Characterization of atypical Erwinia carotovora strains causing blackleg of potato in Brazil


Valmir Duarte, Universidade Federal do Rio Grande do Sul, Departamento de Fitossanidade, Cx. Postal 15100 – CEP 90001-970, Porto Alegre, RS, Brazil (e-mail:


Aims:  To determine the characteristics of bacteria associated with the blackleg disease of potato in Brazil and compare them with species and subspecies of pectolytic Erwinia.

Methods and Results:  Biochemical and physiological characteristics of 16 strains from blackleg-infected potatoes in State of Rio Grande do Sul, Brazil, were determined and differentiated them from all the E. carotovora subspecies and E. chrysanthemi. Pathogenicity and maceration ability of the Brazilian strains were greater than those of E. carotovora subsp. atroseptica, the causal agent of potato blackleg in temperate zones. Analyses of serological reaction and fatty acid composition confirmed that the Brazilian strains differed from E. carotovora subsp. atroseptica, but the sequence of 16S rDNA gene and the 16S-23S intergenic spacer (IGS) region confirmed the Brazilian strains as pectolytic Erwinia. Restriction analysis of the IGS region differentiated the Brazilian strains from the subspecies of E. carotovora and from E. chrysanthemi. A unique SexAI restriction site in the IGS region was used as the basis for a primer to specifically amplify DNA from the Brazilian potato blackleg bacterium in PCR.

Conclusions:  The bacterium that causes the blackleg disease of potato in Brazil differs from E. carotovora subsp. atroseptica, the blackleg pathogen in temperate zones. It also differs from other subspecies of E. carotovora and from E. chrysanthemi and warrants status as a new subspecies, which would be appropriately named E. carotovora subsp. brasiliensis.

Significance and Impact of the Study:  The blackleg disease of potato is caused by a different strain of pectolytic Erwinia in Brazil than in temperate potato-growing regions. The Brazilian strain is more virulent than E. carotovora subsp. atroseptica, the usual causal agent of potato blackleg.


Erwinia carotovora and E. chrysanthemi are the most important of the pectolytic bacteria that cause maceration of plant tissue and disease of many crop plants including potato. Several years ago Hauben et al. (1998) revived the suggestion that the pectolytic bacteria be placed in a separate genus, Pectobacterium, on the basis of 16S rDNA sequences, but the generic epithet, Erwinia, remains the preferred designation in the scientific literature. The E. carotovora species is divided into the five subspecies: atroseptica, carotovora, betavasculorum, odorifera and wasabiae (De Boer and Kelman 2000) some of which should perhaps be elevated to species status (Gardan et al. 2003).

The blackleg disease of potato is caused primarily by E. c. atroseptica in cool temperate climates. E. c. carotovora and E. chrysanthemi may also cause blackleg-like symptoms at high temperatures (<25°C) but economically, E. c. atroseptica is the most important pathogen (Pérombelon and Kelman 1980; Pérombelon 1992). Erwinia c. atroseptica potato isolates from various geographical areas in Canada, the US and western Europe are remarkably uniform in both phenotypic and genetic characteristics (De Boer et al. 1987; Fessehaie et al. 2002). Erwinia c. betavasculorum and E. c. wasabiae also define groups of similar strains and cause soft rot of sugar beet and Japanese horseradish, respectively (Thomson et al. 1981; Goto and Matsumoto 1987). Some strains, not isolated from potato, have been described as atypical E. c. atroseptica because, although they produce acid from α-methylglucoside and reducing substances from sucrose like E. c. atrospetica, they differ in having the ability to grow at 37°C. Some of these strains, isolated from several hosts, including chicory, produce odorous volatile metabolites and are called E. c. odorifera (Gallois et al. 1992).

In a recent study, E. c. atroseptica, E. c. carotovora and E. chrysanthemi were found in 55, 44 and 1%, respectively, of potato plants showing blackleg symptoms in 22 fields of nine counties in Rio Grande do Sul State, Brazil (de Oliveira 2001). The strains associated with blackleg, tentatively identified as E. c. atroseptica, had typical colonies that produced reducing sugars from sucrose, utilized α-methylglucoside, and had no phosphatase activity. However, these bacteria grew at 37°C, produced no PCR product with E. c. atroseptica-specific primers and did not react with a monoclonal antibody specific to serogroup I of E. c. atroseptica. These results, suggesting that the pectolytic bacteria causing blackleg in potato differed from E. c. atroseptica, prompted us to investigate further the characteristics of these strains and their differences from described subspecies of E. carotovora.

