• Open Access

A novel dominant selectable system for the selection of transgenic plants under in vitro and greenhouse conditions based on phosphite metabolism


  • Damar L. López-Arredondo,

    1. Laboratorio Nacional de Genómica para la Biodiversidad, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Irapuato, Guanajuato, México
    Current affiliation:
    1. StelaGenomics Mexico S de RL de CV, Irapuato, Guanajuato, México
    Search for more papers by this author
  • Luis Herrera-Estrella

    Corresponding author
    • Laboratorio Nacional de Genómica para la Biodiversidad, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Irapuato, Guanajuato, México
    Search for more papers by this author

Correspondence (Tel +52 462 166 3008; email lherrera@ira.cinvestav.mx)


Antibiotic and herbicide resistance genes are currently the most frequently used selectable marker genes for plant research and crop development. However, the use of antibiotics and herbicides must be carefully controlled because the degree of susceptibility to these compounds varies widely among plant species and because they can also affect plant regeneration. Therefore, new selectable marker systems that are effective for a broad range of plant species are still needed. Here, we report a simple and inexpensive system based on providing transgenic plant cells the capacity to convert a nonmetabolizable compound (phosphite, Phi) into an essential nutrient for cell growth (phosphate) trough the expression of a bacterial gene encoding a phosphite oxidoreductase (PTXD). This system is effective for the selection of Arabidopsis transgenic plants by germinating T0 seeds directly on media supplemented with Phi and to select transgenic tobacco shoots from cocultivated leaf disc explants using nutrient media supplemented with Phi as both a source of phosphorus and selective agent. Because the ptxD/Phi system also allows the establishment of large-scale screening systems under greenhouse conditions completely eliminating false transformation events, it should facilitate the development of novel plant transformation methods.


In the process of plant genetic transformation, a small fraction of cells exposed to the transformation treatment become stably transformed. Hence, to enable the identification of rare stable transformation events, it is necessary to use highly efficient selectable marker genes that encode proteins that confer a selective advantage to the transformed cell with respect to untransformed cells. In addition, the selectable marker system must allow the selective growth and differentiation of the transformed cell. Dominant selectable marker genes for plant transformation have been developed based on genes that confer resistance to antibiotics or herbicides (Miki and McHugh, 2004) or that confer the ability to metabolize nonmetabolizable agents, such as xylose (Haldrup et al., 1998), mannose (Joersbo et al., 1998) or 2-deoxyglucose (Kunze et al., 2001), among others. However, it has been found that some selection systems are more effective for certain plant species and regeneration systems than others. This is because the sensitivity of plants to selective agents is widely variable between plant species and tissues under selective pressure. Therefore, an ideal selectable marker system should allow resistance to a toxic compound and simultaneously convert it into a compound that is essential for the growth and differentiation of cells in all plant species. Such system should prove universal and decrease or eliminate the risk of recovering false-positive clones that escaped the selection scheme. In this context, phosphite (Phi) represents a promising alternative as a selective agent because it hampers the development of plants (Schroetter et al., 2006; Thao et al., 2008; Ticconi et al., 2001; Varadarajan et al., 2002) but can be converted into an essential nutrient (López-Arredondo and Herrera-Estrella, 2012).

Phi is a structural analogue of phosphate (Pi) that is efficiently absorbed by the Pi transport system and rapidly mobilized by the xylem and phloem of plants (McDonald et al., 2001; Ouimette and Coffey, 1990). Phi has been used as an effective fungicide against oomycetes of the genus Phytophthora, which, together with their capacity to induce plant defence responses, have resulted in better yields for some plant species (Ávila et al., 2011; Carswell et al., 1996; Förster et al., 1998; Moor et al., 2009). These beneficial effects have led to the hypothesis that Phi has also nutritional properties as source of phosphorus (P) for plants. However, numerous reports have demonstrated that plant cells are unable to metabolize Phi, preventing its use as a direct source of P. Moreover, negative effects of Phi treatment on plant growth have been reported for many plant species, including Arabidopsis, tomato, Brassica nigra, pepper and maize (Förster et al., 1998; López-Arredondo and Herrera-Estrella, 2012; Schroetter et al., 2006; Thao et al., 2008; Ticconi et al., 2001; Varadarajan et al., 2002). Phi-mediated negative effects have been observed both under tissue culture conditions and when incorporated into the soil or applied as foliar treatments under field conditions (Ticconi et al., 2001; Varadarajan et al., 2002; Schroetter et al., 2006; Thao et al., 2008; López-Arredondo and Herrera-Estrella, 2012; Förster et al., 1998; Carswell et al., 1996; Moor et al., 2009; Ávila et al., 2011; Carswell et al., 1997).

