Minimal Piggybac Vectors for Chromatin Integration

ENABLING TECHNOLOGIES Minimal piggyBac vectors for chromatin integration

V Solodushko1,2, V Bitko3 and B Fouty1,2,4

We describe novel transposon piggyBac vectors engineered to deliver transgenes as efficiently as currently available piggyBac systems, but with significantly less helper DNA co-delivered into the host genome. To generate these plasmids, we identified a previously unreported aspect of transposon biology, that the full-length terminal domains required for successful plasmid- to-chromatin transgene delivery can be removed from the transgene delivery cassette to other parts of the plasmid without significantly impairing transposition efficiency. This is achieved by including in the same plasmid, an additional helper piggyBac sequence that contains both long terminal domains, but is modified to prevent its transposition into the host genome. This design decreases the size of the required terminal domains within the delivered gene cassette of the piggyBac vector from about 1500 to just 98 base pairs. By removing these sequences from the delivered gene cassette, they are no longer incorporated into the host genome which may reduce the risk of target cell transformation.

Gene Therapy (2014) 21, 1–9; doi:10.1038/gt.2013.52; published online 17 October 2013 Keywords: piggyBac; mutagenesis; stable gene delivery

INTRODUCTION integrated into the host cell genome, they perform no useful Transposon vectors are a proven and viable alternative to viral function. They may potentially increase the risk of cell vectors for stable gene delivery.1–4 Like integrated viruses, transformation due to their retained promoter and enhancer transposons deliver transgenes to target cells in vitro and in vivo activity, however. Attempts to reduce the size of the terminal where they are incorporated into the host genome. Unlike viruses domains to decrease this potential risk of transformation have 9,13 they do not generate an immune response, they have a simpler resulted in a significant loss of transposition efficiency. genome, and are easier to handle. In addition, they can hold a In this paper, we describe a piggyBac vector in which most of significantly larger transgene insert than viruses, in some cases up both terminal domains have been removed from the delivery to 100 kilobases.5 These characteristics make transposons an cassette without a significant loss of transposition efficiency. Only attractive option for gene delivery. two minTR sequences remain within the delivery cassette, both of PiggyBac vectors are one of the most active and flexible which are very short (35 base pairs for 50 minTR and 63 base pairs transposon systems available for the stable transfection of for 30 minTR). The expected negative effect on transposition of mammalian cells.6,7 The wild-type piggyBac transposon is 2472 deleting most of the terminal domains is compensated by the base pairs in length, and is composed of two inverted minimal presence of a transposition-incompetent piggyBac transposon terminal repeats (minTRs), two internal domain sequences and a containing both long terminal domains and the transposase gene transposase-encoding domain8,9 (Figure 1, wild type). Transposase within the helper part of the plasmid. This design contains two catalyzes the excision of the transposon from one DNA different sets of piggyBac sequences each of which has been source (that is, a delivered plasmid) and allows its subsequent modified to serve different functions. The working (transposable) re-integration into another DNA source (that is, the host cell transposon has been truncated to decrease the amount of extra genome). In the majority of piggyBac vectors, the transposase DNA incorporated into the host genome, whereas the helper (non- gene has been removed from the transposon, replaced by transposable) sequence provides the long terminal domains transgenes of interest; the transposase is then usually delivered necessary for efficient transposition of the working transposon. to the cell by a separate plasmid, a technique first demonstrated This design decreases the size of the required terminal domains in insect cells10–12 (Figure 1, delivery and helper plasmids). within the delivered gene cassette from about 1500 to just 98 base The minTRs and internal domains are crucial for the effective pairs which significantly decreases the size of the integrated integration of the transposon into the host genome and together sequence within the host cell genome. (known as terminal domains) consist of more than 700 base pairs each.9,13–16 The 50 terminal domain also serves as a native promoter for transposase expression. During transposition, the RESULTS terminal domains are integrated into the host cell genome, Full-length terminal domain sequences do not need to be in the exclusively at TTAA integration site, alongside the delivered delivered transposon for efficient transposition of piggyBac vectors transgene of interest as part of the transposon.11,14 Therefore, like We hypothesized that most of the terminal domain sequences integrated viruses, they deliver a significant amount of extra DNA could be moved outside of the transposon with minimal loss of to the target cell genome. These sequences (the terminal transposition efficiency. To test this hypothesis, we designed domains) are required for successful transposition, but once several plasmids (Figure 1) and determined their transposition

1Center for Lung Biology, University of South Alabama School of Medicine, Mobile, AL, USA; 2Department of Pharmacology, University of South Alabama School of Medicine, Mobile, AL, USA; 3NanoBio Corporation, Ann Arbor, MI, USA and 4Department of Internal Medicine University of South Alabama School of Medicine, Mobile, AL, USA. Correspondence: Dr V Solodushko, Center for Lung Biology, Department of Pharmacology, MSB 3406, University of South Alabama School of Medicine, Mobile, AL 36688, USA. E-mail: [email protected] Received 1 May 2013; revised 16 July 2013; accepted 27 August 2013; published online 17 October 2013 Minimal piggyBac vectors V Solodushko et al 2

Figure 1. Schematic presentation of vectors. 50 TRmin, 30 TRmin: minimal 50 or 30 terminal repeats (in black); ½ 30 minTR: half of the 30 terminal repeat in plasmid-166 (in black); 50 TD, 30 TD: 50 or 30 full-length terminal domains (including the 50 TRmin or 30 TRmin) (in yellow and black); transgene: delivered gene(s) (in this paper – RFP, (in blue)); transposase: piggyBac transposase gene (in red).

