Tải bản đầy đủ - 0trang
Chapter 2. Population Diversity Groupings of Soybean Bradyrhizobia
JEFFRY J. FUHRMANN
aspects of the soybean symbiosis. Although the inherent complexity of the
soil - soybean environment has certainly contributed to this impasse, ecological research has also been hampered by a persistent uncertainty of how
best to assess the symbiotic diversity present within populations of B.
juponicum. The literature abounds with studies demonstrating diversity
among strains of B. juponicum, but relatively few studies have assessed the
potential symbiotic significance of the resulting groupings. In a recent
review, Bottomley (1992) emphasized that researchers need to develop a
unified approach to the study of rhizobial ecology so as to facilitate comparison and synthesis of experimental results. Clearly, the identification of
coherent strategies for the characterization of indigenous populations of B.
japonicum would constitute a major step toward achieving this goal for the
The objectives of this review are to summarize the present literature
characterizing population diversity in B. juponicum and to draw attention
to established or possible relationships among the various diversity groupings, particularly with respect to the symbiotic performance of the associated strains. Apparent gaps in current knowledge will also be noted as
appropriate. Emphasis will be placed on characteristics that may be considered traditional by current standards, but this simply reflects the relative
depth of the scientific literature involving these groupings. For the same
reason, discussion will primarily address research conducted in North
America, although the literature that is available from other geographical
regions suggests that many of the relationships to be discussed are generally
11. GENOTYPIC GROUPINGS
Research has clearly established that the species B. juponicum is composed of at least two major genetic groupings. Hollis et ul. (198 1) identified
two major DNA homology groups based on DNA - DNA hybridization
studies, and proposed that the species B. juponicum (then Rhizobium
juponicum) be reserved for one of these groups (I/Ia) and that the second
group (11) represented a distinct species. Two highly divergent groups,
consistent with those proposed by Hollis ef ul. (1981), have also been
observed based on molecular genetic studies (Minamisawa, 1990; Stanley
el ul., 1985). In a recent report, Kuykendall ef ul. (1992) analyzed a
number of B. juponicum strains using DNA hybridization probes and
confirmed the groupings proposed by Hollis ef ul. (198 1).
SOYBEAN BRADY RHIZOBIA
As documented in subsequent sections of this review, there is mounting
evidence that the two major DNA homology groupings are highly correlated with many characteristics of soybean bradyrhizobia of potential importance to symbiotic performance. These strong correlations also appear
to permit one to infer the DNA homology grouping to which a given strain
of B. japonicurn belongs based on certain readily obtained phenotypic
characteristics [e.g., Fuhrmann (1990) and Kuykendall et al. (1988)l.
Therefore, the use of phenotypic traits to assign soybean bradyrhizobia
indirectly to established genetic groupings may represent a valuable strategy for initiating characterizations of unidentified soybean bradyrhizobia,
such as those comprising indigenous soil populations.
111. PHENOTYPIC GROUPINGS
Serological reaction is probably the most commonly encountered means
of characterizing soybean bradyrhizobia, particularly for surveys of indigenous B. japonicurn. Similarly, it is often conveniently used as a basis for
comparison with other phenotypic characteristics. The reader is directed to
Vincent ( 1982) for a general introduction and historical perspective on the
use of serological techniques in the Rhizobiaceae.
1. Diversity of Serological Phenotypes
Serological analyses are primarily used to provide empirical groupings of
bradyrhizobia having value for reference purposes. Therefore, serogroups
of B. japonicurn are typically characterized according to their reactions
with antisera produced against strains having an extensive research history.
In North America, the serogroups most commonly referred to are those
derived from strains present in the Rhizobium Culture Collection maintained by the United States Department of Agriculture (Beltsville, Maryland) (Keyser and Griffin, 1987). Much of the present-day framework for
the serological study of soybean bradyrhizobia was established by Damirgi
et al. (1967), Date and Decker (1965), Johnson and Means (1963), and
Koontz and Faber (196 1).