The metabolic capacity to oxidize Phi into Pi has been reported in specific isolates of Escherichia coli, Bacillus caldolyticus, Agrobacterium tumefaciens and Pseudomonas sp. (White and Metcalf, 2007). However, to date, PTXD from Pseudomonas stutzeri WM88, which encodes an NAD-dependent phosphite dehydrogenase that oxidizes Phi into Pi, is the only well-characterized phosphite oxidoreductase (Garcia-Costas et al., 2001; White and Metcalf, 2004). We recently reported the generation of Arabidopsis (ptxDAt) and tobacco (ptxDNt) transgenic plants that are able to efficiently use Phi as the sole source of P (López-Arredondo and Herrera-Estrella, 2012). These plants were produced by expressing the ptxD gene from P. stutzeri WM88 under the control of the CaMV 35S promoter. Because Phi inhibits plant growth by interfering with Pi sensing in plants and because plants expressing the ptxD gene efficiently use Phi as their sole P source, we hypothesized that the ptxD gene could be used as a dominant selectable marker to produce transgenic plants. Here, we report the development of such a system based on the expression of ptxD as a dominant selectable marker and Phi as the selective agent.


Effect of phosphite on Arabidopsis growth

When we previously evaluated ptxDAt and ptxDNt transgenic lines that express the ptxD gene from P. stutzeri WM88 under control of the CaMV 35S promoter, we observed remarkable growth differences between wild-type (WT) and transgenic plants in Phi-containing media (López-Arredondo and Herrera-Estrella, 2012). The growth of WT plants was completely arrested, whereas transgenic lines displayed vigorous growth because they were able to effectively use Phi as a P source. Therefore, we decided to test whether the CaMV 35S::PTXD construct could be used as a dominant selectable marker using Phi as the selective agent. To evaluate the sensitivity of Arabidopsis plants to Phi and determine the minimum and maximum amount of Phi that could be used to select transformed plants, we sowed seeds from WT and three ptxDAt lines, one with high (ptxDAt-3) and two with low (ptxDAt-5 and 7) ptxD expression levels, in solid media containing 0.2, 0.4, 0.6, 0.8, 1, 3 and 5 mm Phi as the sole P source. Under all Phi concentrations tested, we observed an effective inhibition on the growth of WT plants (Figure 1). As previously reported, WT seedlings displayed negative effects induced by Phi, including leaf yellowing, stunted stems, short primary roots and very short or aborted lateral roots (Ávila et al., 2011; Carswell et al., 1996, 1997; Förster et al., 1998; López-Arredondo and Herrera-Estrella, 2012; Moor et al., 2009; Schroetter et al., 2006; Thao et al., 2008; Ticconi et al., 2001; Varadarajan et al., 2002). Even at the lowest Phi concentrations, the growth of WT plants was arrested at the cotyledonary or 4 leaf stage, and plants died 10–12 days after germination (dag). In contrast, ptxDAt plants displayed vigorous growth under all Phi concentrations, with no notable differences between the lines with high or low ptxD expression levels (Figure 1). Additional experiments using a higher Phi concentration (20 mm) showed that line ptxDAt-5 whose ptxD expression is almost 24-fold lower than that of ptxDAt-3 (the line with the highest expression level) is still capable of growing in media containing 20 mm Phi, whereas line ptxDAt-7, which has an expression level 40-fold lower than that of ptxDAt-3, is not (Figure 1 and Figure S1). Selective effects of Phi were also observed for seeds germinated in liquid media supplemented with 0.05, 0.1 and 1 mm Phi, for which the lowest concentration was sufficient to kill WT plants 7–8 dag and to sustain the growth of PTXD plants (Figure S2). These data show that the identification of transgenic plants can be achieved over a wide range of Phi concentrations.

Figure 1.

Evaluation of the sensitivity of Arabidopsis transgenic plants to phosphite on solid media. Transgenic homozygous plants from lines ptxDAt-3, 5 and 7 and WT plants (18 day old) germinated and grown horizontally on 0.1X MS nutrient solid media supplemented with different concentrations of phosphate (Pi) or phosphite (Phi) as a phosphorus (P) source under a photoperiod of 16-h light/8-h darkness and temperature of 22–23 °C. The position of the different transgenic lines and WT control is shown in the first panel on the top left picture.

Selection of Arabidopsis transgenic seedlings using the ptxD as a gene marker and phosphite as a selective agent

As the next step, we tested 0.1, 0.2, 0.4, 0.6, 0.8, 1 and 5 mm Phi as a selective agent for the selection of Arabidopsis transgenic seedlings and compared its efficiency to that obtained using the phosphinothricin acetyl transferase (bar) gene as a dominant selectable marker and phosphinothricin (PPT) as the selective agent. Using a modified floral dip method (Martinez-Trujillo et al., 2004), we inoculated Arabidopsis plants with an Agrobacterium strain harbouring a binary vector containing the CaMV 35S::PTXD gene construct. Because the transferred DNA also contained a NOS::BAR gene construct (Figure S3), it was possible to assess the efficiency of both selectable systems on the same pool of seeds. To evaluate transformation efficiency, we sowed T0 seeds on normal 0.1X MS media supplemented with 20 mg/L PPT or 0.1X MS lacking Pi but supplemented with Phi as both the selective agent and P source. Under all Phi concentrations tested, the growth of most seedlings was rapidly arrested in a similar fashion to that observed for untransformed controls, and only putatively transformed seedlings grew vigorously (Figure 2a, b, Figure S2 and Figure S4). No significant difference (< 0.05) in transformation frequency was observed between the Phi and PPT selectable systems (Table 1).

