• Open Access

Marker-free site-specific gene integration in rice based on the use of two recombination systems


  • Soumen Nandy,

    1. Department of Crop, Soil & Environmental Sciences, University of Arkansas, Fayetteville, AR, USA
    Search for more papers by this author
  • Vibha Srivastava

    Corresponding author
    1. Department of Crop, Soil & Environmental Sciences, University of Arkansas, Fayetteville, AR, USA
    2. Department of Horticulture, University of Arkansas, Fayetteville, AR, USA
    Search for more papers by this author

(Tel 479 575 4872; fax 479-575-7465; email vibhas@uark.edu)


Transgene integration mediated by heterologous site-specific recombination (SSR) systems into the dedicated genomic sites has been demonstrated in a few different plant species. This approach of plant transformation generates a precise site-specific integration (SSI) structure consisting of a single copy of the transgene construct. As a result, stable transgene expression correlated with promoter strength and gene copy number is observed among independent transgenic lines and faithfully transmitted through subsequent generations. Site-specific integration approaches use selectable marker genes, removal of which is necessary for the implementation of this approach as a biotechnology application. As SSR systems are also excellent tools for excising marker genes from transgene locus, a molecular strategy involving gene integration followed by marker excision, each mediated by a distinct recombination system, was earlier proposed. Experimental validation of this approach is the focus of this work. Using FLPe-FRT system for site-specific gene integration and heat-inducible Cre-lox for marker gene excision, marker-free SSI lines were developed in the first generation itself. More importantly, progeny derived from these lines inherited the marker-free locus, indicating efficient germinal transmission. Finally, as the transgene expression from SSI locus was not altered upon marker excision, this method is suitable for streamlining the production of marker-free SSI lines.


Site-specific recombination (SSR) systems carryout recombination between specific DNA sequences without addition or deletion of nucleotides (Van Duyne, 2001). This precision is highly desirable in plant genetic engineering as it minimizes unintended mutations introduced by the transformation process. Most SSR systems are fairly simple involving a single recombinase protein that catalyses recombination between short DNA sequences (Hirano et al., 2011). In addition, these systems are functional in a variety of plant species, making them attractive tools for genetic engineering (Darbani et al., 2007; Gidoni et al., 2008; Ow, 2011; Srivastava and Gidoni, 2010). Although several SSR systems are under development, some efficient systems are already available such as P1 phage Cre-lox and yeast FLP-FRT systems. These systems have been widely used for (a) removing antibiotic selection marker genes from transgene locus (marker excision) and (b) integrating transgene into the predetermined genomic site (site-specific integration or SSI) (Gidoni et al., 2008; Srivastava and Gidoni, 2010; Tuteja et al., 2012). While marker excision technology is widely practiced, SSR-mediated gene integration technology is not being utilized in spite of its distinct ability to generate precise SSI locus and consistent transgene expression (Akbudak et al., 2010; Albert et al., 1995; Chawla et al., 2006; Day et al., 2000; Nanto et al., 2009; Srivastava et al., 2004; Vergunst et al., 1998). One of the reasons for its inapplicability so far is the difficulty in removing marker genes from the SSI locus and recovering marker-free SSI (MF-SSI) lines (Ow, 2002). Marker excision combined with site-specific gene integration requires the use of two SSR systems (Darbani et al., 2007; Fladung and Becker, 2010; Nanto and Ebinuma, 2008; Srivastava and Ow, 2004), and for a long time, only Cre-lox system was popularly used (Gilbertson, 2003). Recent studies have validated additional SSR systems for plant biotechnology, opening the doors for MF-SSI technology (Fladung et al., 2010; Moon et al., 2011; Nandy and Srivastava, 2011; Nanto et al., 2005; Thomson et al., 2009, 2010, 2012; Zhou et al., 2012). The feasibility of MF-SSI has been demonstrated in two previous studies (Akbudak and Srivastava, 2011; Nanto and Ebinuma, 2008). Nanto and Ebinuma (2008) used a combination of R-RS and Cre-lox for transgene integration and marker excision, respectively, for generating MF-SSI plants of tobacco. Using rice callus cultures, Akbudak and Srivastava (2011) showed that SSI locus developed by Cre-lox system can be efficiently converted to MF-SSI locus by employing FLP-FRT system for marker excision. The critical component of this study was the use of thermostable FLP recombinase, FLPe, rather than the wild-type FLP. In both studies, however, the excision recombinase (Cre or FLPe) was introduced by retransformation of SSI lines, which relies on the use of selectable marker genes, thus adding extra steps for removing this newly added marker gene. Marker excisions can also be initiated by manual crosses; however, that would also entail a time-consuming multi-generation screening for developing the final product.

