Journal of General Microbiology (1987), 133, 3451-3466.
Printed in Great Britain
3457
Plasmids and Symbiotic Effectiveness of Representative Phage Types from
Two Indigenous Populations of Rhizobium meliloti
By E. S. P. B R O M F I E L D , * N . P. T H U R M A N , S . T . W H I T W I L L A N D
L. R . B A R R A N
Plant Research Centre, Research Branch, Agriculture Canada, Ottawa, Ontario,
Canada K I A OC6
(Received 7 May 1987; revised 13 June 1987)
Phage types representative of the population of indigenous Rhizobium mezilotiat each of two sites
were evaluated for plasmid content by agarose gel electrophoresis and for symbiotic
effectiveness with Medicago sativa cv. Saranac. Relative to four strains used commercially, 55
and 65 phage types representing these populations showed a high average level of symbiotic
effectiveness; only a single type from one site was relatively ineffective in symbiosis. On the
basis of plasmid number and molecular mass, 160 isolates comprising 45 and 48 types from both
sites were placed in 22 different groups with 17 and 13 groups from the respective sites. The
number of plasmids varied between one and five per isolate with molecular masses ranging from
5 MDa to considerably greater than 267 MDa. Only five isolates lacked a plasmid with mobility
in agarose gels corresponding to that of a reference megaplasmid but instead showed a band of
lesser mobility and therefore greater molecular mass. Phage types, which were divided into
plasmid groups solely on the basis of differences between isolates from each site, may reflect
adaptation of R. meIiIoti to their respective sites. Differences between isolates within certain
phage types due to the presence or absence of a single plasmid May have resulted from genetic
interchange between indigenous R. meliloti. There was no significant correlation between
plasmid number or mass and symbiotic effectiveness or phage sensitivity of the phage types
from either site.
INTRODUCTION
Legumes sown in nitrogen-deficient soils are often inoculated with symbiotically effective
rhizobia with the aim of maximizing crop yield. However, the inoculant strains frequently
occupy only a minority of nodules of the host legume grown in soils harbouring indigenous
populations of rhizobia (Johnson et al., 1965; Roughley et al., 1976; Bromfield et al., 1986). In
these circumstances the amount of nitrogen fixed is largely dependent on the symbiotic
effectiveness of the resident Rhizobium population. Therefore, information on the nature of
these populations is of considerable practical significance.
The extrachromosomal DNA of most fast-growing Rhizobium spp. is known to play an
important role in the symbiotic process (Rolfe & Shine, 1984). A characteristic feature of
Rhizobium meliloti is the presence of plasmid DNA in amounts considered equivalent to that of
the chromosome of Escherichia coli (Prakash & Atherly, 1986). This is principally due to the
presence of megaplasmids, estimated to have molecular masses of about lo00 MDa (Burkhardt
& Burkhardt, 1984), in which genes affecting symbiotic functions are located (Banfalvi et al.,
1981; Rosenberg et a f . , 1981). Apart from these very large plasmids, R. meliloti may harbour
additional plasmids (cryptic) whose functions are largely unknown (Dknarik et al., 198l),
although one report indicated that a cryptic plasmid may influence nodulating competitiveness
(Bromfield et al., 1985).
0001-4174 0 1987 SGM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10
3458
E. S . P . BROMFIELD A N D OTHERS
A number of surveys have considered the symbiotic effectiveness (Gibson et al., 1975;
Hagedorn, 1978; Bottomley & Jenkins, 1983) and plasmid content (Gross et al., 1979; Thurman
et al., 1985)of various indigenous Rhizobium spp., but there is a paucity of information on these
characteristics for native R. meliloti. While such surveys have used collections of isolates from
diverse geographical regions, it is nevertheless important to obtain data for isolates
representative of the overall diversity within a particular resident population (Young et al.,
1987). This can only be achieved by the analysis of large numbers of nodule isolates from
relatively small and well-defined areas. In a previous study (Bromfield et al., 1986), about 2000
isolates of indigenous R.meliloti from cultivars of Medicugo satiua grown in an area (8 x 14.5 m)
at each of two sites (A and B) 2 km apart were characterized on the basis of sensitivity to 16
typing phages. These isolates were subdivided into 55 (site A) and 65 (site B) distinct phage types
representative of the R. meliloti populations at the two sites. The purpose of the present
investigation was to examine the distribution of symbiotic effectiveness and plasmid number
and size among these phage types of R. meliloti.
