A Dark Incubation Period Is Important for Agrobacterium-Mediated Transformation of Mature Internode Explants of Sweet Orange, Grapefruit, Citron, and a Citrange Rootstock

Background Citrus has an extended juvenile phase and trees can take 2–20 years to transition to the adult reproductive phase and produce fruit. For citrus variety development this substantially prolongs the time before adult traits, such as fruit yield and quality, can be evaluated. Methods to transform tissue from mature citrus trees would shorten the evaluation period via the direct production of adult phase transgenic citrus trees. Methodology/Principal Findings Factors important for promoting shoot regeneration from internode explants from adult phase citrus trees were identified and included a dark incubation period and the use of the cytokinin zeatin riboside. Transgenic trees were produced from four citrus types including sweet orange, citron, grapefruit, and a trifoliate hybrid using the identified factors and factor settings. Significance The critical importance of a dark incubation period for shoot regeneration was established. These results confirm previous reports on the feasibility of transforming mature tissue from sweet orange and are the first to document the transformation of mature tissue from grapefruit, citron, and a trifoliate hybrid.


Introduction
Citrus is grown worldwide, is consumed both fresh and processed, and is one of the most economically important fruit crops. Broadly viewed, new citrus rootstock and scion varieties are developed for three general reasons: 1) improve resistance to pest, disease, and environmental problems to which current commercial rootstock and scion cultivars are susceptible; 2) increase fruit/juice yield to improve profitability to growers; and 3) improve fruit/juice quality to remain competitive in the marketplace. The development of new citrus rootstock and scion varieties takes several decades and can span the careers of multiple researchers. One of the reasons for the long development time is an extended juvenile phase in citrus that typically requires at least 5-10 years before flowering. Another reason is that a new selection is tested at multiple field locations, with multiple rootstocks or scions, and over at least four years of fruit production to provide an estimate of productivity.
Plant breeders develop new varieties by introducing genetic variability, generally through controlled hybridizations, into a crop and selecting individuals with useful characteristics. However, there are some serious barriers in citrus biology that must be overcome before progress can be made, and include: the difficulties and expense of working with a tree crop, a long juvenile phase, and many citrus exhibit apomixis and inbreeding depression. These barriers, taken together, make citrus one of the most difficult crops to breed. Genetic engineering is one method to introduce genetic variability into citrus that can potentially address those problems in citrus that have no known viable long-term solution. For example, Huanglongbing (HLB) or citrus greening disease is considered the most serious citrus disease in the world [1]. However, because there is little natural resistance to HLB, it will be difficult to develop resistant varieties through conventional plant breeding methods. Recombinant DNA approaches provide a potential solution for developing citrus trees resistant to HLB and, more generally, many of the serious problems that either threaten or seriously limit citrus production and that have no management or conventional breeding solutions.  Genetic transformation of citrus rootstock and scion varieties is an active area of research and methods to transform citrus have been reported for many of the major citrus types, including sweet orange [2][3][4][5][6][7], mandarin [8], [9], grapefruit [10][11][12][13], lemon [14], [15], lime [16], [17], and important rootstock types [18][19][20]. However, because the primary methods used to transform citrus use juvenile tissue, seedlings or embryogenic cell lines, transgenic citrus trees have the same long juvenile phase as trees grown from seed. To circumvent the juvenile phase requires the transformation of tissue from mature trees.
For plant breeding applications, genetic transformation systems should be 1) sufficiently efficient to produce the required numbers of independent transgenics per transgene to alter the phenotype of the targeted trait(s) while retaining the integrity of the variety and, 2) applicable to a broad range of types, commercial and breeding, within the crop. Our objectives were twofold -first, to determine the effects of factors potentially important (dark incubation, AgNO 3 , GA, basal medium, and growth regulators) for shoot regeneration from mature tissue, a necessary prerequisite for improving the efficiency of mature tissue transformation and, second, to broaden the range of citrus types transformable from mature tissue by including three citrus types not previously transformed including grapefruit, citron, a trifoliate hybrid rootstock, and a high quality and important sweet orange variety (Valencia) not previously reported.