In this study, we developed and validated a strategy for generating first-generation (T0) MF-SSI plants using rice as the model crop. Earlier, we developed a method for transgene integration based on the use of FLPe-mediated FRT × FRT recombination to insert transgene construct into the rice genome. Here, we combined this method with Cre-lox-mediated marker gene excision method to produce marker-free SSI plants. The process is efficient as a number of SSI lines were developed in each attempt and converted to MF-SSI at a rate of ≥50%. Furthermore, expression of the transgene was not altered by marker excision. Therefore, the strategy and the protocol of marker-free SSI described in this study are applicable for crop biotechnology.


Molecular strategy

The goal of the strategy is to incorporate two independent applications of SSR systems into a single platform: (i) transgene insertion into a predetermined genomic locus (SSI) and (ii) marker gene excision from the integration locus (site-specific excision). For this purpose, transgenic rice lines containing a single copy of pNS5 construct were used (Figure 1a). These target lines were developed in the previous study (Nandy and Srivastava, 2011). Briefly, it contains an FRT target site for FLPe-mediated transgene integration and two oppositely oriented loxP sites on either ends of the construct for marker excision from the SSI locus. Target line cells were transformed by co-bombardment of donor construct, pNS7, and pUbiFLPe. Co-bombardment of FLPe gene is a critical step in FLP-FRT-mediated SSI in rice genome as the FLP gene located in pNS5 was found to be nonfunctional (Nandy and Srivastava, 2011). The donor construct pNS7 contains a gene cassette consisting of promoterless marker gene (NPT) and the gene-of-interest (GUS) flanked by two FRT sites (Figure 1b). FLPe-mediated recombination between these two FRT sites in pNS7 will generate a vector-backbone-free donor circle containing a single FRT (Figure 1c). FLPe-mediated integration of the donor circle will generate a precise co-integration structure (SSI) consisting of transcriptional fusion of the selectable marker gene (NPT) with the Ubi promoter in the genomic target site (Figure 1d). As GUS gene in pNS7 is flanked with two oppositely oriented loxP sites, the SSI structure (SSI) will contain four loxP sites that flank marker genes (see Figure 1d: excision fragments 1 & 2). Further, as heat-inducible cre gene is located within the locus, loxP-flanked genes can be removed by inducing Cre activity at any time subsequent to selection of the transgenic clones. Cre activity is expected to remove the excision fragments on either end of GUS gene and generate a marker-free SSI structure (Figure 1e).

Figure 1.

