Genetic diversity and population structure of ...

366

Genetic diversity and population structure of Clavibacter michiganensis subsp. nebraskensis

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF NEBRASKA on 05/16/11 For personal use only.

I.V. Agarkova, P.A. Lambrecht, and A.K. Vidaver

Abstract: Clavibacter michiganensis subsp. nebraskensis (CMN) is a gram-positive bacterium and an incitant of Goss’s bacterial wilt and leaf blight or “leaf freckles” in corn. A population structure of a wide temporal and geographic collection of CMN strains (n = 131), originating between 1969 and 2009, was determined using amplified fragment length polymorphism (AFLP) analysis and repetitive DNA sequence-based BOX-PCR. Analysis of the composite data set of AFLP and BOX-PCR fingerprints revealed two groups with a 60% cutoff similarity: a major group A (n = 118 strains) and a minor group B (n = 13 strains). The clustering in both groups was not correlated with strain pathogenicity. Group A contained two clusters, A1 (n = 78) and A2 (n = 40), with a linkage of 75%. Group A strains did not show any correlation with historical, geographical, morphological, or physiological properties of the strains. Group B was very heterogeneous and eight out of nine clusters were represented by a single strain. The mean similarity between clusters in group B varied from 13% to 63%. All strains in group B were isolated after 1999. The percentage of group B strains among all strains isolated after 1999 (n = 69) was 18.8%. Implications of the findings are discussed. Key words: DNA fingerprinting, AFLP, Rep-PCR, BOX-PCR, Goss’s wilt, corn. Résumé : Clavibacter michiganensis subsp. nebraskensis (CMN) est une bactérie positive à Gram, responsable du flétrissement bactérien de Goss et de la rouille des feuilles du maïs. La structure de la population d’une large collection de souches de CMN (n = 131), recueillies entre 1969 et 2009 à différentes périodes et à différents endroits, a été déterminée par AFLP (sigle anglais de « amplified fragment length polymorphism ») et par BOX-PCR de séquences répétitives. L’analyse des données composites de l’AFLP et des empreintes obtenues par BOX-PCR a révélé la présence de deux groupes possédant une limite de similarité de 60 %, un groupe majeur A (n = 118 souches) et un groupe mineur B (n = 13 souches). Il n’y avait pas de corrélation entre l’agrégation au sein d’un groupe et la pathogénicité d’une souche. Le groupe A contenait deux grappes, A1 (n = 78) et A2 (n = 40), possédant un niveau de liaison de 75 %. L’appartenance au groupe A n’était pas corrélée avec les caractéristiques historiques, géographiques, morphologiques ou physiologiques des souches. Le groupe B était très hétérogène, et huit des neuf grappes étaient représentées par une seule souche. Le niveau moyen de similarité entre les grappes du groupe B variait entre 13 % et 63 %. Toutes les souches du groupe B avaient été isolées après 1999. Les souches du groupe B constituaient 18,8 % de toutes les souches isolées après 1999 (n = 69). L’implication de ces résultats est discutée. Mots‐clés : empreinte d’ADN, AFLP, Rep-PCR, BOX-PCR, flétrissement de Goss, maïs. [Traduit par la Rédaction]

Introduction The bacterium Clavibacter michiganensis subsp. nebraskensis (CMN) (Vidaver and Mandel 1974; Davis et al. 1984) is the causal agent of Goss’s bacterial wilt and blight of maize or corn (Zea mays L.). This disease, also known as leaf freckles and wilt, was first observed in corn in south central Nebraska (Dawson Co.) in 1969 (Vidaver and Mandel 1974). By 1979, the pathogen was identified throughout Nebraska and in the adjacent states of Iowa, Kansas, South Dakota, and Colorado (Vidaver et al. 1981). After genetic sources of resistance in field corn were identified and cultivars partially resistant to Goss’s wilt became available, the disease became sporadic, seen in limited areas and primarily

in susceptible varieties of sweet corn, popcorn, and dent corn hybrids (Ngong-Nassah et al. 1992). In recent years, Goss’s wilt has reemerged, especially in the region bordering western Nebraska, northeast Colorado, and southeast Wyoming. By 2010, CMN has been confirmed in Kansas, Colorado, Wyoming, South Dakota, Iowa, Illinois, Wisconsin, Texas, Indiana, Minnesota, and in Canada (Zhu et al. 1999; Agriculture and Agri-Food Canada 2000; Jackson et al. 2007, 2010; Robertson 2008; Wise and Ruhl 2008; Jackson 2009; Ruhl et al. 2009; Dawson 2010; Heppner 2010; Malvick et al. 2010). The disease exists in two forms: a leaf blight and a systemic vascular wilt. Infection occurs most readily after leaf wounding from wind-driven rains carrying soil or sand and after

Received 29 November 2010. Revision received 3 February 2011. Accepted 10 February 2011. Published at www.nrcresearchpress.com/cjm on 21 April 2011. I.V. Agarkova. Plant Pathology Department, University of Nebraska-Lincoln, 204 Morrison Hall, Lincoln, NE 68583-0900, USA. P.A. Lambrecht. Plant Pathology Department, University of Nebraska-Lincoln, 435 Plant Science Hall, Lincoln, NE 68583-0722, USA. A.K. Vidaver. Plant Pathology Department, University of Nebraska-Lincoln, 406C Plant Science Hall, Lincoln, NE 68583-0722, USA. Corresponding author: A.K. Vidaver (e-mail: [email protected]). Can. J. Microbiol. 57: 366–374 (2011)

doi:10.1139/W11-016

Published by NRC Research Press


Agarkova et al.

