Introduction

In the last decade, many transgenic plants have been obtained by using Agrobacterium-mediated transformation for crop improvement and genetic research. For Agrobacterium-mediated transformation, suppression and elimination of agrobacteria are necessary to minimize the risk of interfering growth and regeneration of the potentially transformed plant tissues, and to reduce the hazard of releasing genetically engineered agrobacteria into the environment. β-Lactam antibiotics such as carbenicillin and cefotaxime have commonly been utilized for this purpose, because this class of antibiotics lyse and kill the bacteria by specifically interfering with the biosynthesis of the prokaryotic peptidoglycan component of the bacterial cell wall by binding to penicillin-binding proteins (PBPs) (Ghuysen 1997; Demain and Elander 1999; Asbel and Levison 2000). On the other hand, stimulating effects on plant cell growth, organogenesis and embryogenesis have also been reported for both carbenicillin and cefotaxime, as listed by Nauerby et al. (1997), as well as for other classes of antibiotics (Pollock et al. 1983; Young et al. 1984; Okkels and Pedersen 1988).

Many β-lactam antibiotics, which are classified into penicillin, cephalosporin, cephamycin, oxacephem, monobactam and carbapenem, are currently being developed to enhance and expand antibacterial activity (Demain and Elander 1999). Recently, several researchers successfully introduced a penicillin derivative, ticarcillin, alone or in combination with a β-lactamase inhibitor, clavulanic acid (Timentin; SmithKline Beecham, Brentford, UK) into the Agrobacterium-mediated transformation procedure for several plant species (e.g., Schroeder et al. 1993; Frary and Earle 1996; Aida et al. 1999). Other β-lactam antibiotics have also been applied to plant genetic transformation, e.g., Augmentin (SmithKline Beecham) (Vergauwe et al. 1996), cefoxitin (mefoxin) (Hammerschlag et al. 1997), cephaloridin (Enomoto et al. 1990) and sulbenicillin (Zhang et al. 2000). Although it is important to understand both the antibacterial activity against Agrobacterium tumefaciens strains and the phytotoxicity of these β-lactam antibiotics in Agrobacterium-mediated transformation, only a few such detailed characterizations have been published (Okkels and Pedersen 1988; Shackelford and Chlan 1996).

In a previous study, we evaluated in vitro antibacterial activities of 12 β-lactams against two common strains of A. tumefaciens, LBA4404 and EHA101, by two parameters—minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) (Ogawa and Mii 2004). In this study, we evaluated in planta antibacterial activities of these β-lactam antibiotics against A. tumefaciens LBA4404 and EHA101, as well as the side effects of the antibiotics on shoot regeneration from tobacco leaf explants. We also discuss the possibilities for application of various β-lactam antibiotics in Agrobacterium-mediated transformation.

Materials and methods

β-Lactam antibiotics

The 12 β-lactam antibiotics used in the present study are listed in Table 1. The quantitative bactericidal activities (MBCs) of these β-lactams against agrobacterial strains reported previously (Ogawa and Mii 2004) are also listed in the table.

Table 1 β-Lactam antibiotics used in the present study. MBC Minimal bactericidal concentration
Full size table

Bacterial strains

Disarmed A. tumefaciens strains LBA4404 (Hoekema et al. 1983) and EHA101 (Hood et al. 1986) were used in the present study without an introduced binary plasmid. Strain LBA4404 has a chromosomal background of an octopine strain Ach5 and carries disarmed helper Ti plasmid, pAL4404. Strain EHA101 has a chromosomal background of a nopaline strain C58 and carries a disarmed helper plasmid pEHA101. Agrobacteria were maintained at −80°C as glycerol stocks, and were cultured routinely on LB (10 g l−1 tryptone, 5 g l−1 yeast extract, 5 g l−1 NaCl, pH 7.0) agar plate at 28°C.

