P u b l i s h i n g Australasian Plant Pathology Volume 30, 2001 © Australasian Plant Pathology Society 2001 A journal for the publication of original research in all branches of plant pathology For editorial enquiries and manuscripts, please contact: Australasian Plant Pathology Editor-in-Chief Dr Eric Cother Orange Agricultural Institute NSW Agriculture, Forest Road Orange, NSW 2800, Australia Telephone: +61 3 6391 3886 Fax: +61 3 6391 3899 Email: ric.cother@agric.nsw.gov.au For general enquiries and subscriptions, please contact: CSIRO Publishing PO Box 1139 (150 Oxford St) Collingwood, Vic. 3066, Australia Telephone: +61 3 9662 7626 Fax: +61 3 9662 7611 Email: app@publish.csiro.au Published by CSIRO Publishing for the Australasian Plant Pathology Society w w w. p u b l i s h . c s i ro . a u / j o u r n a l s / a p p Australasian Plant Pathology, 2001, 30, 145–151 Leifsonia xyli-like bacteria are endophytes of grasses in eastern Australia L. MillsAE, T. M. LeamanB, S. M. TaghaviA, L. ShackelA, B. C. DominiakC, P. W. J. TaylorD, M. FeganB and D. S. TeakleA A Department of Microbiology and Parasitology, The University of Queensland, St Lucia, Queensland 4072, Australia. B Cooperative Research Centre for Tropical Plant Protection, The University of Queensland, St Lucia, Queensland 4072, Australia. C Bureau of Sugar Experiment Stations (BSES), Meiers Road, Indooroopilly, Queensland 4068, Australia; present address: NSW Agriculture, 161 Kite Street, Orange, NSW 2800, Australia D Bureau of Sugar Experiment Stations (BSES), Meiers Road, Indooroopilly, Queensland 4068, Australia; present address: Molecular Plant Genetics and Germplasm Development Group, Department of Crop Production, The University of Melbourne, Parkville, Vic. 3010, Australia. E Corresponding author; email: kahlo@smartchat.net.au Abstract. Bacteria serologically related to Leifsonia xyli ssp. xyli, the causal bacterium of ratoon stunting disease (RSD) in sugarcane, were detected using the fluorescent antibody direct count on filter (FADCF) technique in grasses in eastern Australia. In a survey of 191 grass, sedge and bullrush samples comprising 53 plant species, 90 (47%) of the samples tested harboured bacteria which reacted positively with L. xyli ssp. xyli polyclonal antiserum. A total of 18 grass species was found to be naturally colonised with bacteria serologically related to and morphologically similar to L. xyli ssp. xyli. Grasses colonised by these L. xyli-like bacteria were present in areas both adjacent to, and removed from, sugarcane crops. When L. xyli-like bacteria from Rhodes grass (Chloris gayana) were inoculated into sugarcane, they multiplied at a lower rate than L. xyli ssp. xyli. L. xyli-like bacteria in Rhodes grass were isolated in axenic culture and exhibited growth rates, colony size and pigmentation similar to those of L. xyli ssp. cynodontis, a bacterial pathogen of Cynodon dactylon (couch grass). Further, using a polymerase chain reaction (PCR) test that could differentiate L. xyli spp. xyli from L. xyli spp. cynodontis, the L. xyli-like bacteria infecting Rhodes grass, couch grass and panic grass (Panicum maximum) generated a product of the same size as L. xyli spp. cynodontis. We conclude that L. xyli ssp. cynodontis or closely related bacteria are common endophytes of grasses in eastern Australia. Additional keywords: ratoon stunting disease, serology, PCR, Clavibacter. Introduction Leifsonia xyli subspecies xyli (Davis et al.) Evtushenko, a coryneform bacterium (formerly of the genus Clavibacter, Evtushenko et al. 2000), infects the xylem of sugarcane (Saccharum interspecific hybrids) and causes ratoon stunting disease (RSD; Gillaspie and Teakle 1989) in susceptible sugarcane cultivars. Although sugarcane is the only known natural host of the bacterium, L. xyli ssp. xyli has been experimentally transmitted to 14 alternative hosts, namely elephant grass (Pennisetum purpureum Schum.) (Matsuoka 1971; Steindl 1957), bana grass (P. purpureum × P. glaucum (L.) R. Br.) (Steindl and Teakle 1974), maize (Zea mays L.) (Steindl 1961), panic grass (Panicum maximum Jacq.), para grass (Brachiaria mutica (Forsk.) Stapf), green summer grass (B. milliformis (Presl) Chase), barnyard grass (Echinochloa colonum (L.) Link), blady grass (Imperata cylindrica (L.) Beauv.), red Natal grass (Rhynchelytrum © Australasian Plant Pathology Society 2001 repens (Willd.) C.E. Hubbard), Parramatta grass (Sporobolus capensis (Willd.) Kunth), wild sorghum (Sorghum verticilliflorum (Steud.) Stapf) (Steindl 1957), Sudan grass (Sorghum drummondii (Steud.) Millsp. & Chase), Johnson grass (Sorghum halepense (L.) Pers.) (Liao and Chen 1981) and couch grass (Cynodon dactylon (L.) Pers.) (Liao and Chen 1981). These inoculated hosts remained symptomless. A related bacterium, L. xyli ssp. cynodontis, occurs in, and causes stunting of, Cynodon dactylon (Liao and Chen 1981; Davis et al. 1984). This bacterium has been mechanically transmitted to Sudan grass (Liao and Chen 1981) and maize (Barbehenn and Purcell 1993). L. xyli sspp. xyli and cynodontis share some common cell surface antigens and both react with polyclonal antiserum produced against L. xyli ssp. xyli. They can be distinguished by their higher pathogenicity to their natural hosts, their different 10.1071/AP01003 0815-3191/01/020145 146 L. Mills et al. Table 1. Details of bacterial strains used in this study Bacterial species Leifsonia xyli ssp. xyli L. xyli ssp. cynodontis Clavibacter michiganensis ssp. michiganensis C. michiganensis ssp. insidiosus C. michiganensis ssp. nebraskensis Rathayibacter rathayi R. tritici L. xyli-like L. xyli-like Strain designation T ACM 2272/L1A TB1AB ACM 4950C ACM 4951 ACM 4952 ACM 4954 ACM 3998 Clone 9 Clone 102 Source Reactivity with antiserumA Sugarcane Couch grass Tomato Lucerne Corn Cocksfoot grass Wheat Rhodes grass Rhodes grass + + – – – – – + + A An antiserum to L. xyli ssp. xyli ACM 2272/L1A was used in FADCF tests. Type strain. C ACM = Australian Collection of Microorganisms, Department of Microbiology and Parasitology, The University of Queensland, St Lucia, Queensland 4072. B colony characteristics and growth rates on sugarcane (SC) agar (Davis et al. 1983) and by a polymerase chain reaction (PCR) test (Fegan et al. 1998). Despite the implementation of RSD control measures, some Australian sugarcane farmers are unable to eradicate the disease from their crops. Dominiak et al. (1992) concluded that there might be previously undetected routes of infection, such as transmission from L. xyli ssp. xyli-infected grasses during harvesting of sugarcane fields. This paper reports a survey of grasses in eastern Australia to determine if they were naturally infected by L. xyli ssp. xyli or related bacteria. Methods Bacterial cultures Reference and type strains of L. xyli and related bacteria were received from the Australian Collection of Microorganisms (ACM), The University of Queensland, Australia. In addition, L. xyli-like bacteria were isolated from Rhodes grass (Chloris gayana) in Queensland. Details of these bacterial strains are given in Table 1. L. xyli ssp. xyli and L. xyli ssp. cynodontis were maintained on modified sugarcane agar (MSC) (Croft et al. 1993). C. michiganensis and Rathayibacter spp. were maintained on 523M agar (Riley and Ophel 1992). All cultures were incubated at 28°C. Production of polyclonal antiserum against L. xyli ssp. xyli Antigen production L. xyli ssp. xyli (ACM2271) isolated from sugarcane cultivar Q110 was suspended in 3 mL of sterile saline. The suspension was centrifuged at 15 000 rpm for 5 min and the pellet was washed and resuspended in 1.5 mL PBS (8 g NaCl, 0.2 g KH 2PO4, 2.9 g Na2HPO4.12H2O, 0.2 g KCl, 1 L deionised water, pH 7.4). To the suspension, 1.5 mL of Freund’s incomplete adjuvant was added and the mixture emulsified. A 1 mL volume of the emulsion was injected into the thigh muscle of a rabbit. This injection of freshly prepared bacterial suspension was repeated at fortnightly intervals for 3 months using a freshly prepared bacterial suspension each time. Determination of antiserum titre Approximately 3–5 mL of blood was collected at weekly intervals and stored overnight at 4°C to allow clotting. The serum was separated from the clot by centrifugation at 15 000 rpm for 5 min and the serum was subjected to two-fold dilutions in PBS. A drop of each serum dilution was transferred to a Petri dish and mixed with a drop of concentrated suspension of L. xyli ssp. xyli in PBS. Following a 1 h incubation at room temperature, the Petri dish was placed under a stereo microscope and each of the drops examined for agglutination. The highest dilution of serum at which agglutination occurred was considered to be the titre. After 12 weeks the rabbit was bled out when a titre greater