Turnour-initiating Ability and Nutrition in the Genus

Transcription

J. gen. MicrobioE. (1969), 59, 57-75
Printed in Great Britain
57
Turnour-initiating Ability and Nutrition in the
Genus Agrobacterium
By J . A . L I P P I N C O T T A N D B A R B A R A B. L I P P I N C O T T
Department of Biological Sciences, North western University, Evanston, Illinois
60201,U.S.A.
(Acceptedfor publication 24 July 1969)
SUMMARY
Sixty-nine Agrobacterium strains, including representatives of six current
species, were tested: for colony formation on sucrose + salts medium, 3ketoglycoside production, response on litmus milk medium, infectivity on
carrot-root discs, infectivity and rate of turnour appearance on primary
pinto bean leaves. These tests separated the majority of the strains into
species. Forty-four strains were tumourigenic on pinto bean leaves, ten of
which did not form colonies on a sucrose salts medium and nine of these
showed a rate of tumour appearance typical of auxotrophs. Glutamate was
required for colony formation by seven of the ten auxotrophs, two required
only pantothenate and one strain required either glutamate or pantothenate.
The two A . rubi strains among the ten auxotrophs required nicotinic acid
+ pantothenate + glutamate. Strains that required either pantothenate
or pantothenate + nicotinic acid for colony formation also required these
compounds in liquid medium. None of the strains which required glutamate
for growth on plates required it in liquid medium though glutamate shortened
lag and division times. The growth requirements of the auxotrophs did not
obviously correlate with their specific infectivity. Six of the ten auxotrophs
produced no 3-ketoglycosides, but this response did not correlate with
infectivity or nutritional requirements. Three auxotrophs initiated more
tumours when appropriate growth factors were added to inoculated leaves.
The agrobacteria that induced pinto bean leaf tumours fell into four groups:
I, strains which infected all wound sites with equal efficiency regardless
of wound size; 11, strains which infected all but the smallest wound sites with
equal efficiency; 111, phototrophic strains which infected only larger wound
sites; IV, auxotrophic strains which, in the typical examples, infected only the
larger wound sites. Nineteen of the tumourigenic strains (43%) were restricted in their ability to infect small wound sites and were assigned to
groups I11 and IV. Because loss of the ability to initiate tumours at small
wound sites is due to nutritional limitations imposed on the infecting
bacteria by the wound site medium, these group I11 and IV strains were
nutritionally limited in their ability to infect this host. The large proportion
of strains which showed this defect indicates that variation in the ability
of tumefacient strains to adapt to the wound medium supplied by a host is
quite common and hence represents one major pathway by which pathogenicity within the genus may have evolved.
+
INTRODUCTION
The degree to which a bacterium capable of causing crown-gall is successful in
initiating a tumour can depend upon the interplay between the bacterium and nutritional qualities of the host wound site (Lippincott & Lippincott, 1966). Thus auxo-
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
J . A. LIPPINCOTT A N D B. B. LIPPINCOTT
58
trophs derived from the highly virulent strain ~6 of Agrobacterium tumefaciens have
been characterized by: (I) decreased infectivity on pinto leaves; ( 2 ) ability to induce
only early appearing tumours on this host; (3) increased infectivity when the substances required for their growth are added to inoculated leaves. The demonstration
of a bacterial nutrition phase during the initiation of crown-gall tumours and the
changes in infectivity associated with auxotrophic nutrition suggested that naturally
occurring nutritional defects might account in part for differences in pathogenicity
and host range within the genus Agrobacterium.
Studies of the nutritional requirements of members of this genus were undertaken
in the past to aid in the systematics of the group and to provide information about
the pathogenicity of certain species (Hendrickson, Baldwin & Riker, I 934 ; Sagen,
Riker & Baldwin, 1934; Riker, Lyneis & Locke, 1941; Starr, 1946b). Distinct differences were reported for the nutrition of the several species, but no specific attempt
was made to relate these differences to infectivity characteristics. We have compared
the ability of representative strains of various species of the genus Agrobacterium to
initiate tumours on primary pinto bean leaves, together with the rate at which these
tumours appear, to their capacity to form colonies on a sucrose + salts medium.
Also, all the strains were tested for ability to promote outgrowths on carrot-root discs,
reaction in litmus milk medium and ability to produce 3-ketoglycosides, so that the
infectivity results could be directly related to key taxonomic characteristics of the
different species.
METHODS
Growth of cultures and colony counts. Cultures were grown for 48 hr to the stationary
phase in nutrient broth
yeast-extract liquid medium and plated on yeast-extract
+ nutrient agar as described previously (Heberlein & Lippincott, 1965; Lippincott
& Lippincott, 1966).
Ability to form colonies on sucrose
salts medium. Cultures were diluted to about
1 0 2 and 104viable organisms/o.I ml. with distilled water and each dilution plated in
triplicate on a sucrose
salts medium (Lippincott, Webb & Lippincott, 1965) containing 1.5 yo purified agar (Difco) and on the nutrient agar medium. Colonies were
counted after 3 days (nutrient broth medium) or 7 days (sucrose
salts medium)
at 27' &
Strains were scored negative for growth on sucrose + salts when they
formed no colonies or colonies too small to be counted by using a New Brunswick
plate counter. Some batches of purified agar allowed microcolony formation by
several of the strains that required only glutamate on solid medium. Strains scored
positive generally formed colonies that could be counted at 4 or 5 days, and by 7 days
the colonies were 3 to 5 mm. in diameter.
Litmus milk reaction. Samples of 0-2 ml. (about 109organisms) from 48-hr cultures
were added to 10ml. litmus milk (Difco) in 16 x 150 mm. tubes, mixed well and incubated at 27O & without further mixing or shaking. The cultures were scored weekly,
but only data obtained after 3 weeks are presented. All determinations were done in
duplicate on at least two occasions, using separately grown cultures.
Production of 3-ketoglycosides. The method described by Bernaerts & De Ley
(1963) was used.
Infectivity and rate of tumour appearance on pinto bean leaves. Specific infectivity
+
+
+
+
IO.
IO
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
Agrobacterium infectivity and nutrition
59
was estimated by using the pinto bean leaf bioassay of Lippincott & Heberlein (1965)
and expressed as % relative specific infectivity (% r.s.i.), which is the mean number of
tumours induced/leaf/viable bacterium/ml. inoculum (specific infectivity), relative to
.
dilutions were chosen which were exthe specific infectivity of strain ~ 6 Culture
pected to induce between 10 and 50 tumours/leaf, well within the linear range of the
bioassay. Leaves were inoculated with strain ~6 in each experiment to provide a
standard. The yo r.s.i. reported for each strain is the mean of at least five determinations on different cultures and inoculated at different times. For the 40 strains with yo
r.s.i. values above 3 %, the standard errors expressed as % of each % r.s.i. value
averaged rf: 26% of the yor.s.i. value.
To determine the rate of tumour appearance (Lippincott & Lippincott, 1966), inoculum concentrations were chosen whenever possible to give 50 to 75 tumours/leaf
in 7 days, so that meaningful results could be obtained with early counts. A strain
~6 control was included in each experiment. Many of these data are presented as
the ratio of the tumour counts made after 4 days and 8 days from inoculation for the
test strains, subtracted from the similar ratio for strain ~6 controls. The values reported are the means of at least three separate experiments for each strain, with
separately grown cultures inoculated on plants on different days. The day 4 : day 8
tumour count ratio for strain ~6 in 40 determinations was 44 & 1.5 yo.