hailstorms. Foliar blight lesions are dark, water-soaked streaks with wavy margins parallel to leaf veins. At the margin of lesions, small, dark, discontinuously water-soaked, irregularly shaped “freckles”, characteristic of the disease, are frequently present. A dried, crystalline bacterial exudate may appear as a sheen on blighted leaves. The systemic vascular form of the disease is less common. This phase develops when the pathogen enters the vascular system and moves systemically within the xylem. This form is characterized by orange–yellow discoloration of vascular bundles. The disease can further progress to a stalk rot and cause plant death. Losses from systemic infection can reach up to 50% during severe epidemics (Claflin 1999). The pathogen is a subspecies of the gram-positive bacterium Clavibacter michiganensis (Smith 1910) Davis et al. 1984 belonging to the family Microbacteriaceae Park et al. 1995 emend. Zhi et al. 2009. Clavibacter michiganensis is composed of five subspecies, all of them plant pathogens. The other subspecies cause diseases of alfalfa (C. michiganensis subsp. insidiosus), tomato (C. michiganensis subsp. michiganensis, CMM), potato (C. michiganensis subsp. sepedonicus, CMS), and wheat (C. michiganensis subsp. tessellarius). Clavibacter michiganensis cells are rods of coryneform morphology, with B2g-type cell wall peptidoglycan with the diaminobutyric acid MK-9 as the predominant menaquinone, phosphatidyglycerol and diphosphatidyglycerol as the basic polar lipids, and a high G+C content of 72– 74 mol% (Davis et al. 1984). Little is known about CMN genome organization, and no strains of the pathogen have been sequenced. The molecular mechanism of CMN pathogenicity is unknown. Some strains of CMN carry a circular 51.65 kb plasmid (Gross et al. 1979). Strains of closely related subspecies, namely CMM and CMS, have been sequenced recently (Bentley et al. 2008; Gartemann et al. 2008). Both sequenced pathogens have plasmids: CMM carries two circular plasmids (27.4 and 70.0 kb); CMS has one circular (50.35 kb) and one linear (94.79 kb) plasmid. Information about population structure and diversity of CMN is limited. Studies of pathogen diversity have been done using colony morphology, bacteriophage typing, bacteriocin production, bacteriocin sensitivity, and whole-protein electrophoresis (Vidaver et al. 1981; Smidt and Vidaver 1987). Knowledge of structure and dynamics of the CMN population is necessary to understand the mechanisms of pathogen epidemiology and successful disease management. This information will provide important data for effective strategies in the mitigation of Goss’s wilt, including identifying new sources of disease resistance. The goal of the present study was to characterize the structure of the CMN population using broadly accepted genotypic methods employing the whole bacterial genome, namely amplified fragment length polymorphism (AFLP) and repetitive DNA sequence-based genomic fingerprinting using BOX-PCR (Versalovic et al. 1994; Vos et al. 1995).

Materials and methods Source of strains, growth, and extraction of DNA Strains of CMN used in the study are described in Table 1. Cultures were maintained for routine use by monthly transfer

367

on nutrient broth yeast extract (NBY) agar plates (Vidaver 1967). For long-term storage, cultures were lyophilized and kept at –20 °C. CMN strains were identified based on morphological and physiological characteristics (positive for the Gram stain and negative for the KOH test; and short, pleomorphic rods of coryneform shape) and carbon source utilization (Biolog GP2 microplates) with the MicroLog system 2 database analysis (release 4.01B, Biolog Inc., Hayward, California). For DNA extraction, the bacteria were grown on a rotary shaker in NBY broth. Genomic DNA was isolated using a Gram Positive DNA Purification kit (Epicentre Biotechnologies, Madison, Wisconsin) according to the manufacturer’s protocol. Concentrations of DNA were determined with a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Wilmington, Delaware) standardized to 100 ng/µL with TE buffer (1 mmol/L EDTA in 10 mmol/L Tris–HCl, pH 8.0) and stored at –20 °C. Pathogenicity assay Overnight liquid NBY cultures of CMN were harvested by centrifugation (4000g, 10 min at 4 °C), washed twice, and suspended in sterile 12.5 mmol/L potassium phosphate buffer, pH 7.1, to a final concentration of 1 × 107 to 5 × 107 CFU/mL. Alternatively, inoculum was prepared by suspending log-phase growth from NBY agar (Difco Laboratories, Detroit, Michigan) in potassium phosphate buffer. Concentration and purity of prepared cultures were verified by plating serial dilutions of inoculum aliquots on NBY agar. Prepared inoculum was maintained on ice prior to inoculation. Corn plants (‘Golden Cross Bantam’) were grown in the greenhouse in a steam-pasteurized soil mix to a three- to five-leaf stage of development (2–3 weeks after sowing at 26–28 °C and a 14 h : 10 h (light:dark) photoperiod). The inoculation procedure involved two successive wounds made at right angles to one another. The first wound was made horizontally through the plant stem, 1 cm above soil level with a 26-gauge needle attached to a syringe filled with inoculum. A 5 µL drop of bacterial suspension was formed on the bevel of the needle, and the needle was withdrawn through the stem, depositing the inoculum. The plant was rotated 90° and the needle was reintroduced at the same level perpendicular to the first wound; a 5 µL drop of suspension was formed and the needle withdrawn. Plants were inoculated in triplicate. Inoculation with sterile potassium phosphate buffer and pathogenic CMN served as negative and positive controls, respectively. Symptom development was evaluated 7–10 days after inoculation. AFLP The AFLP procedure was carried out as described previously (Vos et al. 1995) with slight modifications (Agarkova et al. 2006). Briefly, an AFLP template was prepared with a combination of HpaI and EcoRI restriction endonucleases. Amplification was performed using primers EcoRI+0 (5′GAC TGC GTA CCA ATT C) and HpaII+0 (5′-CGA TGA GTC CTC ACC GGA). The EcoRI+0 primer had an infrared fluorescent dye, IRDye 700 (LI-COR Inc., Lincoln, Nebraska). The separation of amplified products was performed on a 6.5% polyacrylamide gel in a LI-COR Long ReadIR Published by NRC Research Press

