Confirmation of ‘Pollen- and Seed-Specific Gene Deletor’ System Efficiency for Transgene Excision from Transgenic Nicotiana tabacum under Field Conditions

The commercial application of genetically modified plants has been seriously impeded by public concern surrounding the potential risks posed by such plants to the ecosystem and human health. Previously, we have developed a ‘pollen- and seed-specific Gene Deletor’ system that automatically excised all transgenes from the pollen and seeds of greenhouse-grown transgenic Nicotiana tabacum. In this study, we conducted seven field experiments over three consecutive years to evaluate the stability of transgene excision under field conditions. Our results showed that transgenes were stably excised from transgenic Nicotiana tabacum under field conditions with 100% efficiency. The stability of transgene excision was confirmed based on PCR, as well as the GUS staining patterns of various organs (roots, leaves, petiole, stem, flower, fruit, and seeds) from transgenic N. tabacum. In six transgenic lines (D4, D10, D31, D56, and D43), the transgenes were stably deleted in the T0 and T1 generations. Thus, the ‘Gene Deletor’ system is an efficient and reliable method to reduce pollen- and seed-mediated unintentional gene flow. This system might help to alleviate the food safety concerns associated with transgenic crops.


Introduction
Sexual crossbreeding has been influential in the genetic improvement of crop species [1]. However, crossbreeding is a time-consuming, laborious process that is limited by reproductive isolation. In contrast, plant gene transfer introduces functional genes directly into host plants, thus eliminating the requirement for reproductive compatibility and allowing rapid directional selection [2]. Thus, plant gene transfer and other genetic modification (GM) techniques are powerful and efficient methods of plant genetic improvement. To date, many important commercial crops have been subjected to GM, including corn, soybean, canola, and cotton [3]. These GM crops are higher in yield and more cost effective than natural crops, as they require less fertilizer, herbicides, labor, and energy [3]. Since the first commercial use of transgenic crops in 1996, the global area devoted to transgenic crop growth has increased~112-fold, reaching 190.4 million hectares in 2019 (http://www.isaaa.org/resources/publications/pocketk/16/default.asp (accessed on 3 January 2022)).
Scientists, environmentalists, politicians, and the public have suggested that transgenes (e.g., genes resistant to bacteria, herbicides, stress, viruses, insects, and diseases) and transgene products might pose health risks to humans and other organisms [4][5][6][7]. It has also been proposed that transgenes might escape from GM crops and integrate into

Gene Deletor Vector Construction
A Gene Deletor vector was constructed to enhance the deletion efficiency of transgene from the targeted tissues. We constructed the Gene Deletor vector by fusing gene sequences from FLP/FRT and CRE/loxP systems and cloning these sequences into the pBIN19 binary vector ("LF" is an abbreviation for the loxP-FRT fusion sequence). The Gene Deletor vector contained the NPTII kanamycin resistance gene, which served as a selectable marker gene. The GUS reporter gene signaled the presence or absence of transgene expression, and FLP recombinase (PAB5) catalyzed site-specific DNA recombination, leading to the excision of the DNA between the two LoxP-FRT sites. The control vector lacked the PAB5 gene for site-specific recombination but contained the GUS::NPTII gene to indicate GUS activity.