than 1:1000 was reached. Purification of immunoglobulin G (IgG) Twenty mL of blood was collected 8 weeks following the initial injection and was centrifuged at 15 000 rpm for 5 min. The serum fraction was withdrawn, diluted 1:10 with sterile deionised water, and 50 mL of saturated ammonium sulfate (76.2 g / 100 mL deionised water) was added and the solution stirred for 45 min at room temperature. The precipitate was collected by centrifugation at 12 000 rpm for 10 min and dissolved in 10 mL of PBS. The suspension was dialysed twice in 1.5 L of half strength PBS overnight. IgG was separated from the other immunoglobulins by elution through a DEAEcellulose column with PBS. Eluted fractions were collected in microfuge tubes and absorbances were determined spectrophotometrically at 280 nm. The approximate IgG concentration was calculated using the formula 1 mg mL–1 IgG = 1.4 absorbance units. Fractions with absorbances greater than 0.5 were retained. Conjugation of IgG with fluorescein isothiocyanate (FITC) The procedure followed was that of Davis and Dean (1984). Specificity of the L. xyli ssp. xyli polyclonal antiserum using the fluorescent antibody direct count on filter (FADCF) technique Bacterial isolates tested included L. xyli ssp. xyli (ACM2272/L1A – type strain), L. xyli ssp. cynodontis (TB1A – type strain), C. michiganensis sspp. michiganensis (ACM4950), insidiosus (ACM4951), and nebraskensis (ACM4952), Rathayibacter rathayi (ACM4954) and R. tritici (ACM3998). A loopful of culture of each bacterium to be tested was suspended in 500 µL of PBS and centrifuged at 15 000 rpm for 2 min. The pellet was resuspended in 200 µL of PBS, 5 µL of FITC-conjugated IgG was added and the suspension was incubated in the dark at 28°C for 30 min. The suspension was diluted in 2 mL of PBS and passed through a 0.2 µm filter membrane (Gelman Sciences, USA). Bacteria were deposited on the membrane and fluorescent-labelled bacteria were detected with PL Fluotar 100× objective on a Leitz Laborlux S microscope with a blue light illumination Liefsonia xyli-like bacteria in grasses Table 2. 147 Grass species in south-east Queensland and northern New South Wales harbouring bacteria serologically related to Leifsonia xyli ssp. xyli as determined by the fluorescent antibody direct count on filter (FADCF) technique Host Brachiaria decumbens Stapf (signal grass) Chloris gayana Kunth (Rhodes grass) Chloris virgata Sw. (feathertop Rhodes grass) Cynodon dactylon (L.) Pers. (couch grass) Cynodon plectyostachyum Pilger (African star grass) Echinochloa colonum (L.) Link (awnless barnyard grass) Eleusine indica (L.) Gaertn. (crowsfoot grass) Eragrostis sp. (love grass) Eragrostis cilianensis (All.) Link ex Lutati (stink grass) Imperata cylindrica (L.) Beauv. (blady grass) Melinus minutiflora P. Beauv. (molasses grass) Panicum maximum Jacq. var. trichoglume Eyles ex Robyns (panic grass) Paspalum dilatatum Poir. (paspalum) Paspalum urvillei Steud. (Vasey grass) Rhynchelytrum repens (Willd.) C.E. Hubb. (red Natal grass) Setaria gracilis Beauv. (pale pigeon grass) Sorghum verticilliflorum (Steud.) Stapf (wild sorghum) Sporobolus caroli Mez (fairy grass) Locality Number of colonised plants per number sampled Brisbane (Qld) Blue knob (NSW) Brisbane (Qld) Bundaberg (Qld) Chuwar (NSW) Gympie (Qld) Kholo Creek (NSW) Kilcoy (Qld) Kunghur (NSW) Mt. Glorious (Qld) North Maclean (Qld) Tumblegum (NSW) Wivenhoe (Qld) Delungra (NSW) Brisbane (Qld) Chinderah (NSW) Bundaberg (Qld) Bundaberg (Qld) Brisbane (Qld) Bundaberg (Qld) Bundaberg (Qld) Bundaberg (Qld) Bundaberg (Qld) Chinderah (NSW) Cudgen (NSW) Cudgen (NSW) Tweed Valley (NSW) Cudgen (NSW) 1/1 1/1 6/11 7/11 1/1 4/5 1/1 1/1 1/1 1/1 8/12 1/1 2/2 3/4 3/6 3/3 1/4 3/6 1/1 4/5 2/3 1/1 3/4 1/1 1/1 1/1 1/1 6/6 Brisbane (Qld) Bundaberg (Qld) Brisbane (Qld) Mt. Glorious (Qld) Brisbane (Qld) 2/4 3/4 3/3 1/1 1/1 Bundaberg (Qld) Brisbane (Qld) Tweed Valley (NSW) Gympie (Qld) 1/2 1/2 1/1 2/2 (12 filter block) (Ernst Leitz Ltd., Germany). A strong uniform fluorescence of cells was deemed a positive reaction with the antiserum. centrifuged at 12 000 rpm for 5 min. The pellet was resuspended in 200 µL of PBS. Serological detection of Leifsonia xyli-like bacteria in plants Inoculation of sugarcane setts Extraction of bacteria from plants Sources of inoculum Grasses, sedges and a bullrush from locations in south-east Queensland and north-east New