DETECTION, ISOLATION, AND PRELIMINARY CHARACTERIZATION OF BACTERIA CONTAMINATING PLANT TISSUE CULTURES

In order to limit the contamination problem in plant tissue cultures experiments on selection of media suitable for detection and isolation of bacteria contaminating plant tissue explants, and preliminary characterization of isolates were made. In the first experiment aiming at detection of bacteria in plant explants four strains representing genera most often occurring at our survey of plant tissue cultures, and earlier isolated and identified (Bacillus, Methylobacterium, Pseudomonas and Xanthomonas) were streaked on five bacteriological media (NA, King B, K, R2A and 523) and on the medium used for plant culture initiation – 1⁄2 MS with milk albumin (IM). All strains grew on all media but on K and IM at the slowest rate and on 523 medium at the fastest. The IM medium proved to be useful for immediate bacteria detection at the initial stage of culture. In the second experiment, aiming at characterization of isolates on the basis of colony growth and morphology 14 strains (Agrobacterium, Bacillus, Curtobacterium, Flavobacterium, Lactobacillus, Methylobacterium – 2 strains Mycobacterium, Paenibacillus, Plantibacterium, Pseudomonas, Stenotrophomonas, Xanthomonas, and species Serratia marcescens) were streaked on five microbiological media: KB, NBY, YDC, YNA and YPGA. All strains grew on all those media but at different rates. The only exception was the strain of Lactobacillus spp., which did not grow on King B medium. This medium allowed the detection of such characteristic traits as fluorescence (Pseudomonas) and secretion of inclusions (Stenotrophomonas). The third experiment was focussed on assessment of the sensitivity of detection of specific bacteria in pure cultures and in plant tissue cultures using standard PCR and BIO-PCR techniques with genus specific primers and 2 methods of DNA isolation. Results showed that the use of Genomic Mini kit enabled an increase of the sensitivity by 100 times as compared to extraction of DNA by boiling. Moreover, the application of BIO-PCR increased sensitivity of detection from 10 to 10 times over the standard PCR. If looking for unknown cultivable bacteria more effective detection seems to be use of microbiological method enabling detection on bacteriological media single cells in the fragments of explants or in wash liquids, in which fragmented explants were shaken.


INTRODUCTION
Bacterial contaminations are a serious problem in plant in vitro cultures, both in commercial plant micropropagation, by making difficult culture initiation, reducing efficiency of multiplication and rooting of shoots, as well as in research laboratories, where contamination can be the causal agent of false results in physiological experiments [1,2,3].The diversity and abundance of genera and species of exo-and endobiotic bacteria accompanying donor plants [4,5] is a major challenge in the sterilization of initial explants, a quick detection of bacteria in the first in vitro passages and a minimization of their adverse effect on shoot multiplication and rooting efficiency.In most cases, bacteria are introduced to the cultures together with initial explants.
In practice, initial explants are only surface sterilized, and thus internally living microorganisms are introduced to in vitro cultures.If symptoms of bacteria colonizing plant tissues appeared within a short time, the contaminated explants should be immediately removed.In case when bacterial growth is very slow or temporarily retarded in plant culture conditions, they remain in a cryptic state and may appear only when the culture conditions will drastically change, for example after delayed subculture, increase of temperature, change of medium composition or due to other factors [6].It is therefore important to use appropriate methods for quick bacteria detection and removal of contaminated cultures before microorganisms' spreading.Our large survey of plant tissue cultures originating from eight laboratories in Poland resulted in obtaining 104 isolates of bacteria, which were assigned by phenotypic tests and 16S rDNA fragments sequencing to 29 genera (data not published).
The aim of present studies was to determine which media are suitable in detecting bacteria inhabiting plant explants.The next task was to assess usefulness of selected media for preliminary characterization of isolated bacteria.The third task was to find a way to identify taxa with genus specific primers in possible minimal population of bacteria, using two PCR techniques and two methods of DNA isolation.

MATERIALS AND METHODS
In the present study, isolates obtained from different plant tissue cultures, identified to the genera , were used.