Addition of nutrients to leaves inoculated with auxotrophic strains. Solutions containing compounds which either promoted or were required for growth of auxotrophic
strains on a sucrose + salts medium were added hourly to leaves inoculated with
these bacteria for the first 12 hr after inoculation (Lippincott & Lippincott, 1966).
Infectivity on carrot roots. Carrot-root slices were used as a second host to estimate
the relative infectivity of different Agrobacterium strains (De Ropp, 1950; Klein &
Tenenbaum, 1955) and because they provided a differential test for the infectivity of
A . rhizogenes (Ark & Thompson, 1961). Carrot roots purchased at local markets were
peeled, wiped with 95% (v/v) ethanol in water and soaked in 0.9% sodium hypochlorite (I vol. commercial Chlorox + 5 vol. distilled water) for I 5 min., rinsed three
times with sterile distilled water and drained between sterile paper towels. Slices about
5 mm. thick were cut from the middle one-third of each carrot with aseptic technique
and placed in Petri dishes (4 slices/dish) containing 30 ml. I % agar in distilled water.
Each strain was tested on 20 slices from at least 10 different roots by pipetting 0.05 ml.
of a 48-hr culture on each disc. Control discs received a similar amount of sterile
growth medium. The inoculated discs were incubated in dim light at 27" rf: 1'; at
least two separate tests were made for each strain.
Tumour and root formation on the discs was scored at 2 and 3 weeks and the discs
observed for possible root formation until they became desiccated. The scores at 3
weeks are presented here. The amount of tumour tissue/disc and the number of discs
showing a positive response were scored. Although the two responses were usually
directly correlated, the amount of tumour tissue was the main factor considered in the
final infectivity ratings. Strains which scored positive for root formation on carrot
discs usually induced roots on more than half of the treated discs; some discs usually
responded by forming IOO or more roots. The few roots which developed from the
proliferation of phloem rays were disregarded, as were occasional roots that developed
from tumourous tissue.
Growth chmacteristics in liquid medium. Cultures (48 hr) of the different strains
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
60
J . A . LIPPINCOTT AND B . B. LIPPINCOTT
Table
I.
Infectivity and physiological characteristics of Agrobacterium strains
Physiology
Infectivity
w
Species
and
strain
Source*
A . tumefaciens
B6
Stonier
I5955 ATCC
TT I I I De Ley (ICPB)
143
De Ley (ATCC)
~ ~ 1 0De7 Ley (ICPB)
ACH-I
Hamilton
TT133 ICPB
T-37
Stonier
~
~
Panagopoulos
2
2
c-58-3 Hamilton
P2
Panagopoulos
C
De Ley (Gemb)
4
De Ley (NCPPB)
398
De Ley (NCPPB)
B 6-806 Stonier
396
De Ley (NCPPB)
11158 ATCC
AG 6
P2UX3gOpOUlOS
H-100
Stonier
AG I
Panagopoulos
17805 ATCC
4452
ATCC
IIBV 7
Stonier
W-I
Manasse
TT6
De Ley (ICPB)
SI
De Ley (IP)
397
De Ley (NCPPB)
~ 6 - 6 Manasse
TT9
De Ley (ICPB)
~6
Hildebrandt
11157 ATCC
~ 6 6Stonier
~
~ 6 6 c Stonier
421v
De Ley (IP)
v-I
Hamilton
11156 ATCC
4720
ATCC
~2
Beardsley
5
De Ley (NCPPB)
13333 ATCC (rhizogenes)
5GLYFE Beardsley
IIBNV6 Stonier
AG 19
Panagopoulos
178
Manasse
B-48
Manasse
SN
Manasse
?11og5 ATCC
?930 De Ley (NCPPB)
A . rubi
I3334 ATCC
I3335 ATCC
Pinto
bean
leaft
4+
4+
4+
4+
4+
4+
4+
4+
4+
4+
4+
3+
3+
3+
3+
2+
4+
4+
4+
4+
3+
3+
3+
3+
3+
3+
3+
3+
3+
2+
2+
2+
2+
2+
2+
2+
2+
2+
+
+-
-
-
-
-
3+
4+
Carrot discs
Tumours
43.
4+
4+
4f
434+
4+
3+
333+
3+
4+
4+
4+
4+
4+
2 f
2+
2+
2+
3+
3+
3+
3+
3+
333+
2+
2 f
3f
3+
3+
333f
3+
2+
23-
t
2+
+
k
+_
-
\
h
I
Growth on q-ketoaycoside
sucrose
producsalts
tion
medium
+
+
+
+
+
+
++
+
++
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
-
+
-
++
+
+-
k
-
-
3f
-
2+
-
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
+
+
+
+
+
++
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+-
-
3
Litmus
milk
colourf
T
T
T
T
T
T
P
T
T
T
P
T
T
T
T
T
T
P
T
P
T
T
T
T
T
T
T
T
T
T
T
T
T
BG
BG
T
T
T
T
T
T
T
T
BG
G
G
ATT
BPUTr
BG
T
--
Table
I
cont.
Infectivity
Species
and
strain
Source*
A . rhizogenes
11325 ATCC
TR I o I De Ley (ICPB)
TR7
De Ley (ICPB)
8196 De Ley (NCIB)
15384 ATCC
TRIO^ De Ley (ICPB)
A. radiobacter
R V ~ De Ley (IP)
TRI
De Ley (ICPB)
s 1005 De Ley (Gemb)
T R ~ De Ley (ICPB)
8 149 De Ley (NCIB)
4718 ATCC
I 3332 ATCC (rhizogenes)
6467 ATCC
?6466 ATCC
A . pseudotsugae
13330 ATCC
13331 ATCC
180
De Ley (NCPPB)
A . gypsophilae
13329 ATCC
B-305 Maas-Geesteranus
Pinto
bean
leaf?
-
-
+
+
Physiology
A
7
7
Growth on 3-keto
sucrose
glycoside Litmus
salts
produo
milk
Roots medium
tion
colour$
+
Carrot discs
Tumours
+
+
+
+
+
+
61
+
+
+
+
+
+
T
T
T
T
T
T
T
BG
PUG
-
+
-
YT
C
* Investigator or source supplying the culture. Abbreviations : ATCC, American Type Culture
Collection, Rockville, Maryland; ICPB, International Collection of Phytopathogenic Bacteria, Department of Bacteriology, University of California, Davis, California; NCIB, National Collection of
Industrial Bacteria, Torry Research Station, Aberdeen, Scotland; NCPPB, National Collection of
Plant Pathogenic Bacteria, Ministry of Agriculture, Fisheries and Food, Hatching Green, Harpenden,
England; IP, Institut Pasteur, Paris, France; Gemb, Institut Agronomique de I’gtat, Gembloux,
Belgium.
TThe infectivity ratings are as follows: 4 + = 85 % or greater % r.s.i.; 3+ = 20-79 % r.s.i.; 2 +
= 5-19% r.s.i.; + = 0-1-4% r.s.i.; - = no tumours observed in any tests and < 0.01% r.s.i.
$Colour of the litmus milk curd zone 3 weeks post-inoculation: T = tan; P = pink; B = blue;
G = grey; Y = yellow; A = Amber; Pu = purple; Tr = transparent; C = cream.
