolarization(9) , expression
of early nodulins(10) , and mitosis in the root
cortex(11, 12) , which, in some cases, leads to nodule
organogenesis(13, 14, 15, 16) . Thus
they appear to play a key role in the establishment of the nodule
symbiosis. Recognition of Nod factors by host plants is relatively
specific, and the nature and position of the Nod factor substitutions
are of crucial importance. It has been proposed that a way for a
bacterium to exhibit a wide host range is to synthesize a large variety
of Nod factors, suggesting that each variety of plant is specifically
triggered by Nod factors possessing a similar set of substitutions. For
example, all Nod factors from bacteria that nodulate soybeans possess a
2-O-methylfucosyl
substitution(8, 12, 16, 17) . A
6-O-sulfate group together with an N-(2E)(9Z)-hexadienoyl substitution seems
specific for nodulation of alfalfa(13, 18) . Rhizobium NGR234, which nodulates more than 60 different
legumes, synthesizes a very complex mixture of Nod factors bearing an N-methyl group and zero, one, or two O-carbamoyl
groups at the non-reducing end together with a
2-O-methylfucose moiety, which may or may not be substituted
by acetate or sulfate(19) . To examine whether a strict
relationship between the structure of Nod factors secreted by bacteria
and their host plant range exists for all symbiotic associations, we
examined the Nod factors from two different bacterial species that
nodulate the common bean (Phaseolus vulgaris). One species is Rhizobium tropici, from which the Nod factor structure has
been described earlier(20) . The species that will be dealt
with in the present paper is Rhizobium etli. R.
etli(21) , previously named Rhizobium leguminosarum bv. phaseoli type I, nodulates the common bean, as does R. tropici. However, the geographical origin and breadth of
the host range of these two species are different. The former, isolated
from Mexican soils, seems to have a narrow host range (21) .
The latter, which originated in South America, has a broad specificity
as it nodulates several tropical legumes and some trees (e.g.Leucaena)(22) . In this paper we show that Nod
factors from R. etli do not share identical structures with
those described for R. tropici and represent a new type of Nod
factor.
Mass spectra were
recorded on an AutoSpec instrument (VG Analytical, Manchester, U.K.)
fitted with a cesium bombardment ion source. The matrix was a mixture
of m-nitrobenzyl alcohol/glycerol (1:1, v/v) spiked either
with 1% trichloroacetic acid in water or with a solution of sodium
iodide (1 mg/ml). The location of double bond of fatty acid was
determined by remote charge fragmentation of the carboxylate anions as
described (25) using a capillary gas chromatograph coupling to
fractionate the fatty acids as pentafluorobenzyl esters, followed by
negative ion ionization and collision-activated dissociation
mass-analyzed ion kinetic energy spectrometry. GC-MS experiments
were performed on a Hewlett-Packard 5989A mass spectrometer in
electronic impact ionization mode. One- and two-dimensional correlation
spectroscopy
Deacetylation of 1 mg of NodRe
factors was achieved by 1 ml of 0.5 N sodium methanolate in
methanol at 30 °C overnight. After acidification with acetic acid,
the evaporated residue was dissolved in 1 ml of water and applied to a
Sep-Pak C
Figure 1:
Comparative analytical C
Figure 2:
A, positive ion fast atom bombardment mass
spectrum of the NodRe factors. Molecular ions are cationizated with the
sodium ion (NaI doped matrix). B, positive ion fast atom
bombardment mass-analyzed ion kinetic energy spectrum of the protonated
ion at m/z 1458.
When NodRe factors were treated with 0.5 N sodium methanolate prior to analysis, the fast atom
bombardment-mass spectrum in the positive ion mode showed
(M+H) The
fragmentation series ended at m/z 440 (or m/z 483), which was attributed to oxenium ions from
the first sugar residue at the non-reducing end. m/z 440 was attributed to a N-methyl-N-vaccenoyl-D-glucosamine (20) . m/z 483, which was 43 mass units
higher, indicated the presence of an additional carbamoyl substitution.
Thus, NodRe factors are a mixture of carbamoylated and
non-carbamoylated molecules, the former representing one-quarter to
one-third of the total molecules. All NodRe factors bore a L-fucose residue linked to the terminal reducing N-acetyl-D-glucosamine together with an additional O-acetyl group.