South Wales were sampled. Plants were removed from the soil with their roots intact. Soil and debris were removed from stalks by washing under running water. Stalks were cut into approximately 5 mm pieces, placed in a 25 mL bottle and immersed in PBS/Azide (500 mL PBS / 0.1 g NaN 3) for 2 h. A 1.5 mL portion of the suspension was transferred to a microfuge tube and Rhodes grass (Chloris gayana) clones 9 and 102 were maintained in a glasshouse at about 25°C and were watered regularly. Each pot was elevated and isolated to prevent cross contamination during irrigation and manipulations. The grasses were shown by FADCF to be infected with L. xyli-like bacteria. Stalks (approximately 200 g) of each clone were cut aseptically into 5–10 mm-long pieces. The cut pieces were added to 500 mL of sterile deionised water and the suspension was kept 148 at room temperature for 2 h. The liquid portion of the suspension was decanted and retained. Sugarcane cultivar Q110 known to be infected with L. xyli ssp. xyli was maintained at the BSES Pathology Farm, Brisbane. Five stalks were pressed to express the juice and the extract was diluted 1:1 with sterile deionised water. FADCF was used to determine the presence of L. xyli ssp. xyli in the extract. The initial inoculum concentration of L. xyli ssp. xyli and L. xyli-like bacteria was approximately 105 cells/ mL. Source of soil and RSD-free sugarcane Twenty stalks of apparently healthy sugarcane variety Q110, maintained at the BSES Pathology Farm, Brisbane, were steam treated at 52°C for 4–5 h (a treatment designed to cure plants of L. xyli ssp. xyli infection; Gillaspie and Teakle 1989). Single node setts were cut using alcoholsterilised cutters. Fibrovascular sap of heated sugarcane stalks was extracted under positive pressure (Croft and Witherspoon 1982) and found by FADCF to be free of L. xyli ssp. xyli. Approximately 100 kg of soil from the BSES Pathology Farm was crudely sieved to remove large pieces of organic matter and steam treated at 52°C for 3 h. L. Mills et al. reacted with L. xyli ssp. xyli cells and exhibited crossreactivity with L. xyli ssp. cynodontis cells. The antiserum did not react using FADCF with cells of C. michiganensis subspp. michiganensis, insidiosus and nebraskensis, Rathayibacter tritici or R. rathayi. PCR for the differentiation of L. xyli ssp. xyli and ssp. cynodontis A PCR assay used to distinguish L. xyli ssp. xyli from ssp. cynodontis (Fegan et al. 1998) was used to amplify DNA from L. xylilike isolates obtained from Rhodes grass. The PCR reaction contained three primers: a conserved forward primer which would anneal to either ssp. xyli or ssp. cynodontis, and two reverse primers which were specific for either ssp. xyli or ssp. cynodontis. The type strains for L. xyli sspp. xyli (ACM2272/L1A) and cynodontis (TB1A) were used as positive controls. The PCR test was also used in conjunction with the FADCF technique to test for the presence of L. xyli ssp. cynodontis in a total of 17 extracts obtained from eight grass species. Presence and geographical location of plants harbouring L. xyli ssp. xyli In a survey of 191 grass, sedge and bullrush samples obtained from 53 plant species, 90 of the samples harboured bacteria that reacted positively using FADCF with L. xyli ssp. xyli polyclonal antiserum. A total of 18 grass species of the family Poaceae was found to be naturally colonised with bacteria serologically related to, and morphologically similar to, L. xyli ssp. xyli . The colonised grass species and the localities from which they were obtained are shown in Table 2. Grass, bullrush and sedge species which did not harbour a natural population of L. xyli-like bacteria were Andropogon gerardii Vitm. (big bluestem), Axonopus compressus P. Beauv. (broadleaf carpet grass), Bothriochloa ewartiana (Domin) C.E. Hubb. (desert blue grass), Bothriochloa insculpta (Hochst. ex A. Rich.) A. Camus cv. Bisset (creeping blue grass), Brachiaria miliiformis (Presl) Chase (green summer grass), Bromus diandrus Roth (great brome), Capillipedium spicigerum S.T. Blake (scented top), Cenchrus ciliaris L. (buffel grass), Cyperus brevifolius (Rottb.) Haask. (Mullumbimby couch), C. difformis L. (rice weed), C. rotundus L. (nut grass), Dicanthium aristatum (Poiret) C.E. Hubb. (Angelton grass), Digitaria ciliaris (Retz.) Koeler (summer grass), D. parviflora (R. Br.) Hughes (small flower finger