§Acid produced very slowly, a distinct pink colour appearing only after 5 weeks growth in this
medium.
were diluted 1/10 with distilled water and 0.2 ml. samples added to nephelometer
flasks containing 20 ml. of the sucrose salts medium + any additions to be tested.
Growth was followed hourly by measuring extinction at 660 nm. with a Bausch and
Lomb Spectronic 20 spectrophotometer. Similar growth measurements were made on
many of these strains with 0 . 5 % glucose instead of sucrose in the above medium,
and also by using the minimal medium of Starr (1946~).These media gave similar
results to those reported in the text, even when such precautions as acid-washing of
the nephelometer flasks, and when washed inocula and much smaller inocula (above
inoculum x I O - ~ ) were used to initiate the cultures.
+
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
62
J. A. L I P P I N C O T T A N D B. B. L I P P I N C O T T
RESULTS
Characteristics of Agrobacterium strains
The sources of the Agrobacterium strains and their infectivity, ability to form colonies on sucrose salts medium, response in litmus milk medium and ability to produce
3-ketoglycosides are shown in Table I . They are listed in descending order according
to their relative infectivity on carrot-root discs and pinto bean leaves or, if nontumourigenic, according to the extent they typify the accounts of their properties in
the current literature. For strains that differed from the response expected for the
species, new cultures were obtained and the tests repeated. The data in Table I are
partially summarized in Table 2 to show the most common responses for each of these
tests for the different species.
+
Table 2 . Characteristic responses of Agrobacterium species
Number of positive strains* which
I
Species
A. tumefaciens
A. rubi
A . rhizogenes
A . radiobacter
A . pseudotsugae
A . gypsophilae
No. of
strains
tested
48
2
6
8
3
2
induce
pinto
bean-leaf
tumours
40
2
2
0
0
0
induce carrot
I
A
form colonies on
sucrose
+salts
medium
produce
3-ketoglycosides
show litmus response
typical of
species
42
38
-
Tumours
Roots
32
2
6
-
2
36
0
0
I
0
6
-
0
0
0
0
0
-8
0
0
.
0
I
8
-
0
0
?
6
-
z
3
?
*The numbers underlined represent a sufficient portion of the strains tested to consider the factor
measured a positive characteristic of the species and those not underlined a negative characteristic.
Agrobacterium tumefaciens. Two strains (I 1095, 930) differed from the common
pattern of this species in all categories and, according to De Ley, Bernaerts, Rassel
& Guilmot (1966), have other differences which indicate that they should not be
considered members of this species or even of the genus. The infectivity of the remaining 46 strains varied considerably; six strains did not initiate tumours on carrotdiscs or bean leaves. The infectivity determinations of strains received from Professor
De Ley were in good agreement with those he has reported using different hosts
(De Ley et al. 1966). In addition to tumours, strains 396 and 398 promoted the formation of roots on carrot discs, 396 being somewhat more active than 398. This was
further tested by picking four colonies of each strain and determining their repsonse
on carrot discs. All showed some root-promoting activity. These responses, however,
were distinctly inferior to those obtained with A . rhizogenes, with both fewer discs
responding and those responding having fewer roots. A few roots were also initiated
on discs inoculated with strains ~6 and ~6-806in some tests.
Discounting strains 11095 and 930, only four of the remaining 46 strains did not
produce 3-ketoglycosides and these four also failed to form colonies on sucrose +
salts medium and differed from the other strains in showing an acid response in litmus
milk. Failure to produce 3-ketoglycosides, therefore, may be only one result of a
series of metabolic differences between these four strains and the other representatives
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
+
63
of A. tumefaciens. Eight of the 46 strains did not form colonies on sucrose
salts
medium, while strain T T was
~
marginaI in this respect. Strain 13333, received as
Agrobacterium rhizogenes, was considered to be an auxotrophic strain of A. tumefaciens because of its infectivity on pinto bean leaves, 3-ketoglycoside production,
litmus milk response and lack of root-inducing ability on carrot discs.
The response of these strains on litmus milk medium showed many variations relative to the rate at which the cultures became alkaline, the rate of litmus reduction,
the presence or absence and extent of proteolysis, and of pellicle formation. These
results are omitted, however, since they were not obviously related to infectivity
differences and were of no aid in distinguishing the separate Agrobacterium species.
The colour of the curd zone 2 to 3 weeks after inoculation was of value generally
(Riker et al. 1930; Huisingh & Durbin, 1967) and provided an additional characteristic to separate the majority of A. tumefaciens strains from those of A. rhizogenes.
The colour of the curd zone differed from the common A. tumefaciens response in
the case of questionable or unusual strains assigned to this species and this pattern
seemed to hold with the other species examined. The designation tan (T) in Table I
includes a wide variety of shades with delicate blue, grey or purple casts. Those
strains listed as giving a different litmus colour were quite distinct from these and
five cases (BG or G ratings) apparently arose from a failure to reduce the litmus.
These five strains were either non-infective or of low infectivity, suggesting that the
metabolic change responsible for this colour difference may be associated with their
decreased infectivity.
Agrobacterium rubi. Both strains of this species were highly infective; strain 13335
promoted the formation of a few roots in one test on carrot discs, similar to the response observed by McKeen (1954). As shown by Starr (1946b), these strains cannot
)
reported previously
grow on a glucose + salts medium and strain 13334 (= T R ~ was
to lack the ability to produce 3-ketoglycosides (De Ley et al. 1966). Too few strains
were available to determine a characteristic litmus milk reaction, although the most
common one is apparently the production of acid and a corresponding pink colour
(Hildebrand, 1940; McKeen, I 954). Thus, a general correlation may exist between
loss of ability to form 3-ketoglycosides and failure to form colonies on minimal
medium and the acid litmus milk response as noted with four strains of A . tumefaciens. Despite such differences, the infectivity of these strains and their similarity
to A. tumefaciens in guanine:cytosine (GC) ratio (De Ley et al. 1966) and DNA base
sequence (Heberlein, De Ley & Tijtgat, 1967) indicate their proper assignment to
this genus.
Agrobacterium rhizogenes. The hairy-root organism induces root formation, often
in association with tumours (Riker et al. 1930). This species is also characterized by
failure to grow on a salts + carbon source medium (Starr, 1946b), its inability to form
3-ketoglycosides (De Ley et al. 1966) and its production of acid in litmus milk medium
(Riker et al. 1930). In several respects, the four strains of A. tumefaciens and the A.
rubi strains which did not form 3-ketoglycosides fit this description, as do the majority
of A . rhizogenes strains. The latter, however, typically induced root formation on
carrot discs. Strains 13332 and 13333, though received as A. rhizogenes, were apparently wrongly classified in view of their failure to induce roots, their production of
3-ketoglycosides and their response on litmus milk medium. Strains I 5834 and
~ ~ 1 were
0 7 weakly tumourigenic on pinto bean leaves and ~ ~ 1 consistently
0 7
pro-
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
64
J . A . L I P P I N C O T T AND B. B. L I P P I N C O T T
duced as many colonies when plated on a sucrose + salts medium as when plated on
nutrient agar. Strain TR 107 was also given a " & " rating in tests of tumourigenicity
by De Ley et al. (1966) although three other strains of A. rhizogenes were scored
negative. Since TR 107 induced copious root formation on carrots and showed an acid
litmus milk response characteristic of the species, A. rhizogenes should not be considered strictly auxotrophic.
