Genome editing in potato plants by agrobacterium-mediated transient expression of transcription activator-like effector nucleases
Genome editing (also known as targeted mutation) has promise for molecular breeding. Compared with the CRISPR/Cas9 system, the transcription activator-like effector nucleases (TALENs) have likely a lesser off-target rate in genome editing. Both a rapid test system for the functionality of designed TALENs and an effective delivery system for introducing the TALENs to plants are critical for successful target mutation. TALENs have usually been tested in protoplasts or introduced to plants with viral vectors. However, plant regeneration from protoplast culture can generate extensive somatic variation. Viral vectors are not always available, and plants edited by these vectors usually require virus elimination. Here, we used a non-viral, Agrobacterium-mediated transient expression approach, to serve both rapid test and effective delivery of TALENs into two vegetatively propagated potato cultivars, Solanum tuberosum Russet Burbank and Shepody. Two TALENs with different molecular weights (22 and 27 aa-repeat modules) were expressed to target two endogenous genes (starch branching enzyme and an acid invertase) by Agrobacterium-mediated infiltration (agroinfiltration) into leaves of potato plants. The infiltrated leaf DNA was analyzed using restriction site loss assay and subsequent DNA sequencing. Deep sequencing of these tetraploid cultivars was also conducted to determine the zygosity at the targeted chromosomal loci. TALENs, with different molecular weights, successfully agroinfiltrated and induced mutations at both targeted loci.
KeywordsNon-transgenic Agroinfiltration Site-specific mutagenesis Polyploid plants Allele specificity Vegetatively propagated plants Somatic genome manipulation Molecular breeding
Sequence-specific nucleases are powerful enzymes/proteins for generating double-strand break (DSB) of nucleic acid (DNA) in living cells. Engineered sequence-specific nucleases, including zinc-finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs) and RNA-guided nucleases (CRISPRs), have all been reported to induce double-strand DNA breaks (DSBs) (Jinek et al. 2012; Puchta and Fauser 2014; Voytas and Gao 2014). A TALEN consists of a DNA-binding domain transcription activator-like effector (TALE) and a FokI endonuclease (Boch et al. 2009; Moscou and Bogdanove 2009; Boch and Bonas 2010; Christian et al. 2010). The TALEN approach has greater efficiency than the ZFN approach (Chen and Gao 2014) and likely a lesser off-target rate than the conventional CRISPR approach (Chen and Gao 2014; De Lange et al. 2014; Fichtner et al. 2014). Wang et al. (2015b) found that TALENs have higher target specificity over CRISPR–Cas9 in human cells. TALENs’ greater accuracy underlines its importance among targeted mutation technologies.
The TALE, a part of the engineered TALEN, was initially described in plant pathogenic Xanthomonas bacteria (Kay and Bonas 2009; Boch and Bonas 2010). Each TALE carries a central DNA-binding domain of tandem, repeated modules with typically no acid residues (102 nt at the gene level) (Boch et al. 2009; Kay and Bonas 2009; Moscou and Bogdanove 2009; Boch and Bonas 2010). These modules are very similar, with variation only at the 12th and 13th amino acids. At the gene level, this two-codon region is hypervariable. These two amino acid residues are called repeated-variable di-residues (RVD); each module has an RVD (Deng et al. 2012; Mak et al. 2012). Different RVDs can bind with different nucleotides of the targeted gene. The four most common di-residues of RVDs include NI, HD, NG and NN which can bind to adenine, cytosine, thymine and either guanine or adenine, respectively. Both di-residue NH and NN can bind to guanine, but NH has greater binding specificity than NN (Cong et al. 2012; Streubel et al. 2012), while NH and NK have similar specificity for binding to guanine (Streubel et al. 2012). The functional FokI endonuclease is a dimer (Kim et al. 1996).
The TALEN approach to genome editing involves protein-based targeting because the TALEN-produce- protein in the cytoplasm is translocated to the nucleus to bind to the targeted site for inducing mutations. Various methods to deliver TALENs into plant cells have been tested. A very effective approach is to insert the designed TALEN gene into the plant genome and let its protein product continuously induce mutations, thereby increasing the effectiveness of genome editing. This stable transformation approach has been successfully used in Arabidopsis thaliana (Christian et al. 2013), rice (Oryza sativa) (Li et al. 2012b; Shan et al. 2013), Brachypodium distachyon (Shan et al. 2013) barley (Hordeum vulgare) (Wendt et al. 2013), cabbage (Brassica oleracea) (Sun et al. 2013), soybean (Glycine max) (Haun et al. 2014) and potato (Solanum tuberosum L.) (Wang et al. 2015a).