 Design of DNA constructs and molecular strategy of marker-free site-specific gene integration. (a) pNS5 target construct contains three transcription units flanked by loxP sites (open triangles): hygromycin phosphotransferase (HPT) gene driven by CaMV 35S promoter, cre gene driven by soybean heat shock (Gmhsp17.5E) promoter (HSP) and FLP gene driven by maize ubiquitin (Ubi) promoter. Target FRTL site (shaded triangle) is incorporated into the leader sequence of FLP gene. RB and LB represent T-DNA borders. (b) Donor vector pNS7 contains a promoterless neomycin phosphotransferase II gene (NPT) and β-glucuronidase gene (GUS) driven by Ubi promoter between FRTR and FRT sites (shaded triangle) in the pBluescript SK backbone (grey line), and the GUS gene is flanked by loxP sites (open triangles). (c) The structure of donor circle that is generated from pNS7 upon FLP-mediated FRTR × FRT recombination. The circle contains the gene construct and FRTR without the vector backbone. (d) Structure of site-specific integration (SSI) locus. Integration of donor circle via FRTL × FRTR recombination traps Ubi promoter to activate NPT gene and forms a unique FRTL+R site along with FRT site (shaded triangles). Site-specific integration structure contains four loxP sites (open triangles) that flank two separate fragments destined for excision (Excision Fragments 1 and 2). (e) Structure of marker-free SSI (MF-SSI) locus. When cre gene in the SSI locus is induced, the loxP-flanked fragments (excision fragment 1 and excision fragment 2) are removed on either side of GUS gene, and marker-free SSI structure is generated. BamHI (B) and EcoRV (R) sites and sizes (in kb) of intervening fragments are shown. Dashed arrows represent transgene–chromosome junction fragments. Shaded bars represent the DNA probe fragments (probe 1–4) used for Southern hybridization. PCR primers (u, f, n, g and r) are indicated. Each transcription unit contains a transcription termination signal of nopaline synthase gene.

Development of site-specific integration lines

Nine transgenic events, based on geneticin™ selection, were obtained by the bombardment of callus derived from two target lines, A and B, with an equimolar mixture of pNS7 and pUbiFLPe consisting of ∼5 μg of each plasmid. Each of these lines was subjected to PCR and Southern blot analysis to determine SSI structure and copy number. EcoRV digestion of the genomic DNA is expected to generate a unique 6.8 kb fragment from the SSI locus, which can be detected on Southern blot with probe 1 (see Figure 1d). With target line A, only two SSI lines were obtained, each of which displayed multi-copy (MC) integration pattern. With target line B, 7 SSI lines were obtained, four of which contained only 6.8 kb band representing single-copy (SC) SSI locus, while the remaining three contained multi-copy (MC) integration pattern (Table 1; Figure S1). Thus, SC lines were produced at the efficiency of 0% and 22%, respectively (Table 1). It may be possible to improve these efficiencies by bombarding a lower amount of plasmid DNA, which results in the integration of a fewer copies of the introduced DNA (Lowe et al., 2009). However, as the selection of SSI lines is based on maintaining Cre activity below a threshold level, it is possible that a higher basal activity in target line A contributed to null efficiency. All SSI lines lacked the co-bombarded FLPe gene (data not shown), indicating exclusion of FLPe expression from stable SSI events (described earlier by Nandy and Srivastava, 2011). It should be noted that mostly monoallelic integrations occurs in the target locus, as a result, SSI lines may also contain an intact target allele.

Table 1.   Generation of site-specific integration lines
Exp. no.TargetNo. of platesGenR linesPrecise SSI lines*Copy numberEfficiency
Single (SC)Multiple (MC)
  1. *Based on PCR for SSI junctions.

  2. Based on EcoRV Southern blot shown in Figure S1.

  3. No. of SC lines/total no. of plates.


Induction of marker excision

Four SC lines (B1, B6, B8 and B9) were chosen for marker excision by heat treatment. The SC lines are appropriate for site-specific marker excision, as only predictable loxP × loxP recombination can occur in these lines. The MC lines, on the other hand, contain two to three copies of the transgene construct; therefore, Cre-mediated recombination in these lines could induce chromosomal rearrangements upon recombination between multiple transgene copies. Two- to 3-week-old regenerated shoots growing in MS/2 media (rooting media) in glass tubes were transferred to 42 °C incubator for 3–4 h to induce Cre activity as described by others (Herzog et al., 2012; Wang et al., 2005; Zhang et al., 2003). The tubes were then transferred to room temperature in full-spectrum light for 2 days before repeating the heat treatment one more time. After 2 weeks of recovery at room temperature, these plants were transplanted into soil and transferred to greenhouse, where they grew for 2–3 weeks before harvesting leaves for the isolation of genomic DNA.

Genomic DNA of each heat-treated plant was first subjected to PCR for marker excision analysis (Figure 1d: primers u, n, g and f) and formation of the distinct marker-free SSI junction (Figure 1e: primers g and r). Both young (4- to 6-cm-long shoots) and old regenerated shoots (8- to 10-cm-long shoots) were subjected to heat treatment as described previously. PCR analysis indicated that marker excision was more efficient on young shoots as no amplification of fragment 1 (NPT) or fragment 2 (FLP) was detected in nine of ten young plants, indicating complete excision (Figure 2a; Table 2). Heat treatment of ‘old’ shoots was effective in removing both fragments in three of ten cases, while in the remaining cases, one of the fragments was still present (partial excision). The nontreated controls, on the other hand, showed the amplification of the expected NPT and FLP fragments (Figure 2a: Frag. 1 and Frag. 2). However, the marker-free SSI footprint (GUS::RB) was detected by PCR in almost all lines including the nontreated controls, indicating background excisions because of the basal expression of HSP::Cre gene (Figure 2a; Table 2).

Figure 2.

 Molecular analysis of marker excision in T0 plants. (a) PCR on genomic DNA isolated from room temperature control (R) or heat-treated young (y) or old (o) regenerated shoots of site-specific integration (SSI) plants (1–20). Excision of fragments 1 and 2 (see Figure 1d) and presence of excision footprint (GUS::RB) were detected using specific primer pairs (in parentheses). The position of each primer in SSI locus is shown in Figure 1d. PCR with target line B and nontemplate control (NTC) is also shown. (b–d) Genomic DNA of one representative heat-treated (H) T0 plant of each line (B1, B6, B8 and B9) analysed along with its room temperature (R) control on EcoRV (b,c) or BamHI (d) Southern blot using specific probe indicated on each blot. Excision of loxP-flanked fragments in SSI structure is indicated in these blot; in addition, excision footprints (LB::GUS::RB) are detected on EcoRV blot with probe 3 (c) and BamHI blot with probe 2 (d).

Table 2.   Marker excision analysis in T0 plants
Site-specific integration (SSI) lineT0 plantPlant agePCR analysis*ExcisionSouthern analysisProgeny analysis
Frag. 1Frag. 2MF-SSI
  1. *PCR using u-n, g-f and g-r primers (see Figure 1d) for Frag. 1, Frag. 2 and MF-SSI, respectively.

  2. Southern blot indicated incomplete excision (see Figure 2b).


Molecular confirmation of marker excision

Based on PCR screening, ten plants were selected for Southern analysis of the SSI locus (see Table 2). Genomic DNA of each of these plants was digested with EcoRV and hybridized with probe 1 (NPT) and probe 3 (GUS) to analyse the excision of the two fragments and formation of the marker-free-SSI locus. No hybridization with NPT probe was observed in the heat-treated (H) plants except in B8 plants, which showed a faint band in heat-treated sample and a strong 6.8-kb band in the nontreated or room temperature (R) controls (Figures 2b and S2a). Absence of NPT fragment indicates excision of fragment 1. In nontreated controls, an additional 5.3-kb band was sometimes visible on Southern blot (see Figures 2b and S2a), which is the predicted size of EcoRV fragment (spanning NPT gene) if it contains an inverted copy of GUS gene. Inversion of Ubi::GUS cassette is expected, as it is flanked by oppositely oriented loxP sites. Next, the blot was hybridized with probe 3 (GUS) to study the excision of fragment 2 and formation of the unique junction (GUS::RB) in the marker-free SSI (MF-SSI) locus. All heat-treated plants either completely lacked or contained a greatly reduced dose of fragment 2 (loxP-flanked FLP gene) as indicated by 1.7 kb band intensity (Figures 2c and S2b). The nontreated controls on the other hand displayed 1.7-kb band, which is consistent with the presence of marker gene in the SSI structure. The excision of fragment 2 is accompanied with the formation of GUS::RB junction (excision footprint), which could be present in two orientations (forward and inverse) because of the presence of two oppositely oriented loxP sites. Thus, two excision footprints of ∼5 and ∼6 kb were observed upon GUS hybridization (Figures 2c and S2b). These excision footprints are also visible in nontreated controls indicating basal HSP::Cre activity. Excision of fragments 1 and 2 from SSI locus is expected to generate a unique LB::GUS::RB excision footprint, which was detected on BamHI Southern blot (Figure 2d). As BamHI cuts outside of MF-SSI locus, a single excision footprint is expected (see Figure 1e). Hybridization of BamHI blot with probe 2 (Ubi) generated a ≥15 kb excision footprint in each heat-treated plant line, while the SSI-specific 5.2-kb band was missing or reduced in intensity, indicating the excision of fragment 1 (Figure 2d). Of all the tested plant lines, only #19 showed a pattern indicative of incomplete excision, as a minor 5.2-kb band was detected (Figure 2d), while the remaining plants displayed a pattern of complete excision. Thus, Southern data confirmed the PCR results except for line #19 in which no amplification for NPT fragment was observed in PCR, while Southern blot indicated incomplete NPT excision (reduced band intensity). This discrepancy may arise from the use of separate DNA samples in the two techniques. Each SSI line may also contain an intact target allele (because of monoallelic insertion into the target locus); therefore, an additional 3.8-kb band is expected in these lines (see Figure 1a). This 3.8-kb band was observed in all plant lines except plant #20 (Figure 2d), suggesting the origin of plant #20 from a hemizygous target cell. In the nontreated controls, the presence of fragments 1 and 2 in SSI locus is revealed by the appearance of a strong 5.2-kb band.