Table 1. Transformation frequencies obtained using the ptxD/Phi system
Selective agentConcentration (mm)Seeds sowedNumber of resistant seedlingsTransformation frequency% ± SE
  1. The use of phosphite (Phi) as selective agent allowed the selection of transgenic plants transformed with the CaMV 35S::PTXD construct. Arabidopsis plants were transformed using the floral dip protocol and the progeny germinated in media containing either phosphinothricin or different concentrations of Phi as selective agents. 15 days after germination, the number of surviving seedlings was counted, and the transformation frequency (calculated as the number of resistant seedlings per total seeds sowed) for each selective media was determined. The average seeds sowed, and number of resistant seedlings and transformation frequency from two independent experiments is presented.

  2. SE, Standard error.

Phosphinothricin0.15,713190.3325 ± 0.017
Phosphite0.15,295190.3587 ± 0.056
0.25,299150.2830 ± 0.018
0.45,306190.3392 ± 0.037
0.65,257170.3233 ± 0.018
0.86,181150.2426 ± 0.016
14,243150.3535 ± 0.023
56,724220.3271 ± 0.029
Figure 2.

Selection of Arabidopsis transformants using the ptxD gene as a selectable marker and phosphite as a selective agent. (a) WT (Col-0) seedlings, (b) T0 seedlings from an Arabidopsis plant transformed by the floral dip method and (c) T1 seeds from a self-pollinated transgenic Arabidopsis plant selected in (b) grown on media lacking phosphate (Pi) and supplemented with 1 mm phosphite (Phi) as a phosphorus (P) source. Plants were photographed 25 dag. (d) Southern blotting was performed to confirm the presence of the ptxD gene. Genomic DNA of homozygous Arabidopsis lines selected using 5 mm Phi was extracted using CTAB and digested with EcoRI, for which the T-DNA has a single restriction site. The pB7WG2D::PTXD expression vector was used as a positive control (C+). (e) Real-time PCR of WT (Col-0) and homozygous transgenic Arabidopsis lines was performed using total RNA and normalized using ACTIN2. Relative expression is presented in log 10 scale (relative to actin expression level). M: 1 Kb ladder molecular weight control.

Putative transgenic seedlings that were able to grow in Phi-containing media had normal morphology and displayed a well-developed root system (Figure S4). Because Phi causes root growth inhibition, this characteristic was used as an additional phenotype for the identification of transgenic seedlings at early stages of development. Although transgenic seedlings were easily identified among the background of untransformed seedlings for all tested Phi concentrations, we observed Phi dose-dependent differences in rosette size and leaf pigmentation (Figure S4). Putative transformed seedlings that were selected and grown with low Phi concentrations (0.1–0.4 mm) showed small rosettes, albeit ones that were clearly distinguishable from untransformed seedlings, and had dark purple leaves, features that are consistent with the response of Arabidopsis to low Pi availability. In contrast, plants selected using higher Phi concentrations (0.6–5 mm) showed large rosettes with green leaves. These data suggest that even low Phi concentrations are sufficient to compromise the growth of untransformed seedlings, but also to support the development of transgenic PTXD seedlings.

To confirm that seedlings selected in Phi-containing media were true transformants and not WT escapes, the presence of the ptxD gene and capacity of the plants to metabolize Phi were assessed. We first assayed for the presence of the complete ptxD gene by PCR in 61 randomly selected T0 plants obtained using different Phi concentrations (6–8 lines per treatment). All tested plants were found to be PCR positive for the presence of the ptxD coding sequence (Figure S5). To confirm that the PCR-positive plants were indeed transgenic, the T1 progeny of these plants was tested for germination and growth in Phi-containing media. We observed that for all tested lines, a large proportion of the seeds were able to germinate and sustain vigorous growth. In most cases, the capacity to metabolize Phi segregated in a 3 : 1 (resistant/susceptible) ratio, suggesting that they had a single active copy of the CaMV 35S::PTXD gene construct, whereas the remaining T1 plants produced a higher proportion of seeds able to metabolize Phi (Figure 2c and Table S1), likely because they contained more than one independent copy of the selectable marker gene. Moreover, the lines identified using Phi as the selective agent were also found to be resistant to PPT, which was expected because of the presence of the NOS::BAR gene in the same T-DNA.

Lines that segregated in a 3 : 1 ratio were selected, and homozygous T3 seed stocks were obtained. Eleven randomly selected homozygous lines were subjected to Southern blot hybridizations (Figure 2d). Total genomic DNA was extracted and digested with the restriction enzyme HindIII, which cleaves the T-DNA once, and then hybridized with a ptxD radiolabelled probe (Figure S3). Hybridization results suggested that all tested lines were indeed transgenic and contained from 1 to 5 independent T-DNA insertions. The expression of the CaMV 35S::PTXD transgene was further confirmed by qRT-PCR analysis, which showed a range of expression levels (Figure 2e).

Taken together, these results demonstrate that the ptxD gene can be used as an effective selectable marker for the transformation of Arabidopsis using the floral dip method and that the transformation frequency obtained is similar to that achieved using the bar gene. The ptxD/Phi system has the additional advantages that the cost of Phi is significantly lower than that of the antibiotics or herbicides commonly used for plant selection, and that Phi is not toxic to humans.