368

Can. J. Microbiol. Vol. 57, 2011


Table 1. Strains used in this study. Strain CN18-1 CN18-5, CN18-6, CN38-1, CN38-4, CN44-1 39, 172, 298*, 313, 716, 2579 NCPPB (311), 2581 NCPPB (type strain), 27822 ATCC (Fur1-3), 721-S CN30-1, CN37-1, CN37-2, CN38-6, CN48-1*, CN50-2, CN51-1 CN28-1 CN4-2 CN15-1 CN49-2, CN62-1 1994 NC+ 1996 CIBA, 1996 NC+, 1996 Pioneer 9910, 9918-1, 9918-3 20032, 20037 31341, 32719 2006 Box Butte 1, 2006 Sheridan 3 2006 Scotts Bluff 3 2007 13, 200710-C, 200711-C, 20071-2, 200712-C, 200713-C, 200715-1, 200715-12, 200715-13, 200715-14, 200715-15, 200715-16, 200715-18, 200715-19, 200715-20, 200715-4, 200715-7, 200715-8, 200715-C, 200716-C, 200720-A, 200722-C, 200723-A, 20075-C 200715-11, 200715-17, 200715-2, 200715-3, 200715-5, 200715-6, 200715-9 200800460, 200800461 200800417 Indiana Fuss/Debus (F16) 200900406, 200900504, 200900407 CN18-4, CNN-1, H.D.363 K293E, NTA*, 313.2 PDG101 CNK-2 CN4-1, CN81-2 CN72-1, CN76-1, CN81-1, CNSD-1, CNSD-2 CN68-1*, CN76-2 CN72-10, CN72-11A, CN72-11B, CN72-37, CN72-3A, CN72-42, CN72-4A, CN72-5A CN72-23A, CN72-25A, CN72-40, CN72-7 CN72-2 atten, CN72-28, CN72-29, CN72-38 200717-C Butcher (G2) 200900798 Texas, Eitzman/Lookabill (F30), Hoff (F78), Schmer/Maser/Refrig (F80) CN104 Eitzman/Carnine (F12), Eitzman/Ross 1 (F38), Eizman/ Ross (F13), Harveson/Loose (F41) Harveson/Schiff-Hall (F32) 20079-B 200900479 20033 200724-C 13424 20047 20038 31491

Pathogenicity† Y Y Y

Year isolated 1969 1970 1971

AFLP group A1 A1 A1

BOX group A1 A1 A1

Composite AFLP–BOX group A1 A1 A1

Y

1972

A1

A1

A1

Y Neg Y Y Y Y Y Y Y Y Y Y

1973 1973 1974 1975 1994 1996 1999 2003 2004 2006 2006 2007

A1 A1 A1 A1 A1 A1 A1 A1 A1 A1 A1 A1

A1 A1 A1 A1 A1 A1 A1 A1 A1 A1 B2 A1

A1 A1 A1 A1 A1 A1 A1 A1 A1 A1 A1 A1

Y

2007

A1

A2

A1

Y Y Y Y Y Y Y Y Y Y Y

2008 2008 2009 2009 1970 1971 1972 1973 1974 1975 1982

A1 A1 A1 A1 A1 A1 A1 A1 A1 A1 A1

A1 A1 A1 A1 A1 A1 A1 A1 A1 A1 A1

A1 A1 A1 A1 A2 A2 A2 A2 A2 A2 A2

Y Neg Y Y Y

1982 1982 2007 2008 2009

A2 A1 A1 A2 A2

A1 A1 A1 A1 A1

A2 A2 A2 A2 A2

NA Y

N/A 2009

A1 A2

B4 B2

A2 B1

Y Y NA Y Y Y Y Y Y

2009 2007 2009 2003 2007 1999 2004 2003 2004

A2 A3 B1 B3 B2 B5 B5 B4 B6

A2 B1 A2 A2 B4 B3 B3 B6 B5

B1 B2 B3 B4 B5 B6 B7 B8 B9

*Strain has a plasmid 51.65 kb. † Y, pathogenic; Neg, not pathogenic; NA, not available.