grass), D. sanguinalis (L.) Scop. (crab grass), Hemarthria altissima (Poir) Stapf & C.E. Hubb. (limpo grass), Heteropogon contortus (L.) P. Beauv. ex Roem. and Schult. (black speargrass), Panicum antidotale Retz. (blue panic grass), Paspalum nicorae Parodi (Brunswick grass), P. notatum Hugge cv. Competidor (Bahia grass), Pennisetum clandestinum Hochst. ex Chiov. (kikuyu), P. glaucum (L.) Pers. (pearl millet), Setaria anceps Stapf (setaria), Sorghastrum nutans (L.) Nash cv. Lometa (yellow Indian grass), Sorghum halepense (L.) Pers. (Johnson grass), Sporobolus pyramidalis P. Beauv. var. pyramidalis (giant rats tail grass), Typha sp. (bullrush), Urochloa mosambicensis (Hack.) Dandy (sabi grass), U. panicoides P. Beauv. var. panicoides (liverweed grass) and Vetiveria zizanioides (L.) Nash (Vetiver grass). It was noted that the size of the pellet obtained from diffusates of these plants was significantly smaller than that of Rhodes grass and thus the population of L. xyli-like bacteria, if present, may have been below the level of detection of the FADCF technique. Results Specificity of L. xyli ssp. xyli polyclonal antiserum Polyclonal antiserum with a homologous agglutination titre of 1:1430 against L. xyli ssp. xyli was produced 3 months after the first injection. Using the FADCF technique, the antiserum Infection in sugarcane Five months after inoculation of sugarcane setts with L. xyli ssp. xyli from sugarcane, fluorescent cells resembling those of this bacterium were detected in fibrovascular extracts of all eight stalks tested. At this time, no fluorescent Inoculation of sugarcane uprights Pots (190 × 210 mm) were washed, rinsed with 70% ethyl alcohol and filled with steamed soil to a depth of approximately 170 mm. Both cut ends of 20 healthy Q110 single-node setts were sprayed until saturated with the appropriate inoculum and the setts incubated for 1 h. The setts were then planted horizontally in the pots with the nodal bud uppermost, with four pots containing five single node setts per treatment. A control treatment was also included in which the setts were left untreated. The pots were maintained in a glasshouse at about 25°C. Two stalks from each pot were sampled at 5 and 8 months following inoculation. Fibrovascular extract was extracted via positive pressure from these stalks and the FADCF technique was used to detect L. xyli ssp. xyli and L. xyli-like bacteria. The extract was deemed positive if one or more bacteria per field were observed in ten microscope fields. The remaining fifth stalk from each pot was sliced longitudinally after 8 months to determine if vascular discoloration had occurred. In vitro isolation of L. xyli-like bacteria Rhodes grass cultivars 9 and 102 which harboured L. xyli-like bacteria were maintained in a glasshouse at about 25°C. Stalks were washed in tap water, blotted dry and cut into pieces 50–60 mm in length. The stalk pieces were immersed in 5% (w/v) sodium hypochlorite and 10% (v/v) Tween 20 for 5 min and rinsed in sterile deionised water. Approximately 10 mm was cut aseptically off a piece from one end and discarded. The sap was expressed from the remaining tissue directly onto MSC agar plates and incubated at 28°C. Liefsonia xyli-like bacteria in grasses Table 3. 149 The number of stalks infected, and the average concentration of fluorescent Leifsonia xyli-like bacteria in extracts, 8 months post-inoculation of sugarcane Q110 with L. xyli ssp. xyli or L. xyli-like bacteria. No. stalks infected Sugarcane juice extract Rhodes grass clone 9 extract Rhodes grass clone 102 extract Control — untreated Av. conc. bacteria/mL No. stalks infected 8/8 0/8 0/8 0/8 3.4 × 108 0 0 0 8/8 4/8 3/8 0/8 Av. conc. bacteria/mL 4.7 × 108 1.2 × 106 0.9 × 106 0 A The extracts from sugarcane (containing L. xyli ssp. xyli) and Rhodes grass (L. xyli-like bacteria) each contained approx. 105 cells/mL bacteria were detected in sugarcane inoculated with the L. xyli-like bacteria from Rhodes grass or in the uninoculated controls (Table 3). Eight months after inoculation with L. xyli ssp. xyli, fluorescent cells resembling those of this bacterium were again detected in 8/8 stalks of sugarcane examined (100% colonisation). However, fluorescent bacteria were also seen in 4/8 and 3/8 sugarcane stalks inoculated with L. xyli-like bacteria from Rhodes grass clone 9 and clone 102, respectively. The fluorescent bacteria were more numerous (approx. 