Assembly PCR synthesis of optimally designed, compact, multi-responsive promoters suited to gene therapy application

Gene therapy has the potential to provide innovative treatments for genetic and non-genetic diseases, with the ability to auto-regulate expression levels of therapeutic molecules so that they are produced locally and in direct response to disease activity. Generating disease responsive gene therapy vectors requires knowledge of the activation profile of transcription factors (TFs) during active disease, in order to assemble binding sites for these TFs into synthetic promoters, which can be appropriately activated by the disease process. In this study, we optimised a PCR random assembly approach to generate promoters with optimal spacing between TF binding sites (TFBSs) and their distance from the TATA box. In promoters with optimal spacing, it was possible to demonstrate activation by individual transcription pathways and either additive or synergistic promoter activation when transfected cells were treated with combined stimuli. The kinetics and sensitivity of promoter activation was further explored in transduced cells and when lentivirus was directly delivered to mouse paws a synthetic promoter demonstrated excellent activation by real-time imaging in response to local inflammation.


Supplementary Methods
Assembly PCR reactions were performed using the G-Storm Thermocycler (model GS482, Gene Technologies Limited, Essex, UK) and all oligonucleotides were manufactured by Sigma-Aldrich with standard desalting purification (listed in Supplementary Table 2). Each forward and reverse TFBSoligonucleotide contained the core binding sites of NF-κB, HIF-1α or AP-1 flanked by annealing sequences of 10 bp, 15 bp or 20 bp. Spacer reverse oligonucleotides contained 5 nt or 10 nt spacer sequences flanked by annealing sequences of 15 bp or 20 bp. The 5'-Stop-NheI primers and 3'-Stop-XhoI contained NheI and XhoI restriction enzyme sites, respectively and annealing sequences of 10 bp, 15 bp or 20 bp.
For the initial assembly (x10 cycle) PCR reaction, an oligonucleotide mix containing 5 µl 100 μM forward TFBS-oligonucleotide, 5 µl 100 μM reverse TFBS-oligonucleotide, 12.8 µl 100 μM 5'-Stop-NheI and 12.8 μl 100 μM 3'-Stop-XhoI oligonucleotides in a final reaction volume of 35.6 μl, was diluted 1:100 in sterile distilled (d)H 2 O. A volume equal to the final reaction volume before dilution was added to the assembly PCR reaction e.g. 35.6 μl diluted oligonucleotide mix was added to 12 μl 5x Phusion HF Buffer (NEB), 6 µl 2 mM dNTPs, 0.5 µl Phusion DNA Polymerase (NEB) and sterile dH 2 O in a final volume of 60 µl. The oligonucleotides were assembled into double stranded PCR products in a ten cycle PCR reaction using the thermocycler program: 95°C for 5 minutes, 10 cycles of 95°C for 30 seconds, 60°C for 45 seconds, 72°C for 45 seconds and an additional extension at 72°C for 2 minutes. The reaction was purified from excess dNTPs and free oligonucleotides using the PureLink ® PCR Purification Kit (Invitrogen Corp., Paisley, UK) to yield the purified double stranded assembled PCR product which served as the template for the amplification PCR reaction (x25 cycles). The amplification reaction comprised 15 μl of assembled PCR product, 10 µl 5x Phusion HF Buffer, 0.5 µl Phusion DNA Polymerase, 5 µl 2 mM dNTPs, 5 μl 10 μM 5'-Stop-Nhe I, 5 μl 10μM 3'-Stop-Xho I oligonucleotides and sterile dH 2 O to a final volume of 50 µl. The PCR products were amplified using the thermocycler program: 95°C for 5 minutes, 25 cycles of 95°C for 30 seconds, 60°C for 45 seconds, 72°C for 45 seconds and an additional extension at 72°C for 5 minutes and then purified by two successive high cut-off PCR purification steps using the Binding Buffer HC provided in the PureLink ® PCR Purification Kit (Invitrogen) to remove the amplification primers and failed PCR products <300bp. The purified PCR products were digested with Nhe I and Xho I and cloned into the compatible sites within pCpGmCMV (see below).