Evaluation of media enabling bacteria characterization.In the experiment aimed at selection of medium enabling preliminary characterization of isolated bacteria, 14 strains, representing the genera most often found in surveyed plant tissue cultures: Agrobacterium, Plantibacterium, and Xanthomonas (each from Sambucus nigra), Bacillus (from Phalenopsis), Flavobacterium (from Daucus carota), Lactobacillus (from Hosta), Methylobacterium (2 strains), Curtobacterium, Mycobacterium, Paenibacillus, Pseudomonas, Stenotrophomonas (each from Rubus ideaus) and Serratia marcescens (from Alocasia), were streaked on the following media: King B (KB), Nutrient Broth-Yeast Extract (NBY), Yeast Dextrose Chalk Agar (YDC), Nutrient Agar supplemented with 0.5% yeast extract (YNA), and Yeast Peptone Glucose Agar (YPGA) [12,13].Observations of colony morphology were made every 24 h for 7 days.Time of colonies appearance, their size, and morphology were recorded.The KB, NA, NBY, YDC, YNA and YPGA are known as universal bacteriological media but R2A and 523 (in both low amount of organic nitrogen, 523 contains sucrose) are used for isolation of bacteria from in vitro plant cultures.K medium has been recommended for isolation of Methylobacteriaceae (without organic nitrogen and with methanol as the only source of carbon) and IM is usually used as the initial medium for introduction of plant explants to in vitro cultures.
Sensitivity of bacteria detection in pure cultures using PCR with genus specific primers.This model experiment aimed at possibility of detection the minimal number of bacteria in a sample.Bacterial suspensions of three strains were prepared by washing off bacterial colonies with sterile water.Strain 87 (Bacillus) was cultivated on NA + 1% sucrose, isolate E (Pseudomonas) and isolate 81b (Methylobacterium) on KB medium.The initial concentration of suspension for strain 87 was 2x10 7 , for 81b -2x10 6 , and for strain E -5x10 7 cfu ml -1 .Each of them was then serially tenfold diluted to final concentration 10 0 cfu ml -1 .From each dilution, DNA was isolated by two methods.1/ One ml of each bacterial suspension was boiled for 10 min, cooled on ice and centrifuged at 14 000 rpm; one μl of supernatant was used in standard PCR reactions.2/ One ml of each bacterial suspension was centrifuged at 14 000 rpm, re-suspended in 100 μl TE buffer and DNA was isolated using Genomic Mini kit for DNA extraction according to manufacturer's instruction (A&A Biotechnology).One μl of DNA (after isolation from each bacterial dilution) was used in standard PCR reactions.
Additionally, to compare detection sensitivity of standard PCR with BIO-PCR (pre-incubation followed by amplification) 100 μl of each dilution of bacterial suspension of each isolate was streaked on appropriate medium, as described above.The plates were incubated at 26 o C for up to 5 days.At respective time, single colonies were counted and then washed off with 3 ml of sterile water; 2 ml of the suspension were used for DNA isolation with the same two methods as described above.
To assess the sensitivity of bacteria detection using standard PCR and BIO-PCR and two different DNA isolation methods the following primer pairs were used: for Bacillus -Bac-200 and Bac-470 [14], for Methylobacterium -Met2F and Met2R [15] and for Pseudomonas -primers Ps-for and Ps-rev [16].Amplifications were performed in 15 μl of a reaction solution containing: 2 μl of DNA, 0.4 U polymerase GoTaq DNA (Promega Corp. USA), 1x buffer GoTaq, 0.2 mM of each DNTP and 1 μM of primers.All amplification reactions were conducted according to the recommendation of the authors with only modification for Bacillus (lowering of annealing temperature by 8 o C).PCRs were carried out in T3000 Biometra thermal cycler.PCR products were separated on a 1.5% agarose gel with 0.5 TBE buffer and visualized by staining with ethidium bromide (0.5 mg l -1 ).
