Molecular differentiation of Pantoea stewartii subsp. indologenes from subspecies stewartii and identification of new isolates from maize seeds




Assays to detect Pantoea stewartii from maize seeds should include differentiation of P. stewartii subsp. stewartii and P. stewartii subsp. indologenes.

Methods and Results

Previously published PCR primers for the identification of P. stewartii subsp. stewartii amplified signals from both subspecies using both conventional and quantitative PCR. In MALDI-TOF mass spectroscopy analysis with the Biotyper software (Bruker), subspecies stewartii and indologenes produced identical score values. Analysis against the Biotyper database produced similar score values for both subspecies. From the subtyping methods provided by the Biotyper software, only composite correlation indexing (CCI) separated both groups. By alignment of 16S rRNA sequences, no subspecies distinction was possible. To develop new techniques for the separation of these subspecies, the partial sequences of several housekeeping genes were compared. The type strains of the two subspecies showed characteristic single-nucleotide polymorphisms (SNPs) in the genes galE, glmS and recA. Other reference strains of P. stewartii subsp. stewartii followed the same nucleotide pattern, whereas known P. stewartii subsp. indologenes strains were different. Based on single-nucleotide polymorphisms in galE and recA, PCR primers were created to separate the subspecies by stepdown PCR analysis. Two putative Pstewartii strains were isolated from imported maize seeds. They were not virulent on maize seedlings, were positive in the indole assay with Kovacs reagent and identified as P. stewartii subsp. indologenes. The subspecies-specific PCR primers confirmed they were subspecies indologenes.


By stepdown PCR, P. stewartii subsp. indologenes can be differentiated from P. stewartii subsp. stewartii.

Significance and Impact of the Study

A possible detection of P. stewartii subsp. stewartii, the causative agent of Stewart's wilt of maize, in plant material by immunological or molecular assays must exclude contamination with P. stewartii subsp. indologenes, which would create false positives in seed tests and affect quarantine measurements.


Pantoea stewartii subsp. stewartii (Pnss) is the causative agent of Stewart's wilt of maize (Zea mays L.) and is classified as quarantine organism in many countries (Pataky and Ikin 2003). Maize varieties, mainly sweet maize and some inbred field maize lines, are affected by the pathogen, which causes leaf blight and wilting. The disease is primarily transmitted by the corn flea beetle (Chaetocnema pulicaria). Seed transmission is also possible at a very low rate (Michener et al. 2002), but it is not considered an important part of the disease cycle. Stewart's wilt has been reported for North America and for Europe in Austria, Greece, Poland, Romania and Russia (Pataky and Ikin 2003), but has not become established in these countries. More than 60 countries place quarantine regulations on maize seed imports from affected areas and surveillance of traded plant material is required to prevent further distribution of the pathogen (Pataky and Ikin 2003). For this reason, it is important to have methods for detection and identification of Pnss that are not subject to false positives caused by closely related Pantoea species (Coplin et al. 2002; Wensing et al. 2010).

Related bacteria belonging to the same genus, such as Pantoea ananatis, can cause diseases of several monocotyledonous plants and even dieback of Eucalyptus trees (Coutinho and Venter 2009). Pantoea stewartii subsp. indologenes (Pnsi) is not virulent on maize, but can cause leaf spots on foxtail millet (Setaria italica) and pearl millet (Pennisetum americanum) (Mergaert et al. 1993). Unknown plant isolates are usually assigned to the genus Pantoea with phenotypic tests such as colony colour and Gram-staining, but more sophisticated methods are needed to assign species and subspecies. To distinguish P. stewartii from the other plant-associated Pantoea species such as Pantoea agglomerans, P. dispersa and P. ananatis, several methods have been described, for example enzyme-linked immunosorbent assay (ELISA) (Lamka et al. 1991), PCR primers from virulence genes (Coplin et al. 2002), TaqMan primers from the cpsD gene (Tambong et al. 2008), ‘miniprimers’ (Xu et al. 2010), specific primers from the pstS-glmS region and MALDI-TOF analysis (Wensing et al. 2010). Species-specific primers and the MALDI-TOF MS analysis of whole cell extracts are especially suitable methods for high-throughput screening. Unfortunately, differentiation between the maize pathogen Pnss and Pnsi, which is not a maize pathogen, has not been possible with these methods, and their use may lead to wrong conclusions for quarantine purposes. Additional time-consuming investigations, for example virulence assays, indole tests or sequencing analysis, have been necessary to discriminate between Pnss and Pnsi.