Inoculation and regeneration of tobacco leaf explants with A. tumefaciens

Bacterial cells were harvested by centrifugation from an overnight culture of agrobacteria, and resuspended in liquid MS medium (Murashige and Skoog 1962) to A600 =0.2 [ca. 5×108 colony-forming unit (cfu) ml−1]. Leaf explants (ca. 7 mm square segments) prepared from in vitro plantlets of tobacco (Nicotiana tabacum L.) ‘Wisconsin 38’ were submerged in inoculum solution for 5 min; after blotting excess inoculum on a sterile paper towel, explants were then placed onto regeneration medium [MS basal medium containing 0.1 mg l−1 1-naphthaleneacetic acid (NAA), 1.0 mg l−1 6-benzylaminopurine (BAP), 30 g l−1 sucrose, 2.5 mM 2-(N-morpholino)-ethanesulfonic acid (MES), pH 5.8, solidified with 8 g l−1 agar]. After 3 days of cocultivation, explants were transferred to regeneration medium supplemented with various concentrations of β-lactams and without selective antibiotics for transformed cells. For EHA101, 2-fold serially increased concentrations were made for each β-lactam and applied at 1–32× MBC (up to 3,200 mg l−1), while moxalactam and meropenem were applied at 1–128× MBC. For LBA4404, 2-fold serial concentrations of cefotaxime (1–32× MBC), cefbuperazone (1–32× MBC), moxalactam (1–8× MBC) and meropenem (1–128× MBC) were applied. Explants were cultured under the 24 h photoperiod at 30–50 μmol m−1 s−1 and at 25°C, and subcultured every 2 weeks onto fresh β-lactam-supplemented regeneration medium. Explants that showed overgrowth of agrobacteria were removed from the plates.

Antibacterial activities of 12 β-lactams in planta

Six weeks after cocultivation, explants were homogenized with 2 ml phosphate-buffered saline (PBS; 0.43 g gl−1 KH2PO4, 1.48 g l−1 Na2HPO4, 7.2 g l−1 NaCl, pH 7.2) using motor and pestle after weighing and counting the number of shoots. Homogenates were appropriately diluted with PBS, and then 100 μl portions of homogenates were spread over LB agar plates. The plates were incubated at 28°C for 3 days or more for colony formation. Explants cocultivated for 3 days (just before antibiotic treatment) were also examined in order to count the number of viable agrobacteria. The number of persisting agrobacteria was defined as cfu per gram fresh weight (FW) of explant.

Effect of β-lactams on shoot regeneration from tobacco leaf tissues

Shoot regeneration from tobacco explants with or without β-lactam treatment was observed 6 weeks after cocultivation. Phytotoxicity was defined as shoot formation rate, which was determined as the percentage of the number of regenerated shoots per explant on antibiotic-treated explants with or without inoculation with agrobacteria, against that on control explants.

Data analysis

In all experiments, five (for EHA101) or six (for LBA4404) explants per treatment were placed individually onto 90×20 mm plates containing 25 ml medium. Each experiment was repeated three times. Significant differences among the data were obtained from one-factor ANOVA followed by Fisher’s PLSD. Arcsin transformation and logarithmic transformation were applied for the data in percentile value and in exponential value, respectively, before statistical analyses, although the data before transformation were represented.

Results and discussion

Suppression of agrobacterial overgrowth on explants inoculated with EHA101 by β-lactams

When tobacco leaf explants were inoculated with A. tumefaciens EHA101 and cultured on media containing β-lactam antibiotics at or above the MBC, overgrowth of EHA101 on the explants and surrounding medium was observed with six β-lactams, i.e., cefotaxime, cefbuperazone, cefminox, moxalactam, flomoxef and meropenem, typically at 1–2× MBC of each antibiotic (Fig. 1). These results are consistent with those obtained by Falkiner (1988, 1990) and Barrett and Cassells (1994), who recommended use of 2–4× MBC of antibiotics to prevent bacterial growth in plant tissue culture. Although meropenem is the most active antibiotic against both LBA4404 and EHA101 in vitro (Table 1; Ogawa and Mii 2004), it allowed agrobacterial overgrowth in 100% of explants inoculated at 1–8× MBC (0.20–1.56 mg l−1). However, meropenem could suppress overgrowth of EHA101 at 6.25 mg l−1 (32× MBC), which is the lowest effective concentration of those evaluated here. The antibacterial activity of many of the β-lactams was slightly affected by factors such as inoculum size, culture medium, and pH of the medium (Fukasawa et al. 1992). Thus, the loss of in planta antibacterial activity of meropenum at relatively high MBC might be due to an interaction between this compound and plant tissues, such as β-lactam permeability or uptake.