efficiency in target cells. The first plasmid contained a delivered promoter (plasmid-200) or a plasmid carrying the entire wild-type cassette encoding the reporter gene, red fluorescent protein (RFP), piggyBac (p3E1.2) failed to substantially increase transposition flanked by 50 and 30 minTRs (plamid-132). In a second plasmid, we efficiency (0.13 and 0.11%, respectively). Transfection with inserted a wild-type piggyBac transposon separated from the RFP plasmid-137, however, resulted in a marked increase in the delivery cassette by 683 and 2466 base-pair linkers (plasmid-137, number of cells stably expressing RFP at 28 days to 3.89%. These Figures 1 and 2b). This construct allowed us to add full-length results suggested that minTR alone are not sufficient to allow terminal domains back into the plasmid without including them plasmid-to-chromatin transposition, but the presence of full- within the RFP-delivered cassette. The presence of piggyBac length terminal domains elsewhere in the plasmid, even if they transposase in the wild-type transposon (driven by its native are located outside of the delivery cassette, will allow successful promoter) eliminated the necessity of using a helper vector to transposition into the host cell genome. Delivery of full-length deliver the transposase. A third plasmid (plasmid-185), also terminal domains in a separate plasmid, as was shown with contained the RFP-delivery cassette, but included an additional p3E1.2, failed to support the transposition. (modified) full-length transposon in which both TTAA integration These results with plasmid-137 did not clarify whether only the sites were mutated (to GTAA) to prevent transposition of the RFP-delivery cassette was integrated into the host cell genome or full-length piggyBac into the host genome. In plasmid-185, the whether the entire fragment, containing both piggyBac transposons, full-length terminal domains of the second transposon were was delivered. Therefore, we tested the integration efficiency of separated from the minTRs of the RFP delivery cassette by two plasmid-185, a plasmid in which the TTAA integration sites flanking linkers of 254 and 3354 base pairs. As the activity of the native the full-length piggyBac vector in the helper part of the plasmid transposase promoter is unpredictable in many mammalian were mutated (to GTAA), preventing its excision from the plasmid cells,17 we replaced the native promoter with an SV40 promoter and thus preventing its integration into the host (other modifica- in this, and subsequent, plasmids (promoters are not shown in tions are described in Figure 2b). Plasmid-185 had significantly Figure 1, refer to Figures 2–4 for plasmid details) to more reliably greater transposition efficiency than plasmid-137 (13.4 versus 3.89% drive expression of the transposase. This replacement necessitated of initially transfected cells at 28 days) (Figure 2a). The transposition a partial duplication of the 50 terminal domain to keep it intact, as efficiency of plasmid-186 (Figure 2), in which the transposase both the 50 and the 30 terminal domains overlap with the promoter was deleted, was reduced to background levels (0.09%) transposase gene9 (plasmid-185, Figure 2b). Although these indicating that both transposase expression and full-length terminal modifications made plasmid-185 more complicated than plas- domains are required for successful transposition. mid-137, the replacement of the native promoter with the SV40 promoter ensured predictable transposase expression and the mutation of the TTAA sites prevented the unwanted excision of Partial truncation of the helper part of piggyBac plasmid yields the second transposon from the plasmid. Plasmid-186 (not shown improved transposition efficiency in Figure 1) was an inactive variant of plasmid-185 that lacked a Although vector-185 showed relatively high integration efficiency, promoter for transposase expression and was used to determine it was a large and complicated plasmid. Keeping the RFP delivery the level of non-specific integration into the host cell. cassette transposon with minTRs unchanged, we tried to truncate We transfected Human Embryonic Kidney (HEK)-293 cells and simplify the helper region of the plasmid to make the entire separately with each plasmid. Two days after transfection, cells vector more compact. First we removed the complete 50 minTR were collected and RFP-positive cells isolated using flow and half of the 30 minTR (including both TTAA sites) from the cytometry. These cells were then monitored for RFP expression helper transposon to disrupt the native 50 terminal domain over 28 days. Initial transfection efficiency was about 90% for all promoter and prevent interaction of transposase with these plasmids. Four weeks after transfection, only 0.07% of cells initially terminal sequences of the helper part of the vector (plasmid-166). transfected with plasmid-132 were RFP positive (Figure 2a). The SV40 promoter was then moved directly in front of the entire Co-transfection of plasmid-132 with a helper plasmid containing helper region of the plasmid to drive the transposase expression only the piggyBac transposase under the control of the SV40 as the native promoter was disrupted and non-functional

Gene Therapy (2014) 1 – 9 & 2014 Macmillan Publishers Limited Minimal piggyBac vectors V Solodushko et al 3

Figure 2. (a) The percentage of RFP-positive HEK-293 cells after their transfection with different minimal transposon vectors (n ¼ 4). We transfected Human Embryonic Kidney-293 cells separately with each plasmid. Two days after transfection, all cells were collected and RFP-positive cells isolated using flow cytometry. These cells were then monitored for RFP expression over 28 days. 132 þ 200 and 132 þ p3E1.2: data for co-expression of plasmid-132 with either helper plasmid-200 or p3E1.2 expressing the piggyBac transposase under control of the SV40 or the transposon native promoter, respectively. (b) Detailed presentation of tested plasmids. All tested minimal piggyBac vectors had a delivered cassette that contained the reporter gene, RFP, flanked by 50 and 30 minTRs. Plasmid-132 contained only this RFP-delivery cassette. Plasmid-137 also had a second wild-type piggyBac transposon separated from the RFP delivery cassette by a 683 base pair linker. Plasmid-185 also had two transposons separated by a 254 base pairs linker in the same plasmid: the RFP-delivered cassette and a modified full-length transposon in which both TTAA integration sites flanking the full-length transposon, but not the RFP-delivery cassette, were mutated (GTAA) to prevent potential transposition of the full-length piggyBac into the host genome. In plasmid-185 we replaced the native promoter with an SV40 promoter to more reliably drive expression of the transposase. To keep closer distance between the RFP-delivery cassette and the helper transposon, only the minimal SV40 promoter was placed upstream of the transposase gene, while the SV40 enhancer was placed downstream. This promoter replacement necessitated a partial duplication of the 50 terminal domain to keep it intact, as both the 50 and the 30 terminal domains overlap with the transposase gene. We also added an additional termination triplet and synthetic polyadenylation signal in front of the SV40 promoter to terminate any production of upstream truncated transposase initiated by residual native promoter activity. Control plasmid-186 (not mentioned in Figure 1) was constructed similar to plasmid-185, but without the SV40 promoter. The lack of promoter prevented transposase expression while retaining all potentially important transposition sequences. 50 TRmin, 30 TRmin: minimal 50 or 30 terminal repeats (in light gray); CMV: cytomegalovirus promoter; RFP: red fluorescent protein; pA: polyadenylation signal (a SV40 polyadenylation signal for the RFP delivered cassette on the left and two structurally different (synthetic) polyadenylation signals for the helper segment in plasmids 185 and 186 on the right; and in plasmid-200); full 50 TD, full 30 TD: 50 or 30 full-length terminal domains; 50 ID, 30 ID: internal 50 or 30 domains that do not overlap with the transposase gene (in gray); SV40prom and SV40enh: SV40 promoter or enhancer; PBase: piggyBac transposase gene; PBase trunc: truncated 50 piggyBac transposase gene with added stop codon in vectors 185 and 186 (this produces a truncated transposase). Black vertical lines indicate non-mutated TTAA integration sites flanking transposition-competent sequences. Arrows indicate the orientation of the operons. Prokaryotic origin of replication and ampicillin resistance gene are not shown. (Vectors are aligned for easier comparison, but distances between delivered cassette and the helper part of the plasmid are not drawn to scale.)