Single-nucleotide polymorphisms (SNPs) have been used for differentiation of related pathogens, especially at the species and subspecies levels, and for phylogenetic analyses (Achtman 2008). SNPs within insertion sequence (IS) elements have been used to distinguish strains within haplotypes of Mycobacterium ulcerans and determine their geographical distribution (Käser et al. 2009). SNPs have also been used to establish correlations between genotypes and pathotypes. For example, a Chlamydia pneumonia SNP was used to differentiate two genotypes using real-time PCR with LNA probes (Rupp et al. 2006). Erwinia species can be distinguished by stepdown (or differential) PCR based on the housekeeping genes gpd and recA (Gehring and Geider 2012b), and E. amylovora strains from different origins can be distinguished using SNPs within galE, acrB and hrpA (Gehring and Geider 2012a).

Several algorithms are available to compare MALDI-TOF MS spectra for bacterial identification (Giebel et al. 2010). While the standard methods are optimized for species identification, other approaches with varying resolution levels are included in the Biotyper software (Bruker). Subtyping can be achieved by differential peak weighting in principal component analysis (PCA) or by differential comparison of spectra parts in a composite correlation index (CCI).

We evaluated differentiation of Pnss and Pnsi based on conserved SNPs in the housekeeping genes recA (encoding recombinase A) and galE (encoding UDP-glucose 4-epimerase) and on subtyping by MALDI-TOF MS protein fingerprints. Specific primers and a stepdown PCR method to discriminate the SNP variants were then developed and optimized with well-characterized strains. Both methods were used to identify these subspecies among uncharacterized isolates from imported maize seed.

Material and methods

Bacterial strains

The bacterial strains used in the assays are described in Table 1.

Table 1. Bacterial strains used in this study
  1. BCCM/LMG, Belgian co-ordinated collections of micro-organisms; CFBP, Collection Française de Bactéries Phytopathogénes; NCPPB, National Collection of Plant Pathogenic Bacteria UK.

Pantoea stewartii subsp. stewartii
DC133From Zea mays, Missouri, 1976, avirulentCoplin et al. (2002)
DC172TSS11T, from Zea mays, Iowa, 1940, avirulentCoplin et al. (2002)
DC283Nalr mutant of SS104Coplin et al. (2002)
EstA1Obtained from EnglandW. Zeller
SS104From Zea mays, Illinois, 1967Coplin et al. (2002)
SW2From Zea mays, Ohio, 1974Coplin et al. (2002)
Pantoea stewartii subsp. indologenes
CFBP 3614TNCPPB 2280T; LMG 2632T, from Setaria italica, India, 1960 
DD54Isolated from Puerto Rican maize seedsThis work
DD59Isolated from Puerto Rican maize seedsThis work
NCPPB 1845LMG 2671, DC551; from pineapple, Hawaii, deposited as P. ananatisD. Coplin
NCPPB 1877LMG 2630, deposited as Pseudomonas cyamopsicola, from Cyamopsis psoralioidesMergaert et al. (1993)
NCPPB 2275LMG 2631, DC552; from Pennisetum americanum, IndiaD. Coplin
NCPPB 2281LMG 2633; deposited as Xanthomonas indica; from Setaria italica, India 1970Mergaert et al. (1993)
NCPPB 2282LMG 2634; deposited as Xanthomonas penniseti; from Pennisetum glaucum, India 1956Mergaert et al. (1993)
Pantoea ananatis
NCPPB 1848Deposited as Pectobacterium carotovorum, Cattleya sp. Brazil 1965 (Robbs ENA439); LMG 2807 
Pantoea agglomerans
EhMB96Isolated from a maize leafLab collection

Strain isolation from maize

Sweet maize seed samples from Central and North America as well as Europe were screened for the presence of Stewart's wilt (P. stewartii). Seed samples (5 g) were soaked in 10 ml 50 mmol l−1 potassium phosphate buffer (pH 7) overnight at 4°C and then crushed in plastic bags. Coarse material was removed by slow centrifugation (10 min at 180 g). In fresh tubes, bacteria were sedimented for 15 min at 7500 g. The pellet was suspended in 200 μl potassium phosphate buffer and the entire suspension spread on LB agar with cycloheximide (50 μg ml−1) or on King's B agar plates.

