Genes and Signals in the Rhizobium-Legume Symbiosis1

Sharon R. Long*
Department of Biological Sciences and Howard Hughes Medical Institute, Stanford University, Stanford,
California 94305

Rhizobium-legume symbiosis begins with two free transmissible plasmid in Rhizobium leguminosarum vi-
living organisms, and ends with an intimate cellular ciae, and Barry Rolfe, Michael Djordevic, and Roger
co-existence. Rhizobium bacteria recognize specific Innes working in Canberra found a parallel situation
plants, provoke development of a root nodule, and in R. leguminosarum trifolii. The groups of Jean Dé-
invade the plant tissue. Eventually, the Rhizobium cell narié in Toulouse and Adam Kondorosi in Szeged
transfers itself into a host cell, surrounds itself with showed that clusters of symbiosis genes in Rhizobium
plant membrane, and arranges a nutrient exchange in meliloti were on incredibly large “megaplasmids,”
which the bacteria brings fixed nitrogen to the plant, over a million bases in size, an exciting discovery that
receiving in turn the sanctuary and sugars that the changed the concept of bacterial genome architec-
plant cell can provide (8,13,17,20). This historical note ture. By contrast, Hauke Hennecke’s group in Zurich
concerning 25 years of Rhizobium plant research will and Gary Stacey’s group in Tennessee defined Bra-
focus mostly on a few stories related to the discovery dyrhizobium japonicum nod genes on the chromosome.
of nod genes and signals. I refer the reader mostly to Bill Broughton and his group showed that Rhizobium
books and reviews for the details of these and of strain NGR234 had an astonishingly broad host
related research stories that are mentioned more range (over 18 genera, including one non-legume)
briefly in the latter part of this review. and with multiple host specificity genes dispersed
around a 500-kb plasmid. Since the cloning of nif and
nod in 1980 to 1981, over 30 different research groups
RHIZOBIUM GENES have contributed to our present understanding of
The Rhizobium-legume symbiosis had attracted se- Rhizobium symbiosis genes through physical cloning,
rious study ever since Beijerinck’s demonstration chromosomal walking, plasmid identification, site-
that bacteria caused nodule formation (13). Consid- directed mutagenesis, and many phenotypic studies
erable progress was made prior to 1975 in studying on diverse plant hosts (4,16,17). The rules of genome
the biochemistry of nitrogen fixation itself. The organization are different for diverse Rhizobium; in
mechanism of nodule formation, however, was the some cases, symbiosis genes are clustered, in other
subject of a great deal of speculation without much cases they are dispersed. In some cases, the genes are
concrete experimental proof. The critical first step on plasmids and can spread at high frequency by
turned out to be the identification of Rhizobium genes, conjugation; in others the genes are scattered among
rather than of plant components. Looking back, we many chromosomes and plasmids; and one case of
can see that experiments before 1980 had little chance symbiosis island transfer has been shown for Meso-
of success, because they analyzed free-living bacteria: rhizobium loti (18). With all of this genomic diversity
In these conditions, most Rhizobium symbiosis genes it is no wonder that systematists have had a field day
would have been silent, and thus symbiosis-related classifying and reclassifying the bacteria (16), some-
properties would have been cryptic. times to the bafflement of the molecular biologists
The first Rhizobium genes for nitrogen fixation (nif) studying the genetics of these species.
and for nodulation (nod) were cloned in the early As the story of nod genes and signals has unfolded,
1980s by Gary Ruvkun and by myself, respectively, described below, comparably deep and interesting
with our colleagues in Fred Ausubel’s laboratory stories have emerged in every aspect of the symbio-
(16), and soon many more nif, nod, and fix (symbiotic sis, and details can be found in a number of recent
fixation) genes were found in laboratories world- reviews. Graham Walker and his colleagues at MIT
wide. Allan Downie, Nick Brewin, and Andrew showed that genes for Rhizobium surface polysaccha-
Johnston at the John Innes Institute found that not rides are required for invasion, although not for early
just genes for nodule formation, but those for host nodulation or host specificity. Through their work
specificity, were tightly clustered with nif genes on a and that in other laboratories, a diverse set of such
components (extracellular polysaccharides, lipopoly-
1
This work was supported by the National Institutes of Health saccharides, and novel types of surface carbohydrate)
(grant no. GM30962), the Department of Energy (grant no. DE– are now known to be important, in some cases as
FG03–90ER20010), and the Howard Hughes Medical Institute. signals that undergo processing from the large Mr
* E-mail [email protected]; fax 650 –725– 8309. form (16). Bacterial genes for invasion and bacterial
Plant Physiology, January 2001, Vol. 125, pp. 69–72,
Downloaded from onwww.plantphysiol.org © 2001by
August 10, 2018 - Published American Society of Plant Physiologists
www.plantphysiol.org 69
Copyright © 2001 American Society of Plant Biologists. All rights reserved.