& 2014 Macmillan Publishers Limited Gene Therapy (2014) 1 – 9 Minimal piggyBac vectors V Solodushko et al 4

Figure 3. (a) The percentage of RFP-positive HEK-293 cells after their transfection with the indicated transposon vectors (n ¼ 4). Transposition data for plasmids 166 and 206 were compared with plasmids 132 and 185 as described above. Two days after transfection with each plasmid, cells were sorted and cells positive for RFP were re-seeded. Cells were monitored for RFP expression for up to 28 days post-transfection as was described. The streamlined plasmid-166 demonstrated greater integration efficiency than vectors 137 and 185. Thirty-two percent of the initially transfected HEK cells stably incorporated and expressed the delivered transgene (RFP) at 4 weeks. The decrease in transposition efficiency in plasmid-206 (relative to plasmid-166) demonstrated that the presence of long internal domains flanking the transposase gene were required for the successful transposition of the piggyBac vector; these long internal domains could be located outside of the transposon, however, as long as minTRs were present within the transposon. (b) Detailed presentation of plasmids 166 and 206. To make plasmid-166, we first removed the complete 50 minTR and half of the 30 minTR (including both TTAA sites) from the helper transposon to disrupt the native 50 terminal domain promoter and prevent interaction of transposase with these terminal sequences of the helper part of the vector. The minimal SV40 promoter was then moved directly in front of the entire helper region of the plasmid to drive the transposase expression as the native promoter was disrupted and non-functional (plasmid-166). SV40 enhancer was located downstream of the helper part similar to plasmid-185. In this configuration, the minimal SV40 promoter in plasmid-166 also functioned as part of the 248 base pairs linker separating the delivered and helper transposons performing a similar function to that of the 254 base pairs linker in plasmid-185. Removing more sequences from the terminal domains of plasmid 166 (leaving a functional transposase gene, but markedly truncated terminal domains in the helper region, resulted in plasmid-206. 50 TRmin, 30 TRmin: minimal 50 or 30 terminal repeats (in light gray); 30 TRmin trunc: internal 37 base pairs fragment of 30 minTR in the helper part of plasmid-166 (in light gray); CMV: cytomegalovirus promoter; RFP: red fluorescent protein; pA: polyadenylation signal (SV40 polyadenylation signal for the RFP delivered cassette on the left and a synthetic polyadenylation signal for helper segment on the right); 50 ID, 30 ID: internal 50 or 30 domains that do not overlap with the transposase gene (in gray); SV40prom and SV40enh: SV40 promoter and enhancer; PBase: piggyBac transposase gene; full 50 TD, full 30 TD: 50 or 30 full-length terminal domains; trunc. 50 TD, trunc. 30 TD: truncated 50 terminal domain by deletion of 50 minTR and truncated 30 terminal domain by deletion of terminal 26 base pairs fragment (terminal part of 30 minTR) in plasmid-166. Black vertical lines indicate non-mutated TTAA integration sites flanking transposition-competent sequences. Arrows indicate the orientation of the operons. Prokaryotic origin of replication and ampicillin resistance gene are not shown. (Vectors are aligned for easier comparison, but distances between delivered cassette and helper part of plasmid are not drawn to scale.)

(Figure 3a, plasmid-166). As the combined effect of two sequential domains could be located outside of the integrated transposon as promoters (in the previous plasmid-185) may have led to long as minTRs were present within the delivered sequence. interference, this move not only eliminated the need for internal polyadenylation signal sequence, but also the requirement to duplicate the sequences in the 50 terminal domain that overlap Minimal transposon vector allows stable gene delivery in multiple with the piggyBac transposase. cell types This streamlined vector-166 demonstrated greater integration After testing multiple transposon plasmids in HEK-293 cells, we efficiency than vector-137 and vector-185. Thirty-two percent of concluded that plasmid-166 had the highest transposition the initially transfected HEK cells stably incorporated and efficiency. As the overall goal of this project was to develop a expressed the delivered transgene (RFP) at 4 weeks (Figure 3a). plasmid in which most of the full-length terminal domains were Removing more sequences from the terminal domains of removed from the RFP-delivery cassette with minimal impairment plasmid-166 (leaving a functional transposase gene, but markedly of transposition efficiency, we tested this plasmid against one in truncated terminal domains in the helper region) significantly which the RFP-delivery cassette consisted of full-length terminal reduced the plasmid’s transposition efficiency (0.72%) (plasmid-206). domains, typical of existing piggyBac vectors. Therefore, we made The differences in transposition efficiency between plasmids a piggyBac plasmid in which the RFP-delivery cassette consisted of 166 and 206 demonstrated that the presence of long internal full-length terminal domains (plasmid-210, Figures 1 and 4b) domains flanking the transposase gene were required for the instead of the minTRs found in plasmid-166. Similar to the design successful transposition of the piggyBac vector; these long internal of plasmid-166, plasmid-210 also contained the transposase