108 cells/mL) in the extracts of infected sugarcane inoculated with L. xyli ssp. xyli than those inoculated, with the L. xyli-like bacteria from Rhodes grass (approx. 106 cells/mL; Table 3). The typical RSD symptoms of a red/orange discoloration of the vascular bundles in nodes of mature stalks, were not apparent when sugarcane inoculated with either L. xyli ssp. xyli or the L. xyli-like bacteria from Rhodes grass was examined after 8 months. In vitro isolation of L. xyli-like bacteria Six L. xyli-like isolates were obtained in pure culture on MSC agar from six Rhodes grass clones of cultivars Pioneer, Topcut or Finecut. The bacteria grew aerobically on MSC agar, producing colonies of approximately 0.5–1 mm diameter in 3–5 days. Colonies were circular with entire margins, were convex and developed yellow pigmentation. Isolation of L. xyli-like bacteria from other grasses was unsuccessful. Identification of L. xyli-like bacteria using PCR The six L. xyli-like strains obtained from Rhodes grass produced a 446 bp amplification product specific for L. xyli ssp. cynodontis (not shown). L. xyli-like bacteria were also detected using both PCR and FADCF in the fibrovascular extract of nine samples of Rhodes grass from which isolates were not obtained and in xylem exudate from Cynodon dactylon (couch grass) and Panicum maximum (panic grass). Discussion The only reported natural host of L. xyli ssp. xyli is sugarcane and the bacterium can be transmitted from diseased to healthy stalks on harvesting equipment (Hughes and Steindl 1955; Taylor et al. 1988). Despite its narrow host range in nature, the RSD bacterium has been shown to infect 14 grass species under experimental conditions (Steindl 1957; Liao and Chen 1981). Dominiak et al. (1992) suggested that grasses are undetected reservoirs of the RSD bacterium and that the bacterium might be transmitted to sugarcane during harvesting. To investigate this possibility we surveyed selected plants in subtropical eastern Australia. The survey of grasses, sedges and a bullrush in south-east Queensland and northern New South Wales using FADCF revealed that, out of 191 samples tested, 90 were positive for bacteria which were serologically and morphologically related to L. xyli ssp. xyli. A total of 18 grass species of the family Poaceae harboured natural populations of L. xyli-like bacteria (Table 2). These grasses were widely distributed in subcoastal areas. Grasses colonised by L. xyli-like bacteria were present in areas both near to, and remote from, commercial sugarcane cultivation areas. For instance, infected Eleusine indica and Paspalum dilatatum plants were found in the Brisbane area, in addition to growing adjacent to sugarcane fields in Bundaberg. Since L. xyli-like infected grasses were not exclusively found in sugarcane regions, there was no evidence that the colonised grasses were always being infected from sugarcane. At 5 months post-inoculation, infectivity studies demonstrated that L. xyli-like bacteria from Rhodes grass were not detected in sugarcane cultivar Q110, and they had increased to only approximately 106 cells/mL of extract at 8 months. Since the limit of sensitivity of the FADCF technique is 104–105 cells/mL, it is assumed that after 5 months the population of L. xyli-like bacteria in sugarcane was below the threshold of detection. In contrast, L. xyli ssp. xyli had multiplied in sugarcane to approximately 108 cells/ mL of extract after 5 months. In addition, while L. xyli ssp. xyli colonised 100% of inoculated sugarcane setts, L. xylilike bacteria from Rhodes grass clones 9 and 102 were detected in only 50% and 37.5% of setts, respectively. While L. xyli ssp. xyli is able to infect non-natural hosts, these infections occur at a lower efficiency and give rise to lower bacterial yields when compared to infections of natural hosts. Davis et al. (1983) observed that L. xyli ssp. xyli did not readily multiply within Cynodon dactylon and Liao and Chen (1981) reported that L. xyli ssp. cynodontis colonised only 22% of Sudan grass-sorghum hybrids, while L. xyli ssp. 150 xyli infected 92% of the plants. It can thus be inferred from this that the L. xyli-like bacteria isolated from Chloris gayana are probably not L. xyli ssp. xyli based on their inability to colonise, and propagate within, sugarcane vascular tissue to the corresponding level exhibited by L. xyli