Development of an activation tagging system for maize

Abstract Activation Tagging, distributing transcriptional enhancers throughout the genome to induce transcription of nearby genes, is a powerful tool for discovering the function of genes in plants. We have developed a transposable element system to distribute a novel activation tagging element throughout the genome of maize. The transposon system is built from the Enhancer/Suppressor (En/Spm) transposon system and uses an engineered seed color marker to show when the transposon excises. Both somatic and germinal excision events can be detected by the seed color. The activation tagging element is in a Spm‐derived non‐autonomous transposon and contains four copies of the Sugarcane Bacilliform Virus‐enhancer (SCBV‐enhancer) and the AAD1 selectable marker. We have demonstrated that the transposon can give rise to germinal excision events that can re‐integrate into non‐linked genomic locations. The transposon has remained active for three generations and events displaying high rates of germinal excision in the T2 generation have been identified. This system can generate large numbers of activation tagged maize lines that can be screened for agriculturally relevant phenotypes.

transforming plants with transposable elements engineered to distribute activation tagging elements. The engineered transposon systems have the advantage over direct transformation in that fewer transformation events need to be generated (Marsch-Martinez, Greco, Van Arkel, Herrera-Estrella, & Pereira, 2002;Qu, Desai, Wing, & Sundaresan, 2008;Schneider et al., 2005;Sundaresan, Qu, Desai, & Wing, 2008) and are especially useful in plants, such as maize, in which transformation is labor intensive.
Several engineered transposon systems have been used in plants (Walbot, 2000). In most of these systems there is a non-autonomous transposon that is introduced into the plant along with a transposase that can excise the transposon and catalyze its integration at a new position; although some systems use an endogenous transposase.
Most of the described systems take advantage of one of the maize transposon systems, Ac/Ds, En/Spm or Mutator (Mu). These transposon systems have different characteristics that influence the types of mutant populations that can be generated. For example, in Ac/Ds systems, the new sites of integration are generally near the site of the original insertion (Dooner & Belachew, 1989;Greenblatt, 1984) while in En/Spm and Mu systems the integration sites for the nonautonomous elements are not usually linked to the original integration sites (Kumar, Wing, & Sundaresan, 2005;Lisch, Chomet, & Freeling, 1995). The transposase of the En/Spm system has been cloned and extensively characterized while the Mu transposase is less well characterized. Non-autonomous Mu transposon systems rely on mutator lines that have an endogenous transposase that facilitates transposition.
We have developed a Spm transposon-based activation tagging system for maize with the goal of generating a large activation tagged population. The goal of generating this population is to have tens of thousands of individual lines with insertions of the activation tagging element at different sites throughout the genome. To generate this population, the activation tagging transposon construct has been introduced into progenitor lines that serve as launch sites.
Lines that undergo high rates of germ-line excision of the transposable element have been identified. To easily distinguish between somatic and germ-line excision events, a seed color system that supports the identification of germ-line excision events was developed.

| Plasmid construction and Agrobacterium transformation
The superbinary plasmid pEPS3004 was constructed for transformation of maize through recombination of pSB1 binary plasmid and pSB11 plasmid derivative-pEPP3028. The pEPP3028 was introduced into Agrobacterium tumefaciens strain LBA4404(pSB1) (Komari et al., 2006;Komori et al., 2007) and the structure of the pSB1:: pEPP3028 co-integrate superbinary plasmid, pEPS3004, was validated by restriction enzyme digestion and DNA sequence analysis of selected segments. The T-DNA in the transformation vector, pEPS3004, contains the elements of the activation tagging sequences; (a) the non-autonomous transposon comprised of the 4X SCBV enhancer, the OsAct1p::AAD1 selectable marker and the terminal inverted repeats (TIR) from the Spm transposon, (b) the transcription factors controlling anthocyanin biosynthesis, ZmGLOB1::B-peru and ZmGLOB1::C1, and (c) the Spm transposase. The 4X SCBV enhancer and the OsAct1p::AAD1 selectable marker were subcloned from pDAB3878 as described in Davies et al. (2014). The ZmGLOB1::B-peru and ZmGLOB1::C1 genes were PCR amplified from the vector pDAB2125 as described in Petolino and Shen (2006). The Spm TIRs and transposase were synthesized at DNA2.0 (now ATUM) according to the description in Kumar et al., 2005 using sequences from GenBank Accession M25427.1.