METHODS
Bacteria andplasmids. A total of 193 isolates of R. melilori, selected to represent the 55 and 65 (88 distinct) phage
types described previously (Bromfield et ol., 1986)from sites A and B respectively, were used in this study. Thirtytwo of these phage types were common to both sites. R. meliloti NRG185 and NRG43 from W. Rice, Research
Station, Beaverlodge, Canada, BALSAC from L. M. Bordeleau, Research Station, Ste Foy, Quebec, Canada, and
SU47, synonymous with RCR2011 (Bromfield, 1984), were standard strains for symbiotic effectiveness tests; all
have been used in commercial inoculants for M.satiua. Plasmids in the following bacteria were used as standards
for estimating molecular masses of R. melilori plasmids. Agrobacterium turnefaciens PD67 was a synthetic strain
from V. N . Iyer, Carleton University, Ottawa, Canada, constructed by insertion of pTiA6 (Knauf & Nester, 1982)
and pPHlJl (Hirsch & Beringer, 1984), with molecular masses of 177 and 35 MDa respectively, into A.
rumejuciens C58 (Casse et al., 1979) containing plasmids of 267 and 117 MDa. Escherichia coli HBlOl (Corbin et
al., 1983)contained either pRK290 (Ditta et al., 1980),molecular mass 13.1 MDa, or pBR322 (Bolivar et a[., 1977),
molecular mass 2.9 MDa.
Ejectioeness rests. Symbiotic effectiveness (nodulation and nitrogen fixation) was assessed using an isolate to
represent each of the 55 and 65 phage types of R . melzloti from sites A and B respectively. Due to the large numbers
involved, site A and site B phage types were tested in separate experiments; no attempt was made to compare
phage types between sites. Seedlings of M . sativa cv. Saranac, from surface-sterilized seed (Vincent, 1970) were
planted in 200 x 30 mm glass tubes (two seedlings per tube) containing vermiculite, each supplied with 40 ml
sterile N-free nutrient solution (Norris & Date, 1976). Two days after sowing, seedlings were inoculated with
0.2 ml cell suspension (about lo8 cells ml-*) of each phage type or standard strain washed from YEM agar slopes
(Bromfield, 1984) in sterile water. Uninoculated controls, with and without 5 mM-KNO, (final concentration)
supplied 12 d after sowing, were included. Five replicate tubes of each inoculation treatment, randomized in
blocks, were maintained in a controlled environment cabinet at 250 p E m-2 s-l (22 "C, 16 h day; 16 "C, 8 h night)
for 6 weeks, after which shoot dry weights and nodule numbers were determined; nodule nitrogenase activity was
measured by acetylene reduction. Variation in symbiotic effectiveness between isolates within phage types was
evaluated using an identical experimental protocol except that six replicates (two plants per tube) of each
inoculation treatment were used.
Plasmid profiles. One hundred and sixty isolates comprising 45 and 48 (68 different) R. meliloti phage types from
sites A and B respectively were analysed for plasmid content using agarose gel electrophoresis according to the
method of Hynes et al. (l986), but modified in that 0.45 % agarose was used and electrophoresis was carried out at
30 V for 30 min and then at I 0 0 V for 4 h. Estimates of the molecular size of R. meliloti plasmids relative to
molecular mass markers were obtained by linear regression of log,, molecular mass (MDa) on log,, mobility using
standard plasmids of 267, 177, I 17 and 35 MDa from A. tumefaciens PD67 and those of 13.1 and 2.9 MDa in E. coli
HBlO1. The megaplasmid of R. meliluri SU47 was used as a standard to detect the presence of plasmids of
equivalent mobility. Each R. meliloti isolate was tested in three separate gels; correlation coefficients from
regression analyses were not less than 0.98.
Treatment ojdata. Since the mass of plasmids exceeding 267 MDa could not be reliably estimated, each was
assigned to one of three classes based on its mobility in agarose gels relative to the megaplasmid of R. meliloti
SU47. The first class consisted of plasmid bands of lesser mobility, the second comprised plasmids of equivalent
mobility and the third included plasmids of greater mobility. To facilitate grouping of R. meliloti isolates according
to plasmid number and masses, unweighted average-linkage clustering (Sneath & Sokal, 1973) was performed
using the GENSTAT package, version 4.04A (Lawes Agricultural Trust, Rothamsted Experimental Station,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10
Plasmids and eflectiveness of R . meliloti
3459
Harpenden, UK). A similarity matrix was calculated using the Gower coefficient (SG;Gower, 1971) for plasmids
of mass 267 MDa or less (means of three measurements) and the Jaccard coefficient (SJ;Sneath, 1957) for
plasmids exceeding 267 MDa, each of which was scored as either 1 (present) or 0 (absent) in one of the three classes
described above. The highest level of similarity at which the three replicates of each isolate grouped together
(94%) was selected for categorization. Data for shoot dry weight, nodule number and nitrogenase activity were
subjected to analyses of variance.
RESULTS
Table 1 shows summarized data for mean effectiveness values of the 55 and 65 representative
phage types of R . meliloti from sites A and B, respectively. All 55 phage types except for type 23
(isolate no. 323) from site A were symbiotically effective since they elicited shoot dry matter that
did not differ significantly from the nitrate control and two or more of the standard strains.