Production and Confirmation of N. tabacum Transgenic Lines
Control and Gene Deletor plasmids were transformed into N. tabacum plants using the Agrobacterium-mediated transformation protocol [30]. We generated 202 independent transgenic lines using the control and Gene Deletor plasmids (Table 1). Table 1. Numbers of plants transformed with the control or 'pollen-and seed-specific Gene Deletor' cassettes and the corresponding transgene excision efficiency in the produced pollen and seeds, based on GUS activity levels in the T1 seedlings grown in the greenhouse. Transgenic and wild-type plants grown in the greenhouse exhibited similar phenotypes and growth patterns. Independent transgenic lines were selected for kanamycin resistance. Transgenic plants were screened based on a histochemical analysis of βglucuronidase (GUS) activity in leaf and root tissues.
2.3. The 'Gene Deletor' System Excised Transgenes from the Pollen and Seeds of Greenhouse-Grown N. Tabacum T1 seedlings germinated from self-pollinated seeds were GUS-stained to determine the transgene excision efficiency of each transgenic line. If the 'Gene Deletor' system functions correctly in a transgenic plant, all DNA (including transgenes) between the two fused loxP-FRT sites should be completely deleted in the pollen and seeds. Therefore, all T1 self-pollinated seedlings should be non-transgenic and should not exhibit GUS activity.
More than 75% of the T1 seedlings of the transgenic plants transformed with the control vector were positive for GUS activity (i.e., exhibited blue staining), following Mendel's law of segregation. This indicated that the transgenes were not excised in the pollen and seeds produced by the control plants, and that some of the control plants had multiple transgenic loci. In contrast, less than 70% of the T1 seedlings transformed with the 'Gene Deletor' vector exhibited GUS activity ( Figure 1A-F) ( Table 1). Thus, transgenes were excised from most of the pollen and seeds produced by some of the lines transformed with the 'Gene Deletor' vector.
Consistent with our previous experiment [29], some transgenic plants with 100% excision efficiency were produced by each transformation with the 'Gene Deletor' cassette. Across all three experiments, we identified five independent lines with 100% excision efficiency (D4, D10, D31, D56, and D43) ( Table 2). By comparing qPCR-calculated relative expressions to our standard curve, we determined that each of these lines possessed a single copy of the transgene. To provide controls for these lines, we also used our standard curve to identify control lines (i.e., lines transformed with the control cassette) in each experiment possessing a single copy of the transgene: C1, C14, and C6 (Table 2). Consistent with our previous experiment [29], some transgenic plants with 100% cision efficiency were produced by each transformation with the 'Gene Deletor' casse Across all three experiments, we identified five independent lines with 100% excision ficiency (D4, D10, D31, D56, and D43) ( Table 2). By comparing qPCR-calculated rela expressions to our standard curve, we determined that each of these lines possesse single copy of the transgene. To provide controls for these lines, we also used our stand curve to identify control lines (i.e., lines transformed with the control cassette) in e experiment possessing a single copy of the transgene: C1, C14, and C6 (Table 2). Table 2. Transgene excision efficiency in the pollen and seeds of transgenic plants grown in a gr house; transgene excision efficiency was determined based on GUS staining of the T1 seedli Lines C1, C14, and C6 (control lines); lines D4, D10, D31, D56, and D43 (lines with "Gene Dele cassette).

Line
Transgene Copies We next verified the transgene deletion efficiencies of the transgenic lines D4, D (generated in Experiment 1), D31, D56 (generated in Experiment 2), and D43 (genera in Experiment 3) (in comparison to the control lines C1, C14, and C6). In the control pla GUS staining indicated that about 75% of the self-pollinated seedlings and about 50% the hybrid seedlings exhibited GUS activity ( Table 2).
These ratios were consistent with Mendel's law of segregation, suggesting that transgenes were not deleted in these control plants. However, across the 5864 self-po nated seedlings and the 4808 cross-pollinated seedlings produced by the transgenic li  We next verified the transgene deletion efficiencies of the transgenic lines D4, D10 (generated in Experiment 1), D31, D56 (generated in Experiment 2), and D43 (generated in Experiment 3) (in comparison to the control lines C1, C14, and C6). In the control plants, GUS staining indicated that about 75% of the self-pollinated seedlings and about 50% of the hybrid seedlings exhibited GUS activity ( Table 2).
These ratios were consistent with Mendel's law of segregation, suggesting that the transgenes were not deleted in these control plants. However, across the 5864 self-pollinated seedlings and the 4808 cross-pollinated seedlings produced by the transgenic lines (D4, D10, D31, D56, and D43), none were positive for GUS activity ( Table 2). This indicated that the transgenes had been completely excised from the pollen and seeds of these lines.