Long

differentiation have been found by direct and indi- flavonoid family of compounds, used by Rhizobium
rect screens (10,16). The bacteria show exquisite as a positive signal, is also the source of many legume
physiological adaptation to the low oxygen environ- phytoalexins, which raises some interesting co-
ment of the nodule, from the production of special- evolutionary questions. Our field’s view of how fla-
ized cytochromes, to the control of nitrogen fixation vonoid signals are used physiologically and ecologi-
genes themselves (6,16). Bacteria also may react to cally may expand, as the plant-microbe field moves
and manipulate host respiration in early stages of in its next era from focus on first-order effects, such
symbiosis via a novel signal, lumichrome (11). Car- as simple transcription activation, to the physiologi-
bon and nitrogen metabolism in the differentiated cal and ecological context (12).
nitrogen fixing bacteria shows amazing new adapta-
tions (16) including the production of “rhizopines”
that may supply specialized nutrition to sibling bac- NOD FACTORS: BACTERIAL CARBOHYDRATES
teria in the environment (16,20). WITH PLANT HORMONE ACTIVITIES
Are there more symbiosis genes to be found? This
is almost certain, especially in light of the elucidation How exactly does Rhizobium cause host-specific
of the first symbiosis plasmid sequence, revealing nodule development? The identification of the nod
new secretion systems and novel genes that respond genes and the elucidation of their regulation was a
transcriptionally to the host (7). The approaching key that unlocked an exciting new room of discover-
complete genome sequence of Rhizobium meliloti and ies: Now, it was possible to trigger symbiotic behav-
other species promises to reveal many more bacterial iors by bacteria grown in culture and to use wild-
genes required for invasion of and function within type versus Nod⫺ strains as controls. Over the period
the host. The big questions will remain: What do from 1986 to 1990, genetics, cell biology, and bio-
these genes do, and how are they regulated? The case chemistry came together to identify a completely
of the nod genes shows an example. new category of signal: the Nod factor (3,9,15). Con-
tributions of many groups, notably Ton van Brussel
and colleagues in Leiden, laid the groundwork show-
PLANT FLAVONOIDS: NEW SIGNALING ROLE ing Rhizobium exudates had effects on plants, de-
FOR A VENERABLE MOLECULAR FAMILY pending on Rhizobium nod gene content and expres-
sion. This came to fruition in 1990 with the work in
The identification of the Rhizobium nodulation Toulouse by the groups of J. Dénarié, G. Truchet, and
genes and subsequent study of their expression J.-C. Promé. Having observed that R. meliloti pre-
showed that nod genes were not expressed in free- treated with flavonoid inducer could cause alfalfa
living cells. My laboratory at Stanford, Allan Downie plants to display nodule-like behaviors, they frac-
and colleagues in the John Innes Institute, Ben tionated the Rhizobium medium and used careful mi-
Lugtenberg’s department at Leiden University, Barry croscopic bioassay of plant reactions to identify spe-
Rolfe at the Australia National University and John cific active fractions, or “Nod factors.” Chemical
Redmond at Macquarrie University in Australia, and analysis revealed the active component to be a novel
several other groups worked to find how these genes lipo-chito-oligosaccharide, based on a chitin oli-
were regulated. Our laboratories exploited reporter gomer backbone, and carrying a sulfate at the reduc-
fusions to show both that the nod genes required ing end of the oligomer. Ben Lugtenberg, Herman
plant inducers to be transcribed and that NodD ap- Spaink, and colleagues in Leiden and Utrecht next
peared to be the transcription activator. found that in R. leguminosarum viciae, host-specific
What exactly are the plant compounds that trigger modifications occur in an N-acyl group on the non-
nod gene expression? The availability of the Rhizo- reducing end residue.