Seed samples included maize imported from Puerto Rico and cv. ‘Seneca Horizon’ sweet maize from US West Coast Seeds (Ladner, British Columbia). Other sweet maize seed samples were supplied by Mark Millard and Kendall Lamkey at Iowa State University (Ames, IA) and David Coplin at The Ohio State University (Columbus, OH).

Selected mucoid, yellow colonies on the isolation plates were further investigated with PCR assays, ELISA tests using P. stewartii-specific antiserum and fatty acid methyl ester analysis (FAME). Two strains with properties of P. stewartii, DD54 and DD59, were analysed in detail.

PCR for sequencing and primer design

Portions of the 16S rRNA, galE and recA genes of Pnss strain DC283 and Pnsi strain CFBP 3614T were amplified with the PCR primers listed in Table 2 and then commercially sequenced. The nucleotide sequences were aligned and used for design of specific primers (Fig. 1). Sequence information from other strains in Table 1 was also obtained with the primers in Table 2.

Table 2. Primers used for sequencing of Pantoea stewartii strains and PCR detection
NameSequence 5′–3′GeneReference
  1. BH1, black hole quencher.

#562ACGTCGTCGTCCTGGATAAT galE From accession nos. AF077292 and AHIE00000000
#85GGTAAAGGGTCTATCATGCG recA Waleron et al. (2008)
#63 (fd2)AGAGTTTGATCATGGCTCAG(16S rRNA)Weisburg et al. (1991)
#338GCACCGAATTGTTCGTTAGG glmS Wensing et al. (2010)
#356CACTGGAGCAATGCAGTAGC glmS Wensing et al. (2010)
#668GCACTCATTCCGACCAC hrpS Coplin et al. (2002)
#357TCGGTAACGGTCGAGTAATG glmS With #356 (see above)
#354GCGCTCAAGCTGAAGGAGAT glmS Wensing et al. (2010)
Figure 1.

Design of primers for the differentiation of Pantoea stewartii subsp. indologenes and P. stewartii subsp. stewartii, based on single-nucleotide polymorphisms (SNPs) within the galE and recA genes, by stepdown PCR. Highlighted are the specific primers for DC283 consistent with the subsp. stewartii, and specific primers for CFBP 3614T consistent with the subsp. indologenes. All primers have an SNP at the 3′-end (underlined). (a) Alignment of part of the galE sequence of strains DC283 and CFBP 3614T containing SNPs specific for each subspecies. (b) Alignment of part of the recA sequence of strains DC283 and CFBP 3614T.

PCR assays

The primers applied in conventional PCR (cPCR) and quantitative PCR (qPCR) for specific detection of P. stewartii at the species level have been described before (Coplin et al. 2002; Wensing et al. 2010). Primers used for analysis of Pnss and Pnsi are listed in Table 3. The assay conditions for stepdown PCR were adapted from Bugert et al. (2003) and have also been described previously (Gehring and Geider 2012a). The PCR runs consisted of a two-step series. The first series (A) consisted of 10 cycles of a denaturation step at 94°C for 20 s, annealing at 68°C for 30 s and elongation at 72°C for 30 s. The next 20 cycles (B) had the same denaturation and elongation steps, but the annealing temperature for step 2 was lowered from 68°C to 63°C. For cPCR, normal Taq polymerase (0·125 units per 25 μl assay) was used, and 15 μl of the reaction product was analysed by electrophoresis in a 1% agarose gel containing ethidium bromide.