Cloning vectors and construction of synthetic promoters
The pGL3smCMV plasmid (33), containing the human small minimal CMV (mCMV) promoter, equivalent to -52 to -14 of the wild type CMV promoter (60) was digested with NheI and AfeI to isolate the mCMV promoter and firefly luciferase reporter gene (Luc + ). The pCpG-mSEAP plasmid (InvivoGen, San Diego, CA, USA) was digested with NheI, SbfI and XbaI to isolate the fragment containing the nuclear matrix attachment regions from the 5'-portion of the human IFN-β gene and β-globin gene. The digested fragments from pCpG-mSEAP and pGL3mCMV were ligated to generate the pCpGmCMV-Luc + construct, which was subsequently digested with NheI and XhoI, to form the pCpGmCMV cloning vector.
Annealed 66 bp-spacer oligonucleotides, containing 5'-XbaI and 3'-XhoI overhangs and an internal NheI restriction site (Supplementary Table 1), were cloned into the compatible NheI/XhoI site within the pCpGmCMV cloning vector and the resulting pCpGmCMV-66 bp plasmid was digested with NheI and dephosphorylated with Calf Intestinal Phosphatase, to form the pCpGmCMV-66 bp cloning vector. The forward and reverse oligonucleotides required for the random ligation cloning method contained the core binding sites of NF-κB, HIF-1α or AP-1 with phosphorylated 5'-CTAG protruding end, compatible to the NheI cleaved end. The pCpG-4bp-composite synthetic promoters were constructed by the random ligation of annealed NF-κB, AP-1 and HIF-1α-oligonucleotides, upstream of the mCMV promoter into the pCpGmCMV-66 bp vector to allow 4 bp space between the TFBSs and 66 bp between the proximal TFBS and the TATA box.
The pCpGmCMV-Xbp plasmids were digested with NheI and XhoI to form the pCpGmCMV-Xbp cloning vectors and the PCR products containing 8xAP-1 or 6xNF-κB were cloned into each cloning vector to allow 55 bp, 60 bp, 66 bp, 70 bp and 74 bp space between the proximal TFBS and the TATA box. The synthetic promoters with 15 bp-60 bp spacing between NF-κB or AP-1 motifs were created by assembling specific combinations of oligonucleotides (listed in Supplementary Table 2). The resulting NheI/XhoI digested PCR products into the pCpGmCMV-5 bp cloning vector, to allow 15 bp-60 bp space between the TFBSs and 66 bp between the proximal TFBS and the TATA box. The clustered composite promoters were assembled by selecting a promoter with the appropriate proximal cluster and digesting it with NheI followed by dephosphorylation and purification. Clusters of 6xNF-κB, 8xAP-1 and 6xHRE were then isolated from the appropriate vectors by digestion with NheI/XhoI, these clusters were then gel purified and two were randomly cloned into the appropriate dephosphorylated vector with the third cluster proximal to the core promoter. The pCpG-20 bpcomposite synthetic promoters were constructed by cloning the PCR products, containing NheI and XhoI cleaved ends, into the equivalent sites within the pCpGmCMV-5 bp cloning vector, to allow 20 bp space between the TFBSs and 66 bp between the proximal TFBS and the TATA box.
All pCpGmCMV plasmid constructs were transformed into chemically-competent GT115 cells and selectively grown in 25 μg/ml Zeocin whilst pGL3mCMV and pRL-CMV (Promega Corp.) plasmid constructs were propagated in chemically-competent DH5α cells and selectively grown in the presence of 100μg/ml carbenicillin. Plasmid DNA was isolated from positive bacterial transformants using the PureLink ® Quick Plasmid Miniprep or HiPure Maxiprep Kits (Invitrogen) and the selected pCpGmCMV-synthetic promoter plasmids were confirmed by DNA sequencing (Genome Centre, QMUL, UK or GATC-Biotech AG, Germany) and/or restriction digestion.