bium nod-lacZ reporter fusions allowed bioassay of In the decade since 1990, a vigorous international
fractions separated by reverse phase chromatogra- enterprise led to isolation and characterization of
phy, followed by spectroscopic analyses to solve the Nod factors from many Rhizobium species. To give
structure of the natural compound. For example, readers an idea of the difficulty and scope of this
Kent Peters and I determined that the active fraction work, it has involved microbiologists, geneticists,
from alfalfa seed exudate was luteolin, a tetrahy- plant cell biologists, physiologists, biochemists, and
droxyflavone. By comparable approaches and by as- analytical chemists from Cuernevaca, Geneva, Gent,
say of available compounds the groups in Leiden, Gif-sur-Yvette, the John Innes Institute, the Univer-
Norwich, Canberra, and Melbourne found other sity of Georgia, Michigan State, the University of
plants export either flavones or flavanones, and the Missouri, Ohio State, Tennessee, and the University
team at Agrigenetics found that the soybean inducer of Utrecht with participation by a number of other
was daidzein, an isoflavone. From this and subse- groups supplying various wild-type and mutant Rhi-
quent work, it was found that each legume produces zobium strains. The outcome of this work showed that
a distinct cocktail of flavonoids and that the quantity diverse Rhizobium all produce Nod factors with a
and spectrum of compounds may vary with the age basic similar structure: a chito-oligo backbone with
and physiological state of the plant (4,12,16). The side groups that include novel modified sugars,
70 Downloaded from on August 10, 2018 - Published by www.plantphysiol.org Plant Physiol. Vol. 125, 2001
Copyright © 2001 American Society of Plant Biologists. All rights reserved.
 Rhizobium-Legume Symbiosis

acetyl or carbamoyl residues, and modified lipids. tion, and exchange and assimilation of nutrients. A
Bioassay on plant hosts demonstrated that the side few highlights follow with references to reviews that
groups provide host specificity for one plant or do better justice to this dynamic field and that point
another. to the primary contributions of many researchers
My laboratory was among several that took a com- whose names could not all be included here.
plementary approach: We asked what were the bio- In addition to morphogenesis itself, plant tran-
chemical activities of the nod gene products known scriptional responses to Rhizobium are striking, from
from previous analysis to be essential for symbiosis? the leghemoglobin genes first cloned by Desh Pal
This approach bore fruit the same year as Nod factor Verma and colleagues (20) to the early nodulins, or
identification, in 1990, with the demonstration by ENODs, identified by Ton Bisseling and colleagues
Julie Schwedock and myself of the first biochemical as Pro-rich sequences with possible cell wall loca-
function for a nod gene enzyme: nodP and nodQ en- tions. Researchers at many laboratories (a partial list
coded an enzyme that activated sulfate to its nucle- includes Wageningen, Versailles, Toulouse, Texas
otide form, APS. It was an exciting moment when we A&M, Sevilla, Ohio State, Minnesota, Leiden, the
all realized that the independent searches for bacte- John Innes Institute, Gif-sur-Yvette, Cuernevaca,
rial compounds with Nod factor activity on the one UCLA, Canberra, Bielefeld, and Aarhus) used both in
hand, and for functions of bacterial genes found only situ hybridization and transgenic plant constructs to
by phenotype on the other, had led to the same place: demonstrate the dramatic transcriptional response of
a molecule with a sulfate side group. In the past plant genes to Rhizobium signals and to correlate
decade, well over a dozen research groups (notably these with the developmental and metabolic changes
including the list above, and research groups in Köln, that characterize symbiotic interactions (2,16). Many
Leuven, and elsewhere) contributed to the demon- of those same research groups and others, including
stration that most Nod proteins have enzymatic ac- groups in Tennessee, Roskilde, Moscow, Marburg,
tivities such as polymerases and N-acyl transferases Dartmouth, and Adelaide, have shown how the sym-
(encoded by common nod genes), and O-sulfonyl, biosome compartment is constructed by targeting of
O-acetyl, N-methyl, and exotic glycosyl transferases novel plant proteins that control exchange between
(encoded by host specific nod genes). These activities the partners, such as the novel ammonium trans-
are consistent with the synthesis of the lipochitooli- porter discovered by Udvardi and Day and col-
gosaccharide Nod factors (for details, see 16). leagues (16,19).