Agarkova et al.


DNA Sequencer (LI-COR model 4200), and electrophoresis data were automatically collected and simultaneously recorded during the run. Repetitive DNA sequence-based genomic fingerprinting using the BOX-PCR BOX-PCR was performed using the primer corresponding to BOX A1 element (Louws et al. 1994, 1998). The primer sequence was 5′-CTA CGG CAA GGC GAC GCT GAC G. The PCR reaction and amplification conditions were as described elsewhere (Smith et al. 2001). The PCR products were resolved on 1.5% agarose gel in 0.5× TBE buffer. Gels were stained with ethidium bromide (0.5 mg/L), and digital images were captured using the ChemiDoc EQ System (BioRad, Hercules, California).

369 Fig. 1. Illustration of homogeneity (top) and heterogeneity (bottom) of amplified fragment length polymorphism (AFLP) banding patterns of Clavibacter michiganensis subsp. nebraskensis (CMN) strains resulting from the EcoRI+0 – HpaII+0 enzyme–primer combination. The DNA fragment size ranges from 50 to 700 bp. CMN strains (from top to bottom): CN76-2, CN81-1, CNSD-1, CN7223A, Butcher (G2), CN72-25A, 2581 NCPPB (type strain), 9918-1, 1996 CIBA, 200711-C, 2579 NCPPB (311), 721-S, CN30-1, 200715-19, 200717-C, 1994 NC+, 1996 NC+, 32719, 172, Eitzman/Carnine (F12), Harveson/Loose (F41), Harveson/Schiff-Hall (F32), 20079-B, 20033, 20038, 13424, 20047, 200900479, 200724C, 31491.

Data analysis Digital images were analyzed with GelCompar II (version 5.1) software (Applied Maths, Kortrijk, Belgium) according to the manufacturer’s instructions. DNA size standards were run alongside every 6–7 strains for proper normalization and alignment of multiple gels: for AFLP LI-COR DNA ladder (50–700 bp) and for BOX-PCR analysis 2-Log DNA ladder (0.1–10.0 kb) from New England Biolabs (Beverly, Massachusetts). After normalization of gels, cluster analysis of resulting banding patterns was performed with Pearson’s correlation coefficient using the Unweighted Pair Group Method Average (UPGMA) (Sokal and Michener 1958). Cluster analysis of composite AFLP–BOX-PCR data sets was carried out using averages from the experiments.

Results Reproducibility of AFLP and choice of restriction enzyme set Preliminary tests were carried out on a panel of 10 CMN strains to identify the most appropriate combination of enzymes and selective primers for AFLP analysis. The following combinations of enzyme pairs were tested: PstI–MseI, PstI–HpaII, EcoRI–MseI, and EcoRI–HpaII, with different combinations of primers with selective nucleotides (0, A, C, G, and T). Comparison of the banding patterns generated by the different enzyme–primer combinations showed that EcoRI+0 – HpaII+0 provided the best resolution for CMN strains (results not shown). This selected combination of primer–enzyme generated about 60 well-defined fragments in a targeted molecular size range (from 50 to 700 bp) suitable for cluster analysis and was chosen for the remainder of the study (Fig. 1). The reproducibility of the AFLP analysis was assessed in two ways: (1) 10 randomly selected strains were analyzed in duplicates (starting from DNA isolation) and (2) each strain was run at least twice (starting from the AFLP amplification procedure and following electrophoresis) to determine proper normalization and alignment of banding patterns from different runs by the GelCompar software. The different runs for each isolate resulted in almost identical banding profiles with slight variations of band intensity within AFLP. Cluster analysis of different runs showed that the similarity between patterns was at least 95%, and the results correlated very well

with previously described reproducibility values (Agarkova et al. 2006). AFLP analysis of CMN strains Cluster analysis of AFLP banding patterns was carried out with three cutoff similarity values: strains with a similarity of >95% were considered clonal isolates, 75% similarity was set for clusters, and 60% similarity for separate groups. The 131 CMN strains were separated into two groups (Table 1; Fig. 2). The majority of strains (124) were clustered into group A. This group had three clusters: A1 (65% similarity, n = 109 strains), A2 (76% similarity, n = 14), and A3 Published by NRC Research Press

370



100

80

60

40

20

Fig. 2. Cluster analysis of Clavibacter michiganensis subsp. nebraskensis banding profiles generated by amplified fragment length polymorphism (AFLP) analysis using the EcoRI–HpaII enzyme combination. The dendrogram was constructed using the unweighted pair group method average (UPGMA) with Pearson’s correlation coefficient. Numbers at the nodes of clusters represent similarity values.