Materials and methods

Strains and biochemical characterization

The strains investigated in this study are listed in Table 1. Sixteen strains of the Brazilian potato blackleg-causing bacterium (BPBB) were chosen from among the 400 isolates with E. c. atroseptica-like characteristics to represent three potato cultivars and two major production areas in the state of Rio Grande de Sol. The bacteria had been isolated in 1999 from potato plants showing blackleg symptoms (de Oliveira 2001). Four of the selected strains (8, 212, 213 and 371) were deposited in the American Type Culture Collection (BAA-416, 417, 418 and 419, respectively). Strains were routinely maintained on nutrient agar (NA).

Table 1.  Strains of the Brazilian potato blackleg bacterium (BPBB) and Erwinia species and subspecies used in this study
Species/subspeciesIdentificationPotato cultivar, region in RS, BrazilReference/source
  1. ATCC, American Type Culture Collection; LMG, Laboratorium Microbiologie Rijksuniversiteit Gent; CFBP, Collection Francaise des Bacteries Phytopathogenes.

BPBB8 (ATCC BAA-416)Elvira, Planaltode Oliveira (2001)/UFRGS 32·338
29Elvira, Planaltode Oliveira (2001)/UFRGS32 ·333
54Baronesa, Planaltode Oliveira (2001)/UFRGS 32·025
101Macaca, Depressao Centralde Oliveira (2001)/UFRGS 32·041
106Macaca, Depressao Centralde Oliveira (2001)/UFRGS 32·064
137Baronesa, Depressao Centralde Oliveira (2001)/UFRGS 32·006
138Baronesa, Depressao Centralde Oliveira (2001)/UFRGS 32·038
142Baronesa, Planaltode Oliveira (2001)/UFRGS 32·402
153Macaca, Planaltode Oliveira (2001)/UFRGS 32·358
200Baronesa, Depressao Centralde Oliveira (2001)/UFRGS 32·031
201Macaca, Depressao Centralde Oliveira (2001)/UFRGS 32·085
205Macaca, Depressao Centralde Oliveira (2001)/UFRGS 32·074
212 (ATCC BAA-417)Elvira, Planaltode Oliveira (2001)/UFRGS 32·401
213 (ATCC BAA-418)Macaca, Depressao Centralde Oliveira (2001)/UFRGS 32·084
219Elvira, Planaltode Oliveira (2001)/UFRGS 32·413
371 (ATCC BAA-419)Macaca, Depressao Centralde Oliveira (2001)/UFRGS 32·062
  Serogroup of potato strain 
E. carotovora subsp. atrosepticaEca 3IS.H. De Boer/3
Eca 6XVIIIR. J. Copeman/E17
Eca 17IS.H. De Boer/17
Eca 19IS.H. De Boer/19
Eca 31IA. Kelman/SR8
Eca 196XXIIS.H. De Boer/196
Eca 198XXR.J. Copeman/E555
E. carotovora subsp. betavasculorumLMG 2398Sugar beetThomson et al. (1981)
LMG 2461Sugar beetThomson et al. (1981)
LMG 2462Sugar beetThomson et al. (1981)
LMG 2464TSugar beetThomson et al. (1981)
  Serogroup of potato strain 
E. carotovora subsp. carotovoraEcc 21XXIXH.P. Maas Geesteranus/139
Ecc 23XVH.P. Maas Geesteranus/162
Ecc 26VH.P. Maas Geesteranus/200
Ecc 51XIVS.H. De Boer/51
Ecc 59XIIIH. P. Maas Geesteranus/257
Ecc 61XS.H. De Boer/222
Ecc 62IXS.H. De Boer/195
Ecc 63IXS.H. De Boer/202
Ecc 65XIVS.H. De Boer/196
Ecc 67XIIA. Kelman/SR162
Ecc 68VIIA. Kelman/SR165
Ecc 71IIIH.P. Maas Geesteranus/226
Ecc 94XVIIR.J. Copeman/E6
Ecc 193XIR.J. Copeman/E193
Ecc 194XIXR.J. Copeman/E103
E. carotovora subsp. odoriferaCFBP 1645CeleryR. Samson (1978)
CFBP 1878TWitloof chicoryR. Samson (1978)
CFBP 1880Witloof chicoryR. Samson (1979)
CFBP 1893CeleryR. Samson (1976)
E. carotovora subsp. wasabiaeEcw SR91THorseradishGoto and Matsumoto (1987)
Ecw SR92HorseradishGoto and Matsumoto (1987)
Ecw SR93HorseradishGoto and Matsumoto (1987)
Ecw SR94HorseradishGoto and Matsumoto (1987)
E chrysanthemiEch 571PotatoH.P. Maas Geesteranus/647

Strains were characterized by standard tests used for pectolytic Erwinia and included a test for pectolytic activity on crystal violet pectate (CVP) medium, acid production from α-methylglucoside, reducing substances from sucrose, growth at 37°C, and erythromycin resistance (De Boer and Kelman 2000). The MICRO-ID kit (Remel Inc., Lenexa, KS, USA) that tests 15 additional metabolic activities was used for further characterization.