With the conjunction of structural determination,
Cellular and tissue rearrangements were studied
bioassay, molecular genetics, and in vitro biosyn-
by microscopic, immunochemical, and biochemical
thetic proof, the Nod factor hypothesis (that nod
analysis: George Truchet in Toulouse, Nick Brewin at
genes encode the synthetic enzymes for host specific
the John Innes Institute, Jan Kijne and colleagues in
lipooligosaccharides) was solidified by the mid-
Leiden, Bob Ridge working in Australia and Japan,
1990s. A new star in the firmament of biochemical
and Kate VandenBosch and Doug Cook at Texas
signals had been discovered, and the fact that it was
such a new, unexpected chemical species was tre- A&M, among a number of other laboratories, found
mendously exciting within and outside of the plant changes in cytoskeletal architecture, cell wall bio-
research community. chemistry, and oxidative metabolism during infec-
But we now know that not all nod genes encode tion (5,16). David Ehrhardt and others in my labora-
enzymes with such activities, a clue that more signal tory discovered membrane depolarization and
surprises may await us. For example, NodO acts to calcium spiking in root hairs, and Hubert Felle in
form ion channels in membranes (4,5). Could the Giessen working with the Kondorosi laboratory at
bacteria be sending other signals in the early nodu- Gif-sur-Yvette, showed that fast ionic changes across
lation stages? Where do the exopolysaccharides fit the cytoplasmic membrane accompany alfalfa treat-
in? Are loci defined in genome projects important? ment by Nod factors (5). Research on calcium signal-
Our assays to detect plant responses now need to be ing, cytoskeletal dynamics, and other aspects of sig-
refined to detect subtle bacterial effects. nal transduction, are now being expanded at the
frontiers by colleagues in Cuernevaca, Leiden, Uni-
versity of Massachusetts, the John Innes Institute,
Toulouse, Wageningen, and beyond, as many new
BEYOND SIGNALS: CELL RESPONSES,
laboratories join this exciting search.
DEVELOPMENTAL BIOLOGY, AND PHYSIOLOGY
For all of these details of “what” happens during
My main narrative has concerned the discovery of nodulation, we are still in the dark about the “how.”
bacterial nod genes and the elucidation of signals that Literally dozens of laboratories, including veterans
control these symbiosis genes. This is the tip of the and newcomers, are now focused on the next set of
iceberg. Symbiosis researchers have used genetics, questions, and the writer of next year’s review will
molecular biology, cell biology, biochemistry, and have much to say about this fast-moving field. What
physiology to produce a wealth of information about signaling pathways do the plants use to transduce
bacterial and plant transcription, cellular organiza- Rhizobium Nod factors into such diverse responses?
Plant Physiol. Vol. 125, 2001 Downloaded from on August 10, 2018 - Published by www.plantphysiol.org 71
Copyright © 2001 American Society of Plant Biologists. All rights reserved.
Long

What is the fate of the Nod factor in the plants? From ture. In a new era of post-genomic study, it is not
Nod factor localization to Nod factor breakdown, only exotic chemistry and gene regulation but basic
from binding proteins that are novel lectins to bind- physiology that may provide many useful clues to
ing sites on plant cell membranes, new work on how follow the thread of molecular signaling, between or
Nod factors interact with plant cells is exciting and within organisms, in the complex fabric of plant
dynamic but still with many more questions than function.
answers (5,15). What happens after the Nod factor
finds its initial target? Possible downstream events LITERATURE CITED
have been inferred from inhibitor and pharmacolog-
ical studies (5,16). Plant hormones such as auxin and 1. Cook DR, Vandenbosch K, De Bruijn, FJ, Huguet T
cytokinin are likely to play a role downstream (8), (1997) Plant Cell 9: 275–281
and ethylene appears to have very early effects (1,16). 2. Crawford N, Leustek T, Kahn M, Long SR (2000) In BB
With many events, many correlations, and many Buchanan, W Gruissem, RL Jones, eds, Biochemistry
components, are these cause or consequence, signif- and Molecular Biology of Plants. American Society of
icant correlations, or minor side effects? To sort this Plant Physiologists, Rockville, MD, pp 786–849
out, our field looks to plant genetics as the key ap- 3. Dénarié J, Debellé F, Promé J (1996) Annu Rev Bio-
proach to sort out what is centrally important in chem 65: 503–535
nodule development (1). 4. Downie JA (1994) Trends Microbiol 2: 318–324
It is thus appropriate to end this bird’s eye research 5. Downie JA, Walker SA (1999) Curr Opin Plant Biol 2:
overview with one specific and exciting new ad- 483–489
vance, the cloning by Stougaard and colleagues of the 6. Fischer HM (1996) Trends Microbiol 4: 317–320
first plant nodulation gene, Lotus japonicus NIN-1, 7. Freiberg C, Fellay R, Bairoch A, Broughton WJ,
encoding a probable transcription factor required for Rosenthal A, Perret X (1997) Nature 387: 394–401
nodule morphogenesis (14). As genes are cloned that 8. Hirsch AM (1992) New Phytol 122: 211–237
correspond to various plant nodulation defects, it 9. Long SR (1996) Plant Cell 8: 1885–1898
will become possible to identify the essential steps in 10. Oke V, Long SR (1999) Curr Opin Microbiol 2: 641–646
plant recognition of its symbiont and to come to the 11. Phillips DA, Joseph CM, Yang GP, Martinez-Romero
evolutionary heart of the matter: why legumes? And E, Sanborn JR, Volpin H (1999) Proc Natl Acad Sci
thus also to the agronomically important corollary, USA 96: 12275–12280
the question asked is can the symbiosis be genetically 12. Phillips DA, Streit WR (1996) In G Stacey, NT Keen,
altered, extended, or improved. eds, Plant Microbe Interactions. Chapman & Hall, New
York, pp 236–271
PERSPECTIVE 13. Quispel ANN, editor (1974) The Biology of Nitrogen
Fixation. North-Holland Press, Amsterdam
The path of the past 25 years has led to elucidation 14. Schauser L, Roussis A, Stiller J, Stougaard J (1999)
of signal exchange primarily through genetic analysis Nature 402: 191–195
and analytical chemistry. But these experimental ap- 15. Schultze M, Kondorosi A (1998) Annu Rev Genet 32:
proaches would not have been possible without care- 33–57
ful and detailed studies of growth, metabolism, and 16. Spaink HP, Kondorosi A, Hooykaas PJJ (1998) The
cell and organ structure. As we look forward we Rhizobiaceae: Molecular Biology of Model Plant-
should recognize that not all signals will be detected Associated Bacteria. Kluwer Academic Publishers,
by following gene transcription changes and that the Dordrecht, The Netherlands
concept of “signal” should be considered in its broad- 17. Stacey G, Evans H, Burris R (1992) Biological Nitrogen
est sense. Consider what we already know: Rhizobium Fixation. Chapman & Hall, New York
evolved not only to detect flavonoids, but to sense 18. Sullivan JT, Ronson CW (1998) Proc Natl Acad Sci
oxygen and carbon dioxide, which are molecular USA 95: 5145–5149
gases of central metabolism. On the plant side, we 19. Udvardi MK, Day DA (1997) Annu Rev Plant Physiol
will doubtless find transcription factors and kinases Plant Mol Biol 48: 493–523
that trigger nodule development, but perhaps we 20. Verma DPS (1992) Molecular Signals in Plant Microbe
will also find genes that affect basic plant architec- Communication. CRC Press, Boca Raton, FL

72 Downloaded from on August 10, 2018 - Published by www.plantphysiol.org Plant Physiol. Vol. 125, 2001
Copyright © 2001 American Society of Plant Biologists. All rights reserved.