Year Cluster isolated

Group

65.1

A1

Group A

represented by a single strain (20079-B). Group A included strains isolated over a 40 year time span (from 1969 to 2009) from various locations. Group B was very heterogeneous and included six clusters: B1, B2, B3, B4, B5, and B6, with most of the clusters containing a single strain. Similarity means between clusters in group B was from 8% to 73%. The similarity values between clusters in group B substantially exceeded the 60% cutoff mean set up for a group. We have decided to represent them as one group despite exceeding this cutoff mean because most of the clusters were represented by a single strain. All strains in group B were isolated from 1999 to 2009. BOX-PCR analysis The BOXA1 primer generated five to nine bands ranging in size from 0.2 to 2.0 kb, suitable for adequate strain subtyping (Fig. 3). The reproducibility of BOX-PCR was assayed in two ways: (1) on a set of 10 strains starting from DNA isolation and (2) all strains were analyzed and run twice starting from PCR reactions. The different experiments for each strain showed identical patterns, and reproducibility was >99%. Cluster analysis of BOX-PCR banding patterns grouped CMN strains into two groups (Table 1; Fig. 4) with a cutoff linkage of 70% that was further subdivided into clusters with a cutoff similarity of 75%. The overwhelming majority of strains (n = 119) belonged to group A that was represented by two clusters: A1 (n = 109) and A2 (n = 10). Strains belonging to group A originated from various locations over a time period of almost 40 years. Group B was formed by very heterogeneous strains (n = 12) and had six clusters. The mean similarity between clusters was from 18% to 64%. All strains in these clusters were isolated during the last 10 years (from 1999 to 2009).

Discussion 61.8

54.2

A2

75.7 52 68.8

49.2

64.9 7.7

56.1 73.1

20079-B 200900479 200724-C 20033 20038 13424 20047 31491

2007 2009 2007 2003 2003 1999 2004 2004

A3 B1 B2 B3 B4 B5 B5 B6

Group B

Goss’s bacterial wilt is a significant bacterial disease of corn in the United States, e.g., Nebraska (Jackson et al. 2007, 2010; Jackson 2009). There are no protective commercial chemicals available for treatment of the disease. The most efficient means of control for plant bacterial pathogens is the use of resistant cultivars. The genetic sources of Goss’s wilt resistance were identified and partially resistant corn hybrids were bred (Ngong-Nassah et al. 1992), which helped limit the damage caused by CMN for years. The disease was sporadic without significant yield losses except for susceptible varieties of popcorn, sweet, and dent corn hybrids (Jackson et al. 2007). Because of the low disease impact for many years, even in Nebraska, less than 25% of seed corn companies screened hybrids for their sensitivity to the disease (Jackson et al. 2007). A rise in Goss’s wilt occurrence for the last 5 years brings up the question as to what may have caused its reemergence. Some phenotypic variability within CMN populations was established by previous studies using bacteriocin sensitivity and production assays and bacteriophage typing (Vidaver et al. 1981; Smidt and Vidaver 1987). In those studies, 85 CMN strains isolated between 1969 and 1979 were typed into eight clusters. The strain grouping did not correlate with Published by NRC Research Press


Agarkova et al.

100

80

60

40

Fig. 4. Cluster analysis of Clavibacter michiganensis subsp. nebraskensis banding profiles generated by BOX-PCR analysis. The dendrogram was constructed using the unweighted pair group method average (UPGMA) with Pearson’s correlation coefficient. Numbers at the nodes of clusters represent similarity values. 20

Fig. 3. Illustration of homogeneity (top) and heterogeneity (bottom) of BOX-PCR banding profiles of Clavibacter michiganensis subsp. nebraskensis (CMN) strains. The DNA fragment size ranged from 0.2 to 2.0 kb. CMN strains (from top to bottom): CNK-2, CNSD-1, K293E, H.D.363, 2581 NCPPB (type strain), CN72-11B, 200713-C, 200900406, CN30-1, CN4-2, CN50-2, Fuss/Debus (F16), Schmer/ Maser/Refrig (F80), 298, 200720-A, CN18-1, 200715-2, 200715-17, 200715-5, 20033, 200900479, Harveson/Schiff-Hall (F32), 20079-B, Eitzman/Ross 1 (F38), Harveson/Loose (F41), Eitzman/Carnine (F12), Eizman/Ross (F13), 13424, 20047, 200724-C, 31491, 20038.

371

Year isolated

Cluster

A1

78.9

Group

Group A

74.3

64.1

A2

75.0

55.6

96.8 88

48.2 76.4

41

95.5 60.4

17.9

either the year of isolation or geographic source (Vidaver et al. 1981). This is the first study on genetic diversity of the Goss’s wilt pathogen using molecular biology methods. We analyzed 131 representative strains of CMN. The collection of strains

20079-B F38 F41 F13 F12 13424 20047 200724-C 31491 20038

2007 2009 2009 2009 2009 1999 2004 2007 2004 2003

B1 B2 B2 B2 B2 B3 B3 B4 B5 B6

Group B

selected for the study was as comprehensive as possible to represent the diversity of the pathogen. Two DNA-based methods were employed: AFLP and repetitive DNA sequence-based genomic fingerprinting, namely BOX-PCR. Both techniques showed a high resolution capacity for subtyping and identifying other C. michiganensis subspecies (de Leon et al. 2009; Kleitman et al. 2008; Louws et al. Published by NRC Research Press


372

76.5

100

80

60

40

Fig. 5. Unweighted pair group method average composite AFLP– BOX-PCR dendrogram was constructed using averages from experiments. Numbers at the nodes of clusters represent similarity values. 20