Table 3. Subspecies-specific primers based on SNPs for differentiation of Pantoea stewartii subsp. stewartii and Pantoea stewartii subsp. indologenes used in stepdown PCR
NameSequence 5′–3′ step 2A/2BTemperature Specificity
DC283galECGACCTGTTTGCCTCTCACT68/63°CFor subsp. stewartii
3614galECGACCTGTTTGCCTCTCACC68/63°CFor subsp. indologenes
DC283recATGACGCTGCAGGTGATTGCT68/63°CFor Subsp. stewartii
3614recATGACGCTGCAGGTGATTGCC68/63°CFor subsp. indologenes

Hot-start Taq polymerase (Ampliqon, Skovlunde, Denmark) was used in the qPCR experiments with an activation time of 20 min at 95°C. A TaqMan probe, SYBR Green or EvaGreen, were used to detect signals for both P. stewartii subspecies. Bacteria were grown as 1 ml overnight cultures in LB medium, and 10 μl were lysed in 1 ml 0·1% Tween 80 by heating for 20 min at 65°C. A 5-μl aliquot of these lysates was used as a template in PCR. The master mix for the PCR assays has been described before (Mohammadi et al. 2009). Analysis was carried out in the CFX96 cycler (Bio-Rad, Hercules, CA, USA).


Sample preparation was performed as described by Sauer et al. (2008). Briefly, bacteria were grown for 24 h at 28°C in 1 ml of Luria–Bertani (LB) broth with 1% glycerol in 2-ml reaction tubes. Cells were harvested by centrifugation, and the pellet washed with 1 ml deionized water to remove residual components of the growth medium. The cells were pelleted again, and then they were resuspended in 0·3 ml water with 0·8 ml ethanol. For lysis, cells were pelleted, air-dried to remove ethanol and resuspended thoroughly in 40 μl 70% formic acid and 40 μl acetonitrile. The cell debris was removed by centrifugation, and clear lysates were stored at −20°C. One to two μl of the extracts were placed on a MSP 96 polished steel target and cocrystallized with the same amount of matrix (saturated alpha-cyano-4-hydroxy cinnamic acid in 50% acetonitrile/2·5% trifluoro acetic acid) for analysis with a Bruker microflex machine.

Biotyper analysis

Protein profiles were derived as an average of 250 spectra and analysed with the Biotyper software in the automation mode (Ver. 2.0, Bruker Daltonics, Bremen, Germany). Pattern analysis was performed against a reference library (Biotyper ver. 3.0). Results were interpreted according to a log-score scheme (Sauer et al. 2008). Values of 2·0 or above represent a high likelihood for positive identification. Each strain identification was repeated at least three times. For subtyping purposes, reference spectra of Pnss strain DC283 and Pnsi strain CFBP 3614T were generated. Subtyping by principle component analysis (PCA) and composite correlation index (CCI) analysis were conducted according to the Biotyper manual with settings according to standard parameters recommended by the supplier.

Plant assays and additional tests

Virulence assays on sweet maize seedlings (Z. mays L. cv. ‘Golden Bantam’), hypersensitive response (HR) assays on tobacco (Nicotiana tabacum L. cv. ‘Samsun’) and the indole test with Kovacs reagent were carried out as previously described (Wensing et al. 2010). All analytical experiments were repeated at least twice.

Polyclonal antiserum from rabbit (Loewe GmbH, Sauerlach, Germany) was used for ELISA tests to detect P. stewartii in maize seeds with immunoassays.


Preliminary characterization of P. stewartii isolates from imported maize

About 50 samples of maize seed imported from Chile, Puerto Rico, USA, Turkey and Serbia were extracted and screened on LB agar for the appearance of yellowish bacterial colonies. The majority of the yellow colonies did not display features of P. stewartii, such as mucoid colonies on CPG agar (Coplin et al. 2002). Two strains from Puerto Rican samples, designated DD54 and DD59, were putative P. stewartii strains and were further characterized for their taxonomic properties to place them in a subspecies. On maize seedlings, both strains were nonpathogenic. In tobacco leaves, they were HR-negative suggesting that they may not be plant pathogens. The Pnsi type strain CFBP 3614T and Pnss strains were positive for HR, and strains NCPPB 1845 and NCPPB 2275 often produced irregular HR lesions, so they were not clearly positive. On the basis of indole production (data given below), these strains were tentatively identified as Pnsi and compared with known Pnsi strains in the studies below.