Strains were also assayed for oxidation of the 95 carbon sources on the GN2 Biolog microplates and an additional 45 carbon sources on the GP microplates (Biolog Inc., Hayward, CA, USA). Bacteria were grown on general medium (BUGM, Biolog Inc.) for 24 h at 24°C. Using sterile cotton swabs, bacteria were transferred from the medium surface to 30-ml test tubes containing 20 ml of inoculating fluid with thioglycolate (Biolog Inc.). The optical density of the suspension was adjusted as recommended by the manufacturer. Microplates were inoculated with 150 μl of suspension per well and incubated at 24°C. Absorbancy was determined at 590 nm after 24 h using Biolog Microlog 2 Workstation and the GN database (release 4·01C) for strain identification.

Biochemical data were evaluated by cluster analysis using the single linkage method (nearest neighbour) using SYSTAT software (Chicago, IL, USA). The data from 67 and 26 carbon sources on GN and GP Biolog plates, respectively, utilized by at least one strain, were included in the analysis. Data were scored zero for negative reaction, 0·5 for borderline reactions, and 1 for positive results.

Pathogenicity and maceration characteristics

To obtain host plants for testing pathogenicity, seed potatoes were planted in 12-cm-diameter pots containing a potting soil substrate (Growing mix #5; Northern Peat Ltd, Berwick, NS, CA, USA). One seed tuber was planted per pot and four pots were used for each bacterial strain tested. Plants were kept in a growth chamber at 21°C, and stems were inoculated 20 days after planting. Inoculation was carried out with sterile toothpicks dipped into 24–48-h-old bacterial colonies and then immediately stabbed into individual potato stems, 5 cm above the soil line, at three stems per pot. Plants were observed for blackleg symptoms for 21 days after inoculation.

To test maceration ability, sterile toothpicks dipped into 24–48-h-old bacterial colonies were stabbed into potato tubers. After inoculation, tubers were sprayed with mineral oil (Nujol) (Lee and Cha 2001) and kept in a humid chamber at 24°C. After incubation for 5 days, the amount of macerated tissue was determined by weighing tubers before and after washing away decayed tissue.

Green peppers, previously disinfected with 70% alcohol and 1% NaOCl for 30 s each, and rinsed with sterile distilled water, were inoculated in the same way as potato tubers, but incubated without oil spray. After 48 h at 24°C in a humid chamber, the diameter of decay lesions were measured at each inoculation locus.

Serological and fatty acid analyses

The monoclonal antibody 4F6 specific for the lipopolysaccharide of serogroup I, the principle serological type of E. c. atroseptica, was tested by ELISA as described previously (De Boer and McNaughton 1987).

Gas chromatography of fatty acid methyl esters (FAME) using the Sherlock Microbial Identification System (MIDI, Inc., Newark , DE, USA) was conducted by R. Phillipe at the Centre for Plant Quarantine Pests, Ottawa, Canada.

DNA sequencing and restriction digestion

The 16S rDNA fragment from selected BPBB strains amplified with the 27f and L1r consensus primers were submitted for sequencing as described (Fessehaie et al. 2002). Sequencing was carried out by Dr L. Wong at the Core Molecular Biology Facility, York University (North York, Ontario, Canada).

DNA from the intergenic spacer (IGS) flanking the 3′ end of the 16S and the 5′ end of the 23S rRNA genes was obtained by the ‘band stab’ amplification technique, in which resolved PCR products, based on the 1491f (5′-GAA GTC GTA ACA AGG TA-3′) and L1r [5′-CA(A/G) GGC ATC CAC CGT-3′] primers, were retrieved directly from an agarose gel with a pipet tip after electrophoresis and re-amplified (Stackebrandt and Goodfellow 1991; Wilton et al. 1997). PCR products were purified using microCLEAN (The Gel Company, San Francisco, CA, USA) and sequenced as described above. Alignments of the sequences were performed using Align Plus 4, version 4·1 (Sci Ed Central, San Francisco, CA, USA).