Agrobacterium radiobacter. The strains of this species in Table I showed the characteristics expected (Hofer, 1941; De Ley et al. 1966), but the responses of strain 6466
make it a questionable member of A. radiobacter.
In an attempt to establish some diagnostic criterion other than infectivity to distinguish Agrobacterium radiobacter from A. tumefaciens, several strains of different
species of agrobacteria were tested on four media reported to differentiate between
them. Growth of the bacteria on the congo red medium of Riker et al. (1930) gave
considerable colour variation from strain to strain, but did not provide a significant
separation. The ability to utilize nitrate and nitrite when grown on the medium of
Sagen et al. (1g34), to produce browning when plated on calcium glycerophosphate
medium (Hendrickson et al. 1934) or to turn media containing ethanol as sole carbon
source acid or alkaline (Sagen et al. 1934) were similarly of no diagnostic value.
Agrobacterium pseudotsugae. The responses of the three strains of this species showed
no obvious relation to those of A. radiobacter, A. rhizogenes, A. rubi or A. tumefaciens.
De Ley et al. (1966) and Heberlein et al. (1967) concluded that strain 180 should not
be a member of the genus Agrobacterium, though this was the only representative of
the species they examined. The infectivity ascribed to this organism is in question
since the original isolates did not initiate tumours when inoculated on Douglas fir,
and attempts to re-isolate a virulent organism from Douglas fir galls in the area of the
original isolations have not been successful (M. Schroth, personal communication).
The value of further studies with these strains relative to the crown-gall problem
appears to be low.
Agrobacterium gypsophilae. Our results with two strains are comparable with our
findings for A . pseudotsugae and show no relation to the other species of Agrobacterium (cf. De Ley et al. 1966; Moffett & Colwell, 1968).
Rate of tumour appearance
The ability of Agrobacterium tumefaciens to infect the smaller wound sites on pinto
bean leaves resulting in tumours that appear after 4 days from inoculation is directly
correlated with the ability of the bacterium to utilize the nutritional medium of this
host for activities essential for tumour initiation (Lippincott & Lippincott, I 966).
Figure I shows that three of the strains which induced pinto bean leaf tumours but
which failed to form colonies on sucrose + salts medium (Table I) gave the rate of
tumour appearance characteristic of auxotrophs.
Tumour appearance rates for other Agrobacterium strains were obtained to determine the significance of rate differences when natural strains were studied and to learn
whether the infectivity of the different strains was in any way correlated. Agrobacterium tumefaciens strain 13333, which was of low infectivity, did not grow on minimal
medium and showed a tumour appearance rate characteristic of auxotrophs (Fig. 2).
The yo r.s.i. of A. tumefaciens strains 11sv7and ~6 - 806 were 35 and 49, respectively,
and the tumour appearance rate of the former (Fig. 3) was nearly identical to that of
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
Strain
Mean of
six strains:
r
I
151
% r.s.i.t
- 15
~~107
+6
I5
v-I
Mean of 19strains:
396
11 I57
-1
+9
+ 2.8
10
9
62
+5
+9
I7
22
20
31
Mean of nine
strains :
-
-
-
-
-
-
-
Subgroup 111-B
+I
+7
B2
w-I
17805
c
H-I00
4-2
+6
+6
+3
+7
SI
Strain
Subgroup 111-A
f
-4
-6
I
~66c
11156
A66~
-5
+o
+6
+5
-2
28
24
58
52
49
44
35
I00
138
I 16
185
231
-
r.t.a. ratio
difference$
A
Tumour appearance group
397
398
4452
B 6-6
A6
I43
B6
TT6
4
B 6-806
TT9
IIBV 7
TTIII
ACH- I
c-58-3
% r.s.i.t
A
\
A
r.t.a. ratio
difference$ Strain
Group I1
Group I
I
-
37
-
I
41
I9
I7
16
6
55
64
I11
% r.s.i.t
Strain
+ 20
+43
-
+20
+20
+ I7
+ 26
+ I5
+ I3
+ 16
+ I5
P2
-
-
-
Mean of ten
strains :
15834
Subgroup IV-C
5
7420
I3333
Subgroup IV-B
I3334
AG I
I3335
AG 6
Subgroup IV-A
m133
3
7
+ 56
I 16
I
2
2
9
+ 32
+-22
+ 57
+ 52
+ 6
+43
- 2
432
204
198
165
81
62
+ 32
I- 22
+ 27
r.t.a. ratio
yo r.s.i.t difference$
A
r
A
r.t.a. ratio
difference$
Group IV
Group 111
*Strains of A. rubi and A. rhizogenes that induce pinto bean leaf tumours are included.
mean number of tumours/leaf/viable cell of A. tumefaciens strain B 6
x I00
t Per cent relative specific infectivity (% r.s.i.) = mean number of tumours/leaf/viable cell of the test strain s
mean number of tumours/leaf at day 4
x IOO for A. turnefaciens strain ~6 minus the comparable ratio for
$ Relative tumour appearance (r.t.a.) ratio difference =
,"
mean number of tumours/leaf at day 8
Q
the test strains. All 4:8 r.t.a. values are a mean of three or more independent determinations. The mean-day tumour count ratio x IOO for strain ~6was 44 +_ 1-5in 40 m
m
determinations.
M
u1
Table 3. Subdivision of strains of Agrobacterium tumefaciens according to their tumour-inducing
ability at small wound sites on pinto bean leaves*
66
J . A . LIPPINCOTT A N D B . B . L IPPIN C O TT
A. tumefaciens strain ~ 6while
,
the latter showed some deviation in the middle of the
tumour appearance curve (Fig. 2). Agrobacterium tumefaciens strains 4452, I I I 56,
B 2 had yo r.s.i. values of 22 or less and induced a greater proportion of the final
number of tumours which appeared during the first 4 days (Fig. 2, 3). These results
suggested there was some correlation between loss of ability to initiate tumours at
small wound sites and low infectivity. Tumour appearance curves for strains A.
tumefaciens I I I 58 and T-37 are shown relative to strain B 6 in Fig. 4. The 6-day tumour
count was taken as IOO % in this graph to illustrate the similarity between the tumour
I
I-
c
+O
2
4
6
8
1
Time after inoculation (days)
Fig. I
0
Fig.
2
Fig. I . Time of tumour appearance on pinto bean leaves inoculated with Agrobacterium
tumefaciens strain ~6 ( 0 ) ;A. tumefaciens strain 4720 (A); A . rubi strain 13334 (0);A . rubi
strain 13335 (0). The pday tumour count for each strain was taken as I O O and
~ ~ the number
of tumours observed at earlier times expressed relative to this value. Each curve is the mean
of two experiments in which each of the four strains was tested.
Fig. 2. Time of tumour appearance on pinto bean leaves inoculated with Agrobacterium
tumefaciens strain ~6 ( 0 ) ;strain 4452 (A); strain ~ 6 - 8 0 6 (0); strain 13333 (0). The
tumour count at day 10 for each strain was taken as I O O and
~ ~ the number of tumours
observed at earlier times expressed relative to this value. Except for strain 13333 (tested
in one experiment only) the curves are the mean of two experiments in which each of the
strains was tested.
appearance curves up to this time, and to show the divergence which occurred after
day 6. Thus, strains I I 158 and T-37, in addition to being highly infective, initiated
tumours which continued to appear for several days after the time new tumours
ceased to appear on leaves inoculated with strain ~ 6 This
. showed they could infect
even smaller wound sites than strain ~ 6 .