1998; Smith et al. 2001) and closely related Rathayibacter toxicus (Agarkova et al. 2006). AFLP is a convenient and reliable fingerprinting technique that has been used to characterize bacterial population structure, taxonomic diversity, and phylogeny of a broad range of bacteria including plant pathogens (Vos et al. 1995). The technique has the ability to discriminate at the species level and can be comparable with DNA hybridization studies for some species of bacteria (Mougel et al. 2002; Rademaker et al. 2000). The other fingerprinting technique employed in this study was based on repetitive DNA sequences (ubiquitous and highly conserved motifs in the genomes of diverse bacteria): the most commonly used repetitive extragenic palindromic (REP) elements, enterobacterial repetitive intergenic consensus (ERIC) sequences, and dispersed-repeat motif (BOX). Collectively, all these assays are referred to as Rep PCR (Louws et al. 1994; Versalovic et al. 1994). AFLP and REP-PCR analyses have some common approaches in the way they generate DNA fragments. They both utilize the entire genomic DNA for generating polymorphic DNA bands. They have similar limitations as well; both of them can exclude from analysis large segments (stretches) of genomic DNA if that portion of DNA does not contain restriction sites for enzymes (AFLP) or repetitive elements (BOX-PCR). The comparative analysis of clusters generated by BOX-PCR and AFLP showed that in general, both methods have similar levels of reliability for subtyping CMN strains. In terms of practicality of the methods, BOX-PCR is less laborious and does not require expensive equipment. On the other hand, AFLP has a higher level of genetic resolution due to the large number of heterogeneous fragments available for numerical analysis. Both AFLP and BOX-PCR cluster analyses produced similar results. They both clustered CMN strains into two analogous groups: a major group A and a minor group B. The minor discrepancies between AFLP and BOX-PCR cluster analysis were in terms of the composition and number of strains in these groups (Table 1). Major group A had 124 strains by AFLP analysis and 119 by BOX-PCR analysis. Group B had seven strains by AFLP-based cluster analysis and 10 strains by analysis of BOX-PCR generated patterns. Five strains (13424, 20038, 20047, 200724-C, 31491) in group B were common for both methods. To overcome some discrepancies between AFLP and BOX-PCR fingerprint analysis a composite data set using averages from the experiments was generated employing features of GelCompar software (version 5.1). The composite data set analysis clustered strains into two groups (group A, n = 118; group B, n = 13) very similar to those in AFLP and BOX-PCR analysis (Fig. 5). The composite data analysis showed that the genome of the majority CMN strains was stable for a relatively long period of time (group A) and started to change in the last 10 years (group B). The similarity values derived from a composite data set were up to 75% for the major group A and up to 17% for the minor group B. The clustering in both groups (including very diverse strains, e.g., 31491 or 20038) was not correlated to strain pathogenicity or other phenotypic traits. There was no correlation in group A relative to strain, origin, history, morphology, or physiology. All strains included in group B were isolated after year 1999. The proportion of group B strains (n = 13)


Year isolated

Cluster

1969- 2009

Group

A1

Group A

73.6

1970-2009 75.1

A2

54.1

96.2 94 89.1 50.7 80.1 63 59.9 57.4 44 55.5 77 13.3

F13 F41 F38 F12 F32 20079-B 200900479 20033 13424 20047 200724- C 20038 31491

2009 2009 2009 2009 2009 2007 2009 2003 1999 2004 2007 2003 2004

B1 B1 B1 B1 B1 B2 B3 B4 B6 B7 B5 B8 B9

Group B

relative to all other strains isolated after 1999 (n = 69) was 18.8%. A possible explanation for the relative genetic homogeneity of CMN populations before 1999 might be its short evolutionary history. Limited diversity within strains collected in Published by NRC Research Press


Agarkova et al.

previous years may also be an indication that the American Midwest is not within the bacterium’s center of origin, and thus, more attention should be given to bacterial pathogens in the corn centers of origin. The appearance of significant genetic variations in some strains isolated after 1999 may be associated with selective pressure exerted by partially resistant cultivars of corn, for instance, or by other types of environmental forces. On the other hand, it is quite possible that corn is not the only nor primary host for CMN. Studies of CMN’s host range demonstrated that besides maize (Zea mays) and teosinte (Euchlaena (syn. Zea) mexicana), the bacterium is also able to cause disease in green foxtail (Setaria viridis), eastern gamagrass (Tripsacum dactyloides), sugarcane (Saccharum officinarum), shattercane (Sorghum bicolor), grain sorghum (Sorghum vulgare), and sudangrass (Sorghum vulgare var. sudanense) (Schuster 1975). Naturally occurring disease in these hosts has not been observed. All the strains analyzed in this study were isolated from diseased corn plants. It may be that CMN can persist asymptomatically on other hosts as part of the normal flora. Unfortunately, neither AFLP nor Rep-PCR analysis can provide any information on genome organization. We know that four strains of CMN (298, CN48-1, CN68-1, and NTA) have a single circular plasmid of 51.3 kb (Bentley et al. 2008; Gartemann et al. 2008). There were no detectable differences between strains with plasmids and those without. That fact might be an indication that there are no EcoRI and HpaII restriction sites on the plasmid. Since the nature of CMN pathogenicity is unknown, it is not possible to relate detected differences among strains to pathogenicity or to the recent disease spread. It is unlikely that the group B strains are responsible for the rise of CMN for the reason that it is a small group. The proportion of this group (n = 13) to all strains isolated after 1999 (n = 69) is only 18.8%. However, the increasing diversity of strains may be associated with increased diversity in corn hybrids or other factors. The results of this work provide a baseline and a foundation for future monitoring of CMN population structure. In general, the CMN population has been stable and not highly diverse for many years. The appearance of new genetically distinct groups of the pathogen may change the situation. Corn breeders need to be aware of diversity among new emerging strains of the bacteria.

