Agrobacterium-mediated transformation of Jatropha curcas leaf explants with a fungal chitinase gene

1 Centro de P&D de Recursos Genéticos Vegetais, Instituto Agronômico, Caixa Postal 28, Campinas, SP, 13075-630, Brazil. 2 Centro APTA Citros Sylvio Moreira, Instituto Agronômico, Caixa Postal 04, Cordeirópolis, SP, 13490-970, Brazil. 3 Instituto Biológico, Secretaria de Agricultura e Abastecimento, Laboratório de Bioquímica Fitopatológica, Av. Conselheiro Rodrigues Alves, 1252, Vila Mariana, 04014002 São Paulo, SP – Brazil. 4 Fundação Oswaldo Cruz, Laboratório de Modelagem de Sistemas Biológicos, CDTS, Av. Brasil, 4365, Manguinhos, 21040-900 Rio de Janeiro, RJ – Brazil.


INTRODUCTION
Considering the importance of renewable energy sources to sustain the global energy matrix, biofuels are among the available alternative to mitigate the current energy crises and climate changes (Divakara et al., 2010;Varshney and Johnson, 2010).In Brazil, nowadays, the blend of 7% biodiesel into fossil oil (B7) is already mandatory; it is scheduled to reach 10%, until 2019 and even 20%, until 2030 (Giersdorf, 2013).The progressive increase in commercial diesel blend on the market has resulted in positive environmental, economic and social impact.However, it will also require urgent oil sources diversification to sustain the growing demand for biodiesel.
Due to its high oil content and quality, Jatropha curcas L., has been proposed as a promising plant source to diversify biodiesel and biokerosene production in tropical climates (Carels, 2013;Sharma, 2011;Li et al., 2007).Despite its potential, J. curcas, also known as physic nut, is a perennial and semi-wild species, which still needs years of selective breeding until the release of a stable cultivar to the world.Plant domestication and genetic improvement process could be sped up by the integration of biotechnology approaches to the conventional breeding process for the introduction of agronomical desirable traits such as oil content and quality, synchronous fruit maturity, early flowering and disease or pest resistance.Although considered as a stress resistant species (Openshaw, 2000;Debnath and Bisen, 2008;Sabandar et al., 2013), physic nut has proven to be highly susceptible to a wide array of pest and disease when it is raised as a monoculture, which may limit its seed and oil yields (Kumar et al., 2013;Sujatha, 2013;Argollo Marques et al., 2013).
The domestication process through selective breeding is a purifying process that leads to the reduction of genetic variability by capturing both additive and nonadditive effects in elite clones via selection and clonal propagation (Carels, 2013).Despite some variations in agro-morphological and chemical properties reported in different studies (Wani et al., 2006;Kaushik et al., 2007;Rao et al., 2008;Argollo Marques et al., 2013), only a limited number of traits has been observed in the J. curcas genotypes evaluated so far.The major part of these traits involves genes engaged in mechanisms of resistance to biotic and abiotic stresses (Argollo Marques et al., 2013).
The genetic base broadening can be obtained by interspecific hybridization and/or biotechnological approaches.The introgression of stress resistance genes as well as other favorable genes found in wild congeneric species into J. curcas may be a promising breeding strategy (Sujatha et al., 2013;Argollo Marques et al., 2013), however, it is time consuming, expensive and limited to the closest wild species (Guidolin, 2003).Furthermore, undesirable genes are inevitably coinherited and need cycles of backcrossing to be eliminated.In turn, the genetic engineering strategy has many advantages such as the shortening of the number of selective breeding steps and the ability to introduce genes for traits that may not be even available within congeneric species (Visarada et al., 2009;Herr and Carlson, 2013;Fu et al., 2015).However, to be successful, the genetic transformation process depends from the availability of an efficient and reproducible in vitro regeneration system.To date, plant regeneration protocols have been developed using various types of explants (Wei et al., 2004;Jha et al., 2007;Deore and Johnson, 2008;Kumar and Reddy, 2010;Kumar et al., 2010;Sharma et al., 2011;Toppo et al., 2012;Franco et al., 2014).Based on these regeneration systems, stable transformation procedures have been developed employing both Agrobacterium tumefaciens (Li et al., 2008;Kumar et al., 2011;Pan et al., 2010;Misra et al., 2012;Fu et al., 2015;Nanasato et al., 2015) or particle bombardment (Joshi et al., 2011;Purkayastha et al., 2010).
Due to little transgene re-arrangement, low copy number and preferential integration into transcriptionally active regions of chromosomes, Agrobacterium-mediated transformation is the most widely used method to generate transgenic J. curcas (Kumar et al., 2013;Fu et al., 2015).Li et al. (2008) first reported J. curcas genetic transformation via A. tumefaciens infection of cotyledon disc using the phosphinothricin herbicide as selective agent, however, the transformation efficiency was low (13%).Recently, Nanasato et al. (2015) improved the rate of cotyledon transformation efficiency to 23.3% using vacuum infiltration for the step of bacterial inoculation and co-culture on filter-paper wicks instead of agar for the step of latency before shoot regeneration.Fu et al. (2015) further increased the rate of transformation efficiency to 56% by optimizing kanamycin concentration in the selective medium and the duration of the selection process.Despite their higher potential of in vitro shoot regeneration, juvenile explants such as cotyledons and hypocotyls may results in a higher rate of unexpected somatic variation (Bhatia et al., 2005).Due to the smaller rate of somatic variation in regenerated shoots obtained *Corresponding author.E-mail: d-argollo@iac.sp.gov.br.
Author(s) agree that this article remains permanently open access under the terms of the Creative Commons Attribution License 4.0 International License from sporofitic explants such as stem and leaves, these tissues should be preferred for transformation (Kumar et al., 2010;Misra et al., 2012).Zong et al. (2010) using leaf explants, kanamycin for selection medium and a 2 days pre-culture period, increased the rate of transformation efficiency to 23.91%.Kumar et al. (2010) using hygromycin as another selective agent with leaf explant and a pre-culture period of 4 days achieved an even greater efficiency of 29%.Despite these efforts, large variations among different Jatropha genotypes in terms of their morphological responses to phytohormones is still observed in vitro and the rate of transformation efficiency of differentiated tissues still varies between 4 and 56% (Basa and Sujatha, 2009;Fu et al., 2015).Thus, the optimization of the regeneration and transformation procedures need to be conducted according to their genetic profile, which makes sense at least for a small number of high-yielding elite genotypes that have been recently selected by conventional breeding programs (Basa and Sujatha, 2009).
Currently, efficient in vitro regeneration protocols were developed using foliar explants for cloning elite genotypes selected under the breeding programs of Instituto Agronômico of Campinas (IAC) in Brazil (Franco et al., 2014).High regeneration rates (~40 shoots per explant) and efficient rooting via micrografting (85%) were obtained for the L4P49 genotype.The micropropagated shoots were considered genetically stable after assessing their ploidy by flow cytometry and their DNA polymorphism by marker-tartrate-resistant acid phosphatase (TRAP), indicating preservation of clonal stability and phenotype uniformity.The subsequent adaptation of this regeneration protocol to an effective and reproducible Agrobacterium-mediated transformation protocol with optimization of some key transformation parameters such as pre-culture period and antibiotics concentration was the purpose of the present report.Then, the combined transformation and regeneration protocol was applied to transfer a fungal chitinase gene into the genome of a J. curcas elite plant to confer fungal disease resistance.
Regenerated shoots were elongated on MS medium supplemented with 6-benzylaminopurine (BAP, 0.3 mg.L -1 ) and IBA (0.1 mg.L -1 ).The pH of all medium was adjusted to 5.8, prior to sterilization (121°C for 20 min).The elongated shoots of 1.5 cm length after 30 days old cultures were used as a source of foliar explants for testing kanamycin resistance for further.
All cultures were maintained at 25 ± 2°C under a 16/8 h (day/night) photoperiod with light provided by cool white fluorescent lamps at an irradiance of 35 to 40 μmol m −2 s −1 .

Agrobacterium strain and binary vector
A cDNA of the ech42 chitinase gene was amplified by quantitative amplification of reversed transcripts by polymerase chain reaction (qRT-PCR) from Trichoderma viride grown on media supplemented with colloidal chitin.The ech42 chitinase gene was initially cloned in pGEM-T Easy (Promega) and completely sequenced.Subsequently, it was sub-cloned in pCAMBIA 1201 using the NcoI and BstEII restriction sites, substituting the uidA (GUS) gene, and transferred to pCAMBIA 2201 using the EcoRI and BstEII sites.
The ech42 gene is under the control of a CaMV 35S promoter in Pcambia 2201 vector and the resistance to kanamycin for transgenic plant selection is conferred by a neomycin phosphotransferase II gene (nptII).The pCAMBIA 2201 vector (Figure 1) was introduced by electroporation into the disarmed A. tumefaciens EHA105 strain.
A. tumefaciens strain containing a binary vector pCAMBIA 2201 were inoculated in solid LB (Miller, 1972) medium supplemented with kanamycin (100 mg.L -1 ) and rifampicin (100 mg.L -1 ) under 28°C for 72 h.A single colony of the transformed bacteria was isolated and inoculated in liquid LB medium supplemented with kanamycin (100 mg.L -1 ).Bacterial cultures were grown at 200 rpm and 28°C for 16 h.The bacterial suspension was centrifuged at 3,000 rpm for 15 min and the cell pellets were re-suspended in sterile distilled water before further use.

Determination of phytotoxic level of antibiotics
The A. tumefaciens EHA105 strain was inoculated in LB medium containing cefotaxime at 0, 300, 400 and 500 mg.L -1 and maintained in biochemical oxygen demand (BOD) at 28°C.After five days, the bacterial growth was evaluated in order to determine the minimum dose necessary for total growth inhibition.
The phytotoxic level of kanamycin was determined in order to find the optimal dose that should be used in selective medium.Two types of untransformed foliar explants were cultured separately on RM including an inhibitory dose of cefotaxime previously identified as 500 mg.L -1 and kanamycin at 0, 25, 50, 75 and 100 mg.L -1 .The first type of foliar explants consisted of leaf sections from shoots grown in vitro (called in vitro leaf explants) and the second was constituted by leaf segment taken from in vivo cuttings, grown in a greenhouse (called in vivo leaf explants).All cultures were maintained at 25±2°C and 16/8 h (day/night) photoperiod.Both antibiotics were filter-sterilized (Millipore ® 0.22 µm pore size) and added to media after autoclaving.The control medium was RM without antibiotic (cefotaxime and kanamycin).To assess the threshold of kanamycin toxicity, an experimental scheme of five plates per treatment and six explants per plate were used.Three repetitions of the test of kanamycin toxicity assessment were performed independently.The assessment of kanamycin toxicity was obtained through the relative number (%) of bleached explants after six weeks of subculture.The rate of shoot regeneration was calculated as the average number of regenerated shoots per explant.
After six weeks, the rate (%) of kanamycin-resistant shoots that overcame 1.5 cm was calculated relative to the total number of inoculated leaf explants.The transformation efficiency was estimated on the basis of one kanamycin-resistant shoot per explant as the ratio between the number of shoots with ech42 amplicons (PCR + ) relative to the total number of inoculated explants.

Confirmation of transgenic shoots by PCR
Transformation was confirmed by PCR amplification using total genomic DNA extracted according to Doyle and Doyle (1990).DNA concentration was measured using a Nano Drop spectrophotometer (NanoVue plus-GE).Specific primers for the ech42 gene (F: 5' CACTTCACCATGTTGGGCTTCCTC 3' and R: 5' GATCTCTAGTTGAGACCGCTTCGG 3') were obtained using Primer3 (Rozen and Skaletsky, 2000) and expected to produce an amplicon of 1.3 Kb.The amplification reaction was conducted on a Bio Rad T100 TM thermocycler in a final volume of 25 μl with 50 ng of each DNA sample in 4.0 μl buffer reaction, 1.2 μl MgCl2 (25 mM), 0.4 μl dNTP mix (2.5 mM of each), 0.5 μl of each primer F and R (10 μM), 0.25 μl of Taq DNA polymerase (Thermo Fisher Scientific).The amplification started at 95°C for 4 min, followed by 30 cycles of 30 s at 95°C, 30 s at 62°C and 2 min at 72°C, followed by a final extension of 7 min at 72°C.PCR products were separated on 1% agarose gel electrophoresis (0.5X TBE, 110 V) and visualized on Alpha Imager HP (Cell Biosciences) using GelRed staining.
Amplifications were performed using the following program: 55°C for 5 min, 95°C for 10 min, 40 cycles of 95°C for 15 s, 60°C for 1 min, followed by a dissociation step (95°C -15 s, 60°C -30 s, and 95°C -15 s) and repeated twice independently with three replicates each.The estimation of ech42 copy number was made according to Mason et al. (2002) by comparing its standard curves plotted using threshold cycle (Ct) values relative to those of JcKASIII in control and transgenic plants.The r1 coefficient (virtual calibrator) was calculated for ech42 using data from all transgenic and control plants, as discussed by Mason et al. (2002).
qRT-PCR reactions were carried out in the same conditions as described earlier.The expression levels were determined using the ∆Ct and 2 ΔΔCt methodology (Bookout, 2003).

Statistical analysis
Experimental data were subjected to Tukey's test (p < 0.05) todetermine the statistical significance of the differences among the means using the Sanest Software (Machado and Zonta, 1995).

Antibiotic effects on explants and shoots in in vitro culture
Sensitivity tests revealed that the Agrobacterium growth was completely inhibited by 500 mg.L -1 cefotaxime but lower concentrations did not prevent its overgrowth; consequently, this dose was used in posterior transformation experiments.
The effect of cefotaxime (500 mg.L -1 ) and increasing kanamycin doses was assessed separately on leaf explants from in vitro and in vivo sources.Cefotaxime seems to stimulate shoots regeneration in both explant kinds.In vitro leaf explant showed higher percentage of shoots production (63%) and shoots per explant (3.9) compared to control (without cefotaxime) in which shoots production was 41.6% and shoots per explant was 1.8.Similarly, the number of shoots per in vivo explants increased from 4.4, in the control, to 13.0 upon cefotaxime addition (Table 1).The phytotoxic level of kanamycin was 100 mg.L -1 in both type of explants with a rate of shoots production almost nil after 30 days of culture: 0.6 shoots per in vitro leaf explant and only 0.2 shoots per in vivo leaf explant.Explants started to bleach after 15 days subculture, resulting in the total loss of chlorophyll (bleaching) in approximately 95 and 84% of in vitro and in vivo leaf explants, respectively (Table 1).After 30 days on kanamycin medium, all shoots became necrotic and started to die.Based on these results, RM was supplemented with 100 mg.L -1 of kanamycin for effective selection of transformed shoots.

Genetic transformation and effect of pre-culture time
EHA strain of A. tumefaciens was efficient for infecting foliar explants of J. curcas superior genotype L4P49  (a) Kan R shoots (kanamycin-resistant shoots): Number of shoots that regenerated on selective medium containing kanamycin (100 mg.L -1 ). (b) Regeneration rate: Kan R shoots/total number of inoculated leaf explants. (c) PCR (+) shoots: Number of independent Kan R shoots (one Kan R shoot from each different explant) with positive result after PCR amplification. (d) Transformation efficiency (%): (number of PCR (+) shoots/total number of inoculated explants) × 100.
using bacterial suspension of 0.8 OD 600 , three days coculture period and suitable selective medium (RM medium supplemented with 500 mg.L -1 cefotaxime and 100 mg.L -1 kanamycin).Leaf explants directly infected with A. tumefaciens (without pre-culture) did not show any transformation and those pre-cultured on RM for two to four days before bacterial infection showed only low regeneration rate with few incompletely developed shoots unsuitable to the transformation efficiency assessment.Pre-culture time larger than six days allowed the onset of callus on the foliar explant edges (Figure 2).The explants pre-cultured on RM for 12 to 15 days prior to cocultivation with A. tumefaciens gave the highest rate of callus formation resulting in larger rate of kanamycinresistant shoot production on selective medium (Table 2).The largest transformation efficiency (50% of explants responded positively to transformation procedure) was observed when explants were pre-cultured on RM for 15 days with an average of seven kanamycin-resistant shoots per explant (Table 2).

Confirmation of transgenic shoots and estimation of transgene copy number
PCR analyses confirmed the presence of a 1.3 kb fragment equivalent to the expected amplicon size of the ech42 gene into independent primary shoots grown on selective medium (Figure 3).
Standard curves of Figure 4 show the reaction efficiency and Ct values of each primer pairs for qRT-PCR assays as well as for calculation of copy number ratios.The evaluated amplification efficiency for the target gene (ech42) and for the endogenous reference gene (JcKASIII) was over 100% (Figure 4).The virtual calibrator (r 1 coefficient) calculated for ech42 as described by Mason et al. (2002) was 0.9.Thus, the estimated ech42 to JcKASIII ratios gave measures of the number of ech42 copies effectively inserted in the J. curcas genome of each transformed plant (Table 3).Single, double and triple copies of the ech42 gene were found in the transgenic lines evaluated.

Measures of gene expression by qRT-PCR
The experimental values of dissociation curves for chitinase and JcKASIII genes overlapped the adjusted regression lines, which demonstrate the reliability of our results.A single peak was observed in the dissociation curves, confirming the amplification of only one fragment   75,37.5,18.75,9.37,and 4.68 ng (Log(DNA)).Reaction efficiency percentages for JcKASIII and chitinase were 112.74% and 105.35%, respectively.with all primers combinations analyzed for both primary shoots #12 and #23.In these two particular cases, it was found out that qRT-PCR allowed the differentiation of transgene expression levels by reference to the actin gene used as a control.Actually, the level of ech42 expression observed in the #12 plant was 2 times larger than that of actin gene whereas it was 168 time larger than this internal control in the # 23 plant (Figure 5).

DISCUSSION
Agrobacterium-mediated transformation requires previous determination of specific antibiotics at suitable concentration to completely inhibit bacterial growth without affecting shoot regeneration.Beta-lactam antibiotics, such as cefotaxime, are the most commonly used antibiotics in plant transformation protocols, since they have a broad spectrum of activity against bacteria and a low toxicity to eukaryotes (Yu et al., 2001).
Although inhibitory effect on shoot regeneration have already been previously reported (Tang et al., 2004;Mendes et al., 2009), stimulatory effects have also been reported for some species such as Zea mays (Danilova and Dolgikh, 2004) and Citrus (Oliveira et al., 2010).Stimulatory effect of antibiotics on plant morphogenesis may be explained by their chemical structure, which may mimic plant growth regulators (Nakano and Mii, 1993), or their degradation by-products, which may generate metabolites with plant growth regulator activity (Mathias and Mukasa, 1987).Several antibiotics such as cefotaxime, augmentin, sporidex, and carbenicillin have been tested in Jatropha transformation procedure using different Agrobacterium strains (Kumar et al., 2010;Misra et al., 2012;Fu et al., 2015).Cefotaxime has been an excellent antimicrobial agent for suppressing Agrobacterium, especially EHA105 and LBA4404 strains (Kumar et al., 2010;Misra et al., 2012;Fu et al., 2015).Even if its inhibitory effect on shoot regeneration has been reported in some studies (Zong et al., 2010), cefotaxime, even at doses as high as 500 mg.L -1 , did not inhibit neither calluses induction nor shoots regeneration on cotyledon explants of J. curcas (Li et al., 2008).The complete elimination of A. tumefaciens EHA105 warrant transformed plant material free of bacterial contamination.In addition, cefotaxime at Franco et al. 2013 500 mg.L -1 was also stimulating the shoot regeneration on in vitro and in vivo leaf.
Considering bacterial concentration, preliminary results indicated that 0.8 OD 600 was the most suitable density for transformation because it promoted the largest number of kanamycin-resistant shoots without having a negative impact on plant cells due to bacterial excess (data not shown).Bacterial inocula with larger OD 600 resulted in a bacterial excess that inhibited callus induction and induced pH changes of the incubation media, which led to a corresponding reduction in transformation efficiency (Yong et al., 2006;Xu et al., 2009;Zong et al., 2010).
The nptII gene that confers kanamycin resistance was used here because of its recognized efficiency as a selectable marker of transformed cells in dicotyledonous plants including J. curcas, which have been found to be hypersensitive to kanamycin (Pan et al., 2010;Kajikawa et al., 2012).According to Tran and Mishra (2015), excessive doses of kanamycin may not only kill untransformed cells, but also inhibit the growth of transformed cells, thereby leading to lower transformation efficiency.Reduction of kanamycin concentration or/and delay selection strategies was necessary for transgenics of almond (Ramesh et al., 2006) and J. curcas (Fu et al., 2015).In other species such as rice (Tran and Mishra, 2015) and some trees (Stevens and Pijut, 2014), kanamycin selection is ineffective because of their natural cell resistance to this antibiotic.
In ) to select J. curcas transformed shoots, we did not find any noticeable kanamycin sensitivity of our J. curcas material at concentrations lower than 75 mg.L -1 independently of explant kind (in vitro or in vivo leaf explants), which suggests that the J. curcas sensitivity to kanamycin is largely depending on its genotype.
Explant pre-culture has been reported to be necessary for transformation of several species with A. tumefaciens (Lawrence and Koundal, 2000;Barik et al., 2005;Xu et al., 2009).Kumar et al. (2010) reported that the transformation rate peaked in J. curcas (18.3%) when leaf explants were pre-cultured in regeneration medium for 4 days.Here, a positive effect of increasing the time of pre-culture was observed on RM until 15 days on kanamycin-resistant shoots production.The pre-culture of explants on media with phytohormones induces mesophyll cells to divide and form callus. Due to their particular dedifferentiation stage typical of highly dividing meristematic cells (Sangwan et al., 1992), these cells are competent for Agrobacterium transformation (Nagl et al., 1997;Guidolin, 2003;Ribas et al., 2011).
qRT-PCR was recently used to assess the copy number of transgenes due to its high sensitivity and also because of its requirement for only very small amounts of genomic DNA (Jha et al., 2011;Casu, 2012;Pinheiro et al., 2014).Transgenic lines showed single, double and triple transgene insertions through qRT-PCR analysis and a difference of gene expression according to the internal control.Interestingly, no positive correlation has been found between copy number and expression of the trait associated to integrated gene(s) (Hobbs et al., 1990).Rather, it has been shown that a larger number of transgene copies resulted in a lower level of their expression (Beltrán et al., 2009;Hadi et al., 2012), while single or double copy integration tends to result in higher expression levels (Flavell, 1994;Pinheiro et al., 2014).
In conclusion, multiple kanamycin-resistant shoots (typically seven per explant) may be produced in 50% of explants using 15 days pre-culture on RM and selection with kanamycin (100 mg.L -1 ).Based on our optimized procedure for the J. curcas accession considered, a chitinase gene was successfully introduced into an elite genotype selected for its high seed and oil yield by the IAC breeding program.Following the complete characterization of the gene insertion pattern as well as biological and field testing, transgenic genotypes will be backcrossed with other valuable genotypes to transfer them the fungal disease resistance trait.Furthermore, it was believed that we succeeded in identifying the key variables that may affect the transformation process according to the elite J. curcas genotype considered.

Figure 2 .
Figure 2. Aspect of explants precultured in regeneration medium (RM) during 6 (a) and 15 days (b).Arrows indicate the onset of callus formation.(c) Shoot regeneration observed after explant co-cultivation with A. tumefaciens.Putative transformed explants were kept on RM supplemented with 100 mg.L -1 of kanamycin.Bar=1 cm.

Figure 5 .
Figure 5. Expression levels of ech42 in two J. curcas transformed shoots: Plant #12 and Plant #23 in relation to the actin gene of J. curcas.

Table 1 .
Effect of antibiotics (cefotaxime and kanamycin) on explants of J. curcas after six weeks of subculture.

Table 2 .
Effect of pre-culture period on shoot regeneration and transformation efficiency.

Table 3 .
Estimated and assumed copy number for chitinase gene in nine PCR + shoots.