Subdivision of tumour-inducing strains based oyt tumour appearance data
Tumour apperance rates were determined for each of the strains of Agrobacterium
in Table I that were tumourigenic on pinto bean leaves. A shortened procedure was used
for most of these determinations. Figures I to 4 show that the day 4: 8 tumour count
ratio for each strain incorporated most of the information contained in these experiments relevant to differences in the rate of tumour appearance. Auxotrophs showed a
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
67
high ratio (in the more extreme cases the day 4: 8 ratio = 100%) while strains such as
I I 158 and T-37 had lower ratios than strain ~6 (=44 %). The results shown in Table
3, along with the % r.s.i. for each of the strains, gave a spectrum of values that varied
from ratios which were less than strain ~6 to those characteristic of the extreme found
for certain auxotrophs (Lippincott & Lippincott, I 966).
A division of the tumour-inducing agrobacteria into four groups based on these
ratios was made by establishing arbitrary divisions which grouped the strains according to the degree to which their nutrition (relative to their nutritional requirements
2140 TJ
i160
?
-
b
-
GI20Y
-
:loo-
m
-
80a
m
5
-
60-
3
+
+
A
-
40I
-
20-
G
G
2
4
6
8
1 0 1 2 1 4
Fig. 3
Fig. 4
Fig. 3. Time of turnour appearance on pinto bean leaves inoculated with Agrubacterium
tumefaciens strain s6 (0);strain ~2 (0); strain 11156(0);strain 11sv7(A). Otherwise see
description for Fig. I.
Fig. 4. Time of tumour appearance on pinto bean leaves inoculated with Agrobacterium
tumefaciens strain ~6 (0);strain T-37 (0);strain I I 158 (A). The 6-day tumour count for
each strain was taken as IOOyo and the counts at other times expressed relative to this value.
Otherwise as in Fig. I.
for tumour initiation) was accommodated at the smaller wound sites of the pinto bean
leaf. The basic criteria for the establishment of these groups were: group I, strains
which had a lower day 4 :8 tumour count ratio than strain B 6 ; group 11, strains with a
;
111, strains with ratios greater than strain
ratio similar to that of strain ~ 6 group
~6 and which initiated colonies on sucrose + salts medium; group IVYstrains which
did not form colonies on sucrose + salts medium. Each group is considered below
in greater detail.
Group I: strains which initiate tumours at smaller wound sites than Agrobacterium
tumefaciens strain ~ 6The
. day 4:8 tumour count ratios for the six strains of Agrobacterium tumefaciens placed in this group (Table 3) were distinctly lower than that of
A . tumefaciens strain ~ 6and
, five of these strains were either of similar or greater
infectivity. The infectivity of strains 15955 and TT 107 was sufficiently greater than
that of strain ~6 to suggest they had additional qualities which made them better
5-2
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
68
J . A . LIPPINCOTT AND B . B . LIPPINCOTT
pathogens for pinto bean leaves. Strain 421v appeared to be defective in the tumour
initiation process at some stage other than the nutritional stage through which the
bacterium must pass since it infected small and large wound sites in the same relative
proportion as the most infective strains, but had a low specific infectivity. Strains with
low day 4 : s tumour count ratios and low specific infectivity should prove useful for
crown-gall studies because their infectivity is apparently not nutritionally limited but
derives from subsequent defects in the tumour initiation process.
Figure 4 shows that the number of tumours on leaves inoculated with two of these
group I strains was not complete until 10 to 14days. When tumour counts were made
on days 10 through 13 after inoculation, the number of tumours induced by strains
I I 158 and T-37 reached a maximum on day 12. Theoretically, the conversion of a
single leaf epidermal cell to the tumourous state, followed by growth at the same rate
as other tumours during the period between inoculation and counting, should give
rise to tumours which first become detectable on days 12 to 13 (Lippincott & Lippincott, 1965). Strains assigned to group I, therefore, may initiate tumours which originate from the conversion and continued multiplication of single leaf cells.
Group 11: strains which initiate tumours over the same range of wound sites and in
similar proportion to Agrobacterium tumefaciens strain B 6 . The day 4 :8 tumour count
ratios at the extremes of this class were six units lower and nine units higher than those
of the corresponding strain ~6 controls (Table 3). The boundaries of this group were
determined in part by the distribution of the tumour count ratios for the different
strains and strengthened by apparent differences in the specific infectivity of the
strains excluded by these divisions. It was desirable to restrict the representatives of
groups I and I11 to strains that clearly differed from strain ~6 and would consequently
provide distinctive characteristics of potential value in further experimentation.
Strains at either extreme of group 11, therefore, may have more in common with
group I or group I11 strains, respectively, than with each other.
The % r.s.i. of the strains in group I1 was partially correlated with the day 4:8
tumour count ratios, decreased infectivity being associated in general with a decreased
ability to initiate tumours at small wound sites. Two of the 19 strains in group I1
were significantly more infective than strain B 6. Thus, strain ACH-I, despite a slight
restriction at smaller tumour sites, and strain C-58-3, appear better adapted for
tumour initiation on pinto bean leaves. Strain 11157 had the lowest yo r.s.i. of the
strains in this class and is apparently restricted in its ability to inititate tumours on
pinto bean leaves, in addition to its somewhat limited ability to infect small wound
sites. Strain 396 showed characteristics expected of a bacterium which is limited in its
ability to carry out one or more of the events in tumour initiation which occur after
the nutritional phase.
Group 111: prototrophic strains which initiate proportionately fewer tumours at small
wound sites than Agrobacterium turnefaciens strain ~ 6Strains
.
assigned to group I11
(Table 3) must develop a nutritional restriction with increasing frequency at progressively smaller wound sites which prevents completion of the tumour initiation
process. These limitations are not readily apparent in studies of colony-forming
ability on sucrose salts medium and consequently their nature is unknown. The
members of group I11 thus occupy an intermediate position between group I1 strains
and those in group IV which are nutritionally limited when studied in vitro.
The lower average infectivity of the strains of group I11 appears to result from the
+
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
69
nutritional limitation imposed by pinto bean leaves. Hypothetically, these restrictions
could be sufficient to prevent the initiation of tumours at all of the smaller wound
sites and strain TR 107 of Agrobacterium rhizogenes, though it forms somewhat questionable tumours (see below), shows this type of response. Most strains placed in
group 111, though of low to very low infectivity, are not suited for studies of the critical
factors in crown-gall formation because their decreased infectivity results from nutritional limitation. The nature of this limitation probably differed from strain to strain,
though all showed a similar rate of tumour appearance, just as adenine, methionine
and asparagine auxotrophs of A . tumefaceins all exhibited similar losses in infectivity
and changes in rate of tumour appearance (Lippincott & Lippincott, 1966). Group
I11 strains of high specific infectivity (e.g. H-100), however, must be highly efficient
in their ability to infect large wound sites. Such strains may prove useful because of
their restricted nature combined with high infectivity.