Sensitivity of bacteria detection in in vitro plant explants artificially contaminated with bacteria.The shoots for the experiment were harvested from the stock cultures "free of cultivable bacteria".The stocks were initiated from microshoots indexed two times, in two subsequent subcultures, for the presence of contamination by placing the basal parts of shoots on two bacteriological media: Nutrient Agar and 523 medium.The shoots, from which bacteria did not grow in both subcultures, were considered as free of cultivable bacteria.The single microshoots were transferred to jars containing perlite saturated with liquid rooting medium appropriate for each plant species and 100 μl of 24 h bacterial suspension was injected in the centre of a jar immediately after shoot transfer.Microshoots of Gerbera x hybrida 'Kormoran' were contaminated with strain 87 of Bacillus spp., Anthurium x andreanum 'Bolero'contaminated with strain E of Pseudomonas putida and Sambucus nigra contaminated with strain of 81b of Methylobacterium lusitanum.After 4 weeks, the roots were cut off, and the shoots transferred to propagation agar medium appropriate for each plant species.Explants showing symptoms of bacteria presence in the form of visible leakage into agar medium or halo around shoot base were aseptically divided into three segments: I -at the base, II -middle part and, III -top.Bulk samples of 100 mg of fragmented tissues from several explants of the same segment of each plant species were shaken in 3 ml of PBS buffer (0.27% Na 2 HPO 4 , 0.04% NaH 2 PO 4 , 0.8% NaCl) at 26 o C for 1h.Two ml of obtained liquids were used for isolation of DNA by boiling or using Genomic Mini kit as described above.
Additionally, to compare detection sensitivity by standard PCR and BIO-PCR (bacteria pre-incubation followed by DNA isolation and amplification) 100 μl of liquid obtained from each plant segment was placed on appropriate medium for each isolate, as described above.After incubation at 26 o C for up to 5 days, the colonies were counted and then washed off with 3 ml of sterile water; 2 ml of the resulting suspension were used for DNA isolation using the same two methods as described above.Detection of bacterial DNA in plant material was conducted using PCR with the genus specific primers and according to conditions described above.
Bacteria detection in apparently healthy in vitro plant explants.Total DNA was individually isolated by Qiagen Plant Mini kit from bulk samples of in vitro shoots of Chrysanthemum x hybrida 'Ludo', Kalanchoë blossfeldiana 'Debbie', Gerbera x hybrida 'Kormoran' and Sorbus aucuparia, which never have shown contamination symptoms at visual inspections.DNA was used for detection of possible bacterial contamination using standard PCR.Universal primers fD1/rP2 for 16S rDNA of bacteria [17] and genus specific primers for Bacillus, Methylobacterium and Pseudomonas (as listed above) were used.The following PCR conditions for primers fD1/rP2 were applied: 94 o C 4 min., 35 cycles: 45 s at 95 o C, 45 s at 55 o C, 90 s 72 o C and 10 min at 72 o C. 25 μl of reaction mixture contained: 25 ng DNA, 0.5 U Tag polymerase (Fermentas), 0.5 μM of each primer, 50 μM each of dNTP and 1.5 mM MgCl 2 .For genus specific primers, the same PCR conditions were applied, as described above.

RESULTS
Growth and colony morphology of bacteria on different media.,The most efficient among six media tested in terms of rate of bacteria growth was medium 523.The IM medium proved to be useful for growth three of the four bacterial strains, although at the slowest rate (colonies were visible 3 -5 days later in comparison to other media used).However, single colonies of Bacillus spp.have not been observed on this medium within 7 days (Table 1; Figs 1-4).
All the five media used allowed growth of 14 bacterial strains representing different genera.NBY was the most favourable medium for growth of Bacillus, Curtobacterium, Lactobacillus, Methylobacterium, Plantibacterium, Xanthomonas and Serratia marcescens strains.The King B medium appeared to be most appropriate for Stenotrophomonas because of secretion of characteristic inclusions and for Pseudomonas because of fluorescent pigment production (Table 2).Lactobacillus growth was possible under mineral oil on four media but not on KB.
Sensitivity of bacteria detection/identification in pure cultures using PCR with genus-specific primers.The sensitivity depended on DNA isolation method (boiling or Genomic Mini kit) and applying pre-incubation of bacteria on bacteriological medium (BIO-PCR).The use of Genomic Mini kit for DNA extraction let to increase sensitivity 100 times for Methylobacterium and 1000 times for Bacillus, as compared to extraction of DNA by boiling.Only the sensitivity of detection/identification of Pseudomonas isolate did not depend on the method of DNA isolation used.Application of BIO-PCR increased the sensitivity by 10 2 to 10 5 times over the standard PCR, independently on the method of DNA isolation used (Table 3, Figs 5-7).