As a way of measuring the genetic diversity of BPBB strains and differentiation from other subspecies, fingerprints of the IGS regions were generated from strains that were isolated from different locations in RS State, Brazil, and characterized in detail. PCR amplifications were carried out as previously reported (Fessehaie et al. 2002). The amplification was carried out with 1 μm each of primers 1491f and L1r, using a thermal regime of 94°C/2 min, 25x (94°C/45 s, 62°C/45 s, 72°C/90 s), and 72°C/10 min. Purified amplicons were digested with RsaI, CfoI, HpyCH4III, and SexAI (BioLabs Inc., New England, CT, USA) and DNA fragments were resolved by gel electrophoresis on 3% agarose gels.

PCR amplification

For each strain evaluated, genomic DNA was extracted using a protocol adapted from De Boer and Ward (1995). Briefly, a loopful of bacterial cells was extracted in 250 μl of extraction buffer (100 mm Tris–HCl pH 8·0, 25 mm EDTA, 1% SDS, and 5 μg proteinase K) and incubated for 3 h at 56°C. Protein components were precipitated with one half volume of ammonium acetate (7·5 m) and removed by centrifugation. DNA was precipitated from the supernatant fraction with isopropanol, washed with 70% ethanol, taken up in 50 μl of ultrapure water, and stored at −20°C.

Primers ECA1f (5′-CGG CAT CAT AAA AAC ACG-3′) and ECA2r (5′-GCA CAC TTC ATC CAG CGA-3′) were used in PCR reactions as described previously (De Boer and Ward 1995) but using a TGradient (Whatman-Biometra, Goettingen, Germany) thermocycler. Primers Y1 (5′-TTA CCG GAC GCC GAG CTG TGG CGT-3′) and Y2 (5′-CAG GAA GAT GTC GTT ATC GCG AGT-3′), selected from a pectate lyase-encoding gene of the Y family (Darrasse et al. 1994), were also used. PCR amplification was carried out as described but with the following thermal regime: 94°C/10 min, 25x (94°C/60 s, 67°C/60 s, 72°C/30 s) and 72°C/10 min. Amplicons were resolved by agarose gel electrophoresis on a 1·5% agarose gel in TBE containing 2·5 μl ml−1 ethidium bromide and documented.

An oligonucleotide (5′-GCG TGC CGG GTT TAT GAC CT-3′), named BR1f, was designed from the IGS region of BPBB based on the restriction enzyme site of SexAI, and used with the primer L1r to detect BPBB strains. PCR amplification was carried out as stated above, with 1 μm each of primer BR1f and L1r, with the following thermal regime: 94°C/2 min, 25x (94°C/45 s, 62°C/45 s, 72°C/90 s) and 72°C/10 min.


Biochemical and physiological tests

Selected phenotypic characteristics that differentiate BPBB from the other subspecies of E. carotovora and E. chrysanthemi are presented in Table 2. Strains of BPBB were pectolytic on CVP, produced acid from α-methyl-glucoside and reducing substances from sucrose, grew at 37°C, and did not utilize psicose, Tween-40 nor Tween-80 as sole carbon source. All strains of E. c. atroseptica tested utilized d-galactonic acid lactone, d-gluconic acid, uridine, and thymidine, while only 12–38% of BPBB strains did so. Moreover, none of the E. c. atroseptica strains utilized d-galacturonic acid, d-saccharic acid, or l-glutamic acid, while 69–75% of BPBB strains utilized these carbon sources.

Table 2.  Selected phenotypic characteristics that differentiate Brazilian potato blackleg bacteria and five other subspecies of E. carotovora and E. chrysanthemi
CharacteristicResponse of strains*
BPBB (n = 16)†Eca (n = 5)Ecc (n = 5)Ecb (n = 1)Eco (n = 1)Ecw (n = 1)Ech (n = 1)
  1. *BPBB, Brazilian potato blackleg bacteria – strains 8, 29, 54, 101, 106, 137, 138, 142, 153, 200, 201, 205, 212, 213, 219, and 371; Eca, E. carotovora subsp. atroseptica– strains 3, 6, 31, 196, and 198; Ecc, E. carotovora subsp. carotovora– strains 21, 51, 59, 63 and 193; Ecb, E. carotovora subsp. betavasculorum– strain LMG 2398; Eco, E. carotovora subsp. odorifera– strain 1878T; Ecw, E. carotovora subsp. wasabiae– strain SR91T; Ech, E. chrysanthemi– strain 571. Percentage of strains showing positive response or (−) negative and (+) positive results.

  2. †Number of strains tested.