In summary, Southern analysis indicated efficient excision of both fragments flanking GUS gene upon heat treatment of the regenerated T0 plants derived from four different SSI lines. Next, the inheritance of marker-free SSI locus was studied by analysing progeny of these plant lines.

Inheritance of marker-free site-specific integration locus

Five T0 plant lines, two each of B1 and B6 and a single line of B8, were chosen for progeny analysis (see Table 2). The B9 line could not be included in this analysis as plant #20 failed to set seeds. Twenty to forty T1 seedlings derived from each of the self-fertilized plants were grown in the greenhouse, and their genomic DNA was tested by PCR for the presence of marker-free SSI locus (Figure 3a). Progeny of each T0 plant (T0–2, 7, 9 and 14) that showed complete marker excision on Southern blot lacked fragments 1 and 2 in SSI locus (Table 3), indicating efficient excisions in B1 and B6 lines. Progeny of B8 line, on the other hand, occasionally contained fragment 1 or 2, which is consistent with the incomplete excision in T0–19 plant (Figure 2b–d). A high percentage of T1 progeny (three-quarter of the population based on genetic segregation ratio) from each T0 line is expected to contain marker-free SSI locus, which was confirmed by PCR for excision footprint (GUS::RB) (Figure 3a). Additionally, absence of fragments 1 and 2 in these plants was confirmed by PCR using primers u-n and g-f, respectively (Figure 3a). However, as parent T0 plants contain hemizygous SSI, and therefore an intact target allele consisting of cre gene, MF-SSI formation could also occur de novo in T1 plants (somatic excisions). Therefore, isolation of T1 lines that segregated from target allele (stable MF-SSI lines) was necessary to confirm inheritance of the MF-SSI locus and rule out somatic excision. Such stable T1 lines were easily identified in each population consisting of a small number of T1 individuals (Table 3). Using PCR with primers u-f, T1 lines lacking intact target allele were identified in each T1 pool (Figure 3a). Each of these stable MF-SSI lines also lacked HSP::cre and 35S::hpt genes (data not shown), indicating full-length excision of fragment 1. Next, a few stable T1 lines were analysed on Southern blot (Figure 3b–c). Presence of marker-free SSI was confirmed by hybridization with probe 2 (Ubi) on BamHI Southern blot. Each T1 line contained MF-SSI locus and also segregated from the target allele, which is indicated by the absence of 5.2 and 3.8 kb bands, respectively. Instead, these lines contain ∼15 kb excision footprint, indicating germinal transmittance of the MF-SSI locus (Figure 3b). The locus structure was further verified on EcoRV Southern blot using three different probes (Figure 3c). Each T1 plant contained the MF-SSI locus in one of the two expected configurations: the original or the inverted copy of the GUS gene, while fragments 1 and 2 were missing (see Figure 3c: probe 1, 3). Absence of the target allele that contains an active cre gene in these lines was verified by hybridization with probe 4. Each of the stable T1 lines lacked 3.5 kb band expected from the intact target allele (see Figure 1a). Thus, a number of stable MF-SSI lines were identified in a small pool of T1 progeny from each heat-treated line.

Figure 3.

 Molecular analysis of T1 progeny. (a) A representative PCR analysis on T1 progeny derived from the selected T0 plants (see Table 2). Genomic DNA was subjected to PCR with specific primer pairs to detect the presence of excision footprint, target allele and excision fragments (fragments 1 and 2) as shown in Figure 1a,d. T1 lines positive for excision footprint, but negative for fragments 1 and 2, and target allele (boxed lanes) represent stable marker-free site-specific (MF-SSI) lines. (b–c) Southern hybridization using BamHI-digested (b) or EcoRV-digested (c) genomic DNA of stable MF-SSI T1 lines (1–4) from heat-treated T0 plants (B1-T0–7 and B6-T0–8) or from room temperature (R) controls. Absence of 5.2 and 3.8 kb bands in BamHI blot and presence of ≥15 kb band indicate the presence of stable MF-SSI locus (b). Similarly, in EcoRV blot, absence of bands upon hybridization with probe 1 and probe 4 and presence of 5–6 kb MF-SSI bands indicate germinal inheritance of MF-SSI locus in T1 lines. T1 line would inherit MF-SSI locus in one of the two possible orientations, generating either 5 or 6 kb band upon hybridization with probe 3 (c).

Table 3.   Progeny analysis
Parent plantNo. of T1 linesPCR analysis of T1 lines*Number of stable MF-SSI T1 lines
Frag. 1Frag. 2MF-site-specific integration (SSI) locusTarget allele
  1. *Excision and genetic segregation analysis for the SSI and target allele.

  2. Indicating the presence of target allele, that is, an active cre gene.

  3. MF-SSI lines that lack target allele (active cre gene).


No alteration in gene expression after marker excision

Transgene expression is often influenced by the genomic location. The underlying control elements could include neighbouring promoters, enhancers or suppressor elements. Even the orientation of transgene on the chromosome (forward or inverse) could affect its expression (Feng et al., 2001). The MF-SSI locus is significantly different from the marker-containing SSI locus, as multiple transcription units consisting of marker genes and recombinase genes are removed. To analyse the effect of their removal on transgene expression, GUS activity was determined in stable MF-SSI lines derived from B1, B6 and B8 and compared with the respective control plants, which contain marker genes and recombinase genes in the SSI locus (non-treated T0 plants). As B1, B6 and B8 are isogenic lines, GUS activity in these lines is expected to be similar. However, GUS activity in T1 population may vary depending on the allelic dosage of GUS gene (Chawla et al., 2006; Srivastava et al., 2004). GUS assay revealed that the three SSI lines, T0–7, T0–14 and T0–19, contained similar GUS activity (Table 4). Additionally, all MF-SSI T1 lines contained GUS activity within twofold range (Table 4). Some T1 lines displayed significantly higher GUS activity (for example B1-T1–19) compared with their siblings, displaying a possible allelic gene dosage effect. However, no effect of marker removal was observed on GUS activity, indicating stability of transgene expression and precise conversion of the locus by Cre-lox recombination.

Table 4.   GUS analysis
Site-specific integration linePlant linesGUS activity (nmol MU/min/mg)
B1T0–739.5 ± 3.5
T1–842.5 ± 4.5
T1–1034.0 ± 3.0
T1–1952.5 ± 4.5
T1–2033.0 ± 3.0
T1–2445.0 ± 6.0
B6T0–1437.0 ± 4.0
T1–728.0 ± 3.0
T1–1148.5 ± 6.5
T1–1947.5 ± 4.5
T1–3029.5 ± 4.5
T1–4048.0 ± 5.0
B8T0–1940.0 ± 2.0
T1–137.0 ± 2.0
T1–242.0 ± 2.0
T1–629.5 ± 1.5
T1–2033.5 ± 2.5

Experimental procedures

Plasmid constructs

Target construct, pNS5, is shown in Figure 1a and described earlier (Nandy and Srivastava, 2011). The donor construct pNS7 was made in this work and shown in Figure 1b. It consists of a promoterless neomycin phosphotransferase gene (NPT II) followed by a GUS gene driven by Ubi promoter. The whole construct is flanked by FRT and FRTR (Figure 1b) and cloned in pBluescript SK backbone. The GUS gene cassette is flanked with oppositely oriented loxP sites. All genes contain transcription termination sequence of nopaline synthase gene (nos 3′). Sequences of FRT sites are described in Senecoff et al. (1988).

Rice transformation

Rice tissue culture and media as described by Nishimura et al. (2006) were used in this work. Scutellar calluses of target lines A and B (Nipponbare background) were generated from mature seeds of target lines—A and B (see Nandy and Srivastava, 2011), which contain a single copy of pNS5 construct. Calluses were induced and maintained on N6D media supplemented with hygromycin (50 mg/L) and kept in the dark at room temperature. Two- to 3-week-old cultures were bombarded with 1-μm gold particles coated with an equimolar mixture of pUbiFLPe and pNS7 (∼5 μg of each plasmid) using standard protocol in PDS 1000/He gene gun (Bio-Rad Inc., Hercules, CA). pUbiFLPe is described earlier (Nandy and Srivastava, 2011). The callus was pretreated on N6D media containing 0.4 m sorbitol for 2 h. Bombarded callus was kept for a week on N6D media and then transferred to the selection media, N6D media containing geneticin™ (100 mg/L) (Invirtogen Inc.). Resistant colonies, scored 2- to 4-week later, were transferred to regeneration media as described by Nishimura et al. (2006).

Molecular analysis

Genomic DNA was isolated from leaves of the plants and subjected to PCR with following primers: u (5′-TCTACTTCTGTTCATGTTTGTG-3′) and n (5′-CTCGATGCGATGTTTCGCTT-3′) to detect the excision fragment 1; g (5′-CACCATCGTCGGCTACAG-3′) and f (5′-CTCAGTGATCTCCCAGATG-3′) to detect the excision fragment 2; and g (5′-CACCATCGTCGGCTACAG-3′) and r (5′-AAACGACAATCTGATCCAAG-3′) to detect the excision locus. All PCRs consisted of 40 cycles of 1 min denaturation at 94 °C, 1 min annealing at 56 °C and 1 min extension at 72 °C followed by final elongation step at 72 °C for 15 min. For Southern hybridization, genomic DNA was digested with EcoRV and BamHI, run on 0.8% agarose gel, transferred to a nylon membrane and hybridized with the P32-labelled DNA probes.

GUS expression analysis

β-Glucuronidase (GUS) activity was detected in the transformed leaves. The GUS assay was performed according to the manufacturer’s protocol using FluorAce™ kit (Bio-Rad Inc).


We tested a strategy for marker-free site-specific gene integration (MF-SSI) in rice based on the use of two recombination systems: FLPe-FRT for site-specific transgene integration and heat-inducible Cre-lox for marker excision (Figure 1). By employing the well-known heat-inducible Cre-lox system (based on the use of soybean heat shock 17.5E gene promoter described by Czarnecka et al., 1992), it was possible to generate marker-free first-generation (T0) transgenic plants that transmit stable MF-SSI locus to the next generation at high rate. The heat shock Cre gene has been reported to function properly in maize, rice and banana (Chong-Pérez et al., 2012; Khattri et al., 2011; Zhang et al., 2003). The strategy involved a single FRT × FRT recombination catalysed by the improved FLPe recombinase (Buchholz et al., 1998), resulting in the integration of the defined gene segment into the dedicated genomic site (Nandy and Srivastava, 2011). Transient FLPe activity, generated by the co-bombardment of FLPe gene, was sufficient to generate SSI structure. Further, the selected SSI lines do not contain FLPe gene because the SSI structure is unstable in the presence of FLPe activity (data not shown, Nandy and Srivastava, 2011). Exclusion of FLPe integration is highly desirable. However, if partial or silenced FLPe copies were integrated, they would be removed through back-crosses—a standard approach for cleaning up the genetic background. To initiate marker excision from the SSI structure, a pair of loxP sites was strategically placed in the target construct, pNS5, and donor construct, pNS7 (see Figure 1a,b). As a result, DNA fragments including marker genes in the SSI locus are flanked by loxP sites, and therefore primed for excision by Cre activity (see Figure 1d). However, in future, a simpler approach for marker excision can be developed that involves placement of all markers in a single fragment. Thus, only two loxP sites, one on each construct (target and donor construct), are sufficient to generate MF-SSI locus. It should be noted that FLP gene in pNS5 is dispensable, and therefore, future target locus could contain only a single loxP site to generate lox-flanked marker genes in SSI locus. Similarly, second loxP site in the donor construct (at the right end) could be removed, which will also fix the chromosomal orientation of the transgene.

Although marker excision process is suitable only for the single-copy SSI lines, two multi-copy SSI lines (A1 and B10) were also subjected to marker excision in this study (Figure S3). Each of these lines contains 2–3 copies of the donor construct including a SSI locus (Figure S1). If extra integrations in these lines are linked to the SSI locus, marker excision process might remove these extra copies without incorporating gross genetic aberrations such as chromosomal deletions, inversions or exchange, which would be expected if the these copies were integrated in a distant locus. Efficient deletion of linked transgene copies and generation of single-copy transgene locus using Cre-lox system has been demonstrated earlier (De Buck et al., 2007; Kumar and Thompson, 2009; Moore and Srivastava, 2006; Srivastava et al., 1999). Heat treatment of A1 and B10 plants resulted in the excision of both flanking fragments (Frag. 1 & 2), leaving only traces of each behind (Figure S3). Furthermore, each line developed MF-SSI locus that was identical on Southern blot to that of the single-copy SSI lines (Figure S3b,c), indicating efficient removal of extra transgene copies and conversion to MF-SSI locus. These plants, however, showed stunted growth and sterility, possibly due to gross genetic mutations induced by loxP × loxP recombination.

In conclusion, an efficient method for marker-free site-specific transgene integration (MF-SSI) in plants was developed that combines the advantages of precision genetic engineering with marker gene excision into a single technology platform. While the strategy for this technology was proposed a number of years ago (Srivastava and Ow, 2004), and its feasibility demonstrated recently (Akbudak and Srivastava, 2011; Nanto and Ebinuma, 2008), a streamlined method suitable for biotechnology was missing. This study developed the MF-SSI method using FLPe-FRT for gene integration and Cre-lox for marker excision, in which particle bombardment was used for delivering transgene construct into plant cells. However, FLPe-FRT and Cre-lox systems can be replaced with alternative SSR systems that display equivalent recombination efficiency, and Agrobacterium can be used for delivering transgene construct. Similarly, co-integration approach used for transgene integration in this study could be replaced with cassette exchange approach demonstrated by others (Li et al., 2009; Louwerse et al., 2007; Nanto et al., 2005). Regardless of these modifications, a comparison of MF-SSI protocol with standard transformation protocol (utilizing particle bombardment or Agrobacterium) shows one critical difference: use of a ‘target line’ that contains a single target site. However, if target lines are available, MF-SSI protocol is comparable to standard transformation methods as it does not require additional efforts except the heat treatment step, but the product is arguably much improved. Further, a larger study is needed to clearly determine MF-SSI efficiency; however, presence of the excisable marker gene in this approach might contribute to somewhat lower efficiencies. Nonetheless, any compromise in the efficiency of the process is paid off by the enhancement of the quality of transgene locus produced.


This project was funded by the Biotechnology Risk Assessment Program (grant# 2010-33522-21715) of the USDA National Institute of Food and Agriculture (NIFA) and Arkansas Division of Agriculture-ABI.