Tuneable endogenous mammalian target complementation via multiplexed plasmid-based recombineering

Understanding the quantitative functional consequences of human disease mutations requires silencing of endogenous genes and expression of mutants at close to physiological levels. Changing protein levels above or below these levels is also important for system perturbation and modelling. Fast design optimization demands flexible interchangeable cassettes for endogenous gene silencing and tuneable expression. Here, we introduce ‘TEMTAC’, a multigene recombineering and delivery system for simultaneous siRNA-based knockdown and regulated mutant (or other variant) expression with different dynamic ranges. We show its applicability by confirming known phenotypic effects for selected mutations for BRAF, HRAS, and SHP2.

Scientific RepoRts | 5:17432 | DOI: 10.1038/srep17432 of interest, regulatory elements and other desired functionalities into a single multifunctional multigene delivery plasmid. This is a viable option if the combination of the modules is sufficiently straight-forward. Addressing this challenge, we previously developed the ' ACEMBL' system to enable fast and flexible generation of multigene delivery constructs by an automatable technique called tandem recombineering (TR) to combine 'Donor' and ' Acceptor' plasmid modules via reversible Cre-loxP fusion in vitro, optionally in high-throughput by robotics 15,16 . We implemented ACEMBL successfully for complex expressions in prokaryotic and eukaryotic cells. However, a useful tool that would afford the means to combine efficient DNA delivery with regulated heterologous expression and efficient silencing in cell-based complementation assays remained elusive to date.
Here, we introduce TEMTAC, a novel system for Tunable Endogenous Mammalian TArget Complementation by multiplexed plasmid-based recombineering. Retaining the original ACEMBL concept, we developed this novel system which affords the means to simultaneously knockdown endogenous targets via RNA interference [8][9][10] and express mutated versions utilizing regulated tetracycline (TET) controlled transcriptional activation 17 . Three modules are combined, one each for (i) shRNA-mediated silencing, (ii) TET-regulated expression of the mutant gene of interest (GOI) that contains silent mutations to forestall its own degradation, and (iii) integration of the complete composite plasmid, which expresses an additional YFP marker gene for identifying transfected cells by fluorescence (Fig. 1a). The three modules can be easily assembled and disassembled in vitro exploiting the Cre-Lox recombination reaction, which can be carried out optionally in high-throughput by robotics applying automated routines we have implemented 16 . The separate modules also work autonomously, for example to test proper functioning of each individual element prior to multigene recombineering. A homing endonuclease (HE)/BstXI based multiplication module 15 is provided to enable engineering within each Donor or Acceptor, if several proteins, for example entire regulatory cascades, are to be complemented concomitantly. As a further option, TEMTAC also affords stable integration into cellular genomes, by providing on the Acceptor module eukaryotic resistance markers and the DNA elements required. We demonstrate the utility of TEMTAC utilizing proteins of the ErbB signalling network, HRAS, BRAF, and SHP2. By using TEMTAC, we silence the proteins of interest, express corresponding mutants, and test these for signalling phenotypes, compellingly validating our approach.

Results
Endogenous gene silencing module. Donor D1 (pMDC-RNAiDual) contains the silencing cassette for knocking down the endogenous gene by RNA interference (Fig. 1b). In one Donor D1, two shRNAs can be bi-directionally co-expressed downstream of H1 and U6 Pol III promoters. This allows the testing of the silencing effects of multiple shRNA hairpin sequences. In addition, previous work has shown that silencing effects are often additive and that combination therefore increases knockdown efficiency, without cumulative off-target effects 18 . We tested the silencing effect of the independent Donor D1 for three proteins. For BRAF and SHP2, we inserted two published siRNA sequences [19][20][21][22] into Donor D1. In the case of RAS we cloned new shRNAs which we designed using online software (http://www.invivogen.com/sirna-wizard). Donor D1 containing the shRNAs was transiently transfected in HEK293 or HeLa cells. After 24 and 48 hours, cells were lysed and endogenous protein levels detected by Western blotting, real time PCR, and deep sequencing. We observed the best silencing effect 48 h post transfection and no (or small) down-regulation using random ('scrambled' scrb) shRNAs as a control ( Fig. 1c and Supplementary Figs. 1 to 4). Incomplete silencing will depend on the efficiency/specificity of the respective shRNA's used. A further aspect is that transfection is transient and the transfection efficiency is small than 80-90% in most cell lines.