Source Plants and Explant Preparation
Citrus paradisi Macf. Ruby Red (grapefruit), C. sinensis (L.) Osbeck Valencia (sweet orange), C. medica (L.) Etrog (citron), and US-942 [24] rootstock (C. reticulata Sunki6Poncirus trifoliata Flying Dragon) adult phase trees (i.e., had produced flowers) maintained under greenhouse conditions were used as the source of internode explants ( Figure 1). New fully expanded shoots (30 cm long) were used as the source of internodes. After removing the leaves, the shoots were surface sterilized by the following sequence -washed in running water for 5 mins, immersed for 15 min in 5 g/L AlconoxH (Alconox, Inc. White Plains, NY) soapy water, 30 sec in 70% ethanol, 30 min in 1% sodium hypochlorite solution and Tween 20 (3 drops/L), and then three rinses in sterile distilled water [25]. Explants approximately 7 mm in length were prepared from the internodes ( Figure 2A) using sterile garden scissors. Internode explants prepared in this manner were used in the dark incubation, shoot regeneration, and Agrobacterium-mediated transformation experiments.

Dark Incubation Experiment
The effect of a dark incubation period on shoot regeneration from internodes of sweet orange, grapefruit, and US-942 was determined. The experiment was designed as a single factor design. Each internode explant was inserted vertically ( Figure 2B) into a Falcon #351007 60615 mm polystyrene dish (Becton Dickinson Labware, NJ, USA) containing 10 mL shoot induction medium (SIM) composed of MS salts [26] (PhytoTechnology LaboratoriesH, Shawnee Mission, KS, USA), glycine (1 mg/L), thiamine-HCl (1 mg/L), pyridoxine-HCl (1 mg/L), nicotinic acid (1 mg/L), zeatin riboside (10 mM) (Gold Biotechnology, St. Louis, MO, USA), 0.8% (w/v) agar (USB Corporation, OH), and the pH reading adjusted to 5.7 with NaOH. Internode explant cultures from sweet orange, grapefruit, and US-942 were incubated in a dark chamber (27uC) for 0, 1, 2, 3, 4, 6, or 8 weeks, then transferred to 16/8 h photoperiod, 35 mmol m 22 s 21 at 27uC for 21 days, and then data were collected (buds/shoots counted). The experiment was repeated and the response at each treatment point was estimated from 9-17 internode explants per replicate. The number of buds/shoots produced was counted and the data analyzed by one-way ANOVA for each citrus type followed by a one-sample t-test that compared the mean of each treatment to 0 (i.e., no buds/shoots produced).

AgNO 3 -GA-PGR-Basal Media Experiment
Four factors considered important or potentially important in shoot regeneration were examined using US-942 and included AgNO 3 (0 to 5 mg/L), GA (0 to 5 mM), PGR (10 mM ZR or 15 mM BA and 10 mM NAA), and basal medium (MS or WPM). A four-factor response surface design was constructed that included two numeric factors, AgNO 3 and GA, and two categorical factors, PGR and basal medium. Design points were selected using Doptimal criterion, modified to include lack-of-fit points, sufficient to estimate a quadratic polynomial. The experimental design included 13 model points, 5 lack-of-fit points, and 5 points to estimate pure error, i.e. 23 design points. The design points are listed in Table 1. The number of shoots regenerated at each design point was estimated from eight 60615 mm culture dishes (184 total), each containing a single internode explant inserted vertically ( Figure 2B) into SIM modified per the specified treatment. Internode explant cultures were incubated in a dark chamber (27uC) for 2 weeks, then transferred to 16/8 h photoperiod, 35 mmol m 22 s 21 at 27uC for 21 days, and then data were collected (buds/shoots counted). Data were analyzed by analysis of variance (ANOVA).