Detection of bacteria using PCR with genus--specific primers in plant explants artificially contaminated.The analysis of bacteria presence in all three segments of Anthurium x andreanum 'Bolero', Gerbera x hybrida 'Kormoran' and Sambucus nigra explants, intentionally contaminated with bacteria, showed decrease in their number in the shoots going from bottom to top (Table 4).The highest differences between segments were found in Bacillus spp./Anthurium complex where in the top segments 125 times less bacteria was found than in a bottom part.However, in case of Pseudomonas putida/Gerbera there was only 8 times less bacteria in the top segments and in Methylobacterium lusitanum/ Sambucus only 3 times less than in a bottom segments, respectively.
Independently of the method of DNA extraction -Genomic Mini kit or boiling, plating of liquids after shaking of tissues in PBS buffer on bacteriological media (BIO-PCR) allowed detection/identification of one cell of Pseudomonas and Methylobacterium isolates.
Moreover, applying pre-incubation of bacteria increased the sensitivity by 100 to 1000 times in comparison to standard PCR (data not shown).However, in case of Bacillus strain the detection was possible only with pre--incubation.
Bacteria detection in apparently healthy in vitro plant explants using PCR with genus-specific and universal bacterial primers.None of specific DNA fragments of Pseudomonas, Bacillus, and Methylobacterium were detected in total DNA isolated using the Qiagen Plant Mini Kit from explants of Kalanchoë, Gerbera, Sorbus and Chrysanthemum that did not exhibit bacterial contamination.On the other hand, PCR analyses of plant DNA with primers fD1/rP2, universal for bacterial 16S rDNA, revealed the presence of bacterial DNA in each sample of plant explants.

DISCUSSION
Our survey of bacteria contaminated plant tissue cultures originating from eight laboratories confirmed earlier reports about the presence of multitude of bacteria associated with plant tissue cultures [4,5].The range of bacterial taxa varies depending on the geographical zone, plant species, conditions in which donor plants grew and on the source of contamination [18].Nevertheless, some bacteria, such as those belonging to genera: Bacillus, Pseudomonas, Methylobacterium, Corynebacterium, Staphylococcus, and Agrobacterium are ubiquitous.
Although disinfection of donor plant materials is always of special importance in the process of mi-cropropagation, usually contaminations are introduced with the initial explants.Bacteria live in the plant vessels, intercellular spaces but also inside plant cells [19], which is the reason of unsuccessfulness of surface sterilization.To overcome this problem initiation of in vitro cultures from as small explants (very short shoot tips or meristems) as possible and detection of cultivable bacteria presence in these small explants as early as possible, is recommended.For this purpose, a shaking of fragmented bases of explants in sterile water or PBS buffer and plating of obtained liquid on microbiological medium, as it was performed in our study, is advisable.It is most effective when bacteria can be detected already on/ in the initial medium, as IM, typical medium for plant culture initiation, with the addition of 0.025% of milk albumin, which, in our experiment, enabled detection at least strains of Methylobacterium, Pseudomonas and Xanthomonas.On this medium, colonies of Pseudomonas were visible after 3 days, Xanthomonas after 4 days, and Methylobacterium after 5 days.Only in case of Bacillus strain single colonies were not observed after streaking on Initiation Medium until 7 th day but weak growth of bacteria only at first lines of streaking was visible.All microbiological media used enabled to observe growth of bacteria but for not all of them single colonies were present.K medium, recommended for Methylobacteriaceae [7] was not useful for our Methylobacterium strain.However, we can underline usefulness of the 523 medium [9], which stimulated growth of colonies of all strains used more than KB or NA media.Nevertheless, a lack of bacteria growth on the bacteriological media does not guarantee the absence of any bacteria in the su-rveyed tissue because some bacteria can be uncultivable permanently or temporarily.For example, Xanthomonas campestris pv.dieffenbachiae (anthurium pathogen) could persist in the latent stage in in vitro shoot explants remaining undetected for one year [20].The bacteriological media used in our experiment were not relevant to complete characterization and identification of bacteria.The exception was King B, which enables visualization of two features -fluorescent pigment characteristic to most Pseudomonas species and secretion of inclusions characteristic to Stenotrophomonas.Although the color of bacterial colonies as the only feature, can be not used for bacteria identification, nevertheless pink color of colonies and their slow growth may suggest the presence of one of the Methylobacterium, claret color and very rapid growth -Serratia marcescens, and yellow color -Xanthomonas or Flavobacterium.