ssp. xyli. While the presence of vascular discoloration in mature nodes of sugarcane has been used to diagnose L. xyli ssp. xyli infection (Hughes and Steindl 1955; Steindl and Teakle 1974; Taylor et al. 1988; Gillaspie and Teakle, 1989), no discoloration or other symptoms of disease were observed in this study. Others have reported asymptomatic infection of sugarcane (Steindl 1961), and discoloration is known to vary among sugarcane clones (Gillaspie and Teakle 1989). It was expected that the susceptible Q110 cultivar used in this study would exhibit symptoms of disease. However, RSD is notably a disease that is observed during adverse growing conditions in the field. These conditions may not have been represented under the experimental conditions used in this study. Thus the lack of observable symptoms precludes the determination of the pathogenic role of L. xyli-like bacteria in RSD of sugarcane. Davis et al. (1988a) reported symptomatic colonisation of sugarcane with bacterial yields of 8.5 × 108 cells/mL from a mature internode, which is similar to the yields achieved in this study (4.7 × 108 cells/mL). The level of colonisation within sugarcane cultivars has also been shown to be an indicator of yield loss (Koike et al. 1982; Davis et al. 1988b). Therefore, the failure of the L. xyli-like bacteria from Rhodes grass to multiply to the level seen in L. xyli ssp. xyli-infected sugarcane cultivar Q110 suggests the L. xyli-like bacteria may be less pathogenic to sugarcane than L. xyli ssp. xyli. Isolates of L. xyli-like bacteria from Rhodes grass in axenic culture exhibited growth rates and colony size and pigmentation similar to L. xyli ssp. cynodontis. L. xyli-like isolates from Rhodes grass generated a PCR product the same size as L. xyli ssp. cynodontis, using a PCR test that can differentiate L. xyli ssp. xyli from ssp. cynodontis (Fegan et al. 1998). Due to the difficulties associated with isolating these fastidious, slow growing bacteria from plants, isolates were not obtained from grasses other than Rhodes grass. It was observed that Rhodes grass harboured a larger population of L. xyli-like bacteria than the other grasses examined and this may have contributed to the relative ease of isolation of these bacteria. As isolates could be obtained only from Rhodes grass, extracts of other grasses were subjected to PCR tests. L. xylilike bacteria were detected by PCR (producing a product the same size as that produced by L. xyli ssp. cynodontis) in Rhodes grass (Chloris gayana), couch grass (Cynodon dactylon) and panic grass (Panicum maximum). The observation that Rhodes grass harboured a larger population L. Mills et al. of L. xyli-like bacteria than other grasses may have also contributed to the low number of grasses that tested positive by PCR. The number of bacteria present in other grasses may have been below the limit of detection of the PCR test used (104 cells/mL; Fegan et al. 1998). We have shown that L. xyli-like bacteria naturally infected some plants of Rhodes grass (Chloris gayana), couch grass (Cynodon dactylon) and panic grass (Panicum maximum) in eastern Australia. L. xyli-like bacteria were present, sometimes in low numbers, in 15 other grasses in eastern Australia. The significance of these endophytes in the health and productivity of their host grasses in Australia remains to be determined. It is possible that these L. xyli-like bacteria could be mechanically transmitted from their grass hosts to sugarcane during harvesting operations. If the results of our inoculation test with the Rhodes grass isolates are applicable to isolates from other grasses, they are unlikely to cause significant disease in sugarcane. However, their cross-reactivity with L. xyli ssp. xyli antiserum in routine screening for RSD in sugarcane could conceivably result in false positive diagnoses. The exact identity of the L. xyli-like bacteria from Rhodes grass and other grasses in eastern Australia has not yet been determined. Although at least some resemble L. xyli ssp. cynodontis in having a faster growth rate and a yellower colony type on MSC agar than L. xyli ssp. xyli, further phenotypic and pathogenicity tests are needed to support evidence of similarity with L. xyli ssp. cynodontis provided by the PCR tests. Acknowledgements The authors thank the Queensland Bureau of Sugar Experiment Stations for their financial assistance. We are