NaBD
Figure 3:
Positive ion fast atom bombardment mass
spectrum of the NaBD
The partially methylated monosaccharides were
reduced with NaBH
Figure 4:
Two-dimensional homonu-clear (
For the D-glucosamine residues, The wild type strain (CFN 42) of R. etli produced
enough Nod factors to enable structural studies without the need of
amplifying the Nod factors production by genetic engineering. The
structural characteristics of the R. etli nodulation factors
are as follows: (i) they have a chitopentameric core; (ii) a
4-O-acetyl-
Figure 5:
A, structural features of the NodRe-V (Me,
Ac-Fuc) and NodRe-V (Carb, Me, Ac-Fuc) factors. B, structure
of the NodRt-V (Me, S) and NodRt-V (Me)
factors.
The structure of NodRe
factors represents a new variation on the Nod factors structure. L-Fucose, as a substituent of chitooligomers, has been
previously found as a minor component of Nod factors from Rhizobium
fredii(8) and Bradyrhizobium
elkanii(16) . However, none of these structures bears an
additional acetyl group on L-fucose. An L-fucose ring
with an acetyl group on O-4 has been found in Rhizobium NGR234, but an O-methyl group was also present on fucosyl
O-2(19) . Carbamoyl and N-methyl groups on the
non-reducing end are more common because they have been found in Rhizobium NGR234(19) , Azorhizobium caulinodans(15) , Bradyrhizobium japonicum(17) , and B. elkanii(16) . The precise location of the carbamoyl
group needs a sufficient amount of material to perform heteronuclear Common bean is also nodulated by R. tropici, a
bacterium with a broader host range than R. etli. When
comparing Nod factors from both bacteria, no common substituents can be
found as NodRt factors are non-fucosylated, non-carbamoylated molecules (20) . Instead, part of them possesses a sulfate group on O-6
of the reducing N-acetyl-D-glucosamine. It has been
found that these sulfated molecules are able to elicit root nodule
organogenesis on bean at 10 Purified Nod factors from R. etli were also found to elicit root nodules on bean in the same
concentration range as sulfate-containing NodRt factors. ( It
thus appears that bean is able to be nodulated by two rhizobial species
that seem to produce Nod factors with structurally different
substitutions. However, purified fractions of Nod factors from
either R. tropici or R. etli seemed to be active on
bean at concentrations higher than those used to induce nodules on
alfalfa by NodRm factors, for example. It remains possible that some
sulfated components, being under the detection level, still contaminate
the fucosylated fractions from NodRe factors, that fucosylated
molecules are present in R. tropici, or that very low
proportions of another common component are present in all tested
fractions. With reversed-phase HPLC, sulfated factors of R.
tropici and acetylfucosylated factors of R. etli have
very different retention volumes. Thus cross-contaminations with an
identical component seemed unlikely. However, because reversed-phase
HPLC separations are very sensitive to the hydrophilic/hydrophobic
balance of the analytes, it cannot be excluded that a minor structural
variation having a weak effect on the activity (such as the acyl chain
length) may induce a dramatic effect on the HPLC retention volumes and
thus permit the elution of analogs of cross-contaminants in the
different fractions. We are currently reinvestigating both strains
to search for impurities that could be a common signal for bean
nodulation.
Volume 270,
Number 11,
Issue of March 17, 1995 pp. 6050-6055
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
)(2, 4, 5) .
Bacterial Strains and Cell Cultures
R. etli strains CFN 42 (wild type strain) and CFN 2001 (wild type strain
cured of its symbiotic plasmid) were grown in a minimal medium in the
same conditions used for R. tropici(20) . Induction of
the bacterial nodulation genes was achieved by using 1.5 µM naringenin (4`,5,7-trihydroxyflavanone, Sigma) or by adding bean
seed exudates that were prepared as follows.Preparation and Testing of Bean Seed
Exudates
After surface sterilization(23, 24) ,
20 g of bean seeds were germinated (in darkness at 28 °C) in 50 ml
of sterile water in a 2-liter Erlenmeyer flask stirred at 180 rpm. The
aqueous solution was collected after 8 h and replaced by 50 ml of
sterile water. This operation was repeated 5 times. The pooled aqueous
exudates were evaporated to dryness, and the residue was dissolved in
20 ml of ethanol. 50 µl, 200 µl, or 1 ml of this solution were
tested for its ability to induce Nod factors synthesis in 50 ml of R. etli radiolabeled cultures.Thin-layer Chromatography and Radioactivity
Assays
The experimental protocols for labeling Nod factors using
[
S]sulfate or [
C]acetate
were identical to those already described for R.