* 24 h, 1-2 mm in diameter, light beige, flat, irregular shaped in the form of elongated tear drops (Fig. 1a) 24 h, 1-2 mm, gray-white, flat ,circular, irregular shaped (Fig. 1b) 24 h, 1 mm, graywhite, elongated, flat, matt, circular, irregular shaped slightly jagged edges (Fig. 1c) 24 h, 2-3 mm creamy, circular shaped with a distinctive shiny hallo around convex, flat rough (Fig. 1d) 120 h lack of single colonies, growth only in the first line of a smear (Fig. 1e and f)

Methylobacterium lusitanum (81b from Sambucus nigra)
96 h 1 mm, bright pink, regular in shape, round and slightly convex with a full edge (Fig. 2a) 96 h, 0.5 mm, small bright pink, regular in shape round and slightly convex with a full edge (Fig. 2b) 96 h, 0.5-1 mm, bright pink, regular in shape, round and slightly convex with a full edge (Fig. 2c) 96 h, 0.5-1 mm, pink, regular in shape, round and slightly convex with a full edge (Fig. 2d) 120 h lack of single colonies growth only in the first line of a smear (Fig. 2e) 120 h, 0.2 mm pink, regular in shape, round and slightly convex (Fig. 2f)

Xanthomonas spp. (78 from Sambucus nigra)
48 h, 0.5-1 mm yellow, round, in shape shiny, little convex (Fig. 4a) 48 h, 0.5-1 mm, small, light yellow, shiny, regular in shape (Fig. 4b) 48 h, 1-2 mm, single, yellow, shiny, circular, regular in shape, little convex (Fig. 4c) 48 h, 2-3 mm bright yellow, shiny, little convex, regular in shape (Fig. 4d) 96 h, 0.5 mm, creamy on K and white on MS medium, flat, regular in shape (Fig. 4e and f) *Given is time (h) and diameter of colonies (mm) when they were recognizable Pseudomonas putida (E)/Gerbera hybrida 1.33x10 5 1x10 5  1.67x10 4   Methylobacterium lusitanum (81b)/Sambucus nigra 1x10 5 0.66x10 5 0.33x10 5  (ThermoScientific, Life Science, Lithuania), 1 -3x10 6 , 2 -3x10 5 , 3 -3x10 4 , 4 -3x10 3 , 5 -3x10 2 , 6 -3x10 1 , 7 -3x10 0 , 8 -3x10 6 , 9 -3x10 5 , 10 -3x10 4 , 11 -3x10 3 , 12 -3x10 2 , 13 -3x10 1 , 14 -3x10 0 , 15 -negative control The highly sensitive method for detection of bacteria in contaminated plant tissue cultures is very important especially at the initial stage when only a small number of bacteria can be present without visible symptoms both on/in explants and in/on media.Results of our study can suggest that detection of cultivable bacteria in washing liquid or plant fragment is more effective using microbiological method (streaking on bacteriological medium) than method based on PCR.A molecular detection of bacteria of known genus using PCR with the genus-specific primers confirmed that the BIO-PCR increased detection level of Bacillus spp.by 10 2 -10 5 times, Methylobacterium by 10 2 -10 4 times and Pseudomonas by 10 2 times, depending on the DNA isolation method.Wang et al. [21] obtained similar results in detection of Xanthomonas albilineans, the pathogen of sugar cane, where BIO--PCR combined with the use of the half selective m--XAM medium was more sensitive than ELISA, DIA and PCR.BIO-PCR technique was most sensitive and reliable for Xanthomonas axonopodis pv.phaseoli and P. syringe pv.phaseolica detection [22,23].P u ł a ws k a and S o b i c z e w s k i [24] recommended this technique in detection of tumorigenic strains of Agrobacterium tumefaciens in soil and Erwinia amylovora in infected plants [25,26].Sensitivity of detection in the above experiments was from 1 to 20 cells per 1 g or 1 ml of a sample.These results suggest that instead of detection specific bacterial DNA, more effective are microbiological methods, including of plating of tissue washings on bacteriological media, because in the molecular method, detectability begins at app. 15 cells whereas on bacteriological medium, it is possible to detect a single bacterium.Moreover, searching for bacterial DNA could give false positive results by amplifying also DNA of bacteria, which do not survive disinfection.