Effects of phosphite on tobacco shoot regeneration

To test whether the ptxD gene could act as a dominant selectable marker in a protocol that involves the regeneration of complete plants from transformed cells, we tested whether Nicotiana tabacum transgenic plants harbouring the CaMV 35::PTXD gene can be obtained using the Agrobacterium-based leaf disc transformation method. We first conducted experiments to determine the effect of Phi on the ability of tobacco leaf explants to produce shoots on MS media supplemented with 2 mg/L BAP to induce shoot formation (shoot-inducing media) and with Phi or Pi as the P source. Therefore, we incubated 0.7- to 1-cm2 explants in media containing 0.2, 1, 5, 7 or 20 mm Phi or Pi (Figure 3). After three weeks of incubation, only the explants under the lowest Phi concentration (0.2 mm) were able to develop sporadic shoots, which grew slowly and soon lost vigour and became brownish. This result is in contrast to explants placed on media containing Pi, which produced abundant green shoots (Figure 3). Importantly, after five weeks, all explants placed in the Phi media turned yellow pale, with necrotic regions, and were not able to generate viable shoots (Figure 4). These data suggests that possible escapes in regenerating shoots and selecting transgenic plantlets can be avoided using Phi concentrations greater than 0.2 mm.

Figure 3.

Effect of phosphite on shoot regeneration in tobacco explants. Tobacco leaf explants were cultivated in solid media lacking a phosphorus (P) source (No P) or containing 1 mm phosphate (Pi) or 0.2, 1, 5, 7 and 20 mm phosphite (Phi) as P source and supplemented with 2 mg/L BAP to induce shoot formation. All explants were incubated at 23 °C with a photoperiod of 16-h light-8-h dark.

Figure 4.

Generation and selection of transgenic tobacco plants using the ptxD gene as a selectable marker and phosphite as a selective agent. Tobacco leaf discs were cocultivated with an Agrobacterium strain harbouring the CaMV 35S::PTXD construct for 48 h and then transferred to shoot-inducing 1X MS media supplemented with 1 mm phosphite. Control leaf discs (a) were unable to regenerate shoots in media containing Phi as the only phosphorus (P) source, whereas leaf discs cocultivated with the Agrobacterium strain containing the CaMV 35S::PTXD construct produced several shoots per leaf disc (b). The use of Phi as both the source of P and the selective agent allows the differentiation (b) and elongation of putative transgenic shoots (c), which were then transferred to rooting media after 8–9 weeks of culture (d). (e) Segregation analysis of the progeny of heterozygous transgenic tobacco plant in media containing Phi as a sole P source.

Regeneration of tobacco transgenic shoots using the ptxD gene as a gene marker and phosphite as a selective agent

Based on these results, we decided to use MS nutrient media supplemented with 1 mm Phi to perform tobacco leaf disc transformation experiments. With this aim, tobacco leaf discs were cocultivated with an Agrobacterium strain that harboured the CaMV 35S::PTXD construct or the empty vector and incubated in selective (1 mm Phi) or control (1 mm Pi) shoot-inducing media. As expected, explants cocultivated with both the ptxD and the empty vector Agrobacterium strains were able to generate abundant shoots in media supplemented with 1 mm Pi. Explants cocultivated with the empty vector strain and incubated in media containing 1 mm Phi became yellowish and were unable to generate shoots (Figure 4a), whereas explants cocultivated with Agrobacterium strain harbouring the CaMV 35S::PTXD construct developed green proliferative sectors, which after 4–5 weeks produced healthy shoots (Figure 4b and c). Upon transfer to fresh Phi-containing media lacking growth regulators, these regenerated shoots rapidly developed into plantlets and produced abundant roots (Figure 4d). Regenerated plants had normal morphology, flowering time and seed production under greenhouse conditions, even when the substrate used for their growth was amended to contain Phi as the sole P source (Figure S6). Eighty-one T1 plants were subjected to PCR analysis to amplify the ptxD gene, of which 100% were PCR positive (Figure S7). The expression of the ptxD gene was further confirmed by qRT-PCR for some of these lines, for which we observed a range of expression levels (Figure S7). Mendelian segregation of the CaMV 35S::PTXD construct in the T2 progeny of transgenic tobacco plants was clearly observed in media containing Phi as the sole source of P (Figure 4e). These data demonstrated that the ptxD/Phi system is effective for regenerating and selecting tobacco transgenic plants from a single transformed cell.

The ptxD/Phi system can be used to select transgenic plants under greenhouse conditions using solid substrates

The selection of transformed plants under sterile conditions is labour-intensive and time-consuming and generally requires expensive selectable agents. Therefore, low-cost greenhouse systems to screen large seed populations for the presence of low-frequency transformation events could facilitate the establishment of new transformation protocols. To test whether transformed seedlings could be identified among a background of nontransformed seedlings under greenhouse conditions, we sowed a 1 : 100 mixture of PTXD tobacco transgenic/WT seed in a solid substrate amended with 120 mg/kg of Phi. Transgenic plantlets could be identified among a background of nontransformed plantlets as early as 5–8 dag (Figure S8a). The transgenic nature of the selected plantlets was corroborated 35 dag, as the selected PTXD transgenic tobacco plants displayed vigorous growth using Phi as P source (Figure 5a). Similar results were obtained by irrigating the substrate with a nutrient solution containing 1 mm Phi as the only P source (data not shown). We also performed high-density selection experiments by directly germinating Arabidopsis T0 seeds in a mixture of sand and vermiculite (1 : 1) amended with Phi at 120 mg/kg. Transgenic Arabidopsis plantlets displayed considerably better growth than nontransformed siblings 10 dag (Figure S8b) and grew vigorously 35 dag (Figure 5b).