Agrobacterium rhizogenes strain TR 107 separated into its own subgroup, 111-B,
was noted earlier for its ability to form colonies on sucrose + salts medium while
retaining the other common characteristics of the species. The tumours induced by this
strain on primary pinto bean leaves are unusual in that many ‘tumours’ observed at
day 4 after inoculation apparently disappear by day 8. Further study is necessary to
determine if outgrowths induced by this strain, as well as by A. rhizogenes strain
15834 considered in group IV, should be considered equivalent to the tumours initiated by A . tumefaciens. In view of the genetic relation between A. rhizogenes and A.
tumefaciens (Heberlein et al. 1967), as well as the nature of the diseases they induce,
some overlap in the morphogenic products formed by different strains of these two
species could be expected.
Group IV: strains which do not form colonies on sucrose + salts medium and commonly initiate significantly fewer tumours at small wound sites than Agrobacterium
tumefaciens strain B 6. The experimentally induced auxotrophs of Agrobacterium tumefaciens studied previously provide the model for this group (Lippincott & Lippincott
1966). Strains 4720, 5 and 13333 exhibited the low infectivity, rate of tumour appearance and nutritional requirements (Table 3) characteristic of the induced auxotrophs
and all produced 3-ketoglycosides; these strains have been placed in a subgroup,
IV-B. The strains placed in subgroup IV-A include the two strains of A. rubi, were
highly infective and did not make 3-ketoglycosides. With two exceptions, they showed
a rate of tumour appearance characteristic of auxotrophs. The third subgroup, IV-C,
was added to accommodate the single strain of A. rhizogenes assigned to this group.
Theoretically, three types of auxotrophs might be expected to deviate from the
nutritional limitation on tumour initiation at smaller wound sites typical of group
IV: (I) those auxotrophs that find their growth requirement fully supplied by the
medium at all wound sites; (2) those auxotrophs with growth requirements that do
not affect the events leading to tumour formation; (3) those auxotrophs that require
little or no nutritionally supported activity for infection. Two of the ten strains
(AG I, AG6) in group IV showed day 4: 8 tumour count ratios typical of group I1
strains, possibly for one of these reasons.
The % r.s.i. values for the two Agrobacterium rubi strains were remarkably high in
view of the nutritional requirements of this species, as were the infectivities of the
other strains in subgroup IV-A. Strains ~ ~ 1 3~23 ,and 13335 in particular appear to
be better pathogens at large wound sites than A. turnefaciens strain ~ 6 Part
. of this
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
J . A. L I P P I N C O T T A N D B. B. L I P P I N C O T T
70
difference, however, may arise from the failure to obtain a comparable population
density in cultures of these strains relative to those of strain ~ 6After
.
48 hr the colony
count of strain 13335, for example, was commonly only 2 to 4 x log/ml. while that of
strain 136 was Iolo/ml. When grown to the lower population density strain ~6 has a
two- to three-fold higher specific infectivity (Lippincott & Heberlein, I 965). The
high infectivity of some of these strains may thus result from physiological differences
associated with lower population density rather than inherent differences in efficiency
of tumour initiation.
Growth requirements of group IV strains
The specific growth requirements of each of these strains was determined, starting
from the observation that all group IV strains could grow on sucrose + salts medium
supplemented with the compounds which Starr (1946b) reported as necessary for the
growth of Agrobacterium rubi. Curiously, half of the group IV strains grew in liquid
sucrose + salts medium with no supplements (Table 4); similar results were obtained
Table 4. Nutritional requirements of group IV Agrobacterium strains*
Growth characteristics on liquid medium
Growth requirements?
r
Strain
A
Agar
medium
4720
5
I3333
B6
IIBNV 6
P or G
P
TTI33
P2
G
G
I3335
AG6
N, p, G
AG I
73334
15834
P
0
G
G
G
N, p, G
G
Minus glutamate
\
I
h
Liquid
Doubling
Lag time$
medium
time (hr)
(hr)
3-Ketoglycoside positive
0
2.6
26
P
8.0
34
4'1
32
p (B)
0
3'2
16
0
2'7
16
3-Ketoglycoside negative
7'5
6.0
p, N(B)
3'4
4'2
0
4'0
0
p, N(B)
4'4
0
16.5
0
0
Plus glutamate ( I mg./ml.)
A
1 1
7
Doubling
time (hr)
Lag time$
(hr)
1-8
3'0
20
28
31
35
27
28
26
56
19
2-0
16
1'5
I1
I2
4'0
I9
1'9
2.5
3'0
1'7
2'4
3'5
5'8
16
27
20
20
23
I9
* Except strain I I B N V ~ , which does not induce tumours, and strain ~ 6which
,
is included for com-
parative purposes.
t The following abbreviations are used: B = biotin (10,ug./ml.); N = nicotinic acid (10,ug./ml.);
P = calcium pantothenate (Io,ug./ml.); G = glutamate (I mg./ml.).
2 Number of hr incubation to reach population density of equiv. 0.100extinction reading at
660 nm.
with Starr's medium (Starr, 1946a) and smaller inocula. Strains 5 and 13333 grown in
liquid medium required pantothenate for growth, A. rubi strains 13334 and 13335
required pantothenate + nicotinic acid and A. rhizogenes strain 15834 grew very
slowly unless glutamate was present. Glutamate decreased both the lag and doubling
times for all of these strains in the liquid medium. When colony formation on solid
medium was used as a growth assay, all but strains 13333 and 5 required glutamate.
This discrepancy between the results on liquid and solid media may be due to differences in the two types of inocula. For the plating experiments 0.1 ml. of either a 1/10"
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
71
or a 1/10'dilution of the cultures was spread over about 30 ml. of agar medium,
whereas the liquid cultures were inoculated by adding 0-2 ml. of a 10-1 dilution of the
starter culture to 20 ml. medium. The use of washed bacteria and much smaller
inocula to initiate liquid cultures, however, did not change the growth requirements
we report for A. rubi strains 13334 and 13335 in Table 4.
Addition of growth supplements to leaves inoculated with nutritionally dejicient
agroba cter ia
The compounds required for the growth of Agrobacterium rubi and A. rhizogenes
as described by Starr (19463) were applied to inoculated leaves in the experiments
shown in Table 5. All the strains tested in these experiments grew well when cultured
on sucrose + salts medium these compounds. These additions increased tumour
initiation by A. rubi strains 13334 and 13335 and A. tumefaciens strain 4720. Tumour
initiation by strain 13333 was not promoted, however, and strains 13332 and 11325
remained non-infective despite these additions. The addition of biotin and glutamate
+
Table 5. Efect on tumour initiation of the addition of supplements for bacterial growth
to leaves inoculated with Agrobacterium species
Strain
Species
Inoculum
Mean numbers of tumours/leaf
A
concentration
I % increase in
(colony count
Plus growth
tumour
ml. x I O ~ )
Control
supplements"
initiation
~6
I3334
13335
A. tumefaciens
A. rubi
A. rubi
0.97
0.46
0.34
B6
4720
I3334
13335
I3333
13332
1I325
A. tumefaciens
2'1
A. tumefaciens
A. rubi
A. rubi
A. tumefaciens
A. radiobacter
A. rhizogenes
I 6.0
0.48
0.76
12'0
54'0
45'0
Experiment I
37'5
41.5
30-2
41.2
48.1
78.6
Experiment 2
97
67
31
41
9.6
I 5.2
44
79
0.7
0.5
0
0
0
0
+ I1
+36
+ 63
0
+ 32
+ 58
+ 80
0
-
-
*A solution containing biotin (10 ,ug./ml.), nicotinic acid (10 pg./ml.), sodium pantothenate (10
pg./ml.) and glutamic acid (1.0 mg./ml.) adjusted to pH 7 was added hourly to the inoculated leaves
for a total of 12 additions beginning 15 min after inoculation.