Acid from α-methyl glucoside1001000++
Reducing substances from sucrose1001000++
Growth at 37°C1000100+++
ONPG Test1000100+++
Utilization of 
 Acetic acid502580+++
 2′-Deoxy adenosine1210040+++
 d-galactonic acid lactone1210080+
 d-galacturonic acid75080+++
 d-gluconic acid1910080+++
 d-glucosaminic acid000+
 d-glucuronic acid000+
 d,l-lactic acid695020+++
 d-malic acid88200++
 3-Methylglusose0   +  
 d-saccharic acid75080++
 l-glutamic acid25080+++
 l-lactic acid94800+++
 Malonic acid000+
 Succinamic acid1000100++++
 Thymidine-5′-monophosphate0   +  

The Biolog system identified all strains of E. c. atroseptica, E. c. betavasculorum and E. chrysanthemi correctly but BPBB, E. c. odorifera and E. c. wasabiae were identified as E. c. carotovora. Differential oxidation (given as percentage of BPBB strains positive/percentage of E. c. carotovora strains positive) of d,l-lactic acid (69/20), glucose-1-phosphate (0/60), l-lactic acid (94/0), α-methyl d-glucoside (81/0), palatinose (94/0), psicose (6/100), inosine (38/100), thymidine (38/100) and Tween-80 (0/60) revealed differences between BPBB and E. c. carotovora (Table 2).

Erwinia c. betavasculorum oxidized 22 carbon sources not metabolized by BPBB including α-cylcodextrin, d-arabitol, dextrin, d-sorbitol, l-proline, N-acetly-d-glucosamine, Tween-40 and Tween-80 (Table 2). But unlike E. c. betavasculorum, BPBB metabolized cellobiose and d-melibiose. Biochemically, BPBB strains were most similar to E. c. odorifera, but differed in being unable to utilize d-arabitol, 3-methylglucose and thymidine-5′-monophosphate (Table 2).

A dendrogram based on the numerical analysis, using the Euclidian distance coefficient and UPGMA, is represented in Fig. 1. Erwinia c. betavasculorum and E. chrysanthemi were distinct, but E. c. odorifera, E. c. wasabiae and E. c. atroseptica were closely related. The Euclidian distances among BPBB strains ranged from 0·139 to 0·275. Strains Ecc 193 and Ecc 59 were in the lower (0·137) and Ecc 51 in the higher range, showing that BPBB strains form an internal group in the heterogeneous and broader group of E. c. carotovora strains.

Figure 1.

Dendrogram based on the numerical analysis, using Euclidian distance coefficient and UPGMA, of 93 utilized carbon sources (borderline = 0·5; positive = 1) of the GN and GP MicroPlates (Biolog Inc.) by the Brazilian potato blackleg bacterium (BPBB), E. carotovora subsp. atroseptica (Eca), E. carotovora subsp. betavasculorum (Ecb), E. carotovora subsp. carotovora (Ecc), E. carotovora subsp. odorifera (Eco), E. carotovora subsp. wasabiae (Ecw) and E. chrysanthemi (Ech)

Pathogenicity and maceration activity

Potato stems inoculated with BPBB strains developed typical blackleg symptoms. However, symptoms of blackleg were visible 3 days after inoculation with BPBB in contrast to stems inoculated with E. c. atroseptica that required at least 7 days to develop symptoms. BPBB strains consistently macerated potato tuber tissue and green pepper fruits more rapidly than strains of E. c. atroseptica in three replications of the experiment (Fig. 2).

Figure 2.

Maceration activity of three strains of the Brazilian potato blackleg bacterium (BPBB) and two of E. carotovora subsp. atroseptica (Eca) on potato tubers and green peppers. Vertical lines show the standard deviation

Serological specificity and fatty acid analyses

In the ELISA test with McAb 4F6, the mean absorbance value (A405) for E. c. atroseptica strains was 0·287 ± 0·140 compared with 0·019 ± 0·010 for 149 BPBB strains and negative controls.

Fatty acid analysis showed that BPBB strains had no striking differences from the other E. carotovora subspecies as do E. c. atroseptica strains. The following six fatty acids, in order of decreasing amounts, were present in all the 46 strains of the six subspecies: hexadecanoid acid (16:0), octadecanoid acid (18:1 w7c), dodecanoid acid (12:0), tetradecanoid acid (14:0), pentadecanoid acid (15:0) and heptadecanoid acid (17:0). Also present in all subspecies but not in all strains were tridecanoid acid (13:0), and two unidentified fatty acids with equivalent chain lengths of 13·957 and 14·502 (Table 3). Along with E. c. betavasculorum and E. c. odorifera, BPBB had lower levels of tridecanoid acid (13:0) than E. c. atroseptica, E. c. carotovora and E. c. wasabiae. In two-dimensional cluster analysis using the FAME program, E. c. atroseptica strains clearly formed a separate cluster from the other subspecies and BPBB, none of which showed distinctly separated groups.