Figure 5.

Selection of Arabidopsis and tobacco transformants using a solid substrate amended with phosphite under greenhouse conditions. (a) Transgenic tobacco seeds mixed with wild-type seeds at 1 : 100 ratio were directly germinated and grown in a mixture of sand and vermiculite (1 : 1) amended with 120 mg/kg of phosphate (Pi) or phosphite (Phi) as phosphorus (P) source. Plants under the different treatments were photographed 35 days after germination, placing pots with the different treatments side by side. (b) Arabidopsis T0 seeds were directly germinated and grown in sand and vermiculite mixture (1 : 1) amended with 120 mg/kg of Pi or Phi as P source. Plant were photographed 35 days after germination. Treatment with no P source (No P) was included as a control.


The simple system we have described, which is based on the metabolism of the reduced P compound Phi, utilizes two important factors: the nutritional requirements of plants for Pi, an essential nutrient, and the inability of plants to metabolize Phi. As we demonstrated, our system can be used to effectively select transformed plant cells under in vitro culture conditions and regenerate plantlets from these cells or to directly select transgenic plants during seed germination under in vitro or greenhouse conditions. Other selectable marker genes have a number of disadvantages, for instance those converting nonmetabolizable compounds into molecules that plant cells can use as a carbon source are only effective for nonphotosynthetic cells, whereas in the case of antibiotic and herbicide detoxification systems, the sensitivity to these compounds widely varies depending upon the plant species to be transformed. In contrast, the ptxD/Phi system is potentially a universal selectable system because all plant species require Pi for their growth and not a single plant species has been reported to be capable of oxidizing Phi into Pi. However, although we do not expect negative effects from Phi on plant differentiation, it will be necessary to test the effect of Phi on the regeneration process of plant species other than tobacco to fully demonstrate the use of the ptxD/Phi system as a universal system for plant transformation. The only requirement for the ptxD/Phi system to be effective is that selection of transformed cells or seedlings must be performed on media containing limiting amounts or completely lacking Pi.

The transformation efficiency we determined using the Phi selectable system was similar to that obtained using the bar gene as a selectable marker. Although in these experiments, the selectable marker genes were expressed from different promoters, it is not expected to have a different efficiency using a weaker promoter to drive the expression of PTXD, because seedlings with very low levels of ptxD expression were able to grow in media containing up to 5 mm Phi as selective agent (Figure 1, Figure 2 and Figure S1).

Untransformed cells or plantlets that escape the effect of selectable agents (commonly known as escape events) are a common problem when antibiotic or herbicides are used as selectable agents. This problem is unlikely to occur with the ptxD/Phi system because Phi must be detoxified and converted into Pi for the plant cell or germinating seed to grow; therefore, the only way of regenerating plants or germinating seeds in media containing Phi as the only P source is that they acquire the capacity to metabolize Phi. In the course of the experiments reported here and additional experiments to search for Phi-resistant Arabidopsis mutants, we have tested more than 50 000 EMS-mutagenized seeds, and not a single escape event has been observed. These results suggest that in contrast to resistance to herbicides, which can be acquired by different mechanisms, including a decrease in herbicide transport or compartmentalization to sequester the herbicide and avoid its toxicity or mutations in the target enzyme, mutations that allow the use of Phi as a sole P source are less likely to occur in plants.

The ptxD/Phi system has several advantages compared with currently used systems: (i) sodium (Na) and potassium (K) Phi salts are readily available from many companies and inexpensive; (ii) Phi salts are innocuous to humans and animals, and therefore, their use does not require special precautions; (iii) Phi salts are water soluble and have high thermo- and photostability, ensuring their persistence as selective agents in both tissue culture and greenhouse conditions; (iv) because all plants need a P source to sustain growth and not a single plant species has been reported to be able to metabolize Phi, the ptxD/Phi system is potentially a universal dominant selectable marker for the production of transgenic plants; (v) the ptxD/Phi system is useful as a selection system under greenhouse conditions using inert substrates or a variety of soils with low Pi content by providing the system with Phi in the soil or the irrigating solution, which would facilitate high-density selection experiments; and (vi) the use of direct selection protocols using trays containing soil or an inert substrate under greenhouse conditions will avoid the time-consuming and expensive step of selection under tissue culture conditions and the individual transfer of each resistant plant into pots for greenhouse growth.

Greenhouse selection systems could be useful to develop transformation systems similar to the Arabidopsis floral dip protocol for other plant species or for selecting transgenic seeds derived from transformation experiments of apical meristems by particle bombardment that produce chimeric plants in which only a few seeds are transgenic and large numbers of seeds must be evaluated.

Finally, it is important to note that of the almost 60 selectable marker systems proposed to date (Miki and McHugh, 2004), only a few, such as the betaine aldehyde dehydrogenase (from spinach) (Daniell et al., 2001), the rstB (from Sinorhizobium fredii strain RT 19) (Zhang et al., 2009) and the AtTPS (from Arabidopsis) (Leymand et al., 2006) genes, which confer salt tolerance and/or osmoprotection, confer a developmental advantage to plants such that in addition to the identification and isolation of transgenics, they are able to cope with biotic or abiotic stresses that reduce the growth and productivity of crops. The presence of the ptxD gene represents not only a method to identify transgenic plants but also a competitive advantage that under field conditions should reduce the application of P-fertilizers and herbicides to control weed growth, as we recently reported (López-Arredondo and Herrera-Estrella, 2012).

Experimental procedures

General reagents

All reagents used in the experiments were from Sigma-Aldrich, unless otherwise stated. For all experiments using solid tissue culture media, agar Plant TC from Phytotechnology Laboratories was used.

Biological material and in vitro growth conditions

Arabidopsis thaliana (Col-0) and Nicotiana tabacum L. cv Xanthi were used to test Phi susceptibility and as background genotypes in transformation experiments. For propagation and plant growth experiments in vitro, Arabidopsis and tobacco seeds were placed in an Eppendorf tube and then surface disinfected as follows: 1 mL 70% (v/v) ethanol was added, and the tube was shaken for 7 min. Then, the solution was discarded, 1 mL 20% (v/v) commercial bleach was added, and the tube was shaken for 8 min. Finally, the solution was discarded, and the seeds were washed four times for 10 min each time using sterile-deionized water.

Hydrated tobacco seeds were used immediately after disinfection or preserved at 15 °C in sterile-deionized water until required for experiments. Hydrated Arabidopsis seeds were stratified at 4 °C for 48 h to promote and synchronize germination. For in vitro experiments, seeds were sown in sterile conditions in Petri dishes or plastic containers containing Murashige and Skoog (MS) nutrient media, as indicated for each experiment. Then, plates and containers were incubated in a plant growth cabinet (PERCIVAL) with controlled conditions and a photoperiod of 16-h light/8-h dark and temperature of 22 ± 1 °C. Petri dishes were placed vertically or horizontally, as required. For propagation purposes, all seeds were germinated in vitro, and then, plantlets were transferred to pots with a sterile mixture of perlite/vermiculite/Canadian peat moss (1 : 1 : 1).

The basic MS medium contained the following: 2.0 mm NH4NO3, 1.9 mm KNO3, 0.3 mm CaCl2 2H2O, 0.15 mm MgSO4 7H20, 5 mm KI, 25 mm H3BO3, 0.1 mm MnSO4 H2O, 0.3 mm ZnSO4 7H2O, 1 mm Na2MoO4 2H2O, 0.1 mm CuSO4 5H2O, 0.1 mm CoCl2 6H2O, 0.1 mm FeSO4 7H2O, 0.1 mm Na2EDTA 2H2O, 10 mg/L inositol, 0.2 mg/L glycine, 0.05 mg/L pyridoxine chlorhydrate, 0.05 mg/L nicotinic acid and 0.01 mg/L thiamine hydrochloride, in addition to the P source, Pi (KH2PO4 or NaH2PO4) or Phi (KH2PO3 or NaH2PO3), which was added as required (KH2PO3 from Wanjie International CAS No. 13977-65-6). The media were adjusted to pH 5.7 and contained 5 g/L sucrose for Arabidopsis media and 15 g/L sucrose for tobacco media, and 10 g/L agar when required.

Experiments using a hydroponic system

To analyse the growth responses of control and transgenic plants, experiments were performed not only in vitro (as described above) but also using a hydroponic system. For these experiments with liquid media, treatments with one source of P (Pi or Phi) at 0, 0.05, 0.1 and 1 mm or with both sources of P at proportions of 0.05/0.05, 0.1/0.1, 1/1 mm/mm were established. Sterile 1-L plastic containers (14 cm in diameter) were filled in with 480 mL of the medium, and a plastic mesh was placed on the liquid surface of each container, in which 50 seeds were sown.

Generation of the CaMV 35S::PTXD construct and transformation of Agrobacterium

To obtain transgenic plants expressing the ptxD coding sequence from Pseudomonas stutzeri WM88, the complete open-reading frame of this gene was amplified from the pWM302 plasmid (a kind gift of Dr. William W. Metcalf) and placed under the control of the CaMV 35S promoter using Gateway® technology.

The Gateway system utilizes the site-specific recombination reactions that enable the bacteriophage λ to integrate and excise itself in and out of a bacterial chromosome (Katzen, 2007). The protocol is based on the presence of two essential components: (i) recombination DNA sequences (att sites) and (ii) enzymes that catalyse recombination reactions. Hence, to generate this construction, the complete coding sequence of ptxD was amplified by PCR using Taq DNA polymerase (Invitrogen) (3 min at 94 °C, 1 min at 94 °C, 50 s at 59 °C, 1 min at 72 °C, 7 min at 72 °C and hold at 4 °C) and the following primers PTXD designed with attB sites:


Amplified PCR fragment were electrophoresed on a 1% agarose/TAE gel and purified using GFX™ PCR DNA and a Gel Band Purification kit (illustra™ GE Healthcare, Buckinghamshire, UK) following the manufacturer's instructions. The resulting ptxD amplification fragment was subcloned into a pGEM-T Easy vector (Promega) following the manufacturer's instructions. Six clones were selected and analysed by restriction analysis and PCRs to ensure the presence of the desired gene, all of which were positive by both assays.

One clone was selected to perform the BP recombination reaction to introduce the ptxD gene into the pDONR221 donor vector. The BP reaction is catalysed by the BP Clonase II enzyme mix, which transfers the DNA fragment of interest, flanked by the attB sites, into a donor vector carrying two attP sites, resulting in an entry clone (pENTR) flanked by two attL sites. Six clones were selected and analysed by restriction analysis and PCRs to ensure the presence of the desired gene, all of which were positive by both assays. Subsequently, one entry clone was selected to perform the LR recombination reaction using the pB7WG2D destination vector, which is catalysed by the LR Clonase II enzyme mix, which transfers the DNA fragment of interest, flanked by two attL sites (in the entry clone) into a destination vector carrying two attR sites, resulting in an expression clone. Transformation of E. coli with this product resulted in six positive clones containing the ptxD gene placed under the control of the CaMV 35S constitutive promoter and terminator. One clone was selected to genetically transform the desired plants.

The pB7WG2D::PTXD expression vector was introduced by electroporation to Agrobacterium tumefaciens strain GPV2260 and used to produce Arabidopsis thaliana (ptxDAt) and tobacco (ptxDNt) transgenic lines. Transformation of Agrobacterium was performed in a BioRad MicroPulser electroporator with the pre-programmed settings for A. tumefaciens (2.20 kV, 1 pulse). After transformation, 400 μL of YEB medium was added to the cells and then incubated at 28 °C with shaking for 1.5 h. Electroporated Agrobacterium cells were plated on YEB plates with antibiotics and grown at 28 °C for 2 days. To verify the presence of the ptxD gene, colony PCR was performed on 10 colonies. All colonies analysed were PCR positive, and one was selected for use in the transformation of plants.

Transformation of Arabidopsis and in vitro selection of transformant seedlings

ptxDAt lines were generated using a modified floral dip transformation protocol (Martinez-Trujillo et al., 2004). An Agrobacterium strain harbouring the pB7WG2D::PTXD expression vector was inoculated in 5 mL of YEB medium supplemented with carbenicillin, rifampicin and spectinomycin at 100 mg/L each and grown to the stationary stage (OD600 approximately 2.0) at 28 °C with shaking at 250 rpm. Then, Agrobacterium cells were harvested from the liquid medium by centrifugation (4 °C and 5 min at 13 400 × g) and washed twice with 0.1X MS with 5% sucrose. The bacterial pellet was then resuspended in infiltration medium consisting of 0.1X MS, 5% sucrose and 0.05% Silwett L-77® (GE, Silicones) to the desired density (OD600 from 0.8 to 2.0). This suspension was dropped onto all unopened floral buds with a micropipette to infect them. Infected plants were kept at high humidity under a dark plastic bag for 10 h. The inoculation procedure was repeated 8–10 times every two days.

Pool seeds (T0) produced by infected plants were collected and screened for resistant seedlings according to the resistant marker of the construct. To select Arabidopsis transgenic plants, two types of experiments were conducted: (i) using phosphinothricin as the selective agent and (ii) using Phi as the selective agent and P source. Disinfected seeds were sown on sterile growth solid medium supplemented with either 20 mg/L phosphinothricin or potassium phosphite at different concentrations. When Phi was used as the selectable agent, potassium phosphate monobasic (KH2PO4), the P source in standard media, was replaced with potassium phosphite monobasic (KH2PO3) from Wanjie International (CAS No. 13977-65-6).

Transformation of tobacco and regeneration of plantlets

To transform Nicotiana tabacum L. cv Xanthi, the basic steps of the leaf disc transformation method (Horsch, 1985) were used. Tobacco leaves were cut into small segments (0.7–1 cm2) and cocultivated with the Agrobacterium culture harbouring the pB7WG2D::PTXD expression vector under darkness and a controlled temperature of 28 °C for 2 days. After this time, explants were washed at least three times with sterile-deionized water and one time with cefotaxime (500 mg/L) and blotted on sterilized filter paper to eliminate excess solution. Explants were then transferred to selection medium (1X MS, 30 g/L sucrose, 2 mg/L BAP, 100 mg/L cefotaxime, 2.5 g/L Gelrite, pH 5.7) supplemented with Phi (1 mm) as the sole P source. Explants were subcultured in fresh medium every 2–3 weeks. After 8–9 weeks under selective conditions, the regenerated shoots were excised from explants and cultured on MS containing Phi (1 mm) with no plant regulators. The successful growth of all lines in Phi-containing media confirmed that all were capable of using Phi as a P source. The presence and expression of the CaMV 35S::PTXD fusion were confirmed in eighty-one and eight randomly selected lines by genomic PCR and qRT-PCR, respectively.