Inducible exogenous mutant expression module. Achieving expression levels that are close to endogenous levels is a crucial prerequisite for analysing the physiological effect of mutations. This is particularly important when studying signalling proteins because of their frequent involvement in downstream activation and feedback loops, which often function at very precise protein levels that need to be maintained meticulously. A major problem in many TET regulated systems is basal expression (without the inducer being present), especially for proteins that are present at markedly low abundance in cells. Furthermore, depending on the protein of interest, different dynamic expression ranges of mutants would be highly desirable to properly recapitulate physiological events. To address this, we created as Donor D2 variants four different TET-related expression system versions with defined properties (Donor D2.1 to D2.4, Fig. 1d and Supplementary Fig. 5).  Integrator module. The final component of TEMTAC is Acceptor A, which provides the integrator cassette that receives the elements from both the silencing and inducible expression modules (Fig. 2b). Acceptor A contains the elements required for Donor integration, construct selection, and transfection visualization by monitoring expression of yellow fluorescent protein (YFP). In addition to transient multigene expression by transfection, Acceptor A can also be used to generate stable cell lines, as it contains the elements required for stable integration of multigene constructions. To this end, two rare homing endonuclease sites, I-SceI and PI-PspI, are included in the Acceptor, between two eukaryotic drug selection markers for efficient linearization with these extremely rare cutters. The two resistance markers are flanking the multigene construction that would result from Cre-mediated Donor D1 and D2 incorporation into Acceptor A, to ensure integration of the complete multigene construction in the resulting stably expressing cell lines by exposure to both drugs simultaneously during selection. Moreover, Acceptor A also contains the flippase recognition target (Frt) site for stable integration into commercially available flippase (Flp)-expressing cells (i.e. Flp-In-293, Invitrogen) if desired. We tested Acceptor A by Cre-LoxP fusing with Donor D2.1-mCherry (containing mCherry as GOI). We observed both a high transfection efficiency transfection (monitored by YFP expression) and moreover also proper TET-regulated expression of the mCherry protein ( Fig. 2c) and of luciferase ( Fig. 2d and Supplementary Fig. 6).
Biological application. We next tested our TEMTAC system with the three oncogenic proteins SHP2, BRAF and HRAS. Similar as observed for luciferase assays, the three target genes could be expressed in a doxycycline-dependent manner to very high levels (Supplementary Figure S7 and S8). In all cases, we silent-mutated the GOI to be expressed exogenously, by altering the nucleotide sequence at the 3'prime end of the positions where the two shRNAs bind, to prevent the induction of the degradation machinery on the exogenous mRNA. We first tested the combined down-regulation of endogenous BRAF with exogenous complementation by a BRAF V600E cancer mutation using a Donor D2.4 variant (pMDS-GOI-TETon-tTS) in HEK293 cells grown in serum-containing medium. Exogenous expression of this oncogenic mutation tends to be higher as compared to wildtype (WT) (C. Kiel et al. submitted, 2015). To rectify this, we took advantage of our TET-regulated system to tune the expression levels of replication derived from phage R6Kγ , rendering their propagation in regular cloning strains dependent on productive Cre fusion with Acceptor A. (b) Donor D1 (pMDC-RNAiDual) is shown in a schematic view. This Donor provides cassettes for multiple shRNA production. (c) shRNA-mediated downregulation of SHP2, HRAS and BRAF after transfection with a Donor D1 producing specific shRNAs. Transfected HEK293 (for HRAS and BRAF) or GH-HEK293 cells were lysed and analysed by Western blotting. (d) Four Donor plasmid variants D2.1 to D2.4 are shown schematically, which realize four distinct dynamic ranges of exogenous protein expression. Abbreviations: Cre, Cre recombinase enzyme; LoxP, imperfect inverted repeat recognized by Cre; GOI, gene of interest; A-D1, fusion of Acceptor A with Donor D1; A-D2, fusion of Acceptor A with Donor D2 (or variants); A-D1-D2, complete fusion of Acceptor A with Donors D1 and D2 (or variants); shRNA, small hairpin RNA sequence; I-SceI, PI-PspI and PI-SceI are homing endonucleases; H1, U6, CMV and CAG are common mammalian active promoters; pA and SV40 are common polyadenylation signals; TRE-pCMVmin, tetracycline response element with minimal CMV promotor; rtTA, tetracycline transactivator with (random) mutagenesis derive Tet repressor part of the transactivator gene; Sp, Cm, Hygr and Zeo denote resistance marker genes for spectinomycin, chloramphenicol, hygromycin and zeocin, respectively; YFP, yellow fluorescence protein; TetR, tet repressor gene; TetR-KRAB, tetracyclinecontrolled hybrid protein of TetR with the KRAB silencing domain of human Kid1; IRES, internal ribosome entry site; Frt, FLP recognition target; ColE1, common colicin E1 derived replication origin; R6Kγ , conditional origin derived from R6Kγ phage.
