INTERNATIONAL JOURNAL of SYSTEMATIC BACTERIOLOGY
April 1974, p, 238-241
Copyright 0 1974 International Association of Microbiological Societies
Vol. 24, No. 2
Printed in U.S.A.
Identification of Agrobacterium gypsophilae Strains NCPPB
179 and NCPPB 1948 as Erwinia herbicola
D. C. GRAHAM and C. E. QUINN
Department of Agriculture and Fisheries for Scotland, Agricultural Scientific Services,
East Craigs, Edinburgh EH12 8NJ, Scotland
On the basis of 56 characters, the organisms known as Agrobacteriurn
gypsophilae NCPPB 179 and NCPPB 1948 are identified as strains of Erwinia
herbicola. These strains did not cause gall formation when inoculated into four
species of Gypsophila or into Lychnis chalcedonica.
In 1934, Brown (3) described a bacterium
which she isolated from galls found at the
unions of stock and scion on plants of
Gypsophila paniculata in the United States. The
organism, named Bacterium gypsophilae, was a
gram-negative rod which was motile by means
of several bipolar flagella. It formed yellow
colonies on beef infusion agar, was facultatively
anaerobic, liquefied gelatin slowly, reduced
nitrate to nitrite, did not produce indole, did
not hydrolyze starch, and produced acid from
glucose, sucrose, maltose, mannitol, and glycerol, but not from lactose. Brown stated that
the organism caused galls when inoculated into
healthy plants of G. paniculata and that it was
also pathogenic t o species of Silene, Dianthus,
Lychnis, and Saponaria. Because of its purported capability to cause plant galls, it was
transferred to the genus Agrobacteriurn by
Starr and Weiss (1 6 ) .
Maas Geesteranus and Barendsen (14) isolated a similar organism from galls on roses in
the Netherlands and stated that pathogenicity
tests showed the organism to be the cause of
the galls. We isolated another yellow-pigmented, gram-negative, rod-shaped bacterium
from galls at the graft union of G. paniculata
cultivar (cv) Bristol Fairy, growing in a local
nursery.
De Ley et al. ( 6 ) studied the deoxyribonucleic acid base composition of many isolates
of agrobacteria and pointed out that the
organism in the National Collection of Plant
Pathogenic Bacteria (NCPPB) named Agrobacteriurn gypsophilae NCPPB 179 had a mol%
G+C value significantly lower than that of A .
tumefaciens, A . rubi, or A . rhizogenes. In
addition, they found that NCPPB 179 was not
pathogenic to Datura strarnonium or Lycopersicon esculentum. Because many of the
phenotypic characters of NCPPB 179 also did
not agree with those of other members of the
genus Agrobacterium but corresponded more
closely with those of members of the family
Enterobacteriaceae, De Ley ( 5 ) suggested that
the organism might be a member of that family,
and, because of its yellow pigmentation, it
might be related to Erwinia herbicola. White
(17) drew attention to the fact that the
organism from rose galls (NCPPB 1948) might
be related to the genus Erwinia.
In view of these possibilities, we carried out
studies to establish the identity of the organisms NCPPB 179 and NCPPB 1948.
MATERIALS AND METHODS
Bacterial strains. A . gypsophilae NCPPB 179 was
originally B. gypsophilae 179 in the collection of the
late W. J. Dowson, who had received a culture of this
strain from M. P. Starr. According to the Catalogue of
Strains (1) of the American Type Culture (ATCC), A .
gypsophilae Dowson 179 is the same organism as A .
gypsophilae ATCC 13329, the latter having been
passed by N. A. Brown to H. J. Conn and thence to M.
P. Starr, who deposited it in the ATCC. ATCC 13329
(=NCPPB 179) has been designated (12) as the type
strain of A . gypsophilae. As mentioned above, NCPPB
1948 is the organism isolated from roses in the
Netherlands (14). Also included for comparison in the
pathogenicity tests were Erwinia milletiae strain JMI
isolated from galls on Milletia floribunda in Japan and
received from M. Goto, the organism isolated by us
from G. paniculata, and two other strains of E.
herbicola, which had been isolated by us from plant
material, The characters of E. milletiae are given by
Graham and Hodgkiss (10) and correspond with those
of E. herbicola; the organism from G. paniculata and
the two isolates from other plant material were
identified as E. herbicola as described by Graham and
Hodgkiss (10).
Methods. For pathogenicity tests, healthy plants of
G. paniculata cv Bristol Fairy were selected at a local
nursery, potted into pots (20-cm diameter), and kept
238
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 00:18:07
239
IDENTIFICATION OF A . G YPSOPHILAE
VOL. 24, 1974
in a greenhouse at 16 to 21 C. When the stems of the
Plants were about 15 cm high, a l l stems on one Plant
were inoculated with isolate NCPPB 179, another
plant was inoculated with isolate NCPPB' 1948, a
third was inoculated with our own isolate from galls,
and the fourth was left as a control. The inoculation
was done by placing a large loopful of bacteria from a
24-h nutrient agar culture (grown at 25 C) on the graft
union and pricking the organism into the tissue with a
sterile needle. The control plant had sterile water
pricked into the graft union instead of bacteria. Soil
was then drawn up around the inoculation points so
that the graft unions were not exposed. Gypsophila
elegans, G. pacifica, G. repens, and Lychnis chalcedonica (grown from seed) were inoculated with each
of the six organisms mentioned above in the same way
at their stem bases when they were about 10 cm high.
Five plants were inoculated with each isolate, and five
were kept as controls.
All plants were covered with plastic bags after
inoculation to prevent quick drying at the inoculation
points; the bags were removed after 48 h. Temperatures were kept at 16 to 21 C and, although the
humidity was not controlled, it was kept high by
regularly watering the plants and spraying down the
floor and benches of the greenhouse. The plants were
grown for 15 weeks and examined periodically for gall
formation.
The methods used to characterize NCPPB 179 and
NCPPB 1948 were the same as those described by
Graham and Hodgkiss (10). Flagella were observed on
cells grown on nutrient agar for 48 h at 20 C with a
JEOL JEMlOOB electron microscope using an accelerating voltage of 80 kV and an initial magnification
of X10,OOO; the cells were negatively stained with a
1% wt/vol potassium phosphotungstate solution. Observations were also made on cells selected for motility
using the Craigie tube technique. Sloppy agar (1%
wt/vol peptone [Oxoid] + 0.5% [wt/vol] meat extract
[Lemco] + 0.3% [wt/vol] agar) was placed to a depth
of about 5 cm in a screw-capped vials containing
Craigie tubes. Inoculations were made onto the agar
surface in the Craigie tube; cultures were incubated at
25 C, and, when growth reached the surface of the
agar outside the tube (usually about 72 h), a loopful
was transferred from the surface to a tube of ordinary
nutrient agar.
gave the same reactions in the following tests.
Positive results were obtained for: fermentative
metabohm of glucose in the O/F (Hugh and
Leifson) test; acid production from glucose,
maltose, trehalose, rhamnose, xylose, sucrose,
arabinose, mannitol, inositol, glycerol, mannose, and salicin; liquefaction of gelatin; utilization of citrate, malonate, acetate, mucate,
galacturonate, and tartrate; production of a
yellow, nondiffusible pigment; production of
nitrite from nitrate; catalase production;
growth in 5 % NaCl broth, and p-nitrophenyl0-D-galactopyranoside hydrolysis. Negative results were obtained for: gas from glucose; acid
production from lactose, raffinose, erythritol,
adonitol, dulcitol, and inulin; production of
indole; pectate gel liquefaction; hydrolysis of
starch; rotting of potato slices; production of
oxidase, lysine, and ornithine decarboxylases,
and arginine dihydrolase, urease, and extracellular deoxyribonuclease; and growth in 10%NaCl
broth and in KCN. The two organisms gave
different reactions in the tests shown in Table
1.
Examination of these two strains under the
electron microscope showed that most cells
were nonflagellated, but both contained some
cells with only a single lateral flagellum and a
very few cells showed several lateral flagella.
After selection of motile organisms using
Craigie tubes, both strains yielded cultures
containing many cells bearing several lateral
flagella (Fig. 1). Brown (3) stated that the
flagellation of Bacterium gypsophilae was bipolar, but, as Graham and Hodgkiss pointed out
(10 ) , it is easy to make mistakes with the light
microscope.
The phenotypic characters described above
are similar t o those of the yellow-pigmented,
fermentative rods from various habitats identified by Graham and Hodgkiss (10) as Erwinia
TABLE 1. Characters by which strains NCPPB
1 79 and NCPPB 1948 diffef'
RESULTS AND DISCUSSION
Although the inoculated Gypsophila and
Lychnis plants were kept for 15 weeks, there
was no evidence of gall formation or any other
pathological response. Knosel (1 1) also found a
strain of A . gypsophilae that was not pathogenic to Gypsophila paniculata. Examination of
plants of G . panicdata cv Bristol Fairy grown
at several nurseries in Scotland showed that
gall-like formation often developed at the graft
unions; these galls appeared t o be callus tissue.
A . gypsophilae NCPPB 179 and NCPPB 1948
contained gram-negative, motile rods which
Character
Symplasmata
Biconvex bodies
Growth at 37 C
Acid on ethanol agar
Methyl red test
Voges-Proskauer test
H, S production
Lipolysis
Gluconate test
Phenylalanine deaminase
a
NCPPB
179
NCPPB
1948
+
+
+
+
+
+
+
-
See Graham and Hodgkiss (10) for test methods.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 00:18:07
240
GRAHAM AND QUINN
INT. J. SYST. BACTERIOL.
philae NCPPB 179 and NCPPB 1948 are strains
of Erwinia herbicola (Lohnis) Dye.
In the Index Bergeyana (4), the authorities
for the name E. herbicola are given as “(Geilinger) Dye,” but Lohnis (13) used the name
Bacterium herbicola for this organism in 191 1,
thus antedating Geilinger’s use of the specific
epithet herbicola by some 10 years. The name
Bacterium herbicola was used by Lohnis in a
Table for the identification of bacteria, and
there is no doubt that the organism referred to
is Bacterium herbicola aureum Duggeli (7), the
name from which the epithet herbicola was
derived. It might be argued that th.e name B.
herbicola was only incidentally mentioned by
Lohnis and therefore was not validly published
(Rule 12c, International Code of Nomenclature
of Bacteria [S]), but this does not seem
convincing because Lohnis obviously intended
the name t o be used for organisms identified by
using the Table, and hence the name cannot be
FIG. 1. Agrobacterium gypsophilae strain NCPPB regarded as having been merely incidentally
1 79; 484 culture on nutrient agar at 20 C.Negatively
stained with potassium phosphotungstate. X 14,000. mentioned. The name Bacillus herbicola was
used as long ago as 1905 by Beijerinck (2), but
the context in which the name was used makes
herbicola. The differences between NCPPB 179 it clear that this was an incidental mention, and
and NCPPB 1948 are well within the range of thus this name was not validly published (see
variability of this group of organisms as Ewing and Fife [ 9 ] for a translation of the
described by Graham and Hodgkiss (10). In relevant part of Beijerinck’s paper). Ewing and
slide-agglutination tests, both organisms agglu- Fife (9) have recently proposed that Erwinia
tinated with antiserum (diluted 1 : 10 with 0.8% herbicola be transferred t o the genus Entero[wt/vol] saline) prepared against a strain of E. bacter as Entero bacter agglomerans (Beijerinck)
herbicola (serotype 2 ; see Muraschi et al. [ 151 ). Ewing and Fife. However, we prefer t o retain
This helped t o confirm the identity of the the organism in the genus Erwinia as Erwinia
organisms.
herbicola until comparative studies of all genera
There is no reason to doubt the authenticity presently placed in the family En terobacteriof A. gypsophilae NCPPB 179. Its history is aceae clarify the taxonomy and nomenclature
known, and its characters correspond with of these organisms.
those given by Brown (3) except with regard to
flagellation (a character which is open to
ACKNOWLEDGMENTS
misinterpretation) and plant pathogenicity. Although NCPPB 179 did not cause gall formaWe thank Thelma F. Muraschi for the gift of
tion in our tests, this does not necessarily mean antisera prepared against strains of Erwinia herbicola,
that NCPPB 179 is a bacterium different from and W. M. R. Laidlaw for the electron micrograph.
Brown’s organism because it is possible that
NCPPB 179 was once pathogenic and has lost
REPRINT REQUESTS
virulence in culture. Furthermore, it seems
unlikely that cultures of the original organism
Address reprint requests to: Dr. D. C. Graham,
have become replaced by a different bacterium Department of Agriculture and Fisheries for Scotland,
having the characteristics of the original orga- Agricultural Scientific Services, East Craigs, Edinburgh
nism excepting plant pathogenicity and flagella- EH12 8NJ, Scotland.
tion. It is noteworthy that NCPPB 1948 and E.
milletiae JMI were also nonpathogenic to
LITERATURE CITED
Gypsophila and Lychnis, but pathogenicity was
not tested on the hosts from which these strains
1. American Type Culture Collection. 1972. Catawere first isolated.
logue of strains, 10th ed. American Type Culture
Collection, Rockville, Md.
It is concluded that Agrobacterium gypso-
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 00:18:07
VOL. 24, 1974
IDENTIFICATION OF A . GYPSOPHILAE
2. Beijerinck, M. W., and A. Rant. 1905. Wundreiz
Parasitismus und Gummiflus bei den Amydaleen.
Zentralbl. , Bakteriol. Parasitenk. lnfektionskr.
Hyg. Abt. 11~366-375.
3. Brown, N. A. 1934. A gall similar to crown gall,
produced on Gypsophila by a new bacterium. J.
Agr. Res. 48:1099-1112.
4. Buchanan, R. E., J. G . Holt, and E. F. Lessel (ed).
1966. Index bergeyana. The Williams & Wilkins
Co., Baltimore, Md.
5. De Ley, J. 1968. DNA base composition and
hybridization in the taxonomy of phytopathogenic bacteria. Annu. Rev. Phytopathol. 6:63-90.
6. De Ley, J., M. Bernaerts, A. Rassel, and J.
Guilmot. 1966. Approach to an improved taxonomy of the genus Agrobacterium. J. Gen. Microbiol. 43:7-17.
7. Diiggeli, M. 1904. Die Bacterien flora gesunder
Samen und daraus gezogener Keimpflanzchen.
Zentralbl. Bakteriol. Parasitenk. Infektionskr.
Hyg. Abt. 12~602614;13:56-63, 198-207.
8. Editorial Board of the Judicial Commission of the
International Committee on Nomenclature of
Bacteria (ed). 1966. International code of nomenclature of bacteria. Int. J. Syst. Bacteriol.
16:459-490.
9. Ewing, W. H., and M. A. Fife. 1972. Enterobacter
agglomerans (Beijerinck) comb. nov. (the Herbi-
10.
11.
12.
13.
14.
15.
16.
17.
241
cola-Lathyri bacteria). Int. J. Syst. Bacteriol.
22:4-11.
Graham, D. C., and W. Hodgkiss. 1967. Identity
of gram negative, yellow pigmented fermentative
bacteria isolated from plants and animals. J . Appl.
Bacteriol. 30: 175-189.
Knosel, D. 1962. Untersuchugen an sternbildenden Bakterien. Arb. Landwirtsch. Hochsch.
Hohenheim 15: 1.
Lessel, E. F. 1962. Bacterial type cultures of the
American Type Culture Collection. lnt. Bull.
Bacteriol. Nomencl. Taxon. 12:71-88.
Lohnis, F. 191 1. Landwirtschaftlich-bakteriologishes Prakticum. Gebruder Borntraeger, Berlin.
Maas Geesteranus, H. P., and H. Barendsen. 1966.
Host plants of Agrobacterium gypsophilae. Neth.
J. Plant Pathol. 72:231-232.
Muraschi, T. F., M. Friend, and D. Bolles 1965.
Erwinia-like microorganisms isolated from animal
and human hosts. Appl. Microbiol. 13:128-31.
Starr, M. P., and J. E. Weiss. 1943. Growth of
phytopathogenic bacteria in a synthetic asparagin
medium. Phytopathology 33:314-318.
White, L. 0. 1972. The taxonomy of the
crow n-gall organism Agrobacrerium tum efaciens
and its relationship to rhizobia and other agrobacteria. J. Gen. Microbiol. 72565-574.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 00:18:07