Nodule numbers per plant did not differ significantly between these phage types and standard
strains except for types 13 and 23, which elicited the most nodules. Data for nitrogenase activity
revealed differences between site A phage types and standard strains similar to those for shoot
dry matter production. With respect to shoot dry matter, all 65 phage types from site B were
effective, differing (P < 0.01) from the uninoculated control but not significantly from any of the
standard strains or plants supplied with combined nitrogen. The majority of site B types were
similar in terms of nodules elicited per plant and there were no significant differences between
any of the phage types or standard strains with respect to nitrogenase activity.
Based on individual phage type means, the distributions of shoot dry weights (percentage of
the values obtained with the four standard strains) of plants inoculated with the 55 (site A) and
65 (site B) types are,shown in Table 2. The distributions representing types from each site
approximated to normality but tended towards a high level of effectiveness relative to the
standard strains (loo%), with means of 86.2% & 1.45% (site A) and 95.4% & 1.47% (site B).
Since phage type 23 was common to both sites and the site A isolate (no. 323) was symbiotically
ineffective while the site B isolate (no. 1731) was effective relative to the standard strains, we
examined the variation in symbiotic properties of a total of six different isolates of this phage
type from both sites (Table 3). The data confirm that isolates within type 23 varied in
effectiveness. Plants inoculated with three of four isolates (58, 114, 323) from site A differed
(P < 0.01) from those inoculated with the remaining isolate (1076) from this site and from those
inoculated with both isolates (1578, 1731) from site B, with respect to shoot dry matter, nodule
number and nitrogenase activity. A further five isolates from each of five frequently occurring
phage types from sites A and B respectively were examined for variation in effectiveness. For
each of 9 of the 10 phage types there were no significant differences between isolates for the
three symbiotic parameters evaluated (data not shown). The exception involved plants
inoculated with two of five isolates from type 38, which differed ( P < 0.01) in terms of nodule
numbers and nitrogenase activity.
Summarized data for the plasmid profiles of the 160 R . meliloti isolates comprising 45 and 48
(68 different) phage types from sites A and B respectively and for their grouping on the basis of
plasmid number and mass are shown in Table 4. These isolates contained between one and five
plasmids with molecular masses varying between 5 and >> 267 MDa. Thirty-five isolates (22% of
the total) from 18 phage types possessed plasmids of molecular mass less than 50 MDa whereas
17 isolates (from eight phage types), all originating from site A, possessed plasmids (excluding
those corresponding to the reference megaplasmid of SU47) with molecular masses greater than
267 MDa. At the 94% level of similarity the 68 distinct phage types from both sites were placed
into 22 different plasmid groups, with 17 and 13 groups representing site A and B phage types,
respectively. Based on chi-squared analysis, the distribution of phage types among plasmid
groups differed between sites ( x 2 = 61.06, P < 0.001). More than 50% of the total isolates from
both sites (43 phage types) were placed in two plasmid groups (1 and 4) consisting respectively of
isolates with only a megaplasmid and those with a megaplasmid plus a plasmid with an average
molecular mass of 155 MDa.
Thirty of the phage types tested were represented by more than one isolate, and of these, each
of 11 types were homogeneous with respect to plasmid profiles and were placed in single plasmid
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10
-
12.4-28.0
7.2-47.211
39-4-73.2
17.8-32'0
10.2-29.4
67.6-9 1 a0
22.2
12.6
55.2
24.8
15.9
78.1
Range
13.0
81.1
26.0
11.3
56.9
25.6
8.0- 1 8.4
72.2-89-6
244-28.2
9.0- 14.0
59.2-62.0
2 1 6-27-8
Four standard strains
I
Mean
Range
1.88
2.1 1
6.05
2-27
2.73
9.25
59
59
71
69
69
SEM
Average
61
Total no.
of means?
NS
P < 0.01
P < 0.01
P < 0.01
P < 0.01
P < 0.01
Significance$
2t
0
A
B
5
0
0
2
61-70
51-60
20
3
71-80
44
20
81-90
18
35
91-100
11
24
101-110
0
I1
111-120
0
5
3
121-130
* Percentage of phage types based on 55 (site A) or 65 (site B) types. Classes (10% intervals) represent shoot dry
weight (mg per plant) expressed as a percentage of the mean of the values obtained with the four standard strains.
These mean values (100%) were 25-6 (site A) and 26.0 (site B).
This value represents a single phage type (type 23).
40-50
Site
Percentage of phage types in shoot dry weight class*
Table 2 . Distributions of values for shoot dry weight of M . sativa cv. Saranac inoculated with
representative phage types of R . meliloti from sites A and B
Mean and range of 55 (site A) or 65 (site B) means.
t No. of means in analyses of variance, each for 10 individual plants from five replicate tubes (two plants per tube).