Table 3.  Relative amount of fatty acids detected in various strains of the Brazilian potato blackleg bacterium and of strains in five subspecies of Erwinia carotovora
Fatty acidBPBB* (N = 14)†Eca (N = 6)Ecb (N = 3)Ecc (N = 15)Eco (N = 4)Ecw (N = 4)
Mean (%)nMean (%)nMean (%)nMean (%)nMean (%)nMean (%)n
  1. *BPBB, Brazilian potato blackleg bacterium – strains 8, 24, 29, 54, 101, 106, 137, 138, 142, 153, 212, 213, 219, and 371; Eca, E. carotovora subsp. atroseptica– strains 3, 6, 19, 31, 196, and 198; Ecb, E. carotovora subsp. betavasculorum– strains 2398, 2461 and 2464; Ecc, E. carotovora subsp. carotovora– strains 21, 23, 26, 51, 59, 61, 62, 63, 65, 67, 68, 71, 94, 193 and 194; Eco, E. carotovora subsp. odorifera– strains 1645, 1878, 1880 and 1893; Ecw, E. carotovora subsp. wasabiae– strains SR91T, SR92, SR93 and SR94.

  2. N = number of strains tested.

  3. n = number of strains with the fatty acid.

12:0 3OH000·463000
Unknown 13·9570·69140·6160·6830·72140·8140·824
Unknown 14·5021·09140·9351·0131·17131·0040·772
17:1 w8c0·50120·6820·4020·89130·5340·574
17:0 w6c0000·63400
17:0 cyclo00·5820·582000
18:1 w7c17·83149·18614·71320·351520·19415·734
12:0/14:04·24 2·40 4·22 4·21 5·03 3·81 
16:0/12:04·22 5·86 4·37 4·10 4·09 4·91 

Sequencing and restriction analysis

Sequence data for the 16S rDNA region of BPBB strains 8, 212, 213 and 371 have been deposited in Genbank as accession numbers AY207086, AY207083, AY207084 and AY207085, respectively. 16S rDNA sequence analysis using different methods (UPGMA, neighbour-joining, minimum evolution, maximum parsimony) revealed a position for BPBB strains among the E. carotovora subspecies in the different trees (data not shown). The scores assigned in a BLAST search confirmed this relationship; the identity varied from 97 to 99% homology with E. carotovora and E. chrysanthemi strains.

PCR amplification of the ITS region generated two fragments (Fig. 3). A summary of percentage matches of global DNA alignment of ITS regions of BPBB strains against five other subspecies of E. carotovora and three other species of Erwinia are shown in Table 4. The sequence alignment of the small region showed the presence of a single copy of the tRNAGlu gene and a single copy of the tRNAIle and tRNAAla genes in the large region. PCR amplification of the ITS of BPBB strains yielded amplicons similar in size to strains of the E. carotovora subspecies but different from E. chrysanthemi (Table 4). Cleavage of these amplicons with restriction enzyme CfoI produced DNA fragments for BPBB similar to E. c. atroseptica, carotovora and odorifera (Fig. 4). The differences among E. c. betavasculorum, wasabiae and E. chrysanthemi were clearly distinguished in the gel. Cleavage of the BPBB ITS region with the restriction enzyme RsaI produced a pattern similar to E. c. odorifera (Fig. 4). The enzyme HpyCH4III cleaved the ITS of all subspecies and E. chrysanthemi but not that of BPBB strains (Fig. 4). The SexAI restriction site, starting at nucleotide position 122 in the small region of the BPBB ITS, was a unique nucleotide target (Fig. 5).

Figure 3.

PCR profiles of the 16S-23S intergenic spacer (IGS) regions of Erwinia sp. using primer set1491f/L1r (lanes 1–7), and primer set BR1f/L1r (lanes 8–14). Lanes 1 and 8, Brazilian potato blackleg bacterium BPBB 212; lanes 2 and 9, E. carotovora subsp. atroseptica Eca 3; lanes 3 and 10, E. carotovora subsp. betavasculorum Ecb 2398; lanes 4 and 11, E. carotovora subsp. carotovora Ecc 21; lanes 5 and 12, E. carotovora subsp. odorifera Eco 1878; lanes 6 and 13, E. carotovora subsp. wasabiae Ecw SR91; lanes 7 and 14, E. chrysanthemi Ech 571

Table 4.  Comparison of the nucleotide sequences of the small and large 16S-23S rRNA intergenic transcribed spacer (IGS) regions of several Erwinia spp. and subspp. to the Brazilian potato blackleg bacterium
*Species/ subspeciesStrainSmall ITS regionLarge ITS region
NCBI accessionLength (bp)Homology (%)Matching nucleotidesNCBI accession†Length (bp)Homology (%)Matching nucleotides
  1. *BPBB = Brazilian potato blackleg bacterium; Eca = E. carotovora subsp. atroseptica; Ecb = E. carotovora subsp. betavasculorum; Ecc =E. carotovora subsp. carotovora; ECC = E. carotovora subsp. odorifera; Ecw = E. carotovora subsp. wasabiae; Ech = E. chrysanthemi; Eam =E. amylovora; Epy = E. pyrifoliae.