Evaluation of previously described PCR assays for the identification of P. stewartii subsp. indologenes

Previously developed PCR assays for identification of Pnss were evaluated for their ability to distinguish subspecies stewartii and indologenes and for identification of the new strains from Puerto Rican maize seeds. Primer pairs designed from the pst-glmS region of P. stewartii strain DC283 (Wensing et al. 2010) were tested against unknowns and known Pnsi strains. The primers #338/#339 and #356/#341 produced the correct signals for the species P. stewartii (Table 4). Assays also included previously described primers from the hrpS gene (Coplin et al. 2002). Primers #668/#669 produced positive signals for most of the Pnsi strains except NCPPB 2282 and DD59. This indicates that the hrpS primers designed for detection of Pnss are not that reliable for detection of all Pnsi strains. P. ananatis strain NCPPB 1848 was negative with all P. stewartii cPCR primers. P. stewartii primers designed for qPCR with SYBR Green and another pair together with a TaqMan probe were also applied to these strains. They produced threshold values of ca. 20, confirming identification of strains DD54 and DD59 as a P. stewartii subspecies.

Table 4. Characterization of P. stewartii strains by PCR and MALDI-TOF analysis of protein patterns from whole cells, by hypersensitive response on tobacco, pathogenicity on maize seedlings and the Kovacs indole test
Assays:cPCR:qPCRMT MSvirHRIndole
#338 /#339#356 /#341#668 /#669Pstew qPCR TMSG #356 #357Biotyper Score for Pnss
glmS glmS hrpS glmS glmS
  1. MT, MALDI-TOF; TM, TaqMan probe (FAM) #353 with primers #340 and #354. The primers are described in ref. Wensing et al. 2010 (glmS region) and (Coplin et al. 2002) (hrpS). vir, virulence assay on maize seedlings (pa, pathogenic; np, nonpathogenic); HR, hypersensitive response on tobacco; (+) irregular lesions. Indole, synthesis tested with Kovacs reagent.

Pantoea stewartii subsp. stewartii
SW2++ 20·219·52·1pa+
EstA1++ 19·920·42·2pa+
Pantoea stewartii subsp. indologenes
CFBP 3614T+++20·621·12np++
NCPPB 1845+++19·920·52·1 (+)+
NCPPB 2275+++2020·21·9 (+)+
NCPPB 1877+++22·624·42·1  +
NCPPB 2281+++22·024·02·0  +
NCPPB 2282++21·925·32·0  +
Pantoea ananatis
NCPPB 1848>40>40  +
Pantoea agglomerans (control)
EhMB96 >40>40np

MALDI-TOF mass spectroscopy for identification of P. stewartii and differentiation of subspecies

In MALDI-TOF MS analysis, most of the Pnsi strains listed in Table 1 had Biotyper scores as high as the known Pnss used as reference (Table 4). Only NCPPB 1848 was different confirming its classification as P. ananatis. For P. stewartii, score values did not result in classification of subspecies.

Subtyping by MALDI-TOF MS fingerprinting produced heterogeneous results. PCA did not result in grouping of isolates according to subspecies (data not shown). In contrast, pattern matching against a CCI matrix based on Pnss DC283, and Pnsi strain CFBP 3614T separated the strains according to the subspecies (Table 5). Score distance (min. 0, max. 1) between the two strains used for creation of the CCI was quite high with 0·92 vs 0·7 for Pnss strain DC283 and 0·75 vs 0·59 for Pnsi strain CFBP 3614T. For other strains, the distance between the two matching scores often was smaller, for example, between 0·66 (Pnss) and 0·63 (Pnsi) for DC172T. Still, the subspecies suggested by the higher CCI score was in line with previous subspecies determinations for most of the strains tested. The new strains from maize imports, DD54 and DD59, tentatively grouped with Pnsi.

Table 5. Differentiation of the Pantoea stewartii subspecies stewartii and indologenes by stepdown cPCR using galE and recA primers from SNPs in subsp. stewartii DC283 and subsp. indologenes CFBP 3614T (Table 3) and by MALDI-TOF composite correlation indexing (CCI) matches
StrainSignal with cPCR primers for galE/recAMALDI-TOF MS CCI values compared with
DC283CFBP 3614TDC283CFBP 3614T
  1. PCR signals with galE/recA primers; nd, not done.

Pantoea stewartii subsp. stewartii
Pantoea stewartii subsp. indologenes
CFBP 3614T−/−+/+0·590·75
NCPPB 1845−/−+/+0·680·69
NCPPB 2275−/−+/+0·670·71
NCPPB 1877−/−+/+0·690·77
NCPPB 2281−/−+/+0·680·81
NCPPB 2282−/−+/+0·580·68

Sequence analysis of the 16S rRNA, recA, galE and glmS genes to locate SNPs that might discriminate P. stewartii subspecies

For confirmation that the newly isolated strains DD54 and DD59 belong to the species P. stewartii, we sequenced their 16S rRNA genes. They clustered with other strains of this species, but this technique does not resolve P. stewartii subspecies (Fig. 2a). Therefore, to distinguish subspecies, portions of the recA, galE and glmS gene sequences were compared.