$ Significance of differences between means based on ratio of between to within mean squares.
5 Means for uninoculated controls with and without combined N were, respectively, 25.4 and 7.6 (site A) and 26.2 and 7.2 (site B).
(1 Only two phage type means (45.6 and 47.2) were greater than 17.0; mean of remaining phage types, 11.3.
Site A phage types
Shoot dry wt
(mg per plant)§
Nodule no. per plant
Nitrogenase activity
(nmol C2H4per plant h-I)
Site B phage types
Shoot dry wt
(mg per plant)§
Nodule no. per plant
N it rogenase activity
(nmol C2H4per plant h-I)
Mean
Phage types*
Data are derived from analyses of variance. The test host was M .satitla cv. Saranac SEM. Standard error of mean; NS, not significant: DF,degrees of
freedom for error.
Table 1. Summarized data ji)r the mean tfliv-titvness iwlues o f 5 5 and 65 R . meliloti phage types .from sites A and B, respectively
272
272
280
232
232
240
DF
w
rA
3:
m
-I
0
W
21
>
Q
r
m
z
2
0
W
%
?
r
A
m
5
w
Plasmids and efectiveness of R . meliloti
346 1
Table 3. Response of M . sativa cv. Saranuc to inoculation with different R . meliloti isolates of
phage type 23
Values are means of 12 individual plants from six replicate tubes (two plants per tube). SED, Standard
error of difference; DF,degrees of freedom for error.
Isolate
no.
Site of
origin
Shoot dry wt
(mg per plant)*
Nodule no.
per plant
Nitrogenase activity
(nmol CzH4per plant h-*)
58
114
323
1076
1578
1731
SU47t
A
A
A
14.0
15.2
11.6
23.4
22.1
22.0
22.7
1-89
45
54.5
52.3
58-2
10.8
8-7
9-0
9.5
7-05
35
48.8
39.5
106.2
88.8
91-8
89.2
14.46
35
A
B
B
-
SED
DF
44-5
Values for uninoculated controls with and without combined N were 21.6 and 8-2, respectively.
SU47 is a standard strain.
Table 4. Grouping of 160 R . meliloti isolates, comprising 4.5 (site A ) and 48 (site B ) phage types,
according to plasmid number and molecular mass
Plasmid
group"
Average plasmid
configurationt
No. of
isolates
Site A phage types
20
4, 12, 14, 15, 19, 29, 40,
1
0, 1, 0
2
3
4
0, 1, 0, 14
0, 1, 0, 81
0, 1, 0, 155
66
5
6
0, 1, 0, 105, 58
0, I, 0, 129, 30
2
17
0, 1, 0, 121, 93
0, 1, 0, 155, 71
0, 1, 0, 207, 49
0, 1, 0, 146, 125
0, 1, 0, 218, 95
0, 1, 0, 135, 99, 84
0, 1, 0, 150, 82, 36
0, 1, 0, 82, 64,28, 14
0, 1, 0, 98, 87, 52, 28
0, 1, 1, 138
0, 1, 1, 142, 122, 28
0, 1, 1, 176, 136, 5
0, 1, 1, 172, 153, 102
1, 0, 0, 161
1, 0, 0, 119, 22
1, 1, 0, 125
13
2
1
6
1
2
5
1
1
7'
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
1
5
5
1
4
1
1
4
1
51, 62, 87
36, 38
1,3,6,8,9,10,16,20,25,
28, 35, 39, 41, 79, 84
53
17, 22,27, 33, 34,42,49,
81, 85
31
80
53
38
5
13, 23, 38
2
42
37
3
33
23
Site B phage types
12, 15, 19, 29, 40, 46,51, 76
40
23, 53, 88
10, 14, 16, 20, 22, 23, 31, 32,
35, 41, 45, 50, 52, 54, 57, 58,
59, 62, 64,66, 73, 75, 85
53
78
9, 47, 49, 68, 74, 76
1
27, 32, 61
56
38, 60
6, 70, 89
38
* Plasmid groups are derived from clustering at the 94% level of similarity using S, and S , coefficients.
t The average plasmid configuration of isolates in each group is given such that the first three values (left to
right) represent the presence (1) or absence (0) of plasmids exceeding 267 MDa in classes one, two or three,
respectively (defined in Methods); any further consecutive values denote the average mass (MDa) of plasmids less
than 267 MDa.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10
3462
E. S. P. B R O M F I E L D A N D OTHERS
Fig. 1. Plasmid grouping of R. meliloti isolates within phage types 38 and 49 from sites A and B.
A . rumefaciens PD67 contains standard plasmids of known molecular mass (MDa). R. meliloti
SU47 contains the reference megaplasmid. Chr, chromosomal band.