In case of searching for specific bacteria using genus specific DNA primers in specimens of plant body, method of DNA extraction is important.The use Genomic Mini kit enables 10 times more sensitive bacteria detection than the extraction by boiling.However, in our study in case of Bacillus isolate the detection was possible only after pre-incubation, which could be related to presence of plant DNA polymerase inhibitors in Anthurium tissues and in shaking washings.
The use of primers universal for amplification of bacterial rDNA, recommended by W e i s b u r g et al. [17], which were applied in this work and in others [27] should be probably not advisable in cases of detection of bacterial DNA in a mixture with plant DNA because the primers amplify also fragments of chloroplast DNA [28].In addition, this method detects all bacterial DNA, also those belonging to uncultivable bacteria and bacteriosomes.To exclude chloroplast DNA amplification other primers, e.g. according to C h e l i u s and T r i p l e t t [29] or B u l g a r i et al. [30], should be used to ensure that obtained PCR products are derived from bacterial templates.On the other hand, there is a question if a plant (in vitro explant) tissue could exist as free of bacteria [1].

CONCLUSIONS
1.The bacteriological media -King B, NBY, YDC, YNA, and YPGA enabled growth and morphological characterization of strains of 13 genera occurring as contaminants in plant tissue cultures, with the exception that Lactobacillus strain did not grow on KB medium, even under mineral oil.Only King B medium enabled the appearance of the characteristic traits as fluorescence of Pseudomonas or secretion of characteristic inclusions by Stenotrophomonas strains.Other strains studied could be not identified even in terms of genus.Pink color of slow growing colonies may eventually imply affiliation to pink Methylobacterium, claret color of fast growing ones to Serratia marcescens and yellow color to Xanthomonas or Flavobacterium.2. The most suitable for visualisation of bacteria present in initial explants or in washings from in vitro explants at the initiation stage was 523 medium.
The IM medium (1/2 MS salts with milk albumin 0.025%) used routinely for initiation of plant tissue cultures can be useful for isolation of bacteria directly at culture initiation.3. Medium K, recommended for Methylobacteriaceae, proved to be useless for the growth of Methylobacterium strain used in this study.4. The BIO-PCR with genus-specific primers can be recommended for detecting/identifying single bacterial cells, but the procedure is extended by 2-5 days.5.At a low bacteria number population, more recommendable for detection of cultivable bacteria is the use of microbiological (placing of explants fragments or washing liquids on bacteriological medium) than molecular method based on DNA markers.
Handling Editor: Elżbieta Weryszko-Chmielewska This is an Open Access digital version of the article distributed under the terms of the Creative Commons Attribution 3.0 License (creativecommons.org/licenses/by/3.0/),which permits redistribution, commercial and non-commercial, provided that the article is properly cited.©The Author(s) 2013 Published by Polish Botanical Society

Table 3
The sensitivity of bacterial DNA detection by PCR depending on technique of DNA isolation and application of biological amplification (BIO-PCR) Bacteria Method of DNA isolation Sensitivity of detection (number of bacterial cells) number of bacterial cells detected in different segments per 1 mg of multishoot tissue Bacteria/plant Segment of multishoot I (at the base) II (middle part) III (top) Bacillus spp.(87)/Anthurium andreanum 1

Table 1
Morphology of bacterial colonies on different media