(the growth requirements reported for A. rhizogenes) (Starr, 19463) to inoculated
leaves at 0.1 mg./ml. did not promote tumour formation by the two strains of A.
rhizogenes (15834, TR 107) which commonly initiate a few tumours on the pinto bean
leaves, and did not produce tumour formation on this host by A. rhizogenes strains
TRIOI and TR7. Strain IIBNV6, an avirulent strain of A. tumefaciens often used in
comparative studies of crown-gall tumour initiation, did not form colonies on sucrose
+ salts medium and was not induced to form tumours by adding these compounds.
Nutritional limitation, therefore, cannot be the sole explanation for the low infectivity
of Agrobacterium strains 13333, 15834 and TR 107 or for the lack of tumour-inducing
ability of strains 1 1 ~ ~I ~1325,
6 , TR 101 and T R on
~ pinto bean leaves.
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
J . A . LIPPINCOTT A N D B . B . LIPPINCOTT
DISCUSSION
The 44 strains of various Agrobacterium species which induced tumours on pinto
bean leaves exhibit different capacities to initiate tumours at small wound sites. They
range from those strains that infect only large wound sites to those that induce
tumours arising from wounds which approach the theoretical size limit, the conversion
of a single leaf-cell. Because the nutritional requirements of the bacterium between
inoculation and the time at which the bacterium becomes dispensable in the tumour
initiation process are directly related to the ability of the bacterium to initiate tumours
at small wound sites (Eippincott & Lippincott, 1966), the differences in the rate of
tumour appearance for these 44 strains must have their origin in bacterial nutritional
effects. Because the results presented here for naturally occurring auxotrophic strains
agree in full with the previous findings for induced auxotrophs, this conclusion appears
well founded.
Other mechanisms may be proposed, however, to account for these results where
strains other than auxotrophic derivatives of a single virulent strain are involved.
Thus, some strains may conceivably require more time to initiate tumours, which
would delay the appearance of tumours. Or the growth rate of tumours induced by
some strains may be less than others, again affecting a general delay in the rate of
tumour appearance. Finally, some strains might find the wound sap medium at large
wound sites to be inhibitory relative to the medium at small wound sites. Each of
these possibilities would yield results which would suggest greater ability to infect
small wound sites than if the nutritional relation between bacterium and wound sap
medium were the only governing factor. There is no evidence as yet, however, to
suggest that one or more of these alternatives apply to any of the 44 strains. For several
strains in each group, the combination of specific infectivity results with daily tumour
counts made over many days is sufficient to rule out these alternatives. We have consequently extended the nutritional explanation to account for all of these differences
because of its consistency with the data at hand and its simplicity.
The infectivity of Agrobacterium species on the pinto bean leaf (and probably
elsewhere) depends upon: (I) survival of the bacterium at a wound site for some
unknown period of time up to a maximum of 24 hr (Lippincott & Lippincott, 1965;
1967); (2) a bacterial nutritional phase which may be necessary for survival as well as
for production of tumour initiation factors (Lippincott & Lippincott, 1966; and
present results); (3) the production of tumour-inducing agents. In addition to the
potential diminution or loss of ability to infect small wound sites which is commonly
associated with auxotrophy, prototrophic strains can be expected which would find
the medium at the smaller pinto leaf wound sites to be nutritionally limiting; those
strains assigned to group 111 apparently belong to this category. The strains with low
specific infectivity classed in group I or 11, however, are probably partially defective
in the production of one or more of the agents essential for tumour initiation. Thus,
the combination of specific infectivity tests with data on rates of tumour appearance
provides a means to detect bacterial strains specificallydefective in stages of the tumour
initiation process subsequent to the bacterial nutrition phase.
Those strains of Agrobacterium placed in groups 111 and IV may be expected t o
show a more limited host range and/or lower infectivity on many hosts than those
in groups I and 11. The natural host range of A . rubi (Pinckard, 1935; McKeen, 1954)
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
73
and A . rhizogenes (Riker et al. 1930) appears to be more limited than that of A . tumefaciens. This may be the case for other strains in groups I11 and IV, though little
information is available. Since about 40 y-, of the tumourigenic strains examined were
placed in groups 111 and IV, genetic changes which alter the growth requirements of
the pathogen apparently constitute a major source from which natural variations in
infectivity and possibly host range and host specificity have evolved in this genus.
Complete loss of pathogenicity by this means appears to be a distinct possibility.
The nutritional requirements of the pathogen, however, must be only one of many
potentially limiting steps in the infection process through which variations in pathogenicity could arise. In groups I and I1 strains of Agrobacterium tumefaciens such as
421v, 396 and 11sv7 showed little evidence of nutritional limitation when compared
, all were much less infective. Also, prototrophic
with A . tumefaciens strain ~ 6 yet
A . radiobacter strains were non-infective though very closely related to A . tumefaciem
(Graham, 1964; De Ley et al. 1966; Heberlein et al. 1967). Although two of the A .
rhizogenes strains consistently induced a few tumours on pinto bean leaves, the cornpounds necessary for their growth on sucrose salts medium did not increase their
infectivity nor did they allow A . rhizogenes strains which were avirulent on pinto bean
leaves to induce tumours. Pathogenicity and host range within the genus Agrobncterium must obviously depend on a complex of variables and their interactions.
The specific infectivity of the bacterial strains in three of the four tumour-appearance
groups shows a general correlation with the group to which they were assigned, those
in group I being most infective and those in group 111least infective. Bacterial strains
with a much higher or lower specific infectivity than the average value, however,
were found within each group. These aberrant strains appear to offer good experimental material for further crown-gall studies.
In agreement with the results of Bernaerts & De Ley (1963, 1967), Kern (I 966) and
De Ley et al. (1966), the ability to produce 3-ketoglycosides was shown to be unrelated
to infectivity within the genus. All the infectious strains which lacked the ability to
produce 3-ketoglycosides, however, also lacked the ability to form colonies on sucrose
+ salts medium, suggesting that loss of the ability to produce 3-ketoglycosides might
be associated with a specific nutritional deficiency. No obvious relation between a
specific nutrient requirement and this response was found, though there appeared to
be a correlation between an acid litmus milk response and the inability to form 3ketoglycosides. Since the defect must arise from a change in carbohydrate metabolism,
the glutamate requirement for colony formation in the 3-ketoglycoside-negative
strains might result from a defect different from that in the 3-ketoglycoside-positive
group IV strains.
The difficulties we have experienced in separating strains of Agrobacterium tumefaciens from A . radiubacter in the absence of infectivity tests have vexed many investigators (Hendrickson et al. 1934; Hofer, 1941; Sagen et al. 1934; Schroth, Thompson & Hildebrand, I 965 ; Huisingh & Durbin, 1967). One possible complicating factor
is the loss of virulence observed after growth of A . tumefaciens under certain culture
conditions (van Lanen et al. 1952a, b). Graham (1964, De Ley et al. (1966) and
Heberlein et al. (1967) have recommended combining the two species based on their
taxonimetric and deoxyribonucleic acid similarities. The student of crown-gall
physiology, however, is left with the dilemma of deciding to what extent current
members of A . radiobacter may be used as prototypes of non-virulent strains of A .