A number of serological methods are currently used to characterize
soybean bradyrhizobia. The most common of these are agglutination
(Date and Decker, 1965; Means et al., 1964; Wollum, 1987), fluorescent
JEFFRY J. FUHRMANN
antibody techniques (Bohlool, 1987), and various forms of the enzymelinked immunosorbent assay (Asanuma et al., 1985; Ayanaba et al.. 1986;
Fuhrmann and Wollum, 1985; Kishinevsky and Jones, 1987). Polyclonal
antisera have been used almost exclusively, although the successful use of
monoclonal antibodies with soybean bradyrhizobia was reported by Velez
er al. ( 1988). Agglutination is historically the most prominent of the three
methods and can probably be considered the standard for comparison. In
reality, however, method selection is based primarily on familiarity, and
there are no published studies providing critical comparisons of the common serological techniques.
Serological studies of indigenous B. juponicum have revealed considerable diversity within and among geographical locations. It has been possible in some instances to correlate the presence of particular serogroups
within a restricted region to soil properties such as pH (Damirgi er al.,
1967; Ham ef al., 1971) or total nitrogen (Bezdicek, 1972). On a regional
basis in the United States, strains aligning with serogroup 123 are especially
prominent in the upper midwest states (Damirgi et al., 1967; Ham et al.,
1971; Kamicker and Brill, 1986; Keyser er al., 1984; Weber el al., 1989),
whereas strains correlating to serogroups 3 1,76, and 94 are common in the
southern and mid-Atlantic states (Caldwell and Hartwig, 1970; Fuhrmann,
1989, 1990 Keyser el al., 1984; Mpepereki and Wollum, 1991; Weber et
al., 1989). Although the prevalence of serogroups 31 and 123 may be a
result of their use in early soybean inoculants, there is no known historical
explanation to account for the high frequencies of serogroups 76 and 94
(Weber et ul., 1989).
2. Correlation with Symbiotic Performance
One common goal of serological characterizations of rhizobia is to identify groups that have practical significance to the management of a particular symbiosis. However, Vincent (1982) cautioned that serological results
should not be indiscriminately extrapolated to other properties such as
symbiotic effectiveness, and warned against allowing the distinction between a serogroup and the corresponding serotype strain to become
blurred. Studies have revealed significant variation within standard strains
of B. japonicum maintained at different laboratories, further suggesting
that generalizations based on serotype strains should be approached with
caution (Mullen and Wollum, 1989). Yet, although many studies have
documented serological diversity within rhizobial populations, relatively
few have assessed the value of the resulting groupings in predicting symbiotic performance.
One common problem with using serology to characterize soybean bradyrhizobia is the presence of strains that are nonreactive with all antisera
tested. The frequency of nonreactive strains is often significant, particularly in soils from the southern and mid-Atlantic United States (Fuhrmann, 1990; Mpepereki and Wollum, 1991). Moreover, there is evidence
that rhizobia in this grouping can be symbiotically diverse (Fuhrmann,
1990). There is clearly a need to identify and document additional reference strains for characterizing indigenous soybean bradyrhizobia.
Serological analyses commonly reveal strains that cross-react with antisera derived from two or more reference strains. The best documented
example of this is the suite of soybean bradyrhizobia that constitute serocluster 123 (serogroups 123, 127, and 129)(Schmidt et al., 1986). Although
related serologically,the serogroups comprising this serocluster are known
to exhibit physiological and symbiotic diversity (Gibson et al., 1971;
Hickey et al., 1987; Sadowsky et al., 1987). Given the possibility that
additional seroclustersexist, currently recognized serogroups may inadvertently serve to mask significant diversity among member strains.
Studies have documented large differences in symbiotic effectiveness
among soybean bradyrhizobia in a single serogroup. Basit et al. (199 1)
found threefold differences in shoot weights and N contents among 37day-old soybean plants nodulated by 34 serogroup 1 10 strains. Other work
found that the mean effectiveness of indigenous isolates within serogroups
was inconsistently related to that of the corresponding serotype strain
(Fuhrmann, 1990). In the later study, grouping the indigenous isolates
according to their reactions with 12 antisera accounted for only 59% of the
variation in N contents of 42day-old soybean plants (Fuhrmann, 1990).