  2. †Sequences were retrieved from the GenBank (NCBI = National Center for Biotechnology Information) database under the accession numbers indicated.

BPBB 212ATCC BAA-417AF448594444100444AF448595484100484
BPBB 213ATCC BAA-418AF44859645194427AF44859749196473
BPBBB 8ATCC BAA-416AF44459245294425AF44859349095470
EcaLMG 2386AF23268744693418AF23428247587435
Ecb AF23268644594422AF23428048693455
EccATCC 15713AF23268444494419AF23428448793457
Eco AF23268045389407AF23427848993458
EcwATCC 43316AF23267944893417AF23427748291443
EchATCC 11663AF23268135659267AF23428749164353
Eam AF29041972643316AF290418103135366
Epy AJ13296942255283
Figure 4.

Scheme representing gels of the CfoI, HpyCH4III and RsaI restriction patterns of the 16S-23S rRNA regions of intergenic transcribed spacer (IGS) of the Brazilian potato blackleg bacterium (BPBB: AF448594, AF448595), Erwinia carotovora subsp. atroseptica (Eca: AF232687, AF234282), E. carotovora subsp. betavasculorum (Ecb: AF232686, AF234280), E. carotovora subsp. carotovora (Ecc: AF232684, AF234284), E. carotovora subsp. odorifera (Eco: AF232680, AF234278), E. carotovora subsp. wasabiae (Ecw: AF232679, AF234277), and E. chrysanthemi (Ech: AF232681, AF234287). Numbers denote accession codes of sequences of small and large IGS regions, respectively, retrieved from GenBank, National Center for Biotechnology Information database

Figure 5.

Homology sequence alignment of the small 16S-23S rDNA intergenic transcribed spacer (IGS) region of Brazilian potato blackleg bacterium (BPBB), E. carotovora subsp. atroseptica (Eca), E. carotovora subsp. betavasculorum (Eca), E. carotovora subsp. carotovora (Ecc), E. carotovora subsp. odorifera (Eco), E. carotovora subsp. wasabiae (Ecw), and E. chrysanthemi (Ech) partially displayed to show the SexAI restriction site (bold) present in BPBB and absent in the others, and the BR1f primer sequence. Sequences were retrieved from the GenBank (National Center for Biotechnology Information) database under the accession numbers indicated

PCR amplification

None of the 16 strains of BPBB produced amplicons in PCR with the E. c. atroseptica primers in contrast to the single 690 bp amplicon obtained with all seven E. c. atroseptica strains tested (data not shown).

We observed amplified fragments using Y primers with 12 strains of BPBB but not with four (8, 212, 213 and 219) of the BPBB strains tested. No amplification of E. c. atroseptica, betavasculorum and wasabiae strains was obtained using these primers (data not shown).

An oligonucleotide (5′-GCG TGC CGG GTT TAT GAC CT-3′), named BR1f, was designed from the IGS region of BPBB, based on the SexAI restriction enzyme site, and used with the primer L1r to amplify DNA from BPBB strains. By using these primers in PCR, a 322-bp fragment was amplified from DNA of all BPBB strains tested but no amplicons were obtained from DNA extracted from the five E. carotovora subspecies or E. chrysanthemi (Fig. 3).


Erwinia c. atroseptica is the major causal agent of potato blackleg in cool and temperate regions of Canada, the US and western Europe (Molina and Harrison 1977; Caron et al. 1979; Persson 1988; Bain et al. 1990). It can be differentiated from all other E. carotovora strains solely on the basis of acid production from α-methylglucoside, production of reducing substances from sucrose, and inability to grow at 37°C (Graham 1972; De Boer et al. 1979). The BPBB strains associated with blackleg in Brazil were clearly different from E. c. atroseptica. Although they produced reducing substances from sucrose and acid from α-methylglucoside, like E. c. atroseptica, they grew at 37°C, their DNA was not amplified in PCR with E. c. atroseptica-specific primers, and did not react with antibodies specific to the major E. c. atroseptica serogroup. Focusing on the production of reducing substances from sucrose and acid from α-methylglucoside as major traits, disregarding the ability to grow at 37°C, resulted in misidentification of Erwinia strains as E. c. atroseptica in the earlier study of Brazilian strains (de Oliveira 2001).