Gene Therapy (2014) 1 – 9 & 2014 Macmillan Publishers Limited Minimal piggyBac vectors V Solodushko et al 5

Figure 4. (a) The percentage of RFP-positive target cells 28 days after transfection with the indicated plasmids (n ¼ 4). Plasmids were tested in the following cells: HEK-293, HeLa, L929 (mouse fibroblasts) and primary rat pulmonary artery smooth muscle cells. Cells were transfected with the corresponding transposon or non-transposon vector. Two days after transfection, cells were sorted and RFP-positive cells re-seeded. Cells were cultured for 28 days and then analyzed for RFP expression. We compared the results of transposition of the minimal (166) and full-length (210) transposon vectors with that of plasmid-186, (a minimal vector, which does not express the piggyBac transposase) and with that of the naked non-transposon plasmid-211 containing the RFP gene. (b) Detailed presentation of plasmids 210 and 211. We made a piggyBac plasmid in which the RFP-delivery cassette consisted of full-length terminal domains (plasmid-210) instead of the minTRs found in plasmid-166. Plasmid-210 also contained the transposase gene within the same construct under the control of the SV40 promoter. The helper part of plasmid-210 was identical to the one in plasmid-206. Similar to plasmid-166, the minimal SV40 promoter functioned as a part of the 248 base pairs linker that separated the delivered and helper transposons; the SV40 enhancer was located downstream of the helper part. The non- transposon plasmid-211 consisted of the same RFP expressing operon and served as a non-specific integration control. 50 TRmin, 30 TRmin: minimal 50 or 30 terminal repeats (in light gray); CMV: cytomegalovirus promoter; RFP: red fluorescent protein; pA: polyadenylation signal (SV40 polyadenylation signal to terminate RFP expression in plasmid-210 and plasmid-211, and two structurally different (synthetic) polyadenylation signals to terminate truncated and full size transposase expression in plasmid-210); 50 ID, 30 ID: internal 50 or 30 repeats that do not overlap with the transposase gene (in gray); SV40prom and SV40enh: SV40 promoter or enhancer; PBase: piggyBac transposase gene; full 50 TD, full 30 TD: 50 or 30 full-length terminal domains; PBase trunc: truncated 50 piggyBac transposase gene with added stop codon in vector 210 (produces truncated transposase) and 30 truncated variant (no product). Black vertical lines indicate non-mutated TTAA integration sites flanking transposition-competent sequences. Arrows indicate the orientation of the operons. Prokaryotic origin of replication and ampicillin resistance gene are not shown. (Vectors are aligned for easier comparison, but distances between delivered cassette and helper part of plasmid are not drawn to scale.) within the same construct under the control of the SV40 promoter. plasmid, not just the transposon, might have contributed to this These plasmids were then tested in the following cells: HEK-293, RFP-positivity. To prove that only the transposon, but not the HeLa, L929 (mouse fibroblasts) and primary rat pulmonary artery entire plasmid, was integrated into the host genome of cells stably smooth muscle cells. We compared the integration efficiency of expressing RFP, we performed PCR on total cell DNA using distinct both these plasmids to that obtained with the non-transposon primers, one that amplified only the RFP-delivery cassette and plasmid-211 (plasmid-211, transposon-independent integration another that overlapped with part of the non-delivered plasmid. control) and with the transposon vector-186 that had the same Because the first 13 base pairs of both terminal repeats and the piggyBac sequences, but did not express transposase due to the following TTAA integration sites plus the next 7 base pairs flanking absence of the upstream promoter. the RFP delivery cassette in plasmid-166 (total 24 base pairs) are As shown in Figure 4b, plasmid-166 and -210 were successfully symmetrical, we used only one primer for each PCR reaction. The transposed in all cell types studied. HEK-293 cells were the most inner primer (TTAACCCTAGAAAGATA) was complementary to transposable, whereas HeLa cells were the most resistant to the common sequence located at both minTRs (single underline) transposition. Although plasmid-210 demonstrated higher trans- and also included a TTAA integration site (double underline) to position efficiency in all cell types, the differences between the which they are flanked in the plasmid or in chromatin. The outer two were not marked. Both plasmid-166 and -210 had significantly primer (GTCGACTTTAACCCTAGAA) straddled the sequences higher integration efficiencies than the non-transposon naked that transitioned between the non-delivered plasmid and the DNA control (plasmid-211) and transposase-deficient vector-186. delivered transposon. This outer primer partially overlapped with the inner primer (single and double underlines). The part of the outer plasmid that coupled with the sequences within the Stably transgene-positive target cells contain only the delivered transposon was 5 base pairs shorter when compared with the transposase sequence inner primer, yet covered an additional 7 base pairs fragment The preceding results demonstrated that the long internal (dotted underline) located outside of the transposon and TTAA domains could be removed from the delivered cassette to other integration site in the plasmid. The outside 7 base pairs fragment parts of the plasmid without significantly impairing transposition at both sides is not transpositioned and therefore exists only in efficiency. However, the spontaneous integration of the entire the plasmid. Therefore, if only the transposon is incorporated into

& 2014 Macmillan Publishers Limited Gene Therapy (2014) 1 – 9 Minimal piggyBac vectors V Solodushko et al 6