Certain soybean genotypes are known to restrict nodulation by particular serogroups or strains of soybean bradyrhizobia (Caldwell, 1966; Caldwell et al., 1966; Vest, 1970; Vest and Caldwell, 1972), although there is
recent evidence that some of these reactions may be only coincidentally
related to serology (Devine et al., 1991). Host genotype-specificnodulation
may have value in altering nodulation competition by displacing undesirable indigenous strains from root nodules in exchange for more effective
indigenous or inoculant strains (Caldwell and Vest, 1968; Devine and
Breithaupt, 1980a; Ishizuka et al., 1991a,b; Kvien et al., 1981; Weiser et
al., 1990). In particular, this possibility has been explored as a means of
developing soybean genotypes that exclude strains belonging to serocluster
123 (Cregan and Keyser, 1986; Cregan et al., 1989a,b; Sadowsky et al..
1987; Schmidt et al., 1986). Indigenous strains in serocluster 123 are
generally considered to be relatively ineffective in N2 fixation, although
support for this contention comes primarily from research conducted with
JEFFRY J. FUHRMANN
the serotype strain USDA 123 (Caldwell and Vest, 1970) or other indirect
evidence (Kvien et af., 1981).
3. Relationship to DNA Homology Groupings
A strong correlation exists between DNA homology groupings in B.
japonicum (Hollis et af., 198 1) and their corresponding serogroups. Documented serogroups in group I/Ia include 6, 38- 1 15, 58, 62, 1 10, 122, and
123, whereas group I1 includes 31, 46, 76, 86, 94, and 130. Additionally,
Keyser and Griffin (1987) listed three serogroups (4, 125 - 126, and 135)
which have apparently not been analyzed as to DNA homology. Keyser
and Griffin (1987) also listed two cross-reaction groups: 1 10- 1 15, in which
both serotype strains are in group I/Ia, and 76- 123, in which the serotype
strains are in groups I1 and I/Ia, respectively. In addition to this 76- 123
grouping, Weber et al. (1989) reported two cross-reaction serogroups (6276 and 94 - 1 10) derived from serotype strains belonging to both group I/Ia
and 11. These appear to be the only reported serological groupings that
bridge the DNA homology groups. It is noteworthy that only group I/Ia
contains serotype strains considered to be highly effective (Keyser and
B. INTRINSIC ANTIBIOTIC
Numerous studies have used intrinsic antibiotic resistance (IAR) as a
means of characterizing rhizobia, including soybean bradyrhizobia, taken
from both established culture collections and indigenous populations
(Eaglesham, 1987). Certainly one reason for its common use is the relative
ease with which IAR can be determined, especially when the bradyrhizobia
of interest are available as pure cultures. The reader is directed to Eaglesham (1987) for general information concerning the use of IAR in rhizobial research.
1. Diversity of
As with serological analyses, characterization of bradyrhizobia by IAR is
used primarily to obtain empirical reference groupings that can be compared with other traits of interest. Although Eaglesham (1987) described a
number of methods for determining IAR, a review of the literature shows
that the one most commonly used is the agar dilution technique in which
antibiotic-amended plates are spotted with cultured bradyrhizobia and
later examined for growth.
Approaches to the use of IAR to characterize bradyrhizobia vary greatly
among laboratories. Most studies have employed several individual antibiotics, each at a single concentration, that concentration being derived from
preliminary studies with a representative subsample of strains [e.g., Hickey
et al. (1987), Kuykendall et al. (1988), and Thompson et al. (1991)l.
Others have used several individual antibiotics, each over a range of concentrations (Cole and Elkan, 1979; Sawada et al., 1990; Young and Chao,
1989), a single antibiotic at many different concentrations (Pankhurst et
al., 1982), or selected combinations of antibiotics (Mueller et al., 1988).
Comparison of IAR results among studies is further complicated by a
lack of standardization of antibiotics employed. The more commonly used
antibiotics for soybean bradyrhizobia are kanamycin, nalidixic acid, rifampicin, and streptomycin, although carbenicillin, chloramphenicol,
erythromycin, neomycin, novobiocin, penicillin, polymyxin, spectinomycin, tetracycline, and vancomycin have also been used by several investigators. Eaglesham (1987) noted that the diversity detected within a given
population is often positively correlated with the number of antibiotics
against which the population is tested. For similar reasons, it may be
desirable to select antibiotics that differ in their modes of antimetabolite
action (Sinclair and Eaglesham, 1984).