Acknowledgements The authors thank Dr. Tamra Jackson, Dr. Robert Harveson, and Kevin Korus for providing many of the strains used in this study.

References Agarkova, I.V., Vidaver, A.K., Postnikova, E.N., Riley, I.T., and Schaad, N.W. 2006. Genetic characterization and diversity of Rathayibacter toxicus. Phytopathology, 96(11): 1270–1277. doi:10.1094/PHYTO-96-1270. PMID:18943965. Agriculture and Agri-Food Canada. 2000. The Canadian Phytopathological Society / Canadian plant disease survey — disease highlights. Agriculture and Agri-Food Canada. Available from http://www.cps-scp.ca/download/cpds_v80.pdf [accessed 11 April 2011]. Bentley, S.D., Corton, C., Brown, S.E., Barron, A., Clark, L.,

373 Doggett, J., et al. 2008. Genome of the actinomycete plant pathogen Clavibacter michiganensis subsp. sepedonicus suggests recent niche adaptation. J. Bacteriol. 190(6): 2150–2160. doi:10. 1128/JB.01598-07. PMID:18192393. Claflin, L.E. 1999. Goss’s bacterial wilt and blight. In Compendium of corn diseases. Edited by D.G. White. American Phytopathological Society, St. Paul, Minn., USA. Davis, M.J., Gillaspie, A.G., Vidaver, A.K., and Harris, R. 1984. Clavibacter: a new genus containing some phytopathogenic coryneform bacteria, including Clavibacter xyli subsp. xyli sp. nov., and Clavibacter xyli subsp. cynodontis subsp. nov., pathogens that cause ratoon stunting of sugarcane and bermuda grass stunting disease. Int. J. Syst. Bacteriol. 34(2): 107–117. doi:10.1099/00207713-34-2-107. Dawson, A. 2010. Goss’s wilt widespread in Manitoba corn [online]. AGCanada.com, 16 Sept. 2010. Available from http://agcanada. ironsolutions.com/Article.aspx?ID=26008 [accessed 11 April 2011]. de Leon, L., Rodriguez, A., Llop, P., Lopez, M.M., and Siverio, F. 2009. Comparative study of genetic diversity of Clavibacter michiganensis subsp. michiganensis isolates from the Canary Islands by RAPD-PCR, BOX-PCR and AFLP. Plant Pathol. 58(5): 862–871. doi:10.1111/j.1365-3059.2009.02117.x. Gartemann, K.H., Abt, B., Bekel, T., Burger, A., Engemann, J., Flügel, M., et al. 2008. The genome sequence of the tomatopathogenic actinomycete Clavibacter michiganensis subsp. michiganensis NCPPB382 reveals a large island involved in pathogenicity. J. Bacteriol. 190(6): 2138–2149. doi:10.1128/JB.01595-07. PMID:18192381. Gross, D.C., Vidaver, A.K., and Keralis, M.B. 1979. Indigenous plasmids from phytopathogenic corynebacterium species. J. Gen. Microbiol. 115: 479–489. Heppner, K. 2010. Goss’s wilt in corn [online]. PortageOnline, 18 Aug. 2010. Available from http://www.portageonline.com/ index.php?option=com_content&task=view&id=18649&Itemid=469 [accessed 11 April 2011]. Jackson, T. 2009. Goss’s wilt expanding in the state [online]. CropWatch — Nebraska crop production & pest management information, 21 Aug. 2009. Available from http://liferaydemo.unl. edu/web/cropwatch/archive?articleID=1095172 [accessed 11 April 2011]. Jackson, T.A., Harveson, R.M., and Vidaver, A.K. 2007. Reemergence of Goss’s wilt and blight of corn to the central High Plains. Plant Health Prog. doi:10.1094/PHP-2007-0919-01-BR. Jackson, T., Timmerman, A.Z., and Rees, J. 2010. Corn diseases update: southern rust confirmed and Goss’s wilt continues [online]. CropWatch — Nebraska crop production & pest managemen information, 6 Aug. 2010. Available from http:// cropwatch.unl.edu/web/cropwatch/archive?articleID=4237646 [accessed 11 April 2011]. Kleitman, F., Barash, I., Burger, A., Iraki, N., Falah, Y., Sessa, G., et al. 2008. Characterization of a Clavibacter michiganensis subsp. michiganensis population in Israel. Eur. J. Plant Pathol. 121: 463– 475. doi:10.1007/s10658-007-9264-z. Louws, F.J., Fulbright, D.W., Stephens, C.T., and de Bruijn, F.J. 1994. Specific genomic fingerprints of phytopathogenic Xanthomonas and Pseudomonas pathovars and strains generated with repetitive sequences and PCR. Appl. Environ. Microbiol. 60(7): 2286–2295. PMID:8074510. Louws, F.J., Bell, J., Medina-Mora, C.M., Smart, C.D., Opgenorth, D., Ishimaru, C.A., et al. 1998. Rep-PCR-mediated genomic fingerprinting: a rapid and effective method to identify Clavibacter michiganensis. Phytopathology, 88(8): 862–868. doi:10.1094/ PHYTO.1998.88.8.862. PMID:18944895. Published by NRC Research Press