Figure 5. (a) PCR of plasmid-166 and genomic DNA using inner (1) and outer (2) primers. DNA from RFP-negative HEK-293 cells, cells stably expressing RFP 28 days after transfection with plasmid-166 (integrated), and HEK-293 cells that were initially RFP positive, but became RFP negative 28 days after being transfected with plasmid-166 (transiently transfected, non-integrated) and DNA from three different clones of HEK-293 cells that stably expressed RFP 65 days after initial transfection with plasmid-166, was analyzed. The PCR amplification of DNA from all cells stably expressing RFP (RFP( þ )) yielded a PCR product equal to the length of the entire transposon (1656 base pairs) only when the inner primer was used. In contrast, amplification of DNA from these cells using the outer primer failed to produce a PCR product indicating that the plasmid flanking sequences were absent. Negative cells, both those that had never been transfected (RFP( À )) and those that had been transiently transfected, but were now negative after 28 days (RFP( þ / À )), demonstrated no PCR product when amplified with either the inner or the outer primer. Plasmid DNA showed PCR products with both the inner and outer primers. plasmid: plasmid-166 DNA; RFP( À ): total DNA from non-transfected HEK-293 cells (negative control); RFP( þ ): total DNA from cells stably expressing RFP (red fluorescent protein) 28 days post transfection (mix population or clones); RFP( þ / À ): total DNA from cells initially RFP positive after transfection, but RFP negative at 28 days; clones A, B and C: RFP positive clones derived from the mixed population of RFP-positive cells (all 65 days post-transfection); MW: molecular weight markers. (b) Quantitative PCR (qPCR) on DNA samples from different HEK-293 cells and plasmid-166 using the inner (1) and the outer (2) primer and normalized to the inner primer (comparison control) and plasmid-166 DNA (n ¼ 3). Using qPCR we demonstrated a 31-fold decrease in the signal intensity with the outer primer between the DNA from a mixed population of RFP-positive cells 28 day after transfection as compared with the inner primer, normalized to the DNA from the plasmid-166. The small, but measurable, product accumulation seen in the mixed population of RFP-positive cells (also visible in Figure 5a) when amplified by the outer primer was likely due to either the prolonged stability of the (non-integrated) vector plasmid in transiently transfected cells or to non-specific integration of the plasmid into the host genome of a few cells. plasmid: double transposon plasmid-166 DNA; RFP( þ ): total DNA from HEK-293 cells stably expressing RFP; mix: mix population of RFP( þ ) cells 28 days post transfection; clones A, B and C: individual clones derived from mix population of RFP-positive cells 65 days post transfection.

the host cell genome, amplification of total cellular DNA with the of the RFP delivery cassette in both PCR products was confirmed outer primer will not generate a PCR product whereas amplifica- by sequence analysis. We then isolated DNA from: RFP-negative tion with the inner primer will. In contrast, if the entire plasmid HEK-293 cells, cells stably expressing RFP 28 days after transfec- had been incorporated into the host cell genome, both the outer tion with plasmid-166 (integrated), and HEK-293 cells that were and the inner primer will generate a PCR product. If only a part of initially RFP positive, but became RFP negative 28 days after being the RFP delivery cassette had been incorporated, no PCR products transfected with plasmid-166 (transiently transfected, non- would be generated with either primer. integrated). In addition, DNA from three different clones of HEK- We first demonstrated that PCR amplification of plasmid-166 293 cells that stably expressed RFP 65 days after initial transfection with either the outer or the inner primer generated a PCR product with plasmid-166, was analyzed. with a similar size (Figure 5a) and an identical rate of accumulation As shown in Figure 5a, PCR amplification of DNA from cells as confirmed by qPCR (data not shown). The complete sequence stably expressing RFP (RFP( þ )) yielded a PCR product equal to the