Large differences also exist among studies with regard to the concentrations used for a particular antibiotic. It is not uncommon for antibiotic
concentrations to differ by more than 10-fold among studies [e.g., Hickey
et al. (1987), Kuykendall et al. (1988), Mueller et al. (1988), and Thomp
son et al. (1991)l. Because inhibition by antibiotics is concentration dependent, it is often unclear whether apparent differences in IAR among
studies represent differences in methodology, actual population differences, or a combination of the two.
Despite the difficulties noted above, a number of studies have demonstrated unambiguous differences among populations by means of IAR. In
particular, Mueller et al. (1 988) determined the IAR of nodule bacteroids
taken from eight cultivars and three locations in South Carolina and found
considerable differences in IAR patterns among the sampling combinations. This study also concluded that, overall, the native bradyrhizobia
displayed high resistance to streptomycin and low resistance to rifampicin,
kanamycin, and nalidixic acid in various combinations. In contrast, bradyrhizobia indigenous to Iowa soils were shown to be very susceptible to a
much lower concentration of streptomycin than that used in the South
JEFFRY J. FUHRMANN
Carolina study (15 versus 500 pg/ml) (Hickey ef al., 1987). Similarly,
Thompson ef al. ( 199 1) reported that bradyrhizobia isolated from soybean
in Thailand were much less resistant to several antibiotics than were the
United States populations studied by Mueller et al. (1988).
2. Correlation with Symbiotic Performance
Possible relationships between IAR phenotype and symbiotic performance have not been adequately explored to allow for meaningful generalizations to be made. Particularly lacking are studies in which IAR group
ings of bradyrhizobia have been subsequently tested for their associated
N,-fixing abilities. Thompson ef al. ( 199 1) concluded that IAR was useful
for separating isolates for subsequent effectiveness testing, but did not
attempt to correlate the two characteristics. However, as described in the
subsequent section, there is evidence that IAR groupings are correlated to
DNA homology in soybean bradyrhizobia (Kuykendall ef al., 1988). Thus,
to the extent that DNA homology is correlated with symbiotic effectiveness, groupings identified by IAR may indirectly indicate N, fixation
The presence of IAR in soybean bradyrhizobia raises the question of its
possible ecological role, especially with respect to saprophytic survival and
nodulation competitiveness. Eaglesham ( 1987) cites a number of studies
that provide indirect support for the hypothesis that IAR may represent an
ecological advantage in many soil systems. Here again, however, there is an
obvious need for additional studies investigating the role of IAR in the
dynamics of bradyrhizobial populations.
3. Relationship to DNA Homology Groupings
Kuykendall ef al. (1988) reported a nearly perfect correlation between
the DNA homology groupings of Hollis ef al. (198 1) and intrinsic resistance to a suite of antibiotics. In general, group I1 strains exhibited high
levels of resistance to carbenicillin ( 5 0 0 pg/ml), chloramphenicol (500
pg/ml), erythromycin (250 pg/ml), nalidixic acid (50 pg/ml), rifampicin
(500 pg/ml), streptomycin (100 pg/rnl), and tetracycline (100 pg/ml),
whereas strains in group I/Ia were susceptible. Support for these findings
can be found in the work of Pankhurst et al. (1982) in which B. japonicum
strains in group I1 (based on serological correlations) were much more
resistant to rifampicin than were strains from group I/Ia. Kuykendall ef al.
(1988) concluded that differentiation based on IAR profiles should prove
valuable in genetic and phenotypic characterizationsof soybean bradyrhizobia, including studies of indigenous populations.
The reduction of protons to hydrogen gas is an integral component of
biological N, fixation (Arp, 1992; OBrian and Maier, 1988):
+ 8H+ + 8e- + 16ATP
+ 16ADP + 16Pi + H,
Although the stoichiometric relationship between N, reduced and H, produced can vary, it is not possible to eliminate entirely energy allocation to
proton reduction. The amount of energy diverted to H, evolution constitutes 25% or more of the total electron flux allocated to N, fixation (Evans
et ul., 1987). Therefore, H2 evolution potentially represents a substantial
inefficiency in the N, fixation process.