grateful to Mr Don Loch of the Queensland Department of Primary Industries for supplying grasses, including Rhodes grass clones 9 and 102. References Barbehenn RV, Purcell AH (1993) Factors limiting the transmission of a xylem-inhabiting bacterium Clavibacter xyli subsp. cynodontis to grasses by insects. Phytopathology 83, 859–863. Croft BJ, Witherspoon JR (1982) Molded unit for positive pressure extraction of ratoon stunting disease bacteria. Sugarcane Pathologists’ Newsletter 28, 33–34. Croft BJ, Teakle DS, Leaman TM (1993) Serological diagnostic tests for ratoon stunting disease in sugarcane. Instruction Manual MN93001. Davis MJ, Dean JL (1984) Comparison of diagnostic techniques for determining incidence of ratoon stunting disease of sugarcane in Florida. Plant Disease 68, 896–899. Davis MJ, Dean JL, Harrison NA (1988a) Distribution of Clavibacter xyli subsp. xyli in stalks of sugarcane cultivars differing in resistance to ratoon stunting disease. Plant Disease 72, 443–448. Liefsonia xyli-like bacteria in grasses 151 Davis MJ, Dean JL, Harrison NA (1988b) Quantitative variability of Clavibacter xyli subsp. xyli populations in sugarcane cultivars differing in resistance to ratoon stunting disease. Phytopathology 78, 462–468. Davis MJ, Lawson RH, Gillaspie AG, Harris RW (1983) Properties and relationships of two xylem-limited bacteria and a mycoplasmalike organism infecting bermuda grass. Phytopathology 73, 341–346. Davis MJ, Gillaspie AG, Vidaver AK, Harris RW (1984) Clavibacter : a new genus containing phytopathogenic coryneform bacteria, including Clavibacter xyli subsp. xyli sp. nov. and Clavibacter xyli subsp. cynodontis subsp. nov., pathogens that cause ratoon stunting disease of sugarcane and bermudagrass stunting disease. International Journal of Systematic Bacteriology 34, 107–117. Dominiak BC, Sinnamon LR, Jones CD, Taylor PWJ (1992) RSD control in Bingera mill area and problems encountered. Proceedings of Australian Sugar Cane Technologists pp. 37–42. Evtushenko LI, Dorofeeva LV, Subbotin SA, Cole JR, Tiedje JM (2000) Leifsonia poae gen. nov., sp. nov., isolated from nematode galls on Poa annua, and reclassification of 'Corynebacterium aquaticum' (Leifson 1962) as Leifsonia aquatica (ex Leifson 1962) gen. nov., nom. rev., comb. nov. and Clavibacter xyli (Davis et al. 1984) with two subspecies as Leifsonia xyli gen. nov., comb. nov. International Journal of Systematic and Evolutionary Microbiology 50, 371–380. Fegan M, Croft BJ, Teakle DS, Hayward AC, Smith GR. (1998) Sensitive and specific detection of Clavibacter xyli subsp. xyli, causal agent of ratoon stunting disease of sugarcane, with a polymerase chain reaction-based assay. Plant Pathology 47, 495–504. Gillaspie AG, Teakle DS (1989) Ratoon stunting disease. In ‘Diseases of sugarcane’, Vol 1. (Eds C. Ricaud, B.T. Egan, A.G. Gillaspie and C.G. Hughes) pp. 59–80. (Elsevier: Amsterdam) Hughes CG, Steindl DRL (1955) Ratoon stunting disease of sugarcane. Queensland Bureau of Sugar Experiment Stations Technical Communication No. 2. Koike H, Gillaspie AG, Benda TA (1982) Cane yield response to ratoon stunting disease. International Sugar Journal 84,131–133. Liao CH, Chen TA (1981) Isolation, culture and pathogenicity to sudan grass of a corynebacterium associated with ratoon stunting of sugarcane and with bermuda grass. Phytopathology 71, 1303–1306. Matsuoka S (1971) Elephant grass, an indicator plant for ratoon stunting virus of sugarcane. FAO Plant Protection Bulletin 19, 110–115. Riley IT, Ophel KM (1992) Clavibacter toxicus sp. nov., the bacterium responsible for annual ryegrass toxicity in Australia. International Journal of Systematic Bacteriology 42, 64–68. Steindl DRL (1957) Host range of the sugarcane ratoon stunting disease virus. The Journal of the Australian Institute of Agricultural Science 23, 238. Steindl DRL (1961) Ratoon stunting disease. In ‘Sugarcane diseases of the World’. (Eds J.P. Martin, E.V. Abbott, and C.G. Hughes) pp. 433–459. (Elsevier, Amsterdam) Steindl DRL, Teakle DS (1974) Recent developments in the identification of ratoon stunting disease. Proceedings of the Queensland Society of Sugar Cane Technologists 41, 101–104. Taylor PWJ, Ryan CC, Birch, RG (1988) Harvester transmission of leaf scald and ratoon stunting disease. Sugar Cane No. 4, pp. 11–14. Received 12 January 2000, accepted 17 January 2001 http://www.publish.csiro.au/journals/app