For cocultivation, the Agrobacterium strain harbouring the pB7WG2D::PTXD was previously grown at 28 °C (250 rpm) for two days in 5 mL liquid YEB medium (Sambrook et al., 1989) supplemented with carbenicillin, spectinomycin and rifampicin (100 mg/L each), harvested by centrifugation (4 °C and 5 min at 13 400 × g) and suspended in 1X MS liquid medium.

Selection of Arabidopsis and tobacco transformants under greenhouse conditions

For tobacco selection, a mixture of transgenic and WT seeds was directly sown on a mixture of sand and vermiculite (1 : 1), which was amended with Pi or Phi at 80 or 120 mg/kg and periodically irrigated with MS 0.1X lacking Pi or only irrigated with 0.1X MS containing 1 mm Pi or Phi. In the case of Arabidopsis, T0 seeds were directly sown on mixture of sand and vermiculite (1 : 1) and treated as above. Seeds were covered with a plastic lid to avoid excessive loss of water.

Analysis of the presence of the ptxD gene by PCR and Southern blot analysis

Genomic DNA was extracted from complete plantlets using the CTAB method (cetyltrimethylammonium bromide) and cleaned using Durapore membranes (MSGVN2250, Millipore) according to the manufacturer's instructions.

For genomic PCR analysis, 100–200 ng of total genomic DNA was used to amplify the complete pxtD coding region using standard amplification conditions (3 min at 94 °C, 1 min at 94 °C, 50 s at 59 °C, 30 s at 72 °C, 7 min at 72 °C and hold at 4 °C) and the following primers: PTXDFW (5′-ATGCTGCCGAAACTCGTTATAACTC-3′) and PTXDRV (5′-TCAACATGCGGCAGGCTC-3′). PCR products were electrophoresed in 1% agarose/TAE gels and observed under UV light.

For Southern blot hybridization analysis, 15 μg of total DNA was digested with EcoRI restriction enzyme (Invitrogen), separated on a 1% agarose/TAE gel and capillary blotted onto a Hybond-N+ nylon membrane (Amersham Pharmacia Biotech). Nucleic acids were covalently fixed to the nylon membrane using a UV cross-linker. A probe corresponding to the first 400 bp of the ptxD gene (Figure S3), starting from the atg and generated by PCR using the following PTXDRT primers (3 min 94 °C, 1 min 94 °C, 50 s 59 °C, 30 s 72 °C, 7 min 72 °C, and hold at 4 °C), was used for the hybridization experiments: PTXDRTFW (5′-ATGCTGCCGAAACTCGTTATAACTC-3′) and PTXDRTRV (5′-CTGCAAGCGATCAGCCATG-3′).

The probe was purified on a 1% agarose/TAE gel, processed using the GENECLEAN® kit (MPBIO) and radiolabelled with [32P] using the Random Primer DNA Labelling System (Invitrogen) according to the manufacturer's specifications. Membranes were blocked using a 5X Denhardt's-based solution, hybridized with the ptxD probe and washed according to standard protocols. Membranes were visualized by phosphor imaging on a Storm 840 Phosphor Imaging System (Molecular Dynamics, Sunnyvale, CA).

Analysis of the expression of the ptxD gene by real-time PCR

Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA) and Qiagen RNA easy columns (Invitrogen). Real-time PCR of the ptxD gene (PTXDRT primers) was performed in an ABI PRISM 7500 real-time thermocycler (Applied Biosystems), and reactions for Arabidopsis Actin 2 (ACT2 primers) and tobacco ActinNt (ACTINNT primers) were utilized for normalization.


The relative quantification (RQ) number for each independent transgenic line was obtained from the equation 2−ΔΔCT, where ΔΔCт represents the subtraction of the CT value of internal control from the CT value of the ptxD gene (PTXDRT primers) (ΔCT(PTXD)–ΔCT(ACTIN)). ΔCT was calculated using the equation [CT(ptxD)*E] - [CT(ACTIN)*E], where E is the PCR efficiency [(10(−1/m))−1] (Livak and Schmittgen, 2001). Expression levels were obtained from at least three technical replicates.

Statistical analysis

The T2 progeny of independent Arabidopsis lines, selected using the ptxD/Phi system, was sown and germinated in Phi-containing media. 10 days after germination, the number of resistant and nonresistant seedlings in segregational analysis was evaluated using the chi-squared test (P < 0.05). Transformation frequencies obtained in Phi and PPT selection experiments were subjected statistical analysis using ANOVA and Tukey's tests (P < 0.05).


We thank Jose Luis Cabrera and Marco Antonio Leyva for technical support; Enrique Ibarra Laclette for Real-time PCR analysis; Verónica Limones for assistance in tobacco transformation; and Erika Alba for the propagation of plants in the greenhouse. We are grateful to William Metcalf for providing plasmid pWM302. This work was supported in part by a grant from the Howard Hughes Medical Institute (Grant 55005946) to L.H-E. DLLA is indebted to CONACyT, México for a PhD fellowship (No. 203571).