tropici(20) .Purification of NodRe Factors
4 liters of culture
medium were extracted twice, first with 1 liter and then with 500 ml of
butanol. The butanol extract was vacuum evaporated to dryness. The
residue was dissolved in water and purified by HPLC on a
semi-preparative C
reversed-phase column (250 
7.5
mm, Spherisorb ODS2, 5 µm, ColoChrom), using a water/acetonitrile
gradient. Detection was done by UV absorbance at 220 nm. The Nod
factor-containing fraction was collected and purified again on an
analytical C reversed-phase column (250 4.6 mm,
Spherisorb ODS1, 5 µm, ColoChrom) using a linear gradient from
80:20, v/v to 40:60, v/v water/acetonitrile in 10 min at a 1 ml/min
flow rate.Analytical Methods
Gas chromatography analyses
were done on a Girdel 30 gas chromatograph equipped with an OV1 bound
capillary column (0.32 mm 30 m, Spiral France). The temperature
gradient was 2 °C/min from 100 to 280 °C.H NMR spectra were measured on a
Brücker AC-300 spectrometer (Karlsruhe, Germany)
using 2 mg of sample dissolved in 0.5 ml of perdeuterated dimethyl
sulfoxide (Sigma).Carbohydrate Determination and Methylation
Studies
Carbohydrate determination was carried out after
complete hydrolysis using 3 N HCl (3 h at 80 °C) followed
by extraction with diethyl ether to remove the fatty acids. Sugars from
the aqueous phase were derivatized as alditol acetates and analyzed by
capillary GC-MS. Alternatively, 2butyl glycosides of individual sugars
were prepared using 3 N HCl in (±)-2-butanol or
(-)-2-butanol (3 h at 60 °C). These glycosides were further
acetylated (acetic anhydride/pyridine mixture (1:1, v/v), 3 h at 40
°C) and analyzed by capillary GC. Permethylation analysis was done
according to the method of Ciucanu and Kerek (26) as described
for NodRt factors(20) . cartridge. After washing with water,
deacetylated NodRe factors were eluted with methanol.Fatty Acid Analysis
Fatty acids extracted from the
acid hydrolysis were either methylated by diazomethane and analyzed by
GC or derivatized as pentafluorobenzyl esters for negative ion GC
coupled to tandem mass spectrometry. Fatty acids (100 µg) were
dissolved into 60 µl of a mixture of dry methanol and acetonitrile
(1:5, v/v) followed by the addition of 2 µl of pentafluorobenzyl
bromide and 2 µl of diisopropylethylamine. After 1 h at room
temperature, reagents were removed by evaporation. The
pentafluorobenzyl esters were dissolved in cyclohexane for GC-MS
analysis (see ``Analytical Methods'') (25) .
Detection of NodRe Factors after nod Gene
Induction
R. etli (CFN 42) and the same strain cured of
the symbiotic plasmid (CFN 2001) were grown in the presence or absence
of naringenin with the addition of sodium
[C]acetate or sodium
[S]sulfate to the medium. Lipophilic compounds
were extracted on C reversed-phase material, eluted with
methanol, and analyzed by reversed-phase thin-layer chromatography.
Autoradiography showed that a spot at R
=
0.53 was detected only in the lane corresponding to an induced culture
of CFN 42 after sodium [C]acetate labeling. When
treated with chitinase, this compound disappeared and a new spot was
observed at R = 0.42, thus suggesting that
the compound has a chitooligomeric backbone. Thus, the compound with a R of 0.53 possesses the characteristics of Nod
factors: a chitooligomeric structure whose biosynthesis is dependent on
the presence of nod genes and flavonoid induction.Purification of NodRe Factors
4 liters of sterile
medium from naringenin-induced culture were extracted with butanol. The
butanol extract was fractionated by HPLC using two consecutive
purifications, first on a C reversed-phase HPLC
semi-preparative column and then on an analytical column; Nod factors
gave a double HPLC peak (Fig. 1). About 1.0 mg of NodRe factors
was purified from a 4-liter culture.