Fig. 2. Plasmid grouping of R. meliloti isolates within phage types 23 and 32 from sites A and B. Other
details as for Fig. 1.
groups whereas 17 of the remaining types were subdivided into two (14 phage types) or three
(three phage types) plasmid groups. The exceptions, types 23 and 38, each represented by six
isolates, were subdivided into four and five plasmid groups, respectively (Figs 1 and 2); type 38
was resistant to all 16 typing phages (Bromfield et al., 1986) and isolates of type 23 varied in
effectiveness (Table 3). From 22 phage types (excluding types 23 and 38) re'presented by isolates
from both sites, 11 types were subdivided into plasmid groups on the basis of differences
between isolates from sites A and B; an example of such a difference is illustrated by isolates of
type 49 (Fig. 1). However, data for all isolates examined (68 different types) indicated no
apparent overall relationship between plasmid grouping and the site from which they
originated. Certain phage types from both sites comprised isolates differing only in the presence
or absence of a single plasmid of mass less that 267 MDa, an example of which is illustrated by
isolates of phage type 32 from site B (Fig. 2). Seven isolates of type 33 were divided into plasmid
groups (6 and 21) with isolates in group 21 being unusual in that they lacked a plasmid of
mobility comparable with that of the SU47 megaplasmid but instead possessed a band of
mobility corresponding to that of the chromosomal band of A. tumefaciens PD67; phage type 3
was divided into plasmid groups 4 (five isolates) and 20 (one isolate) on a similar basis (Fig. 3).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10
Plasmids and eflectiueness of R . meliloti
3463
Fig. 3. Plasmid grouping of R.meliloti isolates within phage types 3 and 33 from site A. Other details as
for Fig. 1 .
Table 5. Distributions of plasmid numbers among representative phage types from two populations
of R. meliloti
Site of
origin
A
B
Percentage of phage types with no. of plasmids*
r
\
A
1
2
3
4
5
24 (15)
15 (11)
36 (45)
49 (49)
30 (30)
25 (32)
2 (2)
9 (7)
8 (8)
2 (1)
* The values are based on 45 (site A) or 48 (site B) phage types. Values in parentheses represent the percentage of
isolates based on 87 (site A) or 73 (site B) isolates.
The symbiotically ineffective isolate (no, 323) of type 23 (plasmid group 22) was also unusual in
that it possessed a band of mobility similar to that of the SU47 megaplasmid as well as a band of
lesser mobility (Fig. 2). Since the six and seven isolates of types 3 and 33 were unusual with
respect to plasmid profiles, we examined their symbiotic effectiveness (shoot dry weight, nodule
number and nitrogenase activity). All isolates were effective and did not significantly differ
from standard strain SU47 for any of the three measurements (data not shown).
Table 5 shows distributions of plasmid numbers among the 45 and 48 phage types from sites A
and B. Both distributions were similar in form and show that phage types with one, two or three
plasmids were the most abundant. Minor deviations between distributions expressed as
percentage phage types and those expressed as percentage isolates are ascribed to variation in
plasmid numbers between isolates within certain phage types.
Neither plasmid number nor total plasmid mass (excluding those exceeding 267 MDa) of
phage types from sites A or B was significantly correlated with symbiotic effectiveness based on
percentage shoot dry matter of the four standard strains, or with phage sensitivity, derived from
a previous report (Bromfield et al., 1986), and expressed as the number of typing phages (from
16) producing lysis.
DISCUSSION
Representative phage types from the two populations of R. meliloti showed a high overall level
of symbiotic effectiveness relative to four strains used commercially. In a previous study, one of
these standard strains (SU47) was used to inoculate M . sativa at sites A and B but was recovered
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10
3464
E . S . P . BROMFIELD A N D OTHERS
from a minority of nodules and did not significantly influence dry matter yield relative to plants
in uninoculated plots (Bromfield et al., 1986). The present data for symbiotic effectiveness
suggest that even if the standard strains were to successfully occupy all nodules of the legume
host at these sites, there would be little benefit in terms of enhanced N 2fixation accruing from
inoculation.
Sherwood & Masterson (1974) and Mytton & Livesey (1983) reported that the choice of test
host genotype can affect estimates of effectiveness of R. leguminosarum biovar trifolii
populations. It is also known that the expression of R. meliloti effectiveness can be influenced by
M . sativa genotype (Gibson, 1962; Mytton et al., 1984). The choice of M. sativa cv. Saranac as
test host in the present investigation was made on the basis that this genotype was one of two
previously used for isolation of R. meliloti at sites A and B (Bromfield et al., 1986) and wherever
possible isolates from this cultivar were used in effectiveness tests.
Thurman et al. (1985) reported that effectiveness and plasmid number were correlated for a
collection of indigenous R. leguminosarum biovar trifolii and implicated an effect of cryptic
plasmids in the symbiotic process. In contrast, our results for R. meliloti indicate a lack of
correlation between either plasmid number or mass and symbiotic effectiveness, possibly
reflecting species differences.