Gene Therapy (2014) 1 – 9 & 2014 Macmillan Publishers Limited Minimal piggyBac vectors V Solodushko et al 7 length of the entire transposon (1656 base pairs) only when the were subsequently degraded with the plasmid. The integrated inner primer was used. Sequence analysis of this PCR fragment unit included only the 35 base pairs 50-end and the 63 base pairs confirmed the presence of the intact transposon with the RFP 30-end, plus the transgene. This is significantly smaller than the operon. Three different RFP-positive clones isolated from a mixed residual (non-essential) DNA sequences left by viral or classical population of RFP-positive HEK-293 cells each showed the same transposon vectors. In addition, neither the 50 nor the 30 piggyBac PCR product demonstrating the presence of the entire transposon. minTRs contain known active promoters or enhancers,12,19 further In contrast, amplification of DNA from these cells using the outer improving the safety profile of these gene delivery vectors. Unlike primer failed to produce a PCR product indicating that the plasmid viral vectors, transgene expression can be terminated by a strong flanking sequences were absent. Negative cells, both those that had polyadenylation signal inside the transposon providing additional never been transfected (RFP( À )) and those that had been protection against unwanted activation of host cell oncogenes. transiently transfected, but were now negative after 28 days This modified piggyBac plasmid is a single-plasmid system which (RFP( þ / À )), demonstrated no PCR product when amplified with makes the vectors potentially useful for in vivo applications. This is either the inner or the outer primer (Figure 5a). Using qPCR we in contrast to many current transposon vectors which commonly demonstrated a 31-fold decrease in the signal intensity with the use two vectors system: one to deliver the transposon and a outer primer between the DNA from a mixed population of second to deliver the transposase.4,6,10–12,20,21 In the wild-type RFP-positive cells 28 days after transfection as compared with the piggyBac transposon, sequences from both the 50 and the 30 inner primer and normalized to the DNA from the plasmid-166 terminal domains overlap with the transposase gene by 4300 (Figure 5b). The small, but measurable, product accumulation seen base pairs on each end. As both the terminal domains and the in the mixed population of RFP-positive cells (also visible in transposase are needed for gene delivery we found that keeping Figure 5a) when amplified by the outer primer was likely due to them together in the helper part of the plasmid made the vectors either the prolonged stability of the (non-integrated) vector plasmid more compact without a change in either transposition efficiency in transiently transfected cells or to non-specific integration of the or transposase expression. In the classical piggyBac systems, the plasmid into the host genome of a few cells. The three clones that optimum ratio of transposase to transposon vector for remained RFP positive after 65 days showed no PCR product with transfections in two-plasmid systems may vary to obtain the the outer primer. Overall, these results demonstrate that only the best transposition efficiency. If needed, a single transposon system transposon, and not the rest of the plasmid, was stably incorporated can be designed to retain this flexibility. For example, stronger, into the genome of the target cell. These results also demonstrate weaker or inducible promoters may be used to drive transposase that the loss of fluorescence in the transiently transfected cells expression depending on the amount required to optimize (RFP( þ / À )) was due to the failure of the transposon to integrate transposition. Moreover, our vectors can be co-transfected with into the host genome, rather than to inactivation of the an additional helper plasmid to further control such ratio. This cytomegalovirus promoter controlling RFP expression. strategy would result in a two-plasmid system to deliver the transposon yet retain the advantage of delivering a transposon that contains only minTRs. DISCUSSION In summary, we demonstrated that although long internal In this report, we describe a novel piggyBac vector in which most domain sequences are required for the successful transposition of of the sequences within the terminal domains have been removed piggyBac vectors, they can be positioned outside of the from the delivery cassette without a significant decrease in transposon and still perform this function, something not transposition efficiency. This is achieved by including a second previously demonstrated. This novel design reduces the amount piggyBac sequence (modified, to make it undeliverable) in the of non-essential DNA incorporated into the host genome from same plasmid. This design decreases the size of the required about 1500 to 98 base pairs and does so without significantly terminal domains within the delivered gene cassette of piggyBac decreasing the integration efficiency of the piggyBac vector. This vectors from about 1500 base pairs13 to just 98 base pairs, the reduction in non-essential DNA may decrease the risk of host cell shortest sequence that allows stable transgene integration for any transformation, thus making this vector safer and more attractive viral or non-viral gene delivery system that has been described to for use in human research. date. By removing these sequences from the delivered gene cassette, they are no longer incorporated into the host genome. This reduction in the length of DNA sequences incorporated into MATERIAL AND METHODS the target cell genome not only decreases the risk of insertional Materials 18 mutagenesis, but also eliminates any apparent or potential Dulbecco’s Modified Eagle Medium, Dulbecco’s Modified Eagle Medium/ promoter or enhancer activity that the terminal domains might Nutrient Mixture F-12 (DMEM/F12), 0.05% trypsin/0.53 mM EDTA and 17,19 exert on host cell oncogenes. L-glutamine were all purchased from Gibco (Grand Island, NY, USA). Fetal Efforts by many investigators have focused on shortening bovine serum (FBS) was purchased from Atlanta Biologicals (Lawrenceville, piggyBac’s terminal domains to decrease the mutagenic possibility GA, USA). FuGENE 6 and FuGENE HD Transfection Reagents were of this transposon and thus improve its safety profile. Although purchased from Roche Diagnostics (Indianapolis, IN, USA). All restriction reductions in the size of the terminal domains were possible, this enzymes, DNA polymerase I (Klenow) and High Efficiency Competent E. coli Cells (NEB 10-beta; Cat. No. C3019H) were from New England BioLabs usually resulted in a marked decrease in the efficacy of transposon 9,13 (Ipswich, MA, USA). Hi-Lo DNA Markers were from Minnesota Molecular, integration into the host genome. We hypothesized that Inc. (Cat. No. 1010, Minneapolis, MN, USA). although the full-length terminal domains were important for efficient transposition, they could be removed from the gene delivery cassette without losing integration efficiency as long as Cells these domains were present somewhere within the plasmid. Rat pulmonary artery smooth muscle cells were isolated and characterized 22 We removed the internal domains from the gene delivery in our cell culture core. HEK-293 (HEK cell line (Cat. No. CRC-1573)), HeLa cassette leaving only the minTRs behind and showed that we (cervical cancer derived human cells (Cat. No. CCL-2)) and L929 (murine aneuploid fibrosarcoma cell line (Cat. No. CCL-1)) were obtained from could stably deliver genes to a number of different cell types with ATCC. Pulmonary artery smooth muscle cells were cultured in DMEM/F12, almost similar efficacy as piggyBac vectors with longer terminal 10% fetal bovine serum, 2 mmol L-glutamine for up to 49 days and used for domains. Only the minTRs and the transgene were integrated into experiments at passages 4–9. Other cells were cultured in Dulbecco’s the host genome, while both the piggyBac transposase and the Modified Eagle Medium, 10% fetal bovine serum, 2 mmol L-glutamine and full-length terminal domains in the helper region of the plasmid routinely passaged after reaching 80% confluency. All cells were grown in

& 2014 Macmillan Publishers Limited Gene Therapy (2014) 1 – 9 Minimal piggyBac vectors V Solodushko et al 8