The 'Gene Deletor' System Stably and Completely Excised Transgenes from the Pollen and Seeds Produced by Transgenic N. tabacum Grown in the Field
The stability and expression of the Gene Deletor system were also analyzed in fieldgrown plants. GUS staining, corresponding to GUS activity, was detected in the presence and its expression in the roots, leaves, petioles, stems, flower stalks, and flowers of both the 'Gene Deletor'-transformed lines (D4, D10, D31, D56, and D43) and the control transgenic lines (C1, C14, and C6) grown in the field. Consistent with this, PCR analysis showed that the FLP gene was present in the genomes of all transgene-excised transgenic lines tested (lines D4, D10, D31, D56, and D43) but not in control lines ( Figure S1). Additionally, GUS staining indicated positive GUS activity in the mature pollen grains, immature seeds, and immature fruits from the control plants (C1, C14, and C6) (Figure 2A,C,E), but not in those from the 'Gene Deletor'-transformed plants (D4, D10, D31, D56, and D43) ( Figure 2B,D,F). This demonstrated that the transgenes had been successfully excised from the pollen and seeds produced by field-grown plants transformed with the 'Gene Deletor' system. genic lines (C1, C14, and C6) grown in the field. Consistent with this, PCR analysis that the FLP gene was present in the genomes of all transgene-excised transge tested (lines D4, D10, D31, D56, and D43) but not in control lines ( Figure S1). Addi GUS staining indicated positive GUS activity in the mature pollen grains, immatu and immature fruits from the control plants (C1, C14, and C6) (Figure 2A,C,E), b those from the 'Gene Deletor'-transformed plants (D4, D10, D31, D56, and D43 2B,D,F). This demonstrated that the transgenes had been successfully excised pollen and seeds produced by field-grown plants transformed with the 'Gene system. Our field experiments, which included thousands of control seedlings and thousands of 'Gene Deletor'-transformed seedlings, showed that 75% of the s nated T1 control seedlings showed GUS activity ( Figure 2E), as did 50% of th control seedlings (Table 3). However, no GUS activity was observed in any self-p T1 'Gene Deletor'-transformed seedlings ( Figure 2F) nor in any hybrid 'Gene transformed seedlings (Table 3). Thus, the transgenes were excised with 100% e in the pollen and seeds produced by lines D4, D10, D31, D56, and D43, even whe in the field under changeable climatic conditions. In contrast, transgenes in contr (lines C1, C14, and C6) followed Mendel's law of segregation. Our field experiments, which included thousands of control seedlings and tens of thousands of 'Gene Deletor'-transformed seedlings, showed that 75% of the self-pollinated T1 control seedlings showed GUS activity ( Figure 2E), as did 50% of the hybrid control seedlings (Table 3). However, no GUS activity was observed in any self-pollinated T1 'Gene Deletor'-transformed seedlings ( Figure 2F) nor in any hybrid 'Gene Deletor'-transformed seedlings (Table 3). Thus, the transgenes were excised with 100% efficiency in the pollen and seeds produced by lines D4, D10, D31, D56, and D43, even when grown in the field under changeable climatic conditions. In contrast, transgenes in control plants (lines C1, C14, and C6) followed Mendel's law of segregation.

Molecular Analysis of Transgene Excision from Pollen and Seeds
If the 'Gene Deletor' system functions correctly in a transgenic plant, the DNA between the two loxP-FRT sites, including the transgenes, should be excised, and only a short T-DNA residue, containing only a flanking sequence with an intact fused loxP-FRT sequence, should remain in the genome (Figure 3). Therefore, only the T-DNA residue, not the transgenes, should be PCR amplified in the seeds and pollen from the T1 progeny of a transgenic line (Figure 3). This implies that if PCR is performed with the primers LF-F and LF-R, a 7.6 kb fragment should be amplified from the T 0 transgenic plants transformed with the 'Gene Deletor' vector, but a 0.2 kb fragment should be amplified from the T1 seedlings germinated from self-or cross-pollinated seeds. Consistent with expectations, 7.6 kb fragments containing 35S-NPT::GUS-3 35S and PAB5-FLP-Tnos were PCR-amplified from the leaf tissues of the transgenic lines D4, D10, D31, D56, and D43 grown in the field, but 0.2 kb fragments (or no fragments) were PCRamplified from each of their self-pollinated T1 seedlings ( Figure S2). Specifically, of the 200-300 self-pollinated T1 seedlings analyzed per transgenic line, about 75% produced a 0.2 kb DNA fragment, while the rest did not produce any specific bands. None of the tested T1 seedlings produced a 7.6 kb fragment (Table 4). Table 4. PCR verification of transgene excision efficiency in the T1 seedlings that were produced by self-pollinating transgenic plants in the field. Line C1 (control line) was transformed with the control cassette; lines D4, D10, D31, D56, and D43 were transformed with the 'pollen-and seed-specific Gene Deletor' cassette. short T-DNA residue, containing only a flanking sequence with an intact fused loxP-FRT sequence, should remain in the genome (Figure 3). Therefore, only the T-DNA residue, not the transgenes, should be PCR amplified in the seeds and pollen from the T1 progeny of a transgenic line (Figure 3). This implies that if PCR is performed with the primers LF-F and LF-R, a 7.6 kb fragment should be amplified from the T0 transgenic plants transformed with the 'Gene Deletor' vector, but a 0.2 kb fragment should be amplified from the T1 seedlings germinated from self-or cross-pollinated seeds.  Each 0.2 kb fragment had the same sequence: a short T-DNA residue containing an intact loxP-FRT fusion sequence. Thus, molecular analysis suggested that the transgenes had been completely excised from the pollen and seeds of the 'Gene Deletor'-transformed transgenic plants grown in field.

Discussion
There has always been public concern regarding inadvertent foreign gene flow and the safety of transgenic foods. Preventing inadvertent gene flow out of transgenic crops is critical, because even very low levels of gene flow can permit the spread of highly adventitious genes. Furthermore, once an external gene becomes established, it is very difficult to prevent its spread [15,31]. Thus, it is important to develop automatic transgene excision systems that completely delete foreign genes from the pollen and seeds produced by transgenic plants. As climatic factors such as light, temperature, rainfall, and wind strongly affect plant growth and development, it is necessary to evaluate the stability and efficiency of any transgene deletion system under field conditions as well as under greenhouse conditions.
Here, we extended our previous study of the 'Gene Deletor' system [29] and showed that transgenic plants with 100% transgene excision efficiency were repeatably produced by transformation with the 'Gene Deletor' vector, in which FLP is controlled by the Arabidopsis PAB5 promoter. More importantly, we showed that these results were replicable under variable climatic conditions over three consecutive years, suggesting that the 'Gene Deletor' system is a reliable automatic transgene excision system. After transgene excision, only a very short foreign DNA fragment, containing the left-and right-border T-DNA sequences, as well as an intact loxP-FRT recognition sequence, remained in the genomes of the pollen and seeds produced by the transgenic plants. This short foreign DNA residue has no function and is not transcribed; therefore, these transgenic plants are environmentally friendly and confer no risk of inadvertent foreign gene flow via pollen and seeds. Thus, plants transformed with the 'Gene Deletor' vector possess the desired traits associated with transgenes but produce non-transgenic pollen and seeds due to automatic transgene excision. Previous studies reported the efficiency of a Cre/LoxP recombination system in both prokaryotes and eukaryotes. These studies showed the effectiveness of this system in the development of transgene free transgenic plants. [32] developed a FLP/LoxP-FRT recombinase system in Escherichia coli, where it was utilized as a gene switch to regulate the gene expression. Similarly, [33] generated selectable marker-free transgenic plants using a Cre/loxP recombination system, which was controlled by −46 minimal CaMV 35S promoter. T0 and T1 transgenic Arabidopsis plants showed marker-free plants harboring only excised constructs in their genomes. These reported studies have not highlighted the effect of any remaining foreign DNA fragments on the genome evolution of species. Consistently, our experimental analysis proved the efficient excision of transgenes without any adverse environmental and phenotypic effects.
The 'Gene Deletor' system may be especially useful for vegetatively propagating species. That is because vegetative propagation methods do not involve pollen or seeds, the transgene cannot escape during propagation, and the 'Gene Deletor' vector prevents transgene escape via subsequent seed and pollen production. This reduces the potential risk of negative environmental impacts due to inadvertent foreign gene flow. Previously, [34] described the use of LEAFY (LFY) controlled promoters to excise the PAB5 marker gene from bananas. The same strategy was implemented with Cavendish bananas, and the Gene Deletor vector driven by the LFY promoter excised 88.5% of the exogenous genes [28].
A modified version of the 'Gene Deletor' cassette may help to prevent foreign gene flow from the pollen and seeds of sexually propagated transgenic crops, such as soybeans, sunflowers, canola, corn, and sorghum, which tend to outcross to wild relatives [2,35]. The introduction of the chemically inducible RNAi-FLP gene into the 'Gene Deletor' cassette, in conjunction with the application of a non-phytotoxic induction agent, may repress FLP gene expression in pollen and seeds [29]. This would prevent the deletion of transgenes from the pollen and seeds and allow the production of certified seed stocks. Without the application of the induction agent, the subsequent generation of sexually produced plants would produce transgene-free pollen and seeds. This system might be especially useful for plants that are genetically modified for patenting purposes.
There are additional applications of the 'Gene Deletor' system. Firstly, non-transgenic seeds or plants could be produced from transgenic plants, allaying consumer concerns over transgenes in food. Secondly, the 'Gene Deletor' technology would allow farmers to replant viable but non-transgenic seeds harvested from transgenic plants. This is in contrast to the controversial 'terminator seed' technology [36], which generates transgenic plants that produce non-viable seeds. Finally, the 'Gene Deletor' system, when used in grain crops, eliminates the need to label and physically separate transgenic and non-transgenic grains after harvest.

Nicotiana tabacum L. Transformation Using the 'Gene Deletor' Vector
Transgenic Nicotiana tabacum plants were generated using the control vector or the "Gene Deletor" vector as described previously [29]. The previously generated "Gene Deletor" vector was comprised of two identical LoxP-FRT (termed "LF") fusion sequences. In between the two LF sequences was located the GUS::NPTII fusion gene, driven by the constitutive CaMV 35S promoter, and the FLP yeast recombinase gene, controlled by the Arabidopsis pollen, ovule, and early embryo specific promoter PAB5 [37] ( Figure S3A). The control vector was identical to the "Gene Deletor" vector, except it was lacking the PAB5-FLP-NOS cassette ( Figure S3B). Agrobacterium tumefacians strain EHA105 was transformed with the "Gene Deletor" or control plasmid as described previously [30].
Wild-type tobacco plants were simultaneously generated in vitro (at 23 • C, with a 14 h light/10 h dark cycle and 50-60% humidity) and used as controls. Flower maturity was determined using the stages proposed by [38]; we modified this scale to include an extra stage (0) before stage 1 ( Figure S4).

GUS Histochemical Staining
GUS staining of transgenic leaf and root tissues was performed by dipping the tissues in GUS staining solution (200 mM NaPO 4 buffer (pH 7.0), 10 mM mercaptoethanol, 10 mM EDTA, 0.1% (w/v) sodium azide, 0.1% (w/v) Triton X-100, and 0.5 mg/mL X-Gluc) and incubating the tissues at 37 • C overnight. After incubation, leaf sections and roots were de-stained in 70% ethanol. The de-staining process was repeated four times prior to visual analysis.

Detection of Transgene Excision Efficiency in Greenhouse-Grown Transgenic Plants
Transgenic seedlings were transferred to the greenhouse. At flower stage 9, five to ten flowers from each transgenic plant were bagged for self-pollination. After six weeks, approximately 300-1500 seeds from each plant were collected and germinated at 25 • C. After two weeks of germination, the efficiency of transgene excision was investigated in the fully developed T1 seedlings using GUS staining, following the protocols described above.