The relatively small proportion of representative phage types from each site which possessed
more than three plasmids might reflect the impact of biotic and environmental selection
pressures against higher plasmid numbers. It is of interest that Thurman et al. (1985) reported a
similar finding for isolates of R. leguminosarum biovar trqolii.
Plasmids of mass less than 50 MDa have been reported to occur very rarely in rhizobia
(Prakash & Atherly, 1986). However, our results showed that 22% of the R. meliluti isolates
examined possessed plasmids of size less than 50 MDa. This moderately high frequency may
reflect our use of representative phage types from two indigenous populations of R. meliloti. A
very small plasmid (molecular mass about 5 MDa) found in isolates of one phage type may be
suitable for the construction of a Rhizobium cloning vehicle.
Gross et al. (1979) reported that the plasmid profiles of indigenous Bradyrhizobium japonicum
strains were not correlated with antibiotic sensitivity, colony morphology, serotype, or
symbiotic effectiveness, and were more discriminating than these four characteristics for strain
distinction. In our study 68 distinct phage types were placed in 22 plasmid groups, indicating
that plasmid profile analysis was less discriminating than phage typing for delineating the
composition of indigenous R. meliloti populations. The lack of correlation between phage
sensitivity and plasmid content and the disparity between the grouping based on each of these
criteria might be explained if it is assumed that determinants for phage sensitivity are encoded
on the R. meliloti megaplasmids, chromosome, or both. In this case, the phage type need not
necessarily reflect the plasmid content of constituent isolates, at least for those containing
plasmids other than the megaplasmids. Evidence lending support to this suggestion includes a
role of Rhizobium cell surface polysaccharides in phage adsorption (Zajac & Lorkiewicz, 1983;
Finan et al., 1985) and the location of genes for exopolysaccharide synthesis on the R. melibti
non-Sym (non-symbiotic) megaplasmid (Finan et al., 1986; Hynes et al., 1986), and chromosome
(Dylan et al., 1986).
It is of interest that the two phage types (23 and 38) which were the most variable with respect
to plasmid profiles were the only types from those tested in which isolates showed differences in
symbiotic effectiveness; type 38 was expected a priori to constitute a potentially heterogeneous
group since isolates of this type are resistant to lysis by 16 typing phages (Bromfield et al., 1986).
Variation in plasmid content within certain phage types (e.g. type 32) was due to the presence or
absence of a single plasmid in isolates with otherwise identical plasmid profiles, perhaps
reflecting possible genetic interchange between indigenous R. meliloti.
In a previous report the diversity and distribution of phage types of indigenous R. meliloti
were shown to differ between sites A and B (Bromfield et al., 1986). The present investigation
similarly shows differences in the distribution and variety of R. meliloti plasmid groups between
these sites, perhaps reflecting the impact of selection pressures imposed by biotic and abiotic
factors differing between sites. Adaptation of R. meliloti to their respective sites was indicated
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10
Plasmids and eflectiveness of R . meliloti
3465
by the observation that certain phage types were divided into plasmid groups solely on the basis
of differences in plasmid content between isolates from each site, and by the fact that all the
isolates which possessed plasmids greatly exceeding 267 MDa (exclusive of those corresponding
to the reference megaplasmid) originated from site A.
The presence of a megaplasmid of mass equivalent to that in the standard strain SU47 is
considered to be a general feature of R. meliloti (Hynes et al., 1986). In the present study, only
five isolates from two phage types lacked a megaplasmid corresponding to that of SU47, but
instead showed a band, in agarose gels, of even greater mass which may represent either a cointegrate of the Sym (symbiotic) and non-Sym megaplasmids (Banfalvi et al., 1985;Hynes et al.,
1986) or chromosomal DNA. A similar very high molecular mass band observed in the
symbiotically ineffective isolate (no. 323) of type 23 may also represent chromosomal DNA or a
co-integrate of either the Sym or non-Sym megaplasmid with a plasmid exceeding 267 MDa,
absent in this isolate but present in all other isolates of the same phage type from site A (Fig. 2);
this possibility could be confirmed by genetic analysis.
Genomic rearrangements are not without precedent in the Rhizobiaceae (Cantrell et al., 1982;
Berry & Atherly, 1984), and may provide an explanation for the observed wide diversity of
plasmid profiles within populations of indigenous R . meliloti. Genetic analysis of representatives
of this diversity may provide a better understanding of the indigenous competitors which affect
inoculant strain establishment and persistence.
Thanks are due to Indu B. Sinha for technical assistance and to L. Lefkovitch and M. S.Wolynetz for statistical
advice. This paper is Plant Research Centre Contribution no. 1061.
REFERENCES
BANFALVI,
Z., SAKANYAN,
V., KONCZ,C., KISS, A.,
DUSHA,I. & KONDOROSI,
A. (1981). Location of
nodulation and nitrogen hation genes on a high
molecular weight plasmid of Rhizobium meliloti.