Fig. 1
figure 1

Frequencies of explants showing agrobacterial overgrowth. Treatments with the same letter are not significantly different according to Fisher’s PLSD. Data for the other six β-lactams are not shown because overgrowth was not observed on explants treated with these antibiotics. ND Not determined

Full size image

In planta bactericidal activities of β-lactams against EHA101

Under the present inoculation conditions, ca. 2.5×109 cfu gFW−1 viable EHA101 cells were detected in the explant after 3 days of cocultivation. Persistence of agrobacteria (up to 2.8×109 cfu gFW−1) was observed in explants cultured on β-lactam-containing agar plates after 6 weeks of culture, even when no agrobacterial overgrowth was observed on explants visually (Fig. 2). Although persisting EHA101 cells in planta decreased to 106–108 cfu gFW−1 over the wide range of β-lactam concentrations tested, bactericidal activity of β-lactams was not complete and was only weakly correlated with antibiotic concentration. Obvious in planta bactericidal activity was observed only for ceftazidime and cefbuperazone, which decreased the agrobacterial survival rate to below 106 cfu gFW−1 at 2× MBC and 4 × MBC, respectively. The poor correlation between concentration and in planta antibacterial activities against EHA101 observed for other antibiotics might be affected by multiple factors, e.g., the physiological state of agrobacteria persisting in plant tissue known as “latent infection” (Hayward 1974; Cooke et al. 1992), the presence of high numbers of agrobacterial cells in inoculated leaf tissue, the stability of the β-lactams at the temperature and light regimes adopted for culture, and the efficiency of uptake of β-lactams from gelled medium to plant tissue.

Fig. 2
figure 2

In planta bactericidal activities of β-lactams against Agrobacterium tumefaciens EHA101. Data represent the number of surviving agrobacteria. Significant differences were detected by ANOVA (P=4.47×10−18). SD and significance of each data point are not shown to minimize complexity of the figure (average SD ±6.38×107 cfu gFW−1)

Full size image

Survival of agrobacteria was observed in both leaf explants and regenerated shoots after 6 weeks of β-lactam treatment. The persistence of agrobacteria in transgenic plants has been reported when carbenicillin, ticarcillin or cefotaxime were used at 125–500 mg l−1 (van der Hoeven et al. 1991; Matzk et al. 1996; Barrett et al. 1997). In addition, the bactericidal activities of β-lactams against A. tumefaciens strains as well as clinical pathogens (Craig and Ebert 1991) are known to exhibit “concentration-independent killing” or “time-dependent killing” (Ogawa and Mii 2004). Therefore, to avoid the risk of releasing genetically engineered Agrobacterium into the environment, β-lactams should be continuously supplied to the culture medium until the elimination of agrobacteria persisting in the plant tissue has been confirmed.

Side effects of β-lactams on plant tissues at concentrations bactericidal to EHA101

At concentrations bactericidal against EHA101, the 12 β-lactams were generally toxic to tobacco tissue and inhibited its organogenesis (Fig. 3), and an inhibitory effect of all β-lactams on shoot regeneration was observed at the concentration that suppressed agrobacterial overgrowth, irrespective of inoculation with EHA101 (Fig. 4). According to the concentration dependency of the inhibitory effects on shoot regeneration, all but two β-lactams could be classified as either: (1) highly phytotoxic drugs (carbenicillin, cefotaxime, ceftazidime, cefbuperazone, cefminox, flomoxef, and aztreonam), or (2) moderately phytotoxic drugs (ticarcillin, moxalactam, and meropenem).