Figure 2.

Dendrogram of partial 16S rRNA (a) and recA (b) sequences from Pantoea species. The outgroup in (a) was Pantoea toletana.

In a previous study (Wensing et al. 2010), we observed characteristic sequence patterns in the recA gene for strains of both P. stewartii subspecies. In a dendrogram, strains DD54 and DD59 were placed on a branch apart from Pnss together with the Pnsi type strain and NCPPB 2275 (Fig. 2b). The position of strain NCPPB 1845 was intermediate, separated from two P. ananatis strains. Alignments of galE sequences showed SNPs typical for Pnsi (Fig. 3a). Switches of C/G and G/A were characteristic in the alignment. A comparable situation occurred for alignments of the glmS region (Fig. 3b) with base changes A/C and T/G. Small deviations were found in both alignments for strain NCPPB 1845, indicating a mismatch within Pnsi nucleotide sequences. These data confirmed the correct classification of P. stewartii DC133 and DC172T (SS11T) as subspecies stewartii, although both strains had lost virulence during storage.

Figure 3.

Sequence alignments for parts of galE gene (a) and the glmS region (b) from Pantoea stewartii subspecies stewartii and indologenes strains. The partial galE sequence starts at position 801 in the gene of P. stewartii subsp. stewartii strain DC283, the partial glmS sequence at position 1435 of the DC283 gene. Strains DC133, DC145, DC407 and SW2 are P. stewartii subsp. stewartii as described by (Coplin et al. 2002). Accession numbers for parts of galE and glmS for strains DC133 (galE or glmS: HG792426/HG792435), DC145 (HG792427/HG792436), DC407 (HG792428/HG792437), SW2 (-/HG792439), EstA1 (HG792429/HG792438), CFBP 3614T (HG792430/HG792440), DD54 (HG792433/HG792443), DD59 (HG792434/HG792444), NCPPB 1845 (HG792431/HG792441) and for NCPPB 2275 (HG792432/HG792442). For DC283, accession number is AF077292 for galE, and the glmS sequence can be deduced from the genomic contig library Accession No. AHIE00000000.

Differentiation of the P. stewartii subspecies by PCR assays

Within the analysed sequences of galE and recA, several conserved SNPs were present that we used to design a primer pair for each subspecies (Fig. 1). These primer pairs could be used for subspecies differentiation by stepdown PCR (Fig. 4, Table 5). Primers specific for Pnss only produced signals with DC283 as a template, and those specific for Pnsi only produced a signal with CFBP 3614T as template (Fig. 4). This approach using stepdown PCR produced positive bands with galE and recA primers from Pnss only for strains of this subspecies, whereas primer pairs specific for CFBP 3614T resulted in positive signals for strains of this subspecies, including strains DD54 and DD59 (Table 5). Accordingly, DD54 and DD59 were confirmed as Pnsi. Strains NCPPB 1845, 1877, 2275, 2281 and 2282 also produced positive signals with galE and recA primers specific for Pnsi.

Figure 4.

Stepdown cPCR to distinguish Pantoea stewartii subsp. stewartii and indologenes. M, size marker (1 kb DNA ladder). For lanes 1–4 Pantoea stewartii subsp. stewartii, DC283 was used as template and for lanes 5–8 P. stewartii subsp. indologenes CFBP 3614T. Primers DC283galE were used for lanes 1 and 5, 3614galE for lanes 2 and 6, DC283recA for lanes 3 and 7, and 3614recA for lanes 4 and 8.

Indole tests

The classification of DD54 and DD59 as Pnsi was further supported by the indole test with Kovacs reagent in which both strains were positive. Known Pnsi type strain CFBP 3614T and strains NCPPB 1845, NCPPB 1877, NCPPB 2275, NCPPB 2282 as well as P. ananatis NCPPB 1848 were also positive, whereas Pnss strains were all negative (Table 4, with summary of additional properties).