Molecular and General Genetics 184, 318-325.
BANFALVI,
Z., KONDOROSI,
E. & KONDOROSI,
A. (1985).
Rhizobium meliloti carries two megaplasmids. Plasmid 13, 129-138.
BERRY,J. 0. & ATHERLY,
A. G. (1984). Induced
plasmid-genome rearrangements in Rhizobium japonicum. Journal of Bacteriology 157, 2 18-224.
BOLIVAR,F., RODRIQUEZ,
R. L., GREENE,P. J.,
BETLACH,
M. C., HEYNEKER,
H. L., BOYER,
H. W. &
FALKOW,
S. (1977). Construction and characterization of new cloning vehicles. 11. A multi-purpose
cloning system. Gene 2, 95-1 13.
B O ~ M L E P.
Y ,J. & JENKINS, M. B. (1983). Some
characteristics of Rhizobium meliloti isolates from
alfalfa fields in Oregon. Soil Science Society of
America Journal 47, 1153-1 156.
BROMPIELD,
E. S. P. (1984). Variation in preference for
Rhizobium meliloti within and between Medicago
sativa cultivars grown in soil. Applied and Environmental Microbiology 48, 1231-1 236.
BROMFIELD,
E. S. P., LEWIS,D. M. & BARRAN,
L. R.
(1985). Cryptic plasmid and rifampin resistance in
Rhizobium meliloti influencing nodulation competitiveness. Journal of Bacteriology 164, 410-413.
BROMFIELD,
E. S. P., SIW, I. B. & WOLYNETZ,
M. S.
(1986). Influence of location, host cultivar and
inoculation on the composition of naturalized populations of Rhizobium meliloti in Medicago sativa
nodules. Applied and Environmental Microbiology 51,
1077-1084.
BURKHARDT,
B. & BURKHARDT,
H. J. (1984). Visualization and exact molecular weight determination of a
Rhizobium meliloti megaplasmid. Journal of Molecular Biology 175, 213-218.
CANTRELL,
M. A., HICKOK,R. E. & EVANS,H. J.
(1982). Identification and characterisation of plasmids of hydrogen uptake positive and hydrogen
uptake negative strains of Rhizobium japonicum.
Archives of Microbiology 131, 102-106.
CASSE,F., BOUCHER,
C., JULLIOT, J . S., MICHEL,M. &
D~NARIE,
J. (1979). Identification of large plasmids
in Rhizobium meliloti using agarose gel electrophoresis. Journal of General Microbiology 113, 229-242.
CORBIN,D., BARRAN,L. R. & DITTA, G. (1983).
Organization and expression of Rhizobium meliloti
nitrogen fixation genes. Proceedings of the National
Academy of Sciences of the United States of America
62, 1159-1166.
DENARIB,J., BOISTARD,P. & CASSE-DELBART,
F.
(1981). Indigenous plasmids of Rhizobium. International Review of Cytology (Supplement) 13, 225-246.
DI-ITA, G., STANFIELD,
S., CORBIN,D. & HELINSKI,
D. R. (1980). Broad host range DNA cloning system
for Gram negative bacteria: construction of a gene
bank of Rhizobium meliloti. Proceedings of the
National Academy of Sciences of the United States of
America 77, 7347-735 1.
DYLAN,T., IELPI,L., STANFIELD,
S., KASHAP,L.,
DOUGLAS,C., YANOFSKY,M., HELINSKI,D. R. &
DITTA,G. (1986). Rhizobium meliloti genes required
for nodule development are related to chromosomal
virulence genes in Agrobacterium tumefaciens. Proceedings of the National Academy of Sciences of the
United States of America 83, 4403-4407.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10
3466
E. S . P. B R O M F I E L D A N D O T H E R S
FINAN,T. M . , HIRSCH,A. M.. LEIGH,J . A., JOHANSEN,
E., KULDAU,A. G., DEEGAN.
S., WALKER,G. C. &
SIGNER,E. R. (1985). Symbiotic mutants of Rhizobium meliloti that uncouple plant from bacterial
differentiation. Cell 40,869-877.
FINAN,T. M.. KUNKEL,B., DE Vos, G. & SIGNER,E.
R. ( I 986). Second megaplasmid in Rhizobium meliloti
carrying exopolyaccharide and thiamine synthesis
genes. Journal of Bacteriology 167, 6&72.
GIBSON, A . H. (1962). Genetic variation in the
effectiveness of nodulation of lucerne varieties.
Australian Journal of Agricultural Research 13, 388399.
GIBSON,A . H., CURNOW.B. C., BERGERSEN,
F. J . ,
BROCKWELL,
J. & ROBINSON,A. C. (1975). Studies of
field populations of Rhizobium : Effectiveness of
strains of Rhizobium rrjfolii associated with Trifolium
suhterraneum L. in south-eastern Australia. Soil
Biology and Biochemistrj 1, 95-1 02.