Fig. 3a,b
figure 3

Effect of β-lactams on shoot regeneration from tobacco leaf explants without inoculation with A. tumefaciens EHA101. The data are arranged by a relative concentration (×MBC), and b absolute concentration (mg l−1) of β-lactams. Significant differences were detected by ANOVA (P=3.36×10−21). SD and significance of each data point are not shown to minimize complexity of the figure (average SD ±18.6%)

Full size image
Fig. 4a,b
figure 4

Effect of β-lactams on shoot regeneration from tobacco leaf explants inoculated with A. tumefaciens EHA101. The data are arranged by a relative concentration (×MBC) and b absolute concentration (mg l−1) of β-lactams. Significant differences were detected by ANOVA (P=3.36×10−21). SD and significance of each data point are not shown to minimize complexity of the figure (average SD ±15.8%)

Full size image

Most of the concentration-dependent phytotoxic β-lactams reduced the shoot regeneration rate to below 50% at 4× MBC, and at 8× or more MBC, all these drugs severely inhibited shoot regeneration, leading to shoot formation rates of less than 10% (Figs. 3, 4). The inhibitory effect generally appeared at concentrations over 100 mg l−1 (Figs. 3, 4). Leaf explants that were treated with these β-lactams at inhibitory concentrations showed chlorosis within 2 weeks and finally showed necrosis, although leaf explants treated with carbenicillin formed calli without showing chlorosis.

In the concentration-independent, moderately phytotoxic, group, both moxalactam and meropenem maintained a relatively high shoot formation rate (approximately 60–80%) over a wide range of concentration, i.e., 1–128× MBC (6.25–800 mg l−1 moxalactam and 0.20–25 mg l−1 meropenem). Ticarcillin-treated explants also showed high regeneration capacity at 1–32× MBC (25–800 mg l−1) and had higher shoot formation rates (73–98%) than moxalactam- and meropenem-treated explants.

The remaining two antibiotics, piperacillin and carumonam, maintained a shoot formation rate of 38–81% at the high concentrations of antibiotics tested: 400–3,200 mg l−1 for piperacillin and 800–3,200 mg l−1 for carumonam, respectively. However, it was not possible to classify carumonam and piperacillin into these two groups because concentrations greater than 3,200 mg l−1 were not examined.

It has been reported that β-lactams have both stimulatory and inhibitory effects on plant cell growth and organogenesis (Nauerby et al. 1997). The β-lactams tested in this study showed no apparent stimulatory effect, although callus formation was observed with carbenicillin and three of the concentration-independent, moderately phytotoxic, antibiotics (ticarcillin, moxalactam and meropenem) without any marked delay or decline in shoot regeneration and growth.

In planta antibacterial activities and phytotoxicities of β-lactams in LBA4404

After evaluation of 12 β-lactams on EHA101, similar experiments were performed on LBA4404 using five β-lactams (carbenicillin, cefotaxime, cefbuperazone, moxalactam, and meropenem) that were highly active against this bacterial strain in vitro (Table 1; Ogawa and Mii 2004). Although all five β-lactams controlled overgrowth of LBA4404 on explants inoculated at 1× MBC for the first 2 weeks, they required 4× or higher MBC to suppress overgrowth during 6 weeks of culture (Fig. 5a). An extremely high concentration of cefbuperazone (256× MBC) was required to inhibit agrobacterial overgrowth. While cefotaxime usually suppressed overgrowth of LBA4404 at 32× MBC, it occasionally failed to suppress overgrowth even at 128× MBC. In planta bactericidal activities of these five β-lactams against LBA4404 were low and concentration-independent, as observed against EHA101. Numbers of persisting LBA4404 cells after β-lactam-treatment were 10-fold to 1,000-fold higher (108–1010 cfu gFW−1) than with EHA101 (Fig. 5b). Amongst five β-lactams tested, only meropenem showed similar high in planta bacterial activity against both LBA4404 and EHA101.