In addition to the above-mentioned biochemical and physiological features, BPBB strains also differed from E. c. atroseptica in their positive ONPG test for B-galactosidase and ability to utilize succinamic acid (Table 2). BPBB strains were similar to E. c. betavasculorum and odorifera in their ability to grow at 37°C, to produce acid from α-methylglucoside, and reducing substances from sucrose. However, they differed from E. c. betavasculorum and E. c. odorifera in their carbohydrate utilization patterns (Table 2).

Fatty acid analysis grouped the BPBB strains with E. c. betavasculorum, carotovora, odorifera and wasabiae, and distinguished them from E. c. atroseptica. Our results confirmed a previous study (De Boer and Sasser 1986) that differentiated E. c. atroseptica strains clearly as a separate cluster from the other subspecies, none of which show distinctly separated groups (Table 3). Erwinia chrysanthemi also presents its own characteristic fatty acid profile, lacking dodecanoid acid (De Boer and Sasser 1986). The usefulness of fatty acid profiles for characterizing soft rot erwinias was limited to identifying E. c. atroseptica and E. chrysanthemi and to differentiating them from the other subspecies of E. carotovora.

While the failure of PCR with the E. c. atroseptica-specific primers to amplify BPBB DNA has already been noted above, PCR amplification with the Y1/Y2 primer set targeting the pel genes was successful for DNA from 12 of 16 BPBB strains. Although we cannot readily explain the failure in amplification of the four BPBB strains other than to speculate on variation in the pectate genes, we note that DNA from several of our E. c. atroseptica strains also were not amplified with the Y1/Y2 primers. However, the failure of DNA from E. c. betavasculourm to be amplified in PCR with these primers is consistent with the published literature (Darrasse et al. 1994).

The 16S rDNA sequence of BPBB was consistent with its identity as a member of the E. carotovora species. BPBB strains were positioned among other strains of the species in dendrograms using various algorithms based on neighbour-joining, maximum parsimony, and unweighted pair-group methods of analysis. The sizes and sequences of the large and small IGS regions of BPBB strains were also similar to those of other E. carotovora subspecies and different from E. chrysanthemi (Fig. 3) (Fessehaie et al. 2002). All the E. carotovora subspecies could be differentiated by restriction fragment length polymorphisms in the IGS region following digestion with appropriate restriction enzymes (Fig. 4). The lack of a HpyCH4III site in the IGS of BPBB strains is particularly noteworthy.

The unique SexAI site in the small IGS region of BPBB strains was a useful target for designing a primer (BR1f) unique to these strains (Fig. 5). The PCR based on the BR1f primer did not amplify DNA from strains representing other E. carotovora subspecies or E. chrysanthemi (Fig. 3). PCR based on the BR1f primer is useful in determining the presence of BPBB-like strains on potato. In preliminary work, a multiplex PCR utilizing primer sets ECA1f/ECA2r and BR1f/L1r was used to show that BPBB-like strains, but not E. c. atroseptica, occur on ostensibly healthy tubers in Brazil while E. c. atroseptica, but not BPBB strains, occur on potato tubers in Canada.

The summation of our results strongly suggests that the BPBB strains form a new subspecies of E. carotovora. Although we lack resources for the DNA–DNA hybridization studies that are evidently required to establish new taxons of bacteria, it is not surprising that new subspecies will be identified among the heterogeneous and poorly defined E. c. carotovora strains. If a new subspecies were to be accepted for the BPBB strains, an appropriate subspecific epithet would be E. carotovora subsp. brasiliensis.

Although the BPBB strains occur in a geographical area that is considered a humid subtropical climate, the temperatures (17–20°C) during the growing season are relatively cool. The association of temperature with the occurrence of BPBB needs to be investigated. In any case, under the temperate conditions in which our pathogenicity and maceration tests were carried out, the BPBB strains were more virulent on potato than the E. c. atroseptica strains (Fig. 2). The high level of virulence shown by these strains, in fact, suggests that some circumspection may be appropriate to curtail their dissemination to other geographical areas.


We would like to express our gratitude to Roger Phillipe at the Centre for Plant Quarantine Pests, Ottawa, Canada, for the fatty acid analysis. We also recognize the importance of the scholarship given by CAPES – Brazilian Federal Agency – to Dr Valmir Duarte to spend his sabbatical working on this research.