WT and V600E to be close to identical (4 ng/ml of doxycycline to express WT and 2 ng/ml of doxycycline to express V600E; Fig. 3a and Supplementary Fig. 9). Higher, but also similar expression levels were obtained by using ratios of 12(WT):20(V600E) and 6:10 ng/ml doxycycline, respectively. When comparing similar protein expression levels, we found in all cases much higher MEK (gene IDs MAP2K1 and MAP2K2) phosphorylation levels, which characterizes this oncogenic mutation 20 . Secondly, we expressed an oncogenic mutant of HRAS (G12V) using donor D2.1 (pMDS-GOI-TETon) in HEK293 cells. Again, higher concentrations of doxycycline were needed for the WT construct to result in similar expression levels as for HRAS (Fig. 3b and Supplementary Fig. 10). As expected, the HRAS G12V mutation resulted in higher ERK (gene IDs MAPK3 and MAPK1) phosphorylation levels. Moreover, we also tested both a phosphatase dead (C459G) and an activating (Noonan syndrome-causing, D61G) mutation of SHP2 (gene ID PTPN11) using Donor D2.1 (pMDS-GOI-TETon-3G) in HEK293 cells that stably express the growth hormone (GH) receptor 23 . SHP2 WT and mutant expression levels were similar at the same doxycycline concentrations. At 24 h after transfection, GH-HEK293 cells were induced with 1 ng/ml doxycycline in serum-free media for 24 hours. Cells were stimulated for 15 min with growth hormone, lysed in SDS buffer, and analysed by Western blotting (Fig. 3c and Supplementary Figure S11). Corroborating previous observations 24 , we found reduced ERK phosphorylation with the phosphatase dead (C459G mutation) and increased ERK signalling with the D61G mutation.

Conclusions
We created and compellingly validated TEMTAC, a novel, highly efficient system for tuneable endogenous mammalian target complementation. TEMTAC exploits multiplexed plasmid-based recombineering, relying on automatable routines for heterologous DNA insertion and Cre-LoxP mediated plasmid fusion in custom-designed Acceptor and Donor plasmid modules, compatible with robotics and high-throughput. TEMTAC enables production of multiple different shRNAs for efficient knockdown of endogenous proteins combined with simultaneous, highly regulated expression of protein variants such as exogenous mutants. Tuning of exogenous proteins to the desired physiological level is achieved in TEMTAC by provision of several different TET-inducible promotor systems covering a large dynamic range. Moreover, it allows for testing hypotheses on how quantitative protein level alterations may impact specific cellular responses. Complex multicomponent signalling systems and even complete metabolic pathways can be analysed by TEMTAC in this way. Furthermore, cancer phenotypes, in which a complete signal cascade is often mutated, can be investigated by TEMTAC, for instance by generating arrays of variants of cascade components that approximate the malignant mutant state.
Our system is by no means restricted to mutation analysis. TEMTAC can also be used for other applications, such as to express alternative spliced isoforms to reveal their cellular function. Likewise, protein variants such as fusion proteins can be substituted for wild-type, fine-tuned at endogenous levels for subsequent comprehensive functional analysis. Tags for efficient tandem affinity purification or fusion proteins of biotin ligase can be included to detect specific endogenous signalling complexes and their dynamic assembly. Moreover, genetic elements for viral encapsulation can be incorporated in TEMTAC, for infecting primary cells to dissect mechanisms that control cell fate. A modified version of the Acceptor can be used in combination with the PiggyBac system for generating stable cell lines 25 (Supplementary Figure S12). Furthermore, the presence of the fluorescent marker YFP enables cells to be efficiently sorted by FACS for subsequent analysis of transfected cells at the single-cell level. We anticipate that a wide range of applications will benefit from TEMTAC for genotype-phenotype analysis in mammalian systems, including genome-wide characterization of alternative splice variants, small nucleotide polymorphisms, mutations under various physiological-, developmental-and cell type-specific conditions in health and disease states.