| Particle bombardment of maize embryos
Maize ears were collected 13 days after pollination (DAP) and surface sterilized with 50% bleach for 15 min, followed by five washes with sterile distilled water. Embryos were isolated using sterile forceps and 30 embryos were placed on a Petri plate containing MS medium (Murashige & Skoog, 1962) with 3% sucrose for 24 hr in darkness at 28°C. On the day of bombardment, the embryos were moved to MS medium containing 12% sucrose and incubated in darkness at 28°C for 4 hr prior to bombardment.
Gold particles (1 μm diameter) were washed with 70% ethanol for 10 min, then rinsed three times with sterile water. The particles were dispensed in 50% glycerol at a concentration of 120 mg/ml. To 50 μl (6 mg) of gold particles, 5 μg of plasmid DNA, 50 μl of 2.5 M CaCl 2 and 20 μl 0.1 M spermidine were added. The reaction (total volume 125 μl) was incubated at room temperature for 10 min with gentle shaking, then for another 10 min without shaking. The DNA coated-gold particles were briefly centrifuged, washed with 150 μl of 70% ethanol and then with 100% ethanol. The final pellet was resuspended in 30 μl of 100% ethanol and subjected to a brief sonication with a Branson 3510 sonicator. A 10 μl aliquot of the gold-particles coated with DNA was spread on macrocarriers (BioRad, Hercules, CA) and used in bombardment assays using a BioRad PDS1000/He system. The embryos were transformed at a target distance of 6 cm using 1,100 psi disks. Following bombardment, the embryos were moved to MS medium containing 3% sucrose and incubated under light (approximately 50 μmoles m −2 s −1 ) for 48 hr at 28°C. Accumulation of anthocyanin pigments was observed under a light microscope.

| Maize transformation
Seeds from the B104 inbred were planted into 4-gallon-pots containing Sunshine Custom Blend ® 160 (Sun Gro Horticulture, Bellevue, WA). The plants were grown in a greenhouse using a combination of high pressure sodium and metal halide lamps with a 16:8 hr Light: Dark photoperiod. To obtain immature embryos for transformation, controlled sib-pollinations were performed. Immature embryos were isolated at 10 to 13 days after pollination when embryos were approximately 1.4 to 2.0 mm in size.
Prior to embryo excision and transformation, maize ears were surface sterilized by immersing them in 50% commercial bleach with Tween 20 (1 or 2 drops per 500 ml) for 10 min and rinsed three times with sterile water. A suspension of Agrobacterium cells containing a superbinary vector was prepared by transferring one or two loops of bacteria grown on YEP (5 g/L yeast extract, 10 g/L peptone, 5 g/L sodium chloride, 15 g/L Bacto Agar) solid medium containing 50 mg/L spectinomycin, 10 mg/L rifampicin, and 50 mg/L streptomycin at 28°C for 3 days or 25°for 4 days into 5 ml of liquid infection medium (MS salts, ISU Modified MS Vitamin stock (1,000×, 2 g/ L glycine, 0.5 g/L each of thiamine HCl and pyridoxine HCl, 0.05 g/L nicotinic acid (Frame et al., 2006), 3.3 mg/L dicamba, 68.4 gm/L sucrose, 36 gm/L glucose, 700 mg/L L-proline, pH 5.2) containing 100 μM acetosyringone. The solution was gently pipetted up and down using a sterile 5 ml pipette until an uniform suspension was achieved, and the concentration was adjusted to an optical density of 0.3 to 0.5 at 600 nm (OD 600 ) using an Ultrospec 10 Cell Density Meter (GE Healthcare/Amersham Biosciences, Piscataway, NJ).
Immature embryos were isolated directly into a micro centrifuge tube containing 2 ml of the infection medium. The medium was removed and replaced twice with 1 to 2 ml of fresh infection medium, then removed and replaced with 1.5 ml of the Agrobacterium solution. The Agrobacterium and embryo solution was incubated for 5 min at room temperature and then transferred to co-cultivation medium, which contained (MS salts, ISU Modified MS Vitamin stock (1,000×, 2 g/L glycine, 0.5 g/L each of thiamine HCl and pyridoxine HCl, 0.05 g/L nicotinic acid, 3.3 mg/L Dicamba, 68.4 gm/L sucrose, 36 gm/L glucose, 700 mg/L L-proline, pH 5.2) containing 100 μM acetosyringone. Co-cultivation incubation was for 3 to 4 days at 25°u nder either dark or 24-hr white fluorescent light conditions (approximately 50 μmoles m −2 s −1 ).
After co-cultivation, the embryos were transferred to a nonselection MS salts, ISU Modified MS Vitamins, 3.3 mg/L Dicamba, 30 gm/L sucrose, 700 mg/L L-proline, 100 mg/L myo-inositol, 100 mg/L Casein Enzymatic Hydrolysate, 15 mg/L AgNO 3 , 0.5 gm/L MES (2-(N-morpholino)ethanesulfonic acid monohydrate; Fischer Scientific, Waltham, MA), 250 mg/L Carbenicillin, and 2.3 gm/L Gelzan ™ , at pH 5.8. Incubation was continued for 7 days at 28°C under either dark or 24-hr white fluorescent light conditions (approximately 50 μmoles m −2 s −1 ). Following the 7-day resting period, the embryos were transferred to selective medium (the MS-based resting medium (above) supplemented with haloxyfop). The embryos were first transferred to selection medium containing 100 nM haloxyfop and incubated for 1 to 2 weeks, and then transferred to selection medium containing 500 nM haloxyfop and incubated for an additional 2 to 4 weeks. Transformed isolates were obtained over the course of