374 Malvick, D., Syverson, R., Mollov, D., and Ishimaru, C.A. 2010. Goss’s bacterial blight and wilt of corn caused by Clavibacter michiganensis subsp. nebraskensis occurs in Minnesota. Plant Dis. 94(8): 1064. doi:10.1094/PDIS-94-8-1064A. Mougel, C., Thioulouse, J., Perriere, G., and Nesme, X. 2002. A mathematical model for determining genome divergence and species delineation using AFLP. Int. J. Syst. Evol. Microbiol. 52: 573–586. doi:10.1099/ijs.0.01956-0. PMID:11931171. Ngong-Nassah, E.N., Carson, M.L., and Wicks, Z.W. 1992. Inheritance of resistance to leaf freckles and wilt caused by Clavibacter michiganense subsp. nebraskense in early maturing maize inbred lines. Phytopathology, 82(2): 142–146. doi:10.1094/ Phyto-82-142. Rademaker, J.L.W., Hoste, B., Louws, F.J., Kersters, K., Swings, J., Vauterin, G., et al. 2000. Comparison of AFLP and rep-PCR genomic fingerprinting with DNA–DNA homology studies: Xanthomonas as a model system. Int. J. Syst. Evol. Microbiol. 50: 665–677. PMID:10758874. Robertson, A. 2008. Goss’s wilt prevalent in western Iowa [online]. Integrated Crop Management News. Available from http://www. extension.iastate.edu/CropNews/2008/0811robertsonjesse.htm [last updated 12 August 2008]. Ruhl, G., Wise, K., Creswell, T., Leonberger, A., and Speers, C. 2009. First report of Goss’s bacterial wilt and leaf blight on corn caused by Clavibacter michiganensis subsp. nebraskensis in Indiana. Plant Dis. 93(8): 841. doi:10.1094/PDIS-93-8-0841B. Schuster, M.L. 1975. Leaf freckles and wilt of corn incited by Corynebacterium nebraskense Schuster, Hoff, Mandel, Lazar, 1972. Research Bulletin No. 270. The Agricultural Experimental Station, IANR, University of Nebraska-Lincoln. Smidt, M.L., and Vidaver, A.K. 1987. Variation among strains of Clavibacter michiganense subsp. nebraskense isolated from a single popcorn field. Phytopathology, 77(2): 388–392. doi:10. 1094/Phyto-77-388.

Can. J. Microbiol. Vol. 57, 2011 Smith, N.C., Hennessy, J., and Stead, D.E. 2001. Repetitive sequence-derived PCR profiling using the BOX-A1R primer for rapid identification of the plant pathogen Clavibacter michiganensis subspecies sepedonicus. Eur. J. Plant Pathol. 107(7): 739– 748. doi:10.1023/A:1011955811847. Sokal, R., and Michener, C. 1958. A statistical method for evaluating systematic relationships. Univ. Kans. Sci. Bull. 38: 1409–1438. Versalovic, J., Schneider, M., de Bruijn, F.J., and Lupski, J.R. 1994. Genomic fingerprinting of bacteria using repetitive sequencebased polymerase chain reaction. Methods Mol. Cell. Biol. 5: 25–40. Vidaver, A.K. 1967. Synthetic and complex media for rapid detection of fluorescence of phytopathogenic pseudomonas: effect of the carbon source. Appl. Microbiol. 15(6): 1523–1524. PMID: 16349789. Vidaver, A.K., and Mandel, M. 1974. Corynebacterium nebraskense, a new, orange-pigmented phytopathogenic species. Int. J. Syst. Bacteriol. 24(4): 482–485. doi:10.1099/00207713-24-4-482. Vidaver, A.K., Gross, D.C., Wysong, D.S., and Doupnik, J.R. 1981. Diversity of Corynebacterium nebraskense strains causing Goss’s bacterial wilt and blight of corn. Plant Dis. 65(6): 480–483. doi:10. 1094/PD-65-480. Vos, P., Hogers, R., Bleeker, M., Reijans, M., van de Lee, T., Hornes, M., et al. 1995. AFLP, a new technique for DNA fingerprinting. Nucleic Acids Res. 23(21): 4407–4414. doi:10.1093/nar/23.21. 4407. PMID:7501463. Wise, K., and Ruhl, G. 2008. Goss’s bacterial wilt confirmed in Indiana [online]. Corny News Network Articles, 16 Aug. 2008. Available from http://www.agry.purdue.edu/ext/corn/news/articles.08/GossWilt-0816.html [accessed 11 April 2011]. Zhu, X., Reid, L., Presello, D., and Woldemarian, T. 1999. Corn pest survey report. Agriculture and Agri-Food Canada Research Branch. Available from http://www.ontariocorn.org/growing/PeststudyAAFC.htm [accessed 11 April 2011].