Agrobacterium-mediated Transformation
Agrobacterium tumefaciens strain EHA-105 strain [27] carrying the p35S GUS INT binary plasmid [28] carrying the marker gene bglucuronidase (GUS) was used in all experiments. Bacterial cultures were made by selecting a single colony from a streaked YEP plate (10 g/L peptone, 5 g/L NaCl, 10 g/L yeast extract, 10 g/L agar, and pH 7), and inoculating a 125 ml flask containing 50 mL of liquid YEP + kanamycin (1 mg/ml) + acetosyringone (40 mM), and culturing on a rotary shaker (225 rpm) at 27uC until the culture achieved an OD 600 of 1.3 to 1.5. The Agrobacterium inoculation solution was prepared by diluting the Agrobacterium culture with SIM to a final OD 600 of 0.5, and then adding acetosyringone to 40 mM.
Approximately one hundred internode explants were prepared at a time. As internode explants were excised they were placed into SIM liquid medium. Once the explants were prepared the SIM medium was discarded and replaced with Agrobacterium inoculation solution for 15 minutes, after which the explants were blotted dry with paper towels. Internode explants were inserted horizontally ( Figure 2B) into 100 x 20 mm culture dishes, 20 explants per dish, containing co-culture medium composed of SIM medium with 40 mM acetosyringone, and incubated at 24uC in the dark for three days.
Following co-culture, explants were removed and triple rinsed with water in a 500 mL glass bottle and then blotted dry with paper towels. Explants were then placed, one explant per dish, into 60 615 mm culture dishes containing SIM +5 mM GA + antibiotics (100 mg/L kanamycin, 250 mg/L vancomycin, and 250 mg/L cefotaxime), incubated in the dark at 27uC for two weeks, transferred to 16   Inc, Tangent, OR). PRO MIX BX soil was steamed at 71uC for 2 h 15 min before use. Rootstocks were grown etiolated at 27uC in the greenhouse for a month. When the rootstock seedlings were approximately 10 cm tall, they were used for micro-grafting with transformed shoots. Before grafting, the seedlings were fed with J Basal Salt Mixture (Murashige & Skoog, PhytoTechnology Laboratories, Shawnee Mission, KS) and also given two eye droplets to each pot of a solution of Sequestrene 138, Iron Chelate Mix, 1.2 g/L (Becker Underwood, Inc., Ames, IA.). GUS assay. Regenerated shoots .2 mm were excised from internode explants and the bottom 1-2 mm of the base of the shoot was excised for assay for GUS activity. The basal sample was placed into a 1.5 microcentrifuge tube containing 200 mL GUS assay buffer [30] and incubated overnight at 37uC. The shoot was inserted into a 100620 mm culture dish containing SIM. Shoots where the basal sample turned blue via the X-gluc histochemical assay were then micrografted.
Micrografting procedure. Regenerated shoots .2 mm were shaped to look like the alphabet ''V'' at the base of the shoot. Rootstocks were cut under the cotyledon and sliced vertically in the middle of the stem with a Double Edge Carbon Steel Breakable Razor Blade (Electron Microscopy Sciences, Hatfield, PA). Trimmed shoots were inserted into the vertical incision. Immediately after grafting, the Cone-tainer pot was covered with a plastic bag and sealed with a rubber band to retain moisture. Grafted plants were incubated for one week in a growth chamber programmed to a 16/8 photoperiod and 27uC and then transferred to the greenhouse.    CAGC39 designed to amplify a 574 bp fragment [31]. PCR was performed in a MJ Research thermocycler and included 34 cycles of denaturation at 94uC for 1 min, annealing at 65uC for 1 min, and extension at 72uC for 2 min. Each sample included 60 ng of DNA and was separated by 100 volt electrophoresis in 1.0% agarose gels, and visualized by ethidium bromide staining. Southern hybridizations. Total DNA was extracted from leaf tissue following the method of Dellaporta [32]. DNA extracts were treated with RNase, phenol-chloroform extracted, and ethanol precipitated; pellets were dissolved in Tris-EDTA and DNA was quantified using a nanodrop spectrophotometer (Thermo Scientific, Wilmington, DE). Each DNA extract was digested with EcoRI; 5 mg of each EcoRI-digested DNA extract was separated by electrophoresis in 0.8% agarose and subsequently transferred to a nylon membrane (Roche Applied Science, Indianapolis, IN) following the manufacturer's instructions. A PCR product was generated using the transformation vector (p35S GUS INT) as template along with the same primers used for confirmation of uidA transgene insertion. The GUS PCR product served as template for the preparation of a DIG-labeled hybridization probe using the DIG DNA Labeling Kit (Roche Applied Science) following the manufacturer's instructions. All subsequent hybridization, stringency washes, and detection steps were carried out using the DIG detection kit (Roche Applied Science) following the manufacturer's instructions.