Certain rhizobia, including some soybean bradyrhizobia, possess a hydrogenase system that catalyzes the oxidative release of energy from H,:
Rhizobia exhibiting this capability have traditionally been referred to as
hydrogen uptake positive (Hup+) and often produce root nodules that
show a complete lack of H, evolution. This recycling of H, and energy is
generally thought to increase the efficiency of N, fixation, and it has been
suggested that the Hup+ phenotype is one criterion that may directly
indicate superior symbiotic effectiveness of soybean bradyrhizobia.
1. Diversity of Hydrogenase Phenotypes
In reviewing the literature regarding hydrogenase phenotypes in soybean
bradyrhizobia, it is important to consider the methods used to detect
hydrogenase activity. Soybean bradyrhizobia have traditionally been categorized simply as Hup- or Hup+ by means of either in vitro or in vivo
techniques. However, recent research has demonstrated the presence of an
additional host-regulated phenotype (Huphr) in which hydrogenase activity is only exhibited by bacteroids while in symbiosis with cowpea [ Vigna
unguiculuta (L.) Walp.] or certain soybean genotypes (van Berkum, 1990;
van Berkum and Sloger, 1991). Differential expression of hydrogenase
between soybean and cowpea had been previously reported for soybean
bradyrhizobia (Keyser et al., 1982). Thus, there are currently three documented hydrogenase phenotypes in soybean bradyrhizobia: Hup +,hydro-
JEFFRY J. FUHRMANN
genase activity expressed in vitro and in vivo with common North American soybean genotypes; Huphr, hydrogenase activity exhibited only in
vivo with certain host genotypes; and Hup-, lack of demonstrable hydrogenase activity under any of the above conditions.
Many studies of soybean bradyrhizobia have relied solely on in vitro
methods for the detection of hydrogenase activity. Some early studies
employed a tritium exchange method (Lim, 1978), but there is evidence
that this method may erroneously indicate the presence of an active hydrogenase system in Hup- strains (Robson and Postgate, 1980; van Berkum et af., 1985). A common alternative technique monitors H, disap
pearance from the headspace of cultures grown either chemotrophically or
heterotrophically under hydrogenase-inducing conditions (van Berkum,
1987). Although there is no evidence that this latter method produces false
positive results, it would apparently fail to detect those bradyrhizobia
possessing the Huphr phenotype (van Berkum, 1990; van Berkum and
Although relatively cumbersome and time-consuming when compared
with in vitro techniques, assays employing intact nodules or bacteroid
suspensions are able to yield definitive information concerning the hydrogenase activity of a particular host -strain combination. Conversely, assays
conducted with a single host genotype, or in the absence of an accompanying in vitro assay, are unable to distinguish between the Hup+ and Hup-hr
phenotypes. Evidence to date, however, indicates that expression of the
Huphr phenotype in soybean may be limited to symbioses formed with
hosts genetically distant from common North American soybean lines (van
Berkum and Sloger, 1991).
A number of studies have examined the frequency of Hup+ strains in
indigenous populations of soybean bradyrhizobia. Early studies assayed
bradyrhizobia taken from nodules collected from throughout much of the
eastern and midwestern United States (Keyser ef al., 1984; Lim et af.,
1981; Uratsu et al., 1982). These studies concluded that the frequency of
Hup+ strains ranged from 20 to 25% overall and from 0 to 90% on a
regional basis. All of these studies were conducted using the tritium exchange method (Lim, 1978) and, therefore, may have overestimated the
frequency of Hup+ phenotypes (van Berkum et al., 1985).
More recently, a survey of soybean bradyrhizobia indigenous to Delaware indicated that 14 of 92 nodule isolates examined (1 5%) were Hup+
(Fuhrmann, 1990). Hydrogenase activity was determined in vitro using a
modified heterotrophic growth technique (van Berkum, 1987) and, therefore, would not have detected the presence of the Huphr phenotype.
Sawada et al. (1989) found that nodules formed by 34% of 85 isolates of
bradyrhizobia indigenous to Japanese soils evolved little or no H, in
symbiosis with soybean, thereby indicating the presence of an active hydrogenase.
2. Correlation with Symbiotic Performance
The literature contains conflicting evidence regarding the symbiotic effect of hydrogenase activity by nodulating rhizobia. Arp (1992) discussed
the mechanisms by which it is thought that the presence of an active Hup
system is beneficial to N,-fixing symbioses:
1. The energy released by H, oxidation can be recycled and used to
produce supplemental ATP for nitrogen fixation or other cellular processes
(Emerich et al., 1979).