| Genomic DNA isolation and PCR amplification of excision sites
Genomic DNA was isolated using Qiagen DNAeasy kits. Polymerase Chain Reaction (PCR) (Mullis et al., 1986) was performed using the forward primer (5′-GTACCTCTTCCTGGAGCACCAG-3′) which hybridized to sequences between the 13,961 bp and 13,982 bp pEPS3004 and the reverse primer (5′-TGTAGAACCCGTCCGTCCGTCCACGTCAG-3′) which hybridizes to sequences between 20,359 bp and 20,383 bp in pEPS3004. PCR products were cloned into a TOPO vector (Invitrogen, Carlsbad, CA) and transformed into Escherichia coli cells. Plasmid DNA was prepared from isolated colonies and sequenced by Sanger dideoxy sequencing (Sanger, Nicklen, & Coulson, 1977).

| Field growth of pEPS3004 transformed plants
In 2010, T1 seed from 68 single copy events containing the T-DNA from pEPS3004 were planted at the DAS field station in Molokai, HI. Prior to planting, the field was prepared by irrigation and application of glyphosate. All field operations followed the DAS Regulated  water to provide 300-400 pounds of nitrogen per acre throughout the lifetime of a plant. Transgenic plants were detasseled prior to pollen shed to assure pollination was from non-transgenic B104. Ears were harvested from individual plants and kept separate by ear.
Kernels were classified as yellow, yellow with purple sectors or purple and number of kernels with each phenotype were tabulated by ear.
In 2011 T2 seed from 21 of the 68 events grown in 2010 was planted at the Gray Research Production field site in Ashkum IL. The field was prepared applying the herbicides Impact ® and Atrazine and fertilized with 150 lbs/acre potash 150 lbs/acre diammonium phosphate and 150 lbs/acre nitrogen. After herbicide and fertilizer application, the field was plowed. All field procedures were performed following DAS Regulated Transgene Standard Operating Procedures (RTSOP). Fifteen-foot rows were planted alternating four rows of transgenic plants and two rows of non-transgenic B104. All transgenic plants were sprayed with Quizalofop (Assure ® II) at the 2-3 leaf stage and detasseled prior to pollen release. Mature ears were harvested, shelled, and sent to DAS.

| Next-Generation Sequencing (NGS) procedures and data analysis
To characterize the site of integration of the T-DNA from pEPS3004 in the genome (launching sites) as well as the site of re-insertion of the transposon, the Sequence Capture-based NGS Event Characterization (EC) pipeline was used following the protocol described in (Guttikonda et al., 2016). Genomic DNA was extracted and sheared to 800 bp fragments, then subjected to hybridization with 50- The standard EC computational analysis pipeline was used to characterize the T-DNA integration site (Guttikonda et al., 2016).
The process relies on T-DNA and genome junction-spanning pairedend reads as well as individual reads that span the junctions. By identifying the junction position in multiple fragments, the genomic location of the T-DNA is determined. Similarly, the re-insertion site of the transposon is determined by identifying paired-end read and reads that span the junctions between the transposon and insertion site.
To confirm the T-DNA and transposon re-insertion sites identified from the EC computational analysis, a parallel analysis was con- Finally, the tag sites were compared with current gene annotations for maize genome (Schnable et al., 2009) to investigate the spatial relationships between tag sites and known genes. In particular, the distances between the tag sites and the closet start codons were calculated and summarized (see Section 3).

| Construct used to develop the activation tagging population
The vector pEPS3004 was used in transformation and transfection experiments described below. This T-DNA vector is comprised of a non-autonomous Spm transposable element containing the 5′ and 3′ terminal inverted repeats (TIRs) of the Spm element (Kumar et al., 2005), four copies of the SCBV enhancer sequence (Braithwaite, Geijskes, & Smith, 2004;Davies et al., 2014) and a plant selectable marker composed of the rice actin1 promoter (OsActin) (McElroy, Zhang, Cao, & Wu, 1990) and the AAD1 gene, which provides resistance to the herbicide haloxyfop (Wright et al., 2010). To mobilize the nonautonomous transposable element, the T-DNA vector also contains the Spm transposase (TPase) gene driven by the maize ubiquitin promoter (UBI) (Toki et al., 1992). A schematic of the T-DNA from this vector is shown in Figure 1 and the sequence of the vector is disclosed in Supporting Information.
In order to observe whether the non-autonomous transposable element has excised in seed tissues, the engineered transposon was inserted in the 5′UTR of the ZmGLOB1::B-peru gene. The B-peru and C1 genes encode transcription factors that regulate anthocyanin biosynthesis (Bodeau & Walbot, 1992;Grotewold et al., 1998;Ludwig, Bowen, Beach, & Wessler, 1990) and their expression has been shown to be sufficient, in certain genetic backgrounds, for the accumulation of anthocyanins in the embryo and aleurone layer of kernels when driven by the ZmGLOB1 promoter (Petolino & Shen, 2006). The B-peru and C1 genes are either absent or not functional in the wild-type B104 inbred, the maize genotype into which the  Six types of fragments were obtained and are summarized in Table 1. These types of mutations have also been observed in other studies characterizing transposon excision sites (Kumar et al., 2005).

| Transposon excision in T1 kernels
The construct pEPS3004 was transformed into maize inbred B104 by Agrobacterium-mediated immature embryo transformation and transformants were selected by resistance to haloxyfop. Primary T0 transformants were grown to maturity and backcrossed to wild-type B104 plants to produce T1 seed. A total of 584 herbicide tolerant plants were generated. Transformants with a low copy number, as determined by TaqMan analysis, were subjected to DNA gel blot analysis.
Three hundred and twenty-two of these plants were confirmed to have the AAD1 selectable marker gene and 311 of these plants produced seed (Supporting Information Table S1. T1 Kernel Phenotypes).    Table S1. T1 Kernel Phenotypes). Figure 5a shows the distribution of seed phenotypes in the T1 ears of the primary transformants.
To determine whether the rate of excision is affected by the copy number of the T-DNA inserts, the data are presented as a function of the copy number. The percentage of events in each phenotypic class is shown in Figure 6. Three hundred and eleven events were examined, 206 with one insert, 80 with two inserts, 18 with three inserts, and 7 with four inserts. In events with a single insert, ears with yellow kernels with purple sectors and only yellow kernels are most commonly observed, ears with yellow and purple kernels, and yellow kernels with purple sectors along with a few purple kernels are seen less frequently. The pattern for plants with two or three inserts is similar except that there are fewer ears showing yellow and purple kernels. The patterns suggest that copy number, up to three copies, does not dramatically affect somatic transposon excision but may have some effect on germinal transposon excision rates since the percentage of ears with just yellow and purple kernels declines as the copy number increases. More than three copies may affect transposon excision as five of the seven events with four copies showed no transposon excision but this is based on a very small sample set.

| Transposon excision in the T2 and T3 generations
To determine whether germinal excision frequencies remained high in the T2 to T3 generation, purple sectored kernels from events exhibiting purple kernel formation in the T1 to T2 generation were grown from the T2 to the T3 generation. Purple sectored kernels from 150 T2 ears from 21 events displaying at least one purple kernel on 30% or more of the ears from the T2 generation were planted in the field and crossed with wild-type B104. Seed from approximately 10,000 T2 plants were recovered and~20 randomly selected ears from each event were analyzed for kernel phenotypes (Supporting Information Table S3 occurrence is unclear but could be the effect of the environment that is known to influence transposon excision (Hashida, Kitamura, Mikami, & Kishima, 2003). Furthermore, the frequency of ears with purple kernels appeared different across different events (Figure 7b).
This suggests that germinal excision is higher in some events than in others and that there may be an influence of the site in the genome from which the transposon launches on the rate of purple kernel formation. because it is highly improbable that half the kernels on an ear would show somatic excision simultaneously early in embryo development.