HPLC chromatograms of the butanol extracts of the culture media of the
wild type R. etli strain CFN 42 and the pSym
strain CFN 2001.
Induction of R. etli Nodulation Genes by Bean Seed
Exudates
Exudates of bean seeds in ethanol were used to induce a
50-ml culture of R. etli after C and S radiolabeling. Reversed-phase thin-layer chromatography
analysis showed the same result as with induction with naringenin: a
compound with a R
of 0.53, labeled only by growing
the bacteria in the presence of sodium
[C]acetate. Using a 4-liter culture induced with
bean seed exudates, the same molecules in the same proportions were
isolated as for the naringenin-induced culture.Constituent Analysis of NodRe Factors
After
complete acid hydrolysis with 3 N HCl for 3 h at 80 °C, N-acetylglucosamine, N-methylglucosamine, and fucose
were detected as water-soluble compounds. GC analysis of the
peracetylated(-)-2-butyl glycosides (27) assigned N-acetylglucosamine and N-methylglucosamine to the D series, whereas fucose was assigned to the L series. A single fatty acid was liberated by acid hydrolysis that
was identified as cis-vaccenic acid both by studying the
decomposition of its carboxylate anion by tandem mass spectrometry (25) and by comparison of the GC retention time of its methyl
ester with an authentic standard.Mass Spectrometry Analysis
In the positive ion
mode, the Nod factor-containing fraction exhibited two
(M+H)
ions at m/z 1458 and
1501. Cationization with the sodium ion resulted in two
(M+Na) ions at m/z 1480 and
1523, respectively (Fig. 2A). Metastable ion spectrum
from the (M+H) ion at m/z 1458 (Fig. 2B) showed the clear fragmentation of the N-acetylglucosamine backbone giving a series of peaks
separated by 203 mass units. The highest peak of this series at m/z 1049 was 409 mass units below the
(M+H) ion. The decomposition of m/z 1501 showed a corresponding series of fragment ions shifted up 43
mass units.
ions at m/z 1416 and
1459. However, all fragmentations remained unchanged. Thus an acetate
group was present near the reducing end. The mass difference between
the first fragment ions (m/z 1049 or 1092) and the
(M+H) ions corresponded to the loss of both L-fucose and N-acetyl-D-glucosamine.Permethylation Studies
Permethylation studies of
NodRe factors were performed in order to confirm the oligochitin core
and to precisely locate the L-fucose residue.
reduction followed by permethylation produced a single compound
with a molecular mass of 1684 Da as determined by fast atom
bombardment-mass spectrometry (Fig. 3). This result indicated
that the acetyl and carbamoyl (28) groups had been lost during
the chemical process and that all the hydroxyl and amide groups had
been methylated.
-reduced and permethylated NodRe
factors. The molecular ion is cationizated with the sodium ion (NaI
doped matrix).
and peracetylated. The partially
methylated alditol acetates were then analyzed by GC-MS. Four compounds
were identified:
1,5-di-O-acetyl-3,4,6-tri-O-methyl-N-acetyl-N-methylglucosaminitol
derived from the glucosamine residue at the non-reducing end of the
molecules;
1,4,5-tri-O-acetyl-3,6-di-O-methyl-N-acetyl-N-methylglucosaminitol
arising from the three internal residues;
4,6-di-O-acetyl-1,3,5-tri-O-methyl-N-acetyl-N-methylglucosaminitol
produced from the reducing glucosaminyl residue; and
1,5-di-O-acetyl 2,3,4-tri-O-methylfucositol provided
by the L-fucose residue. Therefore, this study confirmed the
(1
4) glycosidic linkages between the five N-acetyl-D-glucosamine residues of the NodRe factors
and indicated a (16) glycosidic linkage between the L-fucose residue and the reducing N-acetyl-D-glucosamine.