humidified incubators at 37 1Cin5%CO2 and harvested by 0.05% plasmid (plasmid-200) expressing transposase which in some experiments trypsin/0.53 mM EDTA digestion and counted with Coulter Z1 (Coulter was co-transfected with plasmid-132. Electronics, Luton, Bedfordshire, England). Counts were made in triplicate. Another intermediate version of the full-length piggyBac flanked with normal (TTAA) integration sites and SalI restriction enzyme sites was PCR amplified from the p3E1.2 plasmid using a single primer TATGTCGACTT Vectors and delivery systems TAACCCTAGAAAGATA. This outermost sequence was identical to the flanking sequences of the RFP delivery cassette in all double transposon For plasmid-132, 50 and 30 minTRs were consecutively ligated into a basic vectors and represents the site at which the delivered transposon and the plasmid harboring prokaryotic origin of replication and ampicillin non-delivered plasmid fragments join. The 50 SalI–BglII digestion of this resistance gene as synthetic phosphorylated primers, forming a joined product produced the 50 terminal domain for the transposable part of construct with two outside GTCGACT sequences containing SalI plasmid-210. The 30 PstI–SalI digestion of the same sequence liberated the restriction enzyme site, and a single inside BclI restriction enzyme site. 0 A cytomegalovirus promoter, turboRFP and an SV40 polyadenylation signal 3 terminal domain for the transposable part of plasmid-210. Replacing the minTRs in the transposable part of plasmid-206 with full-length terminal were added sequentially between the 50 and 30 minTRs into the BclI site domains resulted in plasmid-210. Plasmid-210 also contained the yielding control plasmid-132. The plasmid harboring the wild-type 12 transposase gene under the control of the SV40 promoter in the helper piggyBac, p3E1.2 (kindly gifted by Dr A Handler ), was used as a base region similar to plasmid-206. plasmid for generating other sequences containing the transposase gene and/or piggyBac terminal domains. This plasmid was also used as a helper All PCR products used for vector construction were sequenced to plasmid with native terminal domains and co-expressed with plasmid-132 eliminate possible errors in amplified fragments. in some experiments. BssHII–BclI digestion of p3E1.2 liberated a full-length wild-type piggyBac which was ligated into plasmid-132 generating a Quantitative PCR double transposon plasmid, plasmid-137. In this vector the minimal Total DNA was isolated from cells using the DNeasy Blood and Tissue kit (RFP-containing) and the full-length wild-type (containing transposase) (Qiagen, Hilden, Germany; Cat. No. 69504). Identical sequences at both piggyBac transposons were separated by 683 and 3478 base pairs linkers. internal ends of the RFP delivery cassette of the vector-166 as well as their An intermediate version of a full-length piggyBac sequence (for plasmid- flanking (non-delivered) regions in the plasmid allowed us to use a single 185) flanked with mutated integration sites (GTAA instead of TTAA) was PCR primer with either inner or outer primers. The inner primer amplified by PCR from a p3E1.2 plasmid using a single primer (TTAACCCTAGAAAGATA) was complementary to the terminal sequence GTAACCCTAGAAAGATA, which served as both the forward and reverse of the transposon and also included the flanking TTAA integration site. The 0 primer. The 5 BglII digestion of this PCR-amplified wild-type piggyBac in outer primer (GTCGACTTTAACCCTAGAA) straddles the TTAA integration which the integration sites were mutated, produced the first part of the sequence that is incorporated as part of the integrated transposon and a 0 extended piggyBac helper sequence carrying the full-length 5 terminal GTCGACT sequence that is present in the original vector, but is not domain for plasmid-185 and also included a copy of the first 350 base pairs incorporated into the host genome. Both primers generated a nearly fragment of the transposase gene. This fragment was cloned into plasmid- identical product when tested on plasmid DNA. Differences in their ability 132 downstream of the RFP delivery cassette followed by a 254 base pairs to generate a PCR product on harvested chromatin was used to determine linker. A synthetic polyadenylation signal and a minimal SV40 promoter the integration efficiency of the transposon vector-166 by iScript SYBR were added further downstream of this sequence. The same primer, Green RT-PCR kit (Bio-Rad, Hercules, CA, USA; Cat. No. 170-8893, Figure 5b). GTAACCCTAGAAAGATA, paired with the reverse primer, GCGCGCCAC Regular PCR was used to show specificity of each qPCR reaction (Figure 5a). CATGGGTAGTTCTTTAGACGAT, yielded a PCR product on p3E1.2 for the Hi-Lo DNA Markers from Minnesota Molecular, Inc. (Cat. No. 1010, second part of the helper sequence carrying transposase gene (with an Minneapolis, MN, USA) were used to identify the size of PCR products. extra BssHII restriction site and a KOZAK sequence upfront) overlapped DNA sequence analysis of all PCR products was done using multiple 0 with the full-length 3 terminal domain. This part was then cloned primers matching the internal parts of the working transposon. downstream of the minimal SV40 promoter, followed by a synthetic polyadenylation signal and a SV40 enhancer to complete plasmid-185. The linker between the 30 terminal domain of the helper part and the 50 minTR Flow cytometry analysis of the minimal transposon was 3354 base pairs and included the synthetic Cells were transiently transfected with corresponding plasmids (each polyadenylation signal, SV40 enhancer, prokaryotic origin of replication expressing turboRFP with excitation/emission of 553/574 nm) using and ampicillin resistance gene. The lengths of both linkers between the FuGENE 6 or FuGENE HD as transfection reagents. Forty-eight hours after transposable and the helper parts (254 and 3,354 base pairs) were the transfection, the cells were harvested by 0.05% trypsin/0.53 mM EDTA same in plasmids 185, 186, 166, 206 and 210. digestion, washed and re-suspended in cultured medium. RFP-expressing Plasmid-186 (shown in Figure 2 only) was constructed similar to cells were sorted by BD Biosciences FACSAria cell sorter. Selected cells plasmid-185 but lacked both the SV40 minimal promoter and enhancer. were re-seeded and the percentage of RFP-positive cells monitored for up Digestion of the original wild-type piggyBac plasmid p3E1.2 with SphI and to 28 days using BD Biosciences FACCantoII cell analyzer in the University BsiWI, followed by blunting of the fragment with DNA polymerase of South Alabama Flow Cytometry Core. I (Klenow) resulted in a piggyBac sequence lacking an entire 50 minTR and half of the 30 minTR for the helper part of plasmid-166. The 254 base pairs linker and the entire piggyBac helper part of plasmid-185 were replaced Statistical analysis with a minimal SV40 promoter and the SphI-BsiWI truncated fragment of Data are expressed as mean±s.e. Changes in percentage of RFP- the wild-type piggyBac transposon to complete plasmid-166. As the native expressing cells and qPCR data were compared using ANOVA combined transposon promoter activity in plasmid-166 was disrupted, a minimal with Fisher post hoc analysis, with a P-value o0.05 considered significant. SV40 promoter was placed upstream of the entire helper part to drive transposase expression with extended un-translated sequence. This made the helper part of the plasmid smaller and eliminated the need for both CONFLICT OF INTEREST 0 the duplication of the 5 terminal domain and the inclusion of an extra The authors declare no conflict of interest. polyadenylation signal to terminate expression of the truncated transposase (as was necessary for plasmid-185). A minimal SV40 promoter also served as a 248 base pairs linker between the RFP delivered cassette and ACKNOWLEDGEMENTS the helper sequence. PCR amplification on plasmid p3E1.2 using the forward primer We thank Dr Alfred M Handler (U.S. Department of Agriculture) for providing the GCCCGTCTAGATTAGTCAGTCAGAAACAACTTT and the reverse primer p3E1.2 plasmid. This work was supported in part by Grant-in-Aid awards from the ATGCGCGCCACCATGGGTAGTTCTTTAGACGAT resulted in the piggyBac Greater SouthEast Affiliate of the American Heart Association to VS transposase gene fragment for plasmids 206 and 210 beginning with a (12GRNT12070291) and BF (09GRNT2260914). BssHII restriction site and the KOZAK sequence and ending with a stop codon and a XbaI restriction site at the 30 end. For plasmid-206 the piggyBac transposase gene was cloned into plasmid-166 by replacing the REFERENCES SphI-BsiWI truncated helper fragment. SalI deletion of the transposable 1 Meir YJ, Wu SC. Transposon-based vector systems for gene therapy clinical trials: minimal piggyBac unit from plasmid-206 resulted in a separate helper challenges and considerations. Chang Gung Med J 2011; 34: 565–579.