|
While plants derived from purple kernels that produce ears with 50% yellow kernels and 50% yellow kernels with purple sectors are derived from somatic transposon excision events.
To determine the frequency of purple kernels representing germinal and somatic excision events, 20 herbicide-resistant T2 plants derived from purple kernels of ears exhibiting yellow, yellow with purple sectors and purple kernels were grown to maturity and pollinated by non-transgenic B104 to produce seed (one plant died prior to pollination). Table 2  Tissues containing homogenous launch sites with the transposon present indicate that the anthocyanin accumulation in the kernel was due to somatic excision of the transposon prior to development of the aleurone layer. Tissues that contain heterogeneous launch sites also indicate that the anthocyanin accumulation in the kernel that gave rise to the sampled plant was due to somatic excision prior to development of the aleurone layer. Table 3 (Table 3). Events varied from 9% to 30% in the frequency of homogenous empty launch sites.
Since activation tagging has been characterized to be effective on  The colors are just to distinguish the events. Multiple plants from 4 of the 5 events were examined.
the transposon insertions was hard to place because either no pairedend sequencing reads were derived that connected sequences of the transposon and the sequences flanking the insertion site, or that the sequences that flank the insertion site contained only repetitive DNA.
The frequency of placement of re-inserted transposons across events is reported in Supporting Information Table S5. Tpn Re-insert Place Freq. Figure 9 shows transposon re-insertion sites from four events; the data indicate that the location of transposon insertion appears to be independent of launch site.
The location of the transposon relative to the nearest adjacent gene was also determined. It was found that 20 of the 103 transposons mapped (19%) integrated within the coding sequences of a gene and that 56 of 103 (54%) integrated within 10 kbp of a gene, and 27 of 103 (26%) integrated further than 10 kbp away from the closest gene (Supporting Information Table S6. Re-insertion site context).  detected by purple sectors in the aleurone layer of the kernel. Kumar et al. (2005) Weigel et al. (2000) and Marsch-Martinez et al. (2002) but is somewhat less than the 6.4% that Wan et al. (2009) reported.

| DISCUSSION
Future development of the activation tagged population will require the identification of large numbers of germ-line transposition events by screening for purple kernels on ears that also contain yellow kernels with purple sectors. Each of these purple kernels may represent an independent germinal transposon excision and possible re-insertion event. The purple kernels from ears with 50% yellow and 50% purple kernels are most likely due to the same germ-line excision event so the purple kernels from ears that have both purple kernels and yellow kernels with purple sectors will be collected. The T3 plants derived from these purple kernels will be backcrossed to B104 and examined for ear phenotypes. Those that produce a T4 generation with 50% purple kernels and 50% yellow kernels will be kept as germ-line transposition events. As noted above, the transposed activation tagging element is likely not to be genetically linked with the anthocyanin and transposase gene and will segregate independently. In this case, half of the purple kernels and half the yellow kernels from the T4 generation will have the activation tagging element and half will not. When the transposase and the activation tagging element are in the same cell, the transposon may continue to move. To develop lines where the activation tagging element is stable, lines in which the activation tagging element has segregated away from the transposase need to be identified. These lines can be identified as yellow T4 kernels that are herbicide tolerant; the OsActin::AAD1 herbicide tolerance gene is on the activation tagging element. Once these lines have been developed they can be screened for aberrant phenotypes that may be caused by the activation tagging element.
As the earth's population continues to grow, increased food production is needed. Because arable land is not increasing, there is a need to increase the amount of food produced on land currently farmed. Along with enhanced breeding techniques and improved agronomic practices, the discovery of genes that can enhance crop performance is needed. Since no other mutagen efficiently causes the activation of gene expression, the development of an activation tagged population of maize is an unique platform to identify gene function and may lead to the identification of beneficial phenotypes such as tolerance to biotic and abiotic stresses, enhanced nitrogen use efficiency and perhaps increased yield. Creating this activation tagged population in an important crop like maize will avert the problems of translating discoveries made in a model or non-target crops species into maize and may lead to new maize hybrids with enhanced yield.

ACKNOWLEDGMENTS
We would like to acknowledge and thank the Agronomic Traits