The two-dimensional
proton homonuclear spectrum of NodRe factors was used both to locate
the acetyl group and to indicate the anomeric configuration of the
sugar linkages. The anomeric proton of the L-fucose residue
showed a doublet at 5.05 ppm with a coupling constant of 3.7 Hz with
the H-2 proton of the cycle. These data were consistent with an
H NMR Analysis
-linked L-fucose(12, 29) . From this
anomeric proton, a complete assignment of the L-fucosyl
protons was obtained on the two-dimensional spectrum (Fig. 4).
The cross-peak observed from H-1 enabled the location of the H-2 signal
at 3.80 ppm. This proton showed a scalar connection with H-3 at 3.58
ppm (J
= 9 Hz). Connectivity
was also observed through H-3 and H-4 protons (J
= 3.9 Hz), thereby locating
the resonance of H-4 at 4.75 ppm. This downfield signal (more than 1
ppm over the expected value for a non-substituted fucose residue) (12) enabled the acetate group to be located on the O-4 of the
-L-fucose residue. Because of the low coupling constants J
in the L-fucose
residue (0.8-0.9 Hz)(29) , the resulting low amplitude
connectivity peak could not be seen on the
spectrum(12, 30) . The three 6-deoxymethyl protons
(
= 0.95 ppm) correlated with the H-5 at 4.05
ppm.
H-H) NMR spectrum of the NodRe
factors.
(J
= 3.5 Hz at the
reducing end) and 
(J = 9
Hz) anomeric configurations were characterized. Other signals,
classically described for the NMR spectra of nodulation factors, were
also identified (upfield signals not shown).
-L-fucose residue is linked to carbon
6 of the reducing N-acetyl-D-glucosaminyl unit; (iii)
the glucosaminyl residue at the non-reducing end of the molecules is N-methyl-N-vaccenoyl substituted; and (iv)
one-quarter to one-third of these symbiotic signals is O-carbamoylated at their non-reducing end. According to the
nomenclature proposed by Roche and co-workers(31) , we name
these factors NodRe-V (Me, Ac-Fuc) and NodRe-V (Carb, Me, Ac-Fuc),
respectively (Fig. 5A).
H-
C bidimensional NMR. When we tried to
cultivate R. etli on a larger scale (20 liters) to purify
enough NodRe factors for this analysis, we found that the proportion of
carbamoylated molecules was lower than in the first experiments, and
thus the position of the carbamoyl group could not be reliably
assigned.
to 10
M(14) .
)When cultivating R. etli using bean extracts as nod gene inducers, no change in the NodRe factors pattern was
found. In particular, no NodRe factors with a sulfate group could be
detected by sodium [S]sulfate labeling.
))
We thank M. A. Rogel for technical assistance and M.
Bon and M. Vedrenne for recording NMR spectra. We are grateful to J.
Cullimore for helpful reviewing of the manuscript.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  W. D'Haeze and M. Holsters Nod factor structures, responses, and perception during initiation of nodule development Glycobiology, June 1, 2002; 12(6): 79R - 105R. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Laguerre, S. M. Nour, V. Macheret, J. Sanjuan, P. Drouin, and N. Amarger Classification of rhizobia based on nodC and nifH gene analysis reveals a close phylogenetic relationship among Phaseolus vulgaris symbionts Microbiology, April 1, 2001; 147(4): 981 - 993. [Abstract] [Full Text]
|
|
|
|
 |
|
 X. Perret, C. Staehelin, and W. J. Broughton Molecular Basis of Symbiotic Promiscuity Microbiol. Mol. Biol. Rev., March 1, 2000; 64(1): 180 - 201. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Berck, X. Perret, D. Quesada-Vincens, J.-C. Promé, W. J. Broughton, and S. Jabbouri NolL of Rhizobium sp. Strain NGR234 Is Required for O-Acetyltransferase Activity J. Bacteriol., February 1, 1999; 181(3): 957 - 964. [Abstract] [Full Text]
|
|
|
|
 |
|
 S. Philip-Hollingsworth, G. G. Orgambide, J. J. Bradford, D. K. Smith, R. I. Hollingsworth, and F. B. Dazzo Mutation or Increased Copy Number of nodE Has No Effect on the Spectrum of Chitolipooligosaccharide Nod Factors Made By Rhizobiumleguminosarum bv. trifolii J. Biol. Chem., September 8, 1995; 270(36): 20968 - 20977. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |