JOURNAL OF BACTERIOLOGY, July 1967, p. 116-124Copyright © 1967 American Society for Microbiology

Deoxyribonucleic Acid Homology and Taxonomy ofAgrobacterium, Rhizobium, and Chromobacterium

GARY T. HEBERLEIN,1 J. DE LEY, AND R. TIJTGAT

Laboratory for Microbiology, Faculty of Sciences, State University, Gent, Belgium

Received for publication 17 February 1967

Hybridization experiments were carried out between high molecular weight, de-natured, agar-embedded deoxyribonucleic acid (DNA) and homologous, nonem-bedded, sheared, denatured "4C-labeled DNA from a strain of Agrobacteriumtumefaciens and Rhizobium leguminosarum (the reference strains) in the presenceof sheared, nonembedded, nonlabeled DNA (competing DNA) from the same or

different nomen-species of Agrobacterium, Rhizobium, Chromobacterium, andseveral other organisms. Percentage of DNA homology was calculated from theresults. The findings are discussed in relation to previous taximetric studies, presentclassification schemes, and guanine-cytosine content of the DNA. Strains of A.tumefaciens, A. radiobacter, A. rubi, A. rhizogenes, R. keguminosarum, and R. meli-loti exhibited a mean percentage of DNA homology greater than 50 with the tworeference strains. A. tumefaciens, A. radiobacter, and A. rubi were indistinguishableon the basis of DNA homology, with strain variations for this group involving upto 30% of their base sequences. The remainder of the organisms studied fall into atleast six distinct genetic groups: (i) R. (Agrobacterium) rhizogenes, which is morehomologous to R. leguminosarum than to the A. tumefaciens-A. radiobacter group;(ii) R. leguminosarum; (iii) R. meliloti; (iv) R. japonicum, which has a mean DNAhomology of some 38 to 45% with the reference strains; (v) Chromobacterium,which is as genetically remote from the reference strains as, for example, Pseudo-monas; and (vi) A. pseudotsugae strain 180, which has a DNA homology with A.tumefaciens and R. leguminosarum of only about 10%. Since this latter homologyvalue is similar to what was found after hybridizations between the reference strainsand organisms such as Escherichia coli and Bacillus subtilis, A. pseudotsugae shoulddefinitely be removed from the genus.

Bacteria within the genera Agrobacterium andRhizobium have the unique capacity to inducesuch phenomena as prolific root formation,nitrogen-fixing root nodules, and autonomouscrown-gall tumors on many higher plants. Inaddition to their obvious agricultural importance,these bacteria have become very useful tools inattempts to understand better the underlyingmechanisms of plant morphogenesis, phyto-pathogenicity, and host-parasite relationships(12, 15, 18, 26). Surprisingly little, however, isknown about the genetic interrelationships ofthe agrobacteria and rhizobia. The taxonomy ofthese bacteria is also confused. Their classifica-tion is frequently based upon only a few pheno-typic properties, resulting in considerable varia-tion from one text to another (11, 17, 22, 27;

1 Fellow of The Jane Coffin Childs Memorial Fundfor Medical Research. Present address: Departmentof Biology, New York University, University Heights,Bronx, N.Y.

Bergey's Manual). In Bergey's Manual, forexample, where the genera Agrobacterium,Rhizobium, and Chromobacterium are consideredas taxonomically related by virtue of their beinggrouped together in the family Rhizobiaceae, thedistinction among the different species andgenera rests on such qualitative criteria as nitratereduction and pathogenic host range, which atbest represent only a minute portion of the totalbacterial genome. Only recently have attemptsbeen made to clarify the nature and validity ofthe relationship among the bacteria in the abovethree genera by means of more quantitativeapproaches. The need for extensive taxonomicrevision and further study has been emphasizedby the taxim tric analysis of the chromobacteriaby Sneath (24) and of the agrobacteria, rhizobia,and chromobacteria by Graham (13), as well as

investigations by De Ley and co-workers (5, 6, 9)on the guanine-cytosine (GC) content of thedeoxyribonucleic acid (DNA) of these bacteria.

116

Vol. 94, No. IPrinted in U.S.A

on April 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

DNA HOMOLOGY IN RHIZOBIACEAE

Taximetric analysis, though usually based upon100 or more phenotypic properties, still takesinto account only a small portion of the totalgenome of the bacterium (8) and, therefore,does not necessarily represent the actual amountof genetic information that is common amongstrains. Because genetic similarity can be meas-ured by means of the DNA hybridization tech-nique (14, 20) and because such knowledgecould prove useful in clarifying the taxonomy ofthese bacteria and in showing possible correla-tions between infectivity and the presence of a"common DNA part," DNA hybridizationexperiments have been carried out.

MATERIALS AND METHODS

Organisms. Several strains from most of the nomen-species of Agrobacterium, Rhizobium, and Chromo-bacterium were selected on the basis of authenticityand representative GC values, and are listed in Table 1.Those bacteria bearing the designation ATCC werepurchased from the American Type Culture Collection(Washington, D.C.). Strains having the prefixesNCPPB, ICPB, NCTC, USDA, and MRC werekindly donated by the National Collection of PlantPathogenic Bacteria (Ministry of Agriculture, Fish-eries, and Food, Hatching Green, Harpenden, Eng-land), the International Collection of Phytopatho-genic Bacteria (Department of Bacteriology, Univer-sity of California, Davis), the National Collection ofType Cultures (London, England), L. W. Erdman(U.S. Department of Agriculture, Beltsville, Md.),and P. H. A. Sneath (Medical Research Council,Mill Hill, London, England), respectively. StrainsB6, 1.5, 3.2, and S1005 were obtained through thecourtesy of C. Bonnier, Institut Agronomique del'Etat, Gembloux, Belgium. Strain 8196U was iso-lated in this laboratory.

Cultivation of unlabeled bacteria. All bacteria ex-cept those used in the preparation of "4C-labeled DNAwere grown in Roux flasks on agar-solidified mediafor about 2 days at 27 C. The agrobacteria weregrown on a medium containing (w/v): 1% glucose,1% yeast extract (Nederl. Gist- en Spiritusfabriek,Brugge, Belgium), 0.1% (NH4)2SO4, 0.025% KH2PO4,and 2.5% agar. A medium containing (w/v) 1%glucose, 1% yeast extract, 2% CaCO3, and 2.5% agarwas used for growing the chromobacteria. Strains ofR. leguminosarum and R. meliloti were grown on amedium containing (w/v): 1% yeast extract, 0.05%K2HPO4, 0.02% MgSO4-7H20, 0.02% NaCl, 0.02%CaCl2, 0.0002% FeCl3, 2.5% agar. To reduce poly-saccharide production, strains of R. japonicum weregrown on a medium containing (w/v): 0.46% glycerol,0.1% L-arabinose, 0.1% yeast extract, 0.1% K2HPO4,0.13% CaSO4, 0.07% KNO3, 0.0004% FeCl3, 0.036%MgSO4 7H20. The bacteria were harvested withsaline-EDTA (0.15 M NaCl and 0.1 M ethylenedi-aminetetraacetate, pH 8), centrifuged, washed byintermittent centrifugation and resuspension in saline-EDTA to remove polysaccharides, and lyophilized.The dried cells were stored at 4 C until used for thepreparation of DNA.

Cultivation of 14C-labeled bacteria. Klett flasks (1-liter) containing 200 ml of appropriate liquid medium(see above) and 0.2 mc (20 ml) of filter-sterilizeduracil-2-14C (The Radiochemical Center, Amersham,Buckinghamshire, England) were inoculated with ca.106 cells of either A. tumefaciens strain B6 or R.leguminosarum strain 316C1OA. The cultures wereincubated on a reciprocal shaker at 27 C. Growthwas followed with a Klett colorimeter, and at the endof the log phase the bacteria were harvested by cen-trifugation, washed twice with saline-EDTA, andlyophilized. The 14C-labeled DNA was then extractedfrom the dried cells.DNA preparations. All DNA was extracted and

purified according to the procedure of Marmur (21),modified so as to include a deproteinization stepwith pronase (40 ,g/ml; Calbiochem, Los Angeles,Calif.) after ribonuclease treatment. This proceduretypically yielded DNA with a molecular weight be-tween 5 X 106 and 10 X 106, as determined by stand-ard analytical ultracentrifugation techniques. Purified14C-labeled DNA was dissolved in 0.2 X SSC buffer(1 X SSC buffer = 0.15 M NaCl and 0.015 M trisodiumcitrate, pH 7), sheared in a French pressure cell at12,000 psi to an average molecular weight of about3 X 105, denatured by heating for 6 min at 106 C in aglycerol bath, immediately cooled in an ice bath, andadjusted to a final concentration of 50 ,g/ml in 2 XSSC buffer. Purified, precipitated, unlabeled DNAwas dissolved to a concentration of about 5 mg/ml in0.1 X SSC buffer and stored at 4 C with a few dropsof chloroform until used for the preparation of "com-peting DNA" and DNA-agar. Competing DNA wasprepared by shearing and denaturing the DNA asabove and adjusting the SSC buffer concentration to 2.DNA-agar was made according to the procedure ofBolton and McCarthy (3) with the exception that itwas heated at 106 C in 0.1 X SSC instead of at 100 Cbecause of the high melting point of these DNApreparations.

Hybridizations. DNA hybridization was carried outby use of the technique of McCarthy and Bolton (20)as modified by Hoyer et al. (14). In essence, 30 ,ug ofsheared, denatured, "IC-labeled DNA (600 to 800counts per min per ,ug) from either A. tumefaciensstrain B6 or R. leguminosarum strain 316ClOA wasincubated with 300 ,ug of high molecular weight un-labeled, denatured, agar-embedded DNA from thesame strain. Incubation was for 16 hr at 60 C in atotal volume (milliliters) of 2 X SSC buffer amountingto 2 to 2.5 times the weight (grams) of the DNA-agardepending upon the concentration of DNA in theDNA-agar preparation used. The above hybridizationprocedure was carried out both in the presence andabsence of 2,000 ,ug of competing DNA (sheared,denatured, unlabeled, non-agar-embedded DNA)from the same strain or different strains. Appropriatecontrols were run for each new preparation of DNA-agar. After incubation, the hybridization mixture wasadded to a jacketed column having a sintered-glassbase covered by filter paper and glass wool. The DNA-agar in the column was then washed with 2 X SSCbuffer at 60 C to elute the unhybridized fraction of14C-DNA (about 120 ml at a flow rate of 100 ml/hr).

VOL. 94, 1967 117

on April 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

HEBERLEIN, DE LEY, AND TIJTGAT

TABLE 1. DNA base composition and DNA homology among strains of Agrobacterium, Rhizobium, andChromobacterium

Per cent DNA homology with

Organism Strain no. Per cent GCA. tumefaciens R. legumino-

B6 sarum 316C10A

A. tumefaciens

Mean

A. radiobacter

Mean

A. rubi

A. rhizogenes

Mean

A. pseudotsugae

R. leguminosarum

Mvean

R. meliloti

Mean

R. japonicum

Mean

C. lividum

Mean

C. violaceum

Mean

Pseudomonasa fluorescens

ICPBATCC

NCPPBATCCNCPPB

ICPB

ATCCICPB

ICPB

ICPBICPBICPB

NCPPB

USDAUSDA

TT11111.156B6447203978196U

TR-1S10054718TR-6

TR-2

TR-101TR-107TR-7

180

316C10A3F3C1

USDA 3DOa3O1.5

USDA 311b593.2

MRCMRCMRC

NCTCNCTCNCTCNCTC

MBRUGA

9371868586849757

488

60.661.161.461.462.162.2

61.5

61.261.361.862.1

61.6

59.6

61.162.362.8

62.1

67.7

61.462.5

62.0

62.362.5

62.4

64.064.8

64.4

66.165.571.4

67.6

66.166.967.067.1

66.8

62.4

7771100708610092

85

93708287

83

70

585544

52

18

5252

52

4048

44

4038

39

352538

33

28353433

33

40

59396552495270

55

62565853

57

59

715964

65

6

10084

92

5265

59

3845

42

363438

36

42403430

37

38

118 J. BACrERIOL.

on April 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

DNA HOMOLOGY IN RHIZOBIACEAE

TABLE 1-Continued

Per cent DNA homology with

Organism Strain no. Per cent GCA. tumefaciens R. legumino-

B6 sarum 316C10A

Acetobactera liquefaciens 9505 64.3 31 41

Azomonasa macrocytogenes 8200 58.6 28 48

Escherichiaa coli B 52.5 14 19

Bacillusa subtilis BQ2 47.7 3 10

aExperiments by Park (Ph.D. Thesis, State University of Gent, Gent, Belgium, 1966), carried outin conjunction with the present investigation.

The 14C-labeled DNA that was bound to the highmolecular weight DNA in the agar was then eluted bywashing with 90 ml of 0.01 X SSC buffer at 75 C. Thedegree of hybridization was determined by comparingthe amount of 14C-label which had annealed to theagar-bound DNA with that which was washed free.The percentage of homology between the DNA of thereference strains and the DNA from the various otherstrains (competing DNA) was calculated from theirrelative effect upon the amount of "4C-labeled DNAwhich, in their presence, hybridized with the agar-bound reference DNA. DNA homology values weretypically reproducible to within 1 to 6%. Hybridiza-tions were usually repeated two to three times, and thevalues given represent the mean of the separate deter-minations. The competition procedure was chosenbecause it eliminated the error associated with makingDNA-agar from different DNA preparations.

Radioactivity determintations. Disintegrations werecounted as previously described by De Ley and Fried-man (7). The correction factor for self-absorption in2 X SSC buffer under the conditions used was 14.4%;for 0.01 X SSC buffer, quenching was negligible.GC content. DNA base composition was deter-

mined by thermal denaturation as described by De Leyand co-workers (9, 10).

RESULTS AND DISCUSSIONThe different bacteria used in this study,

along with their GC content, and DNA homologywith A. tumefaciens strain B6 and R. legumino-sarum strain 316C10A are listed in Table 1.With the exception of A. pseudotsugae and R.

japonicum, the agrobacteria and rhizobia haveabout the same range of GC content, mainly,59 to 63% of their DNA. R. japonicum and thechromobacteria typically have a GC value ofaround 64 and 66%, respectively. Far moreextensive studies by De Ley and co-workers(6, 9) suggest that these are representative values.The genetic relationships among the different

strains studied are depicted in Fig. 1. The differ-ence between A. tumefaciens and A. radiobacterresides mainly in the tumor-forming capacity ofthe former. It is impossible to distinguish between

these two strains on the basis of DNA homology.Strains of A. radiobacter exhibit the same rangeof homologies to strain B6 as do the pathogenicstrains and, therefore, should have the samespecies designation. Such a move would be inaccordance with the proposals of Graham (13)and De Ley et al. (6), which were made after anexamination of the phenotypic and molecularproperties of these bacteria.Although definite conclusions cannot yet be

drawn about the taxonomic status of A. rubifrom the single strain studied, it would appear onthe basis of its homology with the referencestrains that it, too, should be placed in the samespecies with A. radiobacter and A. tumefaciens.The limited criteria upon which the speciedesignation of A. rubi is based has already beenemphasized (6). The belief that bacteria withinthis species are pathogenic only to plants in thegenus Rubus has been disproven a number oftimes both in this laboratory (6) and in others(18, 25).The DNA from strain 180 of A. pseudotsugae

was only slightly similar to that of the referencestrains, its GC content was high, and its patho-genicity was questionable. These observationsclearly indicate that strain 180 is not a validAgrobacterium strain and that it should be re-moved from this genus. This agrees very wellwith the consensus (16; Second InternationalConference on Phytopathogenic Bacteria, Lisbon,1967) that the taxonomic status of A. pseudotsu-gae should be reinvestigated and revised.

Unexpectedly, strains of A. rhizogenes show ahigher degree of homology with R. leguminosarumthan with A. tumefaciens. This finding is inagreement with recent studies by Kersters (per-sonal communication) on the particulate dehy-drogenases of bacteria in the Agrobacterium-Rhizobium group. A. rhizogenes, like the rhizobia,contain a greater abundance of particulate en-zymes active on saccharides than does the A.

119VOL. 94, 1967

on April 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

HEBERLEIN, DE LEY, AND TIJTGAT

radiobacter-A. tumefaciens group. Although A.rhizogenes and R. meliloti exhibit about thesame degree of DNA homology with R. legu-minosarum, the question of whether they aresimilar by virtue of the same portion of theirgenomes is presently under investigation.The three species of Rhizobium fall within

distinct genetic groups in their relation to the R.leguminosarum reference strain, thus confirmingtheir species status and the findings of Graham(13) and De Ley (9).

Io

0-l

50[

In contrast to the other valid agrobacteria andrhizobia, strains of R. japonicum exhibit onlyabout as much DNA homology with both refer-ence strains as do such genera as Acetobacter,Azomonas, and Pseudomonas. The implicationsof these findings are discussed later in the text.

C. lividum and C. violaceum are also far re-moved genetically from the bulk of the Agrobac-terium-Rhizobium group. They should thereforebe reclassified in a manner which does not exag-gerate their actual relationship to the agrobac-

Agrobacterium Rhizobium

_ _

IAzomonas

mocrocytogenes8200LAcetobacter

liquefociensChromobocterium

lividum GA,MB,RU

'Agrobocterium"pseudotsugoe

Bocillus subtilis

UI-

Rhizobium joponicum-/ 3.2, 31ib59

I" Pseudomonos fluorescens 488Chromobocterium violoceum8684,8685,9377,9757

Escherichio coli B

OL I

FIG. 1.!Schematic representation of the genetic relationships of the different strains of bacteria examined withAgrobacterium tumefaciens strain B6 and Rhizobium leguminosarum strain 316C1OA (heavy lines). All data weretaken from Table 1. Those organisms having a mean of less than 50% of their DNA homologous to both strainsB6 and 316CIOA are plotted as the average of their mean percentage of homology with both strains. Those strainshaving a mean percentage ofDNA homology greater than 50% with either strain B6 or strain 316CIOA are plottedagainst the strain to which they are most homologous. The point at which strain B6 and strain 316CIOA are plottedas separate lines (52%) represents the mean percentage of homology that the "B6-like" strains share with the"316CIOA-like" strains and vice versa.

120 J. BACTERIOL.

on April 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

DNA HOMOLOGY IN RHIZOBIACEAE

teria and rhizobia as is presently the case inBergey's Manual. We suggest that they be re-moved from the family Rhizobiaceae. Whetherthis family grouping has any real meaning hasalready been questioned (6).The Rhizobiaceae is one of the few groups of

bacteria in which both taximetric analysis andDNA homology studies have been carried outand, thus, provides an interesting opportunity tocompare the results obtained by these ratherdifferent techniques. Figure 2 shows a plot of themean percentage of taximetric similarity againstthe mean percentage ofDNA homology. Greaterresolution was obtained by the hybridizationtechnique, which expands the differences over a67% homology range, as compared with taxi-metric analysis, which extends only over a 38%similarity range. The farther removed two strainsare from each other genetically, the less agree-ment there is between both methods. At leasttwo explanations might account for this discrep-ancy. First, taximetric analysis involves prob-ably not more than one-fifth of the genome; andsecond, hybridization experiments will show lackof homology among cistrons which code forenzymes producing similar phenotypic proper-ties, by virtue of a common active center, but

80[

o60

Q40

20

0 20 40 60 80

MEAN % TAXIMETRIC SIMILARITYx: with A. tumefocienso: with R leguminosarLn

t00

FIG. 2. Relationship between DNA homology andtaximetric similarity of bacteria within the Rhizo-biaceae. The mean percentage of similarity between thedifferent taxospecies and either Agrobacterium tume-faciens or Rhizobium leguminosarum was calculatedfrom Graham's (13) data. These values are plottedagainst the mean percentage ofDNA homology betweenthe same species and either A. tumefaciens strain B6or R. keguminosarum strain 316CIOA. The latter dataare taken from Table 1.

differing in amino acid sequence or content (orboth).With the exception of one nomen-species,

there is a good qualitative correlation betweenthe results obtained by DNA hybridization andtaximetric analysis in this particular case. Thevery high taximetric similarity observed betweenstrains of R. meliloti and A. tumefaciens was notreflected in their DNA homology. This highvalue might be a consequence of the particularproperties selected for analysis and, if so, wouldbe expected to diminish as more properties areexamined.

It may be noticed from Fig. 2 that percentageof DNA homology is not higher than percentageof phenotypic similarity. There is thus no indi-

TABLE 2. Comparison between the orthodox and aproposed classification scheme for the agrobac-

teria and rhizobia based upon DNA homol-ogy, taximetric similarity (13), DNAbase composition (6, 9), and flag-

ellation (6, 9)

Classification according to Proposed classificationBergey's Manual

Rhizobium Rhizobium Frank, 1889leguminosarum leguminosarum (Frank)

Frank, 1889 (type spe-cies)

trifolii leguminosarumphaseoli leguminosarummeliloti meliloti Dangeard, 1926japonicum japonicum (Kichner)

Buchanan, 1926lupini japonicumcowpea group japonicum

Agrobacterium Rhizobiumradiobacter radiobactera (Beijerinck

and van Delden) Pri-bram, 1933

tumefaciens radiobacter var. tumefa-ciens

rubi radiobacter var. tumefa-ciens

rhizogenes rhizogenes (Riker et al.)comb. nov.

pseudotsugae Taxonomic position un-known

gypsophilae Taxonomic position un-known (6)

stellulatum Pseudomonas? (6)a Graham's (13) combination of R. radiobacter

with variety tumefaciens is justified according toRule 19, International Code of Nomenclature ofBacteria and Viruses. Indeed, the species Bacillusradiobacter was validly and legitimately describedby Beijerinck and van Delden (1), and the de-scription was extended by L6hnis (19) before thedescription of Bacterium tumefaciens by Smithand Townsend (23).

R leguminosw&ne,A. tumefocins ,A.rodiobocter X

////

R. meliloti tA. radiobocteA. tumefociensQ. leguminosorum.t

RQ meliloti AlRjoponicumaC. vioponicurn'C. violoceum

XC. vioIcceum X

VOL. 94, 1967 121

on April 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

HEBERLEIN, DE LEY, AND TIJTGAT

-

0a)

C)

a)

>,I5-4

a)0

--o

o >o _*--

So

oo) 0

So)_- S. .-..tn Go c

c£c o E

u)

0 0

cs.- 0 od t0 a-0 °0 .

-o6

00~ ~ ~C,

0 CZ~~

EnrA)z-'6

-,C C: -- OX° o

U 0 0

,C) _ CzC;

*-AC~~~~~~~~~~~~~~5I11 -4

, ;

::~~~~~~:.oC0 5--

C,OU0°°

0I 0 z

0 sf~~~~~~~~~C oiJ

tn,j050Nc,)

i,

0

Ia u SVn c/a

_

0-

+U, *-A0

OIO0

en dL

122 J. BACrERIOL.

C._

5Z5bc

-o

-

"S00t

0

*-

04

55,*-

L.-c4i

Qn

-a

:Z

C.

on April 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

DNA HOMOLOGY IN RHIZOBIACEAE

cation that homologous, phenotypically non-expressed DNA (unused genes) would be presentin these chromosomes. Similar conclusions havebeen reached for Pseudomonas (10).The DNA hybridization technique, because it

takes into account the entire genome, offers asensitive and quantitative approach to the classi-fication of organisms. It seems feasible, there-fore, that it should be used to clarify the presentlyill-defined concept of bacterial genus and species.It can be seen from the present study that authen-tic strains for the same nomen-species (see A.tumefaciens, Table 1 and Fig. 1) display at leasta 70% DNA homology. In view of the resultswith the A. radiobacter and A. rubi strains, itseems that this degree of homology would be asuitable boundary for a genetic species. In acomprehensive study of DNA homology amonga number of Xanthomonas strains (8), a similarvalue was suggested for the limit of the geno-species.

Figure 1 presents the taxonomist with an inter-esting nomenclatorial problem of where to drawthe genus line. If the line is drawn between the50 to 70% homology level, then Agrobacteriumremains a genus independent from Rhizobium,but A. rhizogenes and R. meliloti also have tobecome separate genera, and a fortiori R. japoni-cum. It seems more realistic to draw the line atabout 50% homology. This latter value wouldalso correspond with that which was found to bea suitable genus boundary for Pseudomonas(10). Table 2 gives a proposed classificationscheme that would be in line (except for R.japonicum) with the above findings. For compari-son, the orthodox classification as it stands inBergey's Manual is also included. In Table 3 aregiven some of the distinguishing phenotypiccharacteristics that might be used for identifyingthe different geno-species proposed in Table 2.

According to the above proposal, the sub-polarly flagellated, slow-growing rhizobia, ex-emplified here by the two R. japonicum strains(see Table 1), should be incorporated into aseparate new genus. Their relatively high per-centage of GC, typical physiological features,and the results of numerical analysis (13) allsupport such a move. However, we feel that thisimportant step should be backed up by hybridi-zation experiments involving more strains. Tem-porarily, therefore, we prefer to keep the sub-polarly flagellated nodule-inducers in the genusas R. japonicum.The group of bacteria producing nodules on

the roots of Leguminosae and fixing free nitrogenduring this symbiosis has been designated by fourdifferent generic epithets: Phytomyxa Schroeter,1886; Rhizobium Frank, 1889; Rhizobacterium

Kirchner, 1896; and Rhizomonas Orla-Jensen,1909 (4). The type species R. leguminosarum isthe same for all of them. Rhizomonas Orla-Jen-sen, 1909 is already a nomen generum rejiciendum(4). According to the rules of the InternationalCode of Nomenclature of Bacteria and Viruses,Phytomyxa Schroeter, 1886 should be used asthe earliest and only legitimate name. However,only Rhizobium Frank, 1889 (val. pub., illeg.)has been internationally accepted and has comeinto common use. Phytomyxa and Rhizobac-terium (illeg.) have been ignored and not em-ployed.To avoid disadvantageous changes in the

nomenclature of this genus, by the strict appli-cation of the rules, we propose under provision2 for Exception to The Rules: (i) that PhytomyxaSchroeter, 1886 be made a nomen generumrejiciendum; (ii) that Rhizobium Frank, 1889[type species: R. keguminosarum (Frank) Frank,1889] be made a nomen generum conservandum;and (iii) that Rhizobacterium Kirchner, 1896 bemade a nomen generum rejiciendum. The opinionof the Judicial Commission of the InternationalCommittee on Bacteriological Nomenclature isrequested.DNA hybridization data to date indicate that

the agrobacteria and rhizobia are geneticallymore related to the Pseudomonadales than to theEubacteriales. Thus, in addition to changes atthe genus and species level, a change in thehigher classification of these bacteria may also bein order. If supergeneric classifications are to bemaintained at all in the future, and if they are tohave significant genetic or evolutionary meaning,it would seem prudent to wait for more data toaccumulate before proposing any new groupingsthat may otherwise prove to have been madeprematurely.

ACKNOWLEDGMENTSThis investigation was supported by a grant from

The Jane Coffin Childs Memorial Fund for MedicalResearch to G. T. Heberlein, and by research and per-sonnel grants from the Belgian Nationaal Centrumvoor Biochemie en Molekuul Biologie to J. De Ley.

LITERATURE CITED1. BEIJERINEK, M. W., AND A. VAN DELDEN. 1902.

Ueber die Assimilation des freien Stickstoffsdurch Bakterien. Centr. Bakteriol. Parasitenk.Abt. II 9:3-43.

2. BERNAERTS, M. J., AND J. DE LEY. 1963. A bio-chemical test for crown gall bacteria. Nature197:406-407.

3. BOLTON, E. T., AND B. J. MCCARTHY. 1962. Ageneral method for the isolation of RNA com-plementary to DNA. Proc. Natl. Acad. Sci. U.S.48:1390-1397.

4. BUCHANAN, R. E., J. G. HOLT, AND E. F. LESSEL,

VOL. 94, 1967 123

on April 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

HEBERLEIN, DE LEY, AND TIJTGAT

JR. 1966. Index Bergeyana. The Williams &Wilkins Co., Baltimore.

5. DE LEY, J. 1964. Effect of mutation on DNA com-position of some bacteria. Antonie van Leeu-wenhoek J. Microbiol. Serol. 30:281-288.

6. DE LEY, J., M. BERNAERTS, A. RASSEL, AND J.GUILMOT. 1966. Approach to an improvedtaxonomy of the genus Agrobacterium. J. Gen.Microbiol. 43:7-17.

7. DE LEY, J., AND S. FRIEDMAN. 1965. Similarityof Xanthomonas and Pseudomonas deoxyribo-nucleic acid. J. Bacteriol. 89:1306-1309.

8. DE LEY, J., I. W. PARK, R. TIJTGAT, AND J. VANERMENGEM. 1966. DNA homology and taxon-omy of Pseudomonas and Xanthomonas. J. Gen.Microbiol. 42:43-56.

9. DE LEY, J., AND A. RASSEL. 1965. DNA basecomposition, flagellation, and taxonomy of thegenus Rhizobium. J. Gen. Microbiol. 41:85-91.

10. DE LEY, J., AND J. VAN MUYLEM. 1963. Someapplications of deoxyribonucleic acid base com-position in bacterial taxonomy. Antonie vanLeeuwenhoek J. Microbiol. Serol. 29:344-358.

11. DOWSON, W. J. 1957. Plant diseases due to bac-teria. Cambridge Univ. Press, Cambridge,England.

12. FRED, E. B., I. L. BALDWIN, AND E. McCoy. 1932.Root nodule bacteria and leguminous plants.Univ. of Wisconsin Press, Madison.

13. GRAHAM, P. H. 1964. The application of com-puter techniques to the taxonomy of the rootnodule bacteria of legumes. J. Gen. Microbiol.35:511-517.

14. HOYER, B. P., B. J. MCCARTHY, AND E. T.BOLTON.1964. A molecular approach in the systematicsof higher organisms. Science 144:959-967.

15. KLEIN, R. M., AND G. K. LINK. 1955. The etiologyof crown gall. Quart. Rev. Biol. 30:207-277.

16. KNOSEL, D. 1962. Untersuchungen an sternbild-enden Bakterien. Arb. Landwirtsch. Hochsch.Hohenheim 15:1-63.

17. KRASS1LNIKOV, M. A. 1959. Diagnostik der Bak-terien und Actinomyceten. DDR GustavFischer Verlag, Jena, Germany.

18. LiPPINCOTr, J. A., AND G. T. HEBERLEIN. 1965.The quantitative determination of the infec-tivity of Agrobacterium tumefaciens. Am. J.Botany 52:856-863.

19. LOHNIS, F. 1905. Beitrage zur Kenntnis derStickstoffbakterien. Centr. Bakteriol. Para-sitenk. Abt. II 14:582-604.

20. MCCARTHY, B. J., AND E. T. BOLTON. 1963. Anapproach to the measurement of genetic related-ness among organisms. Proc. Natl. Acad. Sci.U.S. 50:156-164.

21. MARMUR, J. 1961. A procedure for the isolation ofdeoxyribonucleic acid from microorganisms. J.Mol. Biol. 3:208-218.

22. PRfVOT, A. R. 1961. Traite de systematique bac-terienne. Dunod, Paris, France.

23. SMITH, E. F., AND C. 0. TOWNSEND. 1907. Aplant tumor of bacterial origin. Science 25:671-673.

24. SNEATH, P. H. A. 1957. The application of com-puters to taxonomy. J. Gen. Microbiol. 17:201-226.

25. STARR, M. P. 1946. The nutrition of phytopatho-genic bacteria. II. The genus Agrobacterium.J. Bacteriol. 52:187-194.

26. VALERA, C. L., AND M. ALEXANDER. 1965. Nodu-lation factor for Rhizobium-legume symbiosis.J. Bacteriol. 89:1134-1139.

27. WILSON, J. K. 1944. Over five hundred reasons forabandoning the cross-inoculation groups of thelegumes. Soil Sci. 58:61-69.

124 J. BACTERIOL.

on April 25, 2020 by guest

http://jb.asm.org/

Dow

nloaded from