+
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
J. A. L I P P I N C O T T AND B. B. L I P P I N C O T T
74
tumefaciens, and whether strains previously described as non-virulent forms of A .
tumefaciens are strains of A . radiobacter. This confusion may have been present in
the A . radiobacter strains used by Graham (1964), De Ley et al. (1966) and ourselves,
with the results that a specific group of differences was not observed. To be certain
of having a non-virulent strain of A . tumefaciens for physiological studies, therefore,
non-virulent prototrophic strains isolated from a genetically marked virulent strain
treated with a mutagen are to be recommended.
This investigation was supported by Public Health Service research grant CA-05387
from the National Cancer Institute. The authors are indebted to Mary Kenny and
Ingebjorg Kvam for excellent technical assistance.
REFERENCES
ARK,P. A. & THOMPSON,
J. P. (1961). Detection of hairy root pathogen, Agrobacterium rhizogenes,
by the use of fleshy roots. Phytopathology 51,69.
BERNAERTS,
M. & DELEY,J. (1963). A biochemical test for crown gall bacteria. Nature, Lond. 197,406.
M. & DE LEY,J. (1967). Mechanism of the 3-ketolactose test for Agrobacterium. Arch.
BERNAERTS,
Mikrobiol. 56, 8I.
DE LEY,J., BERNAERTS,
M., RASSEL,
A. & GUILMOT,
J. (1966). Approach to an improved taxonomy
of the genus Agrobacterium. J. gen. Microbiol. 43,7.
DE ROPP.R. S. (1950). The comparative growth promoting action of indole-3-acetic acid and Agrobacterium tumefaciens. Am. J. Bot. 78, 352.
GRAHAM,
P. H. (1964). The application of computer techniques to the taxonomy of the root-nodule
bacteria of legumes. J. gen. Microbiol. 35, 5 I I.
HEBERLEIN,
G. T. & LIPPINCOTT,
J. A. (1965). Photoreversible ultraviolet enhancement of infectivity
in Agrobacterium tumefaciens. J. Bact. 89, 1511.
R. (1967). Deoxyribonucleic acid homology and taxonomy
HEBERLEIN,
G. T., DELEY,J. & TIJTGAT,
of Agrobacterium, Rhizobium, and Chromobacterium. J. Bact. 94, I 16.
HENDRICKSON,
A. A., BALDWIN,
I. L. & RIKER,A. J. (1934). Studies on certain physiological characters of Phytomonas tumefaciens, Phytomonas thizogenes and Bacillus radiobacter. J. Bact.
28, 597.
HILDEBRAND,
E. M. (1940). Cane gall of brambles caused by Phytomonas rubi n. sp. J. agric. Res.
61, 685.
HOFER,A. W. (1941). A characterization of Bacterium radiobacter (Beijerinck and Van Delden)
Lohnis. J. Bact. 41,193.
HUISINGH,D. & Durbin, R. D. (1967). Physical and physiological methods for differentiating among
Agrobacterium rhizogenes, A. tumefaciens, and A. radiobacter. Phytopathology 57,922.
KERN,H. (1966). Zum Nachweis der Virulenz bei Agrobacterium tumefaciens. Arch. Mikrobiol. 53,gz.
KLEIN,R. M. & TENENBAUM,
I. L. (1955). A quantitative bioassay for crown-gall tumour formation.
Am. J. Bot. 4,
709.
VAN LANEN,
J. M., BALDWIN,
I. L. & RIKER,A. J. (1g5za). Attenuation of crown gall bacteria by
cultivation in media containing glycine. J. B a t . 63,715.
VAN LANEN,
J. M., RIKER,A. J. & BALDWIN,I. L. (I952b). The effect of amino acids and related
compounds upon growth, the virulence and enzyme activity of crown gall bacteria. J. Bact.
63,723.
LIPPINCOTT,
B. B. & LIPPINCOTT,
J. A. (1966). Characteristics of Agrobacterium tumefaciens auxotrophic mutant infectivity. J. Bact. 92, 937.
LIPPINCOTT,
J . A. & HEBERLEIN,
G. T. (1965). The quantitative determination of the infectivity of
Agrobacterium tumefaciens.Am. J. Bot. 52, 856.
LIPPINCOTT,
J. A. & LIPPINCOTT,
B. B. (1965). Timing of events in crown-gall tumour development
on pinto bean leaves. Devl Biol. 12, 309.
LIPPINCOTT,
J. A. and LIPPINCOTT,
B. B. (1967). Time required for tumour initiation by Agrobacterium
tumefaciens on pinto bean leaves. Nature, Lond. 213, 596.
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06
75
LIPPINCOTT,
J. A., WEBB,J. H. & LIPPINCOTT,
B. B. (1965). Auxotrophic mutation and infectivity of
Agrobacterium tumefaciens. J. Bact. 90, I I 55.
MCKEEN,W. E. (1954). A study of cane and crown-galls on Vancouver Island and a comparison of
the causal organisms. Phytopathology 44, 65 I.
MOFFETT,
M. L. & COLWELL,
R. R. (1968). Adansonian analysis of the Rhizobiacea. J. gen. Microbiol. 51,245,
J. A. (1935). Physiological studies of several pathogenic bacteria that induce cell stimulaPINCKARD,
tion in plants. J . agric. Res. 50, 933.
RIKER,A. J., LYNEIS,
M. M. & LOCKE,S. B. (1941). Comparative physiology of crown gall, attenuated
crown gall, radiobacter and hairy root bacteria. Phytopathology 31, 964.
W. M., WRIGHT,W. H., KEITT,G. W. & SAGEN,H. E. (1930). Studies on
RIKER,A.J., BANFIELD,
infectious hairy root of nursery apple trees. J . ngric. Res. 41,507.
SAGEN,H. E., RIKER,A. J. & BALDWIN,
I. L. (1934). Studies on certain physiological characters of
Phytomonas tumefaciens, Phytomonas rhizogenes and Bacillus radiobacter. J . Bact. 28, 57 I.
SCHROTH,
M. N., THOMPSON,
J. P. & HILDEBRAND,
D. C. (1965). Isolation of Agrobacterium tumefaciens-A. radiobacter group from soil. Phytopathology 55, 645.
STARR,M. P. (1946a). The nutrition of phytopathogenic bacteria. I. Minimal nutritive requirements
of the genus Xanthomonas. J. Bact. 51, 131.
STARR,M.P. (1946b). The nutrition of phytopathogenic bacteria. 11. The genus Agrobacterium. J .
Bact. 52, 187.
IP: 78.47.27.170
On: Fri, 14 Oct 2016 09:23:06

Turnour-initiating Ability and Nutrition in the Genus

Transcription

Similar documents

Chryseomonas luteola - International Journal of Systematic and

Moraxella: Classification and Taxonomy

SCIENTIFIC NOTE

Read Article - DeSimone Research Group

Foreword - touchONCOLOGY

The orthopedic-oncology service - Hôpital Maisonneuve

Phycology / Microbiology

from u