Dark Incubation Experiment
The effect of a dark incubation period on the number of shoots regenerated was determined for sweet orange, grapefruit, and US-942; citron was not tested. For all three citrus types shoots were only observed with a dark incubation period; no shoots were observed for the 0 week dark treatment. The SIM medium used was based on results from prior experimentation that determined the effects of various cytokinins and auxins on shoot regeneration from juvenile epicotyl explants of sweet orange and grapefruit [33]. Shoots were counted three weeks after the explants were moved from dark to light. An ANOVA was conducted for each citrus type and showed that the effect of a dark incubation period on the number of shoots regenerated was significant (Table 2); the means 6 S.D. are shown in Table 2 and Figure 3. A one sample ttest was used to determine which dark incubation treatments for each citrus type had shoot numbers greater than 0. The one sample t-test and 0 values were used because no shoots were regenerated in the 0 week dark treatment. For sweet orange, 2-4 weeks in the dark produced a significantly greater number of shoots than 0; for grapefruit and US-942 it was 2-8 weeks. The three week treatment was not run with US-942.  The number of shoots was counted three weeks after the explants were moved to light. Shoot number for US-942 ranged from 0 to 9.6 ( Table 1) and indicated that AgNO 3 , GA, PGR, and basal medium affected shoot number (Figure 4). A summary of the ANOVA, lack-of-fit test and three R 2 statistics for quality are presented in Table 3. A reduced quadratic polynomial model was selected (p,0.0001). Data was transformed with a log10 function as the Box Cox plot identified a violation of the normality assumption. Residual and model diagnostics were within acceptable limits. The lack-of-fit test was not significant (p = 0.7198) and indicated that additional variation in the residuals could not be removed with a better model. R 2 , R 2 adj and R 2 pred statistics ranged from 0.98-0.95. Main effects were the primary affecters of shoot regeneration. PGR had the single largest effect on the number of shoots regenerated (F Value = 382; p,0.0001) and was due to the large positive effect of zeatin riboside. The type of basal medium was also highly significant (F Value = 67; p,0.0001), though much less so than PGR, with more shoots regenerated on MS medium than WPM. The two-way interaction effects of GA and PGR with basal medium were significant. Quadratic effects (i.e., curvature of the response) were significant for GA. Based on these results the treatment selected for use in the transformation experiment was MS +5 mM GA +10 mM ZR (replicated design point #9/#23).

Transformation Experiments
Twenty-one transformation experiments were conducted with the four citrus types (Tables 4-7) and resulted in a total of 30 transgenic shoots identified by GUS staining, including 15 US-942, 5 Valencia sweet orange, 9 Etrog citron, and 1 Ruby Red grapefruit. Transformed trees were obtained by micrografting ( Figure 5) from sweet orange, citron and US-942-the grapefruit graft was lost. Transgenic plants obtained from micrografting were reconfirmed by the expression of GUS in leaf discs ( Figure 5), PCR amplification of the uidA gene fragment and Southern blot analysis using a uidA gene fragment probe ( Figure 6).