Confirmation of Transgene Excision in the T1 Generation of Greenhouse-Grown Transgenic Plants
Additional seeds (at least 20,000 per line) from T1 transgenic lines (showing negative GUS staining) were selected and germinated for further confirmation. Lines producing no blue-stained T1 seedlings during this second round of testing potentially had a transgene excision efficiency of 100%. All subsequent experiments were performed using these transgenic lines.
Confirmed transgenic T1 plants were then self-and cross-pollinated. For crosspollination, the immature anthers were removed from wild-type and transgenic flowers at stage 9, and mature pollen grains collected from transgenic plants were placed on the stigmas of the emasculated wild-type flowers. Simultaneously, pollen grains collected from wild-type plants were placed on the stigmas of the emasculated transgenic flowers. Seeds produced using self-and cross-pollination were collected separately and germinated. The developed seedlings were GUS-stained as described above.
We then measured the relative expression of FLP in each mixture using quantitative real-time PCR (qPCR). Primer sequences for pFLP-F and pFLP-R were designed (5 -CAAGAAAACCAGCTGTGACAAGCCTTAAAC-3 and 5 -CGAGTTCTGCCTCTTTG TGAGTCTCAATAG-3 ). Each qPCR reaction (20 µL) contained 2× SoSo Fast EvaGreen Supermix (BioRad, Hercules, CA, USA), 5 µM of each primer, and 30 ng of genomic DNA, with nuclease-free water added to make the final volume. The thermal cycling conditions were as follows: 98 • C for 3 min, followed by 40 cycles of 98 • C for 5 s and 60 • C for 20 s. All qPCRs were performed using a Bio-Rad CFX96/C1000 Real-Time System. The L25 gene, which has an amplification efficiency approximately equivalent to that of FLP [40], was used as an internal reference gene. Next, the standard curve was generated, with FLP copy numbers plotted against the ∆Ct value. The FLP copy number was calculated using Bio-Rad CFX MANAGER software (http://www.bio.rad.com (accessed on 3 June 2019). Three technical replicates of each experiment were performed for each transgenic line (100% transgene excision efficiency) and control line (0% transgene excision efficiency).

Transgene Excision in Field-Grown Transgenic Plants
Vegetatively propagated greenhouse transgenic lines were cultivated in the field between May and September for three years to determine transgene excision efficiency under field conditions. In total, seven independent field experiments (using subsets of the control and transgene-excised transgenic lines) were conducted over the three years. In each field experiment, the roots, leaves, petioles, stems, flower stalks, and flower organs of each plant at flower stage 1 were stained for GUS activity. The FLP gene was also PCR amplified from the leaves of the young plants to test whether the transgenes were present. Transgene excision in the pollen and seeds of the F1 plants was tested using GUS staining, as described above.
We tested the transgene excision efficiency of each control and transgene-excised transgenic line in the field experiments by analyzing GUS activity in seedlings produced via self-pollination and reciprocal crossing. Mature pollen grains and immature seeds were collected from the young fruits of the T1 seedlings at 18 days after pollination (DAP) and GUS stained to test for transgene excision in the pollen and seeds.

Molecular Verification of Transgene Excision in the Transgenic Plants
For transgene excision verification, genomic DNA was isolated from wild-type and transgenic plants in the T 0 and T1 generations. A primer pair specific to the T-DNA sequences outside the two loxP-FRT sites was designed (pLF-F (5 -TGCAAGGCGATTAAGTT GGGTAAC-3 ) and pLF-R (5 -ACCATTATTGCGCGTTCAAAA GTC-3 )). PCRs were performed on a S1000 Thermal Cycler (Bio-RAD, Hercules, CA, USA) with Primestar Taq polymerase (TaKaRa, San Jose, CA, USA). Each 25 µL reaction mixture include 30 ng genomic DNA as a template. The PCR cycling conditions were as follows: initial denaturation at 98 • C for 5 min; 10 cycles of touch-down PCR (denaturation at 98 • C for 5 s, annealing at 55-65 • C for 15 s, and extension at 72 • C for 3.5 min); 30 cycles of regular PCR (denaturation at 98 • C for 10 s, annealing at 55 • C, and extension at 72 • C for 2 min), and a final extension at 72 • C for 5 min. Then, 1 µL of PCR product was used as the template for a second PCR with the same primers and the following cycling conditions: initial denaturation at 98 • C for 3 min; followed by 35 cycles of 98 • C for 5 s, 60 • C for 10 s, and 72 • C for 3.5 min; and a final extension at 72 • C for 10 min. The PCR products were resolved using electrophoresis in a 1% agarose gel and photographed using the Bio-Rad Gel Doc system (Bio-RAD, Hercules, CA, USA). About 10-15 PCR amplicons per transgenic line were sequenced to analyze the composition of the amplified DNA fragments.  Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: All data generated or analyzed during this study were included in this published article and its Supplementary Information files.

Conflicts of Interest:
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.