2. In aerobic N,-fixing systems such as those involving rhizobia, recovery of the energy released by H, oxidation is coupled to 0, consumption
via the electron transport chain. This 0, consumption may serve to help
protect the nitrogenase enzyme from 0,-induced inactivation (Emerich et
3. Hydrogen oxidation may reduce H,-induced inhibition of nitrogenase by decreasing the partial pressure of the gas in the nodule (Rasche and
Additionally, it has been suggested that Hup+ nodules may have a greater
functional longevity than Hup- nodules (Zablotowicz et al., 1980) and
that the Hup+ phenotype may be an ecological advantage to soybean
bradyrhizobia existing chemoautotrophically in low 0, environments,
such as aggregate interiors (Ozawa et al., 1989; Viteri and Schmidt, 1989).
Many studies have investigated the symbiotic response of the soybean
plant to nodulation by wild-type or nonisogenic mutant strains differing in
Hup phenotype. These studies have reported both beneficial (Albrecht et
al., 1979; Hanus et al., 1981; Zablotowicz et al., 1980) and neutral (Basit et
al., 1991; Hume and Shelp, 1990; Kimou and Drevon, 1989) responses to
nodulation by Hup+ bradyrhizobia relative to Hup- strains. In a greenhouse study, Fuhrmann (1990) found that the Hup+ phenotype was either
beneficial or neutral to the soybean symbiosis, depending on the accompanying colony morphology of the isolate under examination. This same
study found that hydrogenase phenotype was a relatively poor indicator of
the symbiotic effectiveness of indigenous soybean bradyrhizobia when
compared with groupings based on serology. However, interpretation of all
of these studies is hampered by possible confounding effects of other
differences between the two groups other than hydrogenase phenotype.
JEFFRY J. FUHRMANN
Other investigations have employed isogenic strains of soybean bradyrhizobia differing only in Hup expression so as to eliminate possible confounding effects. The weight of these experiments indicates that hydrogenase activity is beneficial to the symbiosis (Arp, 1992; Evans et al., 1987;
Hungria et al., 1989), although detrimental effects were observed in a
short-term study using hydroponic plant culture (Drevon el al., 1987).
Evans et al. (1987) proposed a number of precautions that should be
observed in studies of hydrogenase effects: strains isogenic except for Hup
phenotype should be used, the Hup+ phenotype should be stable and
vigorous, the hydrogenase should be effectively coupled to ATP formation,
and the plants should be grown to maturity under normal environmental
3 . Relationship to DNA Homology Groupings
Hydrogenase phenotype in soybean bradyrhizobia has been shown to be
strongly correlated to DNA homology groupings. Essentially all Hup+
strains appear to belong to group I/Ia, although not all strains in this
grouping are Hup+ (Fuhrmann, 1990; Minamisawa, 1990). One apparent
exception to this is a Hup+ strain from serogroup 31, which would be
expected to align with group I1 based on its serological classification (Hollis
et al., 1981; van Berkum et al., 1985; van Berkum, 1990). In contrast, all
strains reported to exhibit the Huphr phenotype are from serogroups
correlating to group 11, but, again, not all group I1 strains have the Hup-hr
phenotype (van Berkum, 1990; van Berkum and Sloger, 1991).
Dissimilatory nitrate reduction (DNR) refers to the use of the anionic
nitrogen oxides (NO; and NO;) as alternative terminal electron acceptors
in the absence of oxygen, thereby permitting respiration to occur under
anaerobic conditions. Denitrification specificallyrefers to the dissimilatory
reduction of one or both anionic oxides to gaseous nitrogen compounds
(NO, N,O, or N2) (OHara and Daniel, 1985). Although denitrifying ability has been detected in other members of the Rhizobiaceae, its relatively
frequent occurrence in B. japonicum has resulted in a correspondingly
greater number of studies for the species. Additionally, many strains of B.
japonicum are capable of nitrate respiration, i.e., the reduction of NO? to
NO? without further reduction to gaseous products (OHara and Daniel,
1985). Although technically not denitrification, nitrate respiration is often