Discussion
Because shoot regeneration is an essential prerequisite for Agrobacterium-mediated transformation experiments, the basic approach was to first identify factors important for promoting shoot regeneration from internode explants from adult phase citrus trees, and then to produce transgenic trees using the identified factors and factor levels for inducing shoots. To a large extent, the citrus types and numbers of explants used in the various experiments was subject to the type and quantity of plant material available at the time. Internode explants were inserted vertically for the shoot regeneration experiments for convenience; because shoots only developed on the upper surface, their formation and development were more easily monitored than explants placed horizontally. Horizontal placement was required for transformation experiments to expose tissue to the selective antibiotic kanamycin.
Extensive preliminary experiments on the effects of various growth regulators on shoot regeneration resulted in a small number of internode explants developing small shoot primordia that never developed into shoots (unpublished data). Because these  experiments utilized only a 16/8 photoperiod (i.e., no dark incubation), the effect of a dark incubation period was considered. For juvenile citrus tissue, a dark incubation period enhanced shoot regeneration in sweet orange [34] and lime [16], but was inhibitory in citrange [35]. For mature citrus tissue, the effect of a dark incubation period has not been determined. Of the five studies that report transformation of mature citrus tissue, four included a dark incubation period [9], [21][22][23] as part of the protocol. The single study that did not include a dark incubation period [4] utilized adventitious buds that formed directly from the wounded regions on seedlings where the growing tip and axillary buds were removed. Of the four studies that report shoot regeneration from mature citrus tissue, three include dark incubation [36][37][38] and one did not [39]. Though a dark incubation period is often included in transformation and shoot regeneration protocols, the periods are variable and range from 15 [21] to 50 days [38]. These results suggested genotype or condition-specific effects and, therefore the importance of testing a dark incubation period under local conditions when initiating a citrus tissue culture project. We observed that a dark incubation period was essential for shoot regeneration from sweet orange, grapefruit, and US-942; citron was not tested. When internode explants were incubated in the dark, they developed numerous buds and shoots; no shoots were regenerated from any of the citrus types without a dark incubation period. We also observed that shoot regeneration generally declined after a 2 week dark period, but this may not be true for transformation. Cervera et al. [9] extended the dark incubation period from 2-4 weeks to 5-6 weeks to promote more callus formation, with the intent of improving transformation efficiency; however, the specific effect of this change on transformation was not determined. The cytokinin zeatin riboside was used in these dark experiments as it was the most effective cytokinin for shoot regeneration from juvenile epicotyl explants of 'Valencia' sweet orange and worked well for 'Duncan' grapefruit [33]; it was selected on the assumption that it might also work well with mature tissue. To the best of our knowledge, all citrus mature tissue shoot regeneration and transformation experiments utilize BAP alone or in combination with NAA to induce shoots. To further improve shoot regeneration, the individual and combined effects of four additional factors identified as potentially important were determined. The protocol for this experiment included a two week dark incubation period per the results from the dark incubation study. AgNO 3 , an ethylene inhibitor, was selected because it can sometimes enhance shoot regeneration [40][41]. Because its effect on shoot regeneration was weak, AgNO 3 was not included in the shoot regeneration medium used for transformation. GA was selected as a potential enhancer of shoot regeneration and growth [36], [43][44][45] and, because its effect was small but positive on shoot regeneration, GA was included in the shoot regeneration medium used for transformation. BA + NAA was compared to ZR as BA + NAA are two growth regulators commonly used together to regenerate shoots from mature citrus explants [9], [22], [23], [46]. MS and WPM were selected as they have both been used in citrus tissue culture [9], [36], [46]. The use of ZR was the most important factor that affected shoot regeneration and indicates the importance of using the appropriate cytokinin. The level of ZR used in this experiment was 10 mM, a level found to be useful for juvenile explants [33] and therefore potentially not optimal for mature internode shoot regeneration. In addition, because the growth regulator treatments were fixed concentrations, their effects on mature tissue shoot regeneration over a range of proportions and concentrations could not be determined. To fully explore these effects would require a 3-component (ZR-BA-NAA) mixtureamount experimental design. Given the large effect of ZR, additional research to optimize growth regulator types, proportions, and concentration for mature tissue may provide further improvements. This argument also applies to the comparison of MS and WPM where little is learned, apart from which formulation works better, about the proportions and concentrations of specific mineral nutrients and their effects on shoot regeneration.
Transgenic plants were obtained by Agrobacterium-mediated transformation of internode explants from mature greenhouse grown trees of sweet orange, grapefruit, citron, and a trifoliate hybrid rootstock. The results confirm previous reports on the feasibility of transforming mature tissue from sweet orange and are the first to document the transformation of mature tissue from grapefruit, citron, and a trifoliate hybrid. Variability was high and was presumably due to the various relatively uncontrolled environmental and physiological effects. For example, the transformation experiments were conducted over an eight-month period where plant material was harvested from greenhouse-grown trees exposed to light levels and seasonal temperatures that varied widely. Also, the type and condition of the shoot flush can affect regeneration frequencies [21]; internode explants were taken from stems harvested from trees from which shoot flushes were previously harvested. These factors and their interaction on transformation efficiency are essentially unknown, not accounted for, and therefore part of the observed variability.
Transformation efficiency was greatest for US-942. This may be because trifoliate hybrids are some of the most transformable citrus types. However, the AgNO 3 -GA-PGR-Basal medium experiment was conducted only on US-942 and the identified conditions for good shoot regeneration were then applied to all the citrus types. To what extent this effect was due to the inherent responsiveness of trifoliate hybrids vs. using conditions specifically optimized for US-942 is unknown.
In this study we have quantified the effects of a dark incubation period and the medium components AgNO 3 , GA, type of PGR, and basal medium on shoot regeneration from internode explants from adult phase greenhouse-grown citrus trees. Under our conditions, a dark incubation period was essential and GA, zeatin riboside, and MS basal medium enhanced shoot regeneration.
The assumption that treatments that enhanced shoot regeneration would also enhance transformation was made and four diverse citrus were transformed using a single protocol. Because transgenic plants were obtained from these four citrus types, with little control of the initial physiological status of the trees, suggests 1) the potential importance of optimizing protocols to local conditions and, 2) that further improvements in transformation efficiencies across a broad range of citrus types are likely.