Construction of lentiviral vectors expressing luciferase
The HIV-1 based, 3'LTR self-inactivating (SIN) lentiviral plasmid pLV.CMVenh.gp91.eGFP.cHS4 (Addgene plasmid 30471; Barde et al., 2011) was digested with BamHI and SalI to release the GFP gene and create the LV.CMV cloning vector. The forward and reverse PCR primers; 5'-GATGAGCAGGATCCCATGGAAGACG-3' and 5'-ATGTACGCGTCGACTCTAGAATTAC-3', containing BamHI and SalI restriction enzyme sites (underlined) respectively, were used to amplify the luciferase gene from pCpGmCMV-Luc + . The PCR product was digested with BamHI and SalI and cloned into the corresponding site within the pLV.CMV cloning vector to incorporate the luciferase gene and create the pLV.CMV.Luc + plasmid. The CMV promoter and the 5'-portion of the luciferase gene were removed from the LV.CMV.Luc + plasmid by digestion with PmeI and BstBI to create the LV.Luc + cloning vector. The candidate synthetic promoters and the 5'-portion of the luciferase gene were amplified from their respective pCpG-20 bp-composite promoter constructs using the forward and reverse PCR primers; 5'-GTCGGATTATCGATGCTAGCGTGCC-3' and 5'-ACTCGTAGTTCGAAGTACTCAGCGT-3', containing ClaI and BstBI restriction enzyme sites (underlined), respectively. The mCMV promoter and the 5'-portion of the luciferase gene were amplified from pGL3mCMV using the forward PCR primer; 5'-GTCGGATTATCGATGCGTGCTAGC-3' and the above reverse primer. All PCR products were digested with ClaI and the 5'-end of the DNA was blunted by incubation with Klenow fragment of DNA polymerase and subsequently digested with BstBI. The PCR products were cloned into the PmeI/BstBI within the LV.Luc + cloning vector to restore the luciferase gene and incorporate the candidate synthetic promoters or the mCMV negative control promoter to create LV.2.Luc + , LV.9.Luc + and LV.mCMV.Luc + transfer plasmids, respectively. The pUCL-Luc + plasmid (61) was digested with EcoRI and incubated with Klenow to generate 5'-blunt ended DNA, which was digested with BstBI to release the constitutive SFFV promoter and the 5'-portion of the luciferase gene. The resulting fragment was cloned into the compatible PmeI/BstBI site within the LV.Luc + cloning vector to create the LV.SFFV.Luc + positive control transfer plasmid.

Production of lentiviruses and transduced cell lines.
Human 293T cells (9 x 10 6 ) were seeded in complete DMEM medium in a 15 cm 2 tissue culture dish and transiently transfected with three plasmids using polyethyleneimine (PEI) the following day: 18μg of transfer plasmid (above), 18μg of packaging plasmid pCMVΔR8.2 (62), 4μg vesicularstomatitis virus glycoprotein (VSV-G) envelope plasmid pMD.G (63) were added to OptiMEM medium (Invitrogen) in a final volume of 1ml. Polyethyleneimine (PEI; Sigma Aldrich) was diluted by combining 200μl PEI and 800μl of OptiMEM medium and the DNA:PEI complex was formed by incubating the DNA solution with the diluted PEI for 10 minutes at room temperature. A final concentration of 25 μg/ml chloroquine was added to the cells one hour prior to transfection, after which the transfection mix (2 ml) was added in a drop-wise manner to the 293T cells and incubated at 37°C in 5% CO 2 overnight. The cell medium was replaced with complete DMEM medium the following day and after 72 hours post-transfection, the cell medium containing the packaged lentiviruses were collected and filtered through 0.45μM filters to remove cell debris. The lentiviruses were harvested and concentrated by ultracentrifugation in a Beckman XL-90 Ultracentrifuge (Beckman Coulter, CA, USA) at 23,000 rpm, 4°C for 2 hours and the lentivirus pellets were resuspended in DMEM medium and stored at -80°C. The Lenti-X™ p24 Rapid Titer Kit (Clontech Laboratories, CA, USA) was used to quantify the p24 antigen and therefore, the total viral titre, in the concentrated lentivirus preparations as per the manufacturer's instructions. The titres obtained were in the range 3.3x10 7 to 3x10 8 IFU/ml Oligonucleotide Sequences (

Supplementary Results
Supplementary Figure 1. Schematic representation of assembly PCR and electrophoresis of obtained products. The stages of synthetic promoter synthesis by assembly PCR include assembly, amplification, and restriction digest with NheI and XhoI prior to cloning into vector pCpGmCMV, these steps are schematically represented in A. The effect of annealing sequence length on the annealing and amplification products are illustrated in B and C respectively and the effect of increasing the 'stop' oligonucleotide concentration in the annealing reaction on assembly and amplification products is shown in D and E respectively . Figure 2. Reduced function of optimally arranged singularly responsive promoters following multiple stimulation. The responsiveness of each promoter to stimulation was analysed by 1 way ANOVA (within each promoter) which showed significant differences. A post-hoc Šidak test was then performed to determine when the activity with combined stimulation (grey bars) was significantly (p<0.05) lower than the highest activation caused by an individual stimuli this is indicated by .