Gene Therapy (2014) 1 – 9 & 2014 Macmillan Publishers Limited Minimal piggyBac vectors V Solodushko et al 9 2 Li SD, Huang L. Non-viral is superior to viral gene delivery. J Control Release 2007; 13 Li X, Harrell RA, Handler AM, Beam T, Hennessy K, Fraser Jr. MJ. piggyBac internal 123: 181–183. sequences are necessary for efficient transformation of target genomes. Insect 3 Kawakami S, Higuchi Y, Hashida M. Nonviral approaches for targeted delivery of Mol Biol 2005; 14: 17–30. plasmid DNA and oligonucleotide. J Pharm Sci 2008; 97: 726–745. 14 Elick TA, Bauser CA, Fraser MJ. Excision of the piggyBac transposable element 4 Nakanishi H, Higuchi Y, Kawakami S, Yamashita F, Hashida M. piggyBac trans- in vitro is a precise event that is enhanced by the expression of its encoded poson-mediated long-term gene expression in mice. Mol Ther 2010; 18: 707–714. transposase. Genetica 1996; 98: 33–41. 5 Li MA, Turner DJ, Ning Z, Yusa K, Liang Q, Eckert S et al. Mobilization of giant 15 Elick TA, Lobo N, Fraser Jr. MJ. Analysis of the cis-acting DNA elements required piggyBac transposons in the mouse genome. Nucleic Acids Res 2011; 39: e148. for piggyBac transposable element excision. Mol Gen Genet 1997; 255: 605–610. 6 Wilson MH, Coates CJ, George Jr. AL. PiggyBac transposon-mediated gene transfer 16 Li X, Lobo N, Bauser CA, Fraser Jr MJ. The minimum internal and external in human cells. Mol Ther 2007; 15: 139–145. sequence requirements for transposition of the eukaryotic transformation vector 7 Wu SC, Meir YJ, Coates CJ, Handler AM, Pelczar P, Moisyadi S et al. piggyBac is a piggyBac. Mol Genet Genomics 2001; 266: 190–198. flexible and highly active transposon as compared to sleeping beauty, Tol2, and 17 Cadinanos J, Bradley A. Generation of an inducible and optimized piggyBac Mos1 in mammalian cells. Proc Natl Acad Sci USA 2006; 103: 15008–15013. transposon system. Nucleic Acids Res 2007; 35:e87. 8 Cary LC, Goebel M, Corsaro BG, Wang HG, Rosen E, Fraser MJ. Transposon 18 Meir YJ, Weirauch MT, Yang HS, Chung PC, Yu RK, Wu SC. Genome-wide target mutagenesis of baculoviruses: analysis of Trichoplusia ni transposon IFP2 insertions profiling of piggyBac and Tol2 in HEK 293: pros and cons for gene discovery and within the FP-locus of nuclear polyhedrosis viruses. Virology 1989; 172: 156–169. gene therapy. BMC Biotechnol 2011; 11: 28. 9 Zhuang L, Wei H, Lu C, Zhong B. The relationship between internal domain 19 Shi X, Harrison RL, Hollister JR, Mohammed A, Fraser Jr MJ, Jarvis DL. Construction sequences of piggyBac and its transposition efficiency in BmN cells and Bombyx and characterization of new piggyBac vectors for constitutive or inducible mori. Acta Biochim Biophys Sin (Shanghai) 2010; 42: 426–431. expression of heterologous gene pairs and the identification of a previously 10 Li X, Heinrich JC, Scott MJ. piggyBac-mediated transposition in Drosophila mel- unrecognized activator sequence in piggyBac. BMC Biotechnol 2007; 7:5. anogaster: an evaluation of the use of constitutive promoters to control trans- 20 Kahlig KM, Saridey SK, Kaja A, Daniels MA, George Jr AL, Wilson MH. Multiplexed posase gene expression. Insect Mol Biol 2001; 10: 447–455. transposon-mediated stable gene transfer in human cells. Proc Natl Acad Sci USA 11 Fraser MJ, Ciszczon T, Elick T, Bauser C. Precise excision of TTAA-specific 2010; 107: 1343–1348. lepidopteran transposons piggyBac (IFP2) and tagalong (TFP3) from the 21 Yusa K, Rad R, Takeda J, Bradley A. Generation of transgene-free induced baculovirus genome in cell lines from two species of Lepidoptera. Insect Mol Biol pluripotent mouse stem cells by the piggyBac transposon. Nat Methods 2009; 6: 1996; 5: 141–151. 363–369. 12 Handler AM, McCombs SD, Fraser MJ, Saul SH. The lepidopteran transposon 22 King J, Hamil T, Creighton J, Wu S, Bhat P, McDonald F et al. Structural and vector, piggyBac, mediates germ-line transformation in the Mediterranean fruit functional characteristics of lung macro- and microvascular endothelial cell fly. Proc Natl Acad Sci USA 1998; 95: 7520–7525. phenotypes. Microvasc Res 2004; 67: 139–151.

& 2014 Macmillan Publishers Limited Gene Therapy (2014) 1 – 9