GOWER,J . C. (1971). A general coefficient of similarity
and some of its properties. Biometrics 27, 857-874.
GROSS, D. C., VIDAVER.
A. K . & KLUCAS,R . V. (1979).
Plasmids, biological properties and efficacy of
nitrogen fixation in Rhizobium japonicium strains
indigenous to alkaline soils. Journal of General
Microbiology 114, 257-266.
HAGEDORN,C. (1978). Effectiveness of Rhizobium
trjfolii populations associated with Trifoliurn suhterraneum L. in Southwest Oregon soils. Soil Science
Socieg of America Journal 42, 4 4 7 4 5 I .
HIRSCH.P. R. & BERINGER,
J . E. (1984). A physical
map of pPHlJl and pJB4Jl. PIasmid 12, 139-141.
HYNES,M . F., SIMON,R., MULLER,P., NIEHAUS,K . ,
LABES,M. & PUHLER,A. (1986). The two megaplasmids of Rhizobium meliloti are involved in the
effective nodulation of alfalfa. Molecular and General
Genetics 202, 356-362.
JOHNSON.H . W., MEANS, U. M. & WEEER,C. R .
(1965). Competition for nodule sites between strains
of Rhizobium juponicum applied as inoculum and
strains in the soil. Agronomy Journal 57, 179-185.
KNAUF.V. C. & NESTER,E. W . (1982). Wide host
range cloning vectors: a cosmid clone bank of an
Agrobacterium Ti plasmid. Plasmid 8, 45-54.
MYTTON,L. R. & LIVESEY,C. J . (1983). Specific and
general effectiveness of Rhizobium trifolii populations from different agricultural locations. Planr and
Soil 73. 299-305.
MYTTON.L. R., BROCKWELL,J. & GIBSON, A. H.
(1984). The potential for breeding an improved
lucerne-Rhizobium symbiosis. 1. Assessment of
genetic variation. Euphytica 33, 401-410.
NORRIS, D. 0. & DATE,R. A. (1976). Legume
bacteriology. Commonwealth Bureau of Pastures and
Field Crops Bulletin no. 51, p. 134.
PRAKASH,R. K . & ATHERLY,A . G . (1986). Plasmids of
Rhizobium and their role in symbiotic nitrogen
fixation. International Review of Cytology 104, 1-23.
ROLFE, €3. G . & SHINE, J . (1984). RhizobiumLeguminosae symbiosis : the bacterial point of view.
In Genes Involved in Microbe-Plant Interactions, pp.
106-109. Edited by D. P. S. Verma & T. Hohn. New
York : Springer-Verlag.
ROSENBERG.
C., BOISTARD,P., DENARIE,
J . & CASEDELBART
( I98 1 ). Genes controlling early and late
functions in symbiosis are located on a megaplasmid
in Rhizobium meliloti. Molecular and General Genetics
184, 325-333.
ROUGHLEY,
R. J . , BLOWES,W. M . & HERRIDGE,D. F.
(1976). Nodulation of Trifolium subterraneum by
introduced rhizobia in competition with naturalized
strains. Soil Biology and Biochemistry 8, 403-407.
SHERWOOD,M. T. & MASTERSON,C. L. (1974).
Importance of using the correct test host in assessing
the effectiveness of indigenous populations of Rhizobium trifolii. Irish Journal of Agricultural Research 13,
101-108.
SNEATH,P. H. A. (1957). The application ofcomputers
to taxonomy. Journal of General Microbiology 17.
201-226.
SNEATH,P. H. 8c SOKAL,R. R. (1973). Numerical
Taxonomy. The Principles and Practice qf Numerical
Classificafion.San Francisco: Freeman.
THURMAN,N. P., LEWIS,D. M. & JONES, D. G. (1985).
The relationship of plasmid number to growth, acid
tolerance and symbiotic efficiency in isolates of
Rhizobium trvolii. Journal of Applied Bacteriology 58,
1-6.
VINCENT,J . M. (1970). A Manua1.for the Practical Study
of the Root-nodule Bacteria (IBP Handbook no. 15).
Oxford : Blackwell Scientific Publications.
YOUNG,J. P. W., DEMETRIOU,
L. & APTE,R. G. (1987).
Rhizohium population genetics : enzyme polymorphism in Rhizobium leguminosarum from plants and
soil in a pea crop. Applied and Encironmentul
Microbiology 53, 397-402.
ZAJAC, E. & LORKIEWICZ,
Z . ( 1 983). Role of 0-acetyl
groups in the lipopolysaccharide receptor for Rhizobium phage I P. Journal of General Mirrobiology 129,
31 7 1-3 176.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 06:34:10