Identification of strains at the subspecies level is important for P. stewartii, because only Pnss causes Stewart's wilt, but Pnsi is commonly isolated from maize seed and could create false positives. Therefore, for quarantine purposes, a simple and rapid test is needed to distinguish these subspecies when contamination with Pnsi occurs. To test imported seeds, bacteria are commonly cultured from seed washes and examined for the presence of yellow, mucoid colonies, which are then subjected to further tests, such as Gram-staining, oxidation/fermentation, motility, HR and pathogenicity. A method based on ‘miniprimer’ PCR was able to differentiate both subspecies (Xu et al. 2010). In this assay, different short primers were used for amplification of multiple genes and generated a banding pattern that allowed classification. However, the evaluation of complex banding patterns takes time and experience, so this method is not practical for rapid mixed culture analysis.

Simple methods such as species-specific PCR and MALDI-TOF MS-based protein fingerprinting allow a fast distinction of P. stewartii from other Pantoea species, but in this study and a previous one (Wensing et al. 2010), we found that their resolution was not sufficient for subspecies identification. Primer-binding sites in a more conserved genetic region tend to be similar between subspecies for standard PCR. We previously developed a PCR method for determination of a diagnostic SNP-type that relies only on standard PCR equipment and is applicable to mixed culture analysis, which we used to differentiate Erwinia species (Gehring and Geider 2012b). This method relies on highly conserved SNPs within housekeeping genes. A primer is designed with an SNP in the 3′ position and PCR specificity obtained in a two-step protocol, using an increased annealing temperature in the first 10 cycles.

The SNPs identified in this study from the galE and recA genes of Pnss and Pnsi were applied as subspecies-specific primers. PCR specificity was verified with reference cultures, and the protocol applied to determine the subspecies of two new Pantoea isolates, which we obtained from Puerto Rican maize seed imports. The identification of these strains with the novel SNP PCR assays was then confirmed by a combination of nucleotide sequencing of 16S rRNA genes, biochemical assays and pathogenicity tests. With reference cultures, subtyping by CCI analysis of MALDI-TOF MS data resulted in correct clustering for most spectra, yet the absolute difference observed in score values for both subspecies can be small. In most cases, the higher score indicated the correct subspecies, but for isolates such as DC172T individual spectra showed a similar score for the two subtypes.

The type strain for Pnsi was isolated in India from foxtail millet. Other strains (NCPPB 2275, NCPPB 2281, NCPPB 2282) came from related grasses, pineapple (NCPPB 1845) and guar bean (NCPPB 1877), which is an atypical host. Pnsi isolates seem to be derived from a variety of host plants and could be a common contaminant of maize seeds. The SNP-based stepdown cPCR method used in this study should facilitate standard identification procedures for P. stewartii subspecies. For quarantine testing, a prescreening PCR assay for the presence of P. stewartii was described by Wensing et al. (2010) using primers designed from the pstS-glmS region. While further verification by pure culture isolation and identification by MALDI-TOF MS fingerprinting and subsequent CCI analysis can also be used to discriminate between both subspecies, further methods should be used for confirmation. The PCR with SNP-based primer pairs from the recA and galE genes developed in this study is suitable to determine the subspecies of P. stewartii isolates and can even be applied to mixed cultures. Known Pnsi strains from NCPPB were positive with primers #668/#669 (hrpS), except for NCPPB 2282 (Table 4). In classifying strains DD54 and DD59 as Pnsi, even though they are unable to cause an HR on a nonhost plant, we have extended the description of Pnsi to presumably nonpathogenic strains. Similarly, in a population of E. pyrifoliae, a portion of the strains were not able to induce HR (Jock et al. 2003). This deficiency was due to a mutation in the hrpL gene. PCR primers from the Pnss hrpS region detected strain DD54 but not DD59.

PCR assays based on SNPs can also be used for differentiation of other micro-organisms (Gehring and Geider 2012b). For distinction of P. stewartii subspp. stewartii and indologenes, we provide here a valuable tool for molecular separation of these subspecies to decide about possible contamination of maize imports with the Stewart's wilt pathogen.


We thank S. Zimmermann, Hygiene Institute Heidelberg, for access to the Bruker Microflex mass spectrometer.

Conflict of interest

No conflict of interest declared.