TEMTAC plasmid design and preparation. Plasmids Donor D1 (pMDC-RNAiDual) and Donor
D2 (pMDS-GOI) were created based on the original plasmids of the ACEMBL system for protein complex expression in prokaryotic hosts 15 . Briefly, origins of replication, resistance markers, the loxP sequence, the multiplication module consisting of matching homing endonuclease site and BstXI were retained from the original ACEMBL Donors and the former protein expression cassettes exchanged by the novel production modules for producing shRNAs (D1) or the protein variant of interest (D2), respectively. These novel modules were designed based on commonly used mammalian active expression cassettes for shRNA or genes of interest and synthesized by a commercial supplier (GenScript Corporation, Piscataway, NJ USA). Synthetic fragments were inserted into the DNA plasmid backbones using standard restriction/ligation cloning. In Donor D1 (pMDC-RNAiDual), shRNAs are bidirectionally co-expressed using specific polymerase III promoters H1 and U6, respectively, as previously described 24 . A series of Donor D2 plasmids was designed to enable tunable expression by adapting commercially available tetracycline-dependent induction (TET) modules (Clontech laboratories, Mountain View, CA USA). Thus, Donor D2.1 (pMDS-GOI-TETon3G) comprises a TET module derived from Tet-On 3G; Donor D2.2 (pMDS-GOI-TEToff) a TEToff module; Donor D2.3 (pMDS-GOI-TETon-PGK) is derived from D2.1, however with the promotor for rtTA exchanged to a minimal promotor (PGK); Donor D2.4 (pMDS-GOI-TETon-tTS) is derived from D2.1 however comprises a transcriptional silencer (Tet-tTS).
Plasmid Acceptor A was designed de novo from scratch. All DNA elements (YFP expression cassette, resistance markers, multiplication module, Frt site and homing endonuclease sites for linearization, see Microscopy. HEK293 cells were seeded on 6 well plates and grown to 60% confluency in normal growth medium. Cells were transfected with 2.5 μ g of the fully Cre-assembled A-D1-D2 composite plasmid containing mCherry as the GOI (containing siRNA scramble) using lipofectamine 2000 (Invitrogen). 24 hours after transfection, cells were supplemented with indicated amounts of doxycycline (Sigma-Aldrich, Carlsbad, CA USA). 24 h later YFP and mCherry fluorescence was imaged on an inverted DMI-6000 Leica wide-field fluorescent microscope equipped with a Leica DFC 350FX camera with a 20× objective.
Amplicon sequencing/MiSeq. RNA  The resulting amplicons were used for determination of cycle numbers for the second PCR with the TruSeq dual barcoding primers. The primary PCR product was amplified using NEBNext 2x High Fidelity PCR Master Mix (New England Biolabs), with 200nM primer concentration, and 0.1x SYBR Green I (SIGMA Aldrich; S9430) monitoring the amplification in a Roche LightCycler 480. The cycle number for the final library PCR was chosen to get a sufficient amount of product without reaching the plateau phase of the PCR. Then 2.5 μ l of the primary PCR product was used for amplification with the NEBNext Master Mix and barcode primers (forward adaptor primers: 5′ -AAT GAT ACG GCG ACC ACC GAG ATC TAC AC [8 bp barcode] ACA CTC TTT CCC TAC ACG ACG CTC TTC-3′ ; reverse adaptor primers: 5′ -CAA GCA GAA GAC GGC ATA CGA GAT [8 bp barcode] GTG ACT GGA GTT CAG ACG TGT GCT CTT C-3′ ) but without SYBR Green, for the number of cycles determined from the qPCR. PCR products were purified using AMPure XP beads and concentration was determined on a Bioanalyzer DNA 1000 chip. Samples were sequenced using MiSeq v3 chemistry with 130 cycle single reads in a MiSeq sequencer (Illumina). Flanking reads were subtracted by mutant reads to calculate the reads corresponding to the endogenous expression level.