Fig. 5a–c
figure 5

In planta antibacterial activities of four β-lactams against A. tumefaciens LBA4404 and phytotoxicities against tobacco leaf explants. a Effect of β-lactams on overgrowth of agrobacteria. Columns followed by the same letter are not significantly different according to Fisher’s PLSD. b Effect of β-lactams on survival of agrobacteria. Significant differences were detected by ANOVA (P=4.39×10−3). SD and significance of each data point are not shown to minimize complexity of the figure (average SD ±1.93×109 cfu gFW−1). c Effect of β-lactams on shoot regeneration from tobacco leaf explants inoculated with or without LBA4404. Significant differences were detected by ANOVA (P=3.41×10−34). SD and significance of each data point are not shown to minimize complexity of the figure (average SD ±14.6%)

Full size image

In the absence of LBA4404, shoot formation rate was 60% or higher (Fig. 5c). As shown in prior experiments, the same high shoot formation rate was also obtained when β-lactams were applied at low phytotoxic concentration ranges (Fig. 3). It is interesting to note that shoot formation rate was reduced by inoculation with LBA4404 (Fig. 5c) but not with EHA101 (Figs. 3, 4) when the explants were exposed to either carbenicillin, cefotaxime or moxalactam at concentrations that completely suppressed agrobacterial overgrowth. This phenomenon might be related to the difference in number of persisting agrobacteria in explants after inoculation with these two bacterial strains as described above. Occasional overgrowth of LBA4404 on explants treated with 128× MBC cefotaxime may be explained by the large number of persisting agrobacteria as well as by the concentration-independent bactericidal activity of this β-lactam.

Conclusions

One of the most important requirements for Agrobacterium-mediated transformation is the selection of suitable antibiotics that do not inhibit growth, organogenesis or embryogenesis of plant tissues at concentrations effective for control of agrobacteria. Based on the results presented here, we conclude that meropenem and moxalactam are the most effective antibiotics for Agrobacterium-mediated plant genetic transformation of tobacco leaf explants. Meropenem exhibited the highest antibacterial activities against EHA101 and LBA4404 while also allowing a high level of shoot formation rate at rates of ≥6.25 mg l−1. Therefore, it could be utilized for Agrobacterium-mediated transformation studies with strains such as LBA4404, EHA101, and probably other C58 derivatives. Moxalactam also appears to be suitable for Agrobacterium-mediated transformation, since it showed effects similar to those of meropenem on both agrobacterial strains and plant tissues. Appropriate concentrations of moxalactam are 25 mg l−1 for EHA101 and 6.25 mg l−1 for LBA4404. Carbenicillin and cefotaxime, which are currently utilized in many laboratories, appear to be less optimal for Agrobacterium-mediated transformation of tobacco leaf explants. However, optimization of the concentration of these β-lactams could improve transformation efficiencies. Based on these results, the optimal concentration of carbenicillin is at least 50 mg l−1 for LBA4404 or EHA101 and, for cefotaxime, 25–200 mg l−1 for LBA4404 and 100–200 mg l−1 for EHA101. Since side effects of antibiotics may vary depending on plant species and explant sources (Nauerby et al. 1997), preliminary experiments would be required for other antibiotic-plant combinations. The MBCs established in our previous study (Ogawa and Mii 2004) as well as the results presented here could be utilized as a baseline.

Incorporating β-lactams such as meropenem and moxalactam into Agrobacterium-mediated transformation procedures appears to be a simple and effective strategy for achieving high transformation efficiencies while minimizing persistent Agrobacterium growth. Although the novel, highly active, drugs described in this study are more expensive than conventionally used β-lactams, such as penicillin derivatives, lower concentration requirements will help minimize additional expense.