In the case of diploid cereals, these inserted genes can be removed by genetic segregation in the progeny. However, this genetic segregation approach cannot be used in vegetatively propagated crops such as potato, because sexual reproduction would totally change the clone genotype. In addition, transgenic approaches have a strong negative public perception. A transient expression system that does not insert genes into the crop genome, but transiently produces proteins in the cytoplasm, may find greater public acceptance.
Two transient expression systems tested in plant genome editing are the viral vector approach (Baltes et al. 2014) and DNA uptake by protoplasts (Shan et al. 2013; Zhang et al. 2013; Clasen et al. 2015; Nicolia et al. 2015). Coexpression of two TALEN plasmids by agroinfiltration into tobacco plants has been used to test the efficiency of TALENs at the plasmid level, but the study did not address whether the endogenous gene was mutated (Mahfouz et al. 2011; Li et al. 2012a). Protoplast culture is time consuming and likely induces a greater extent of random somatic variation than micropropagation (Li et al. 1988b), while the agroinfiltration approach may result in faster regeneration from leaf explants.
Success has been reported in TALEN-induced genome editing in various diploid species including Arabidopsis, tobacco (Nicotiana tabacum), barley, cabbage, Brachypodium and rice (Carroll 2014), as well as recently in soybean (diploid-like in chromosome pairing) (Haun et al. 2014). Potato is one of the most important vegetable and food crops in the world. Research is required to effectively mutate genes on targeted potato genome loci. However, potato is a highly heterozygous autotetraploid, with up to four alleles at each locus. It is unknown whether TALENs can target the alleles selectively or nonselectively.
In this study, we designed TALENs to target two endogenous genomic loci (two genes) of two potato cultivars (Russet Burbank and Shepody) and transiently delivered the constructs to potato leaves through agroinfiltration. Endogenous genomic DNA mutation was confirmed at both loci using a restriction site loss assay, DNA clone sequencing and potato genomic deep sequencing.
Materials and methods
Potato genes targeted
Two potato genes targeted in this study were represented by the following genes/cDNAs: 1,4-alpha-glucan branching enzyme gene (SBE1, GenBank accession Y08786) (Khoshnoodi et al. 1996) and StvacINV2 (GenBank accession XM_006355428) (Liu et al. 2011). BLASTn search was performed against the potato reference genome through the RefSeq Genome Database (refseq_genomes) at NCBI (http://www.ncbi.nlm.nih.gov/). The potato reference genome had identical sequences to these two originally published gene sequences. Since this study mainly focused on whether the agroinfiltration method could induce mutation and not on the biochemical function of the genes, we tested one locus (target site) per gene.
Two potato cultivars, Shepody and Russet Burbank, were used for the Agrobacterium tumefaciens-mediated transient expression experiments. In vitro plantlets were transplanted into a computerized greenhouse and grown for 4 weeks (before flowering) in a controlled environment (16 h light/8 h cool white fluorescent; 23/18 °C day/night cycle; 60% relative humidity).
TALEN gene design and expression vector construction
Description of engineered TALENs and target sites
Left TALE modules (RVD)
Right TALE modules (RVD)
RVDs of left TALE modules
RVDs of right TALE modules
NH NH NI NI HD NG NG NI NG NH NI NG NH HD NI NH NH NH NH HD
NI HD HD NI NI HD NI NG HD NI NI NG NG NI NG NG NH NH
NH NI NG NI NG NG NH NH HD HD NI NI NH NH NG NG HD NG HD NI
HD HD HD NI HD NI NI HD NG NG NI NG NH NH NG NI HD HD HD NG
Agrobacterium tumefaciens-mediated transient expression in potato
All the recombinant vectors were introduced into A. tumefaciens strain GV3101 by the freeze–thaw method (Hofgen and Willmitzer 1988). A single colony of the Agrobacterium strain was cultured in 5 ml of LB liquid medium (10 g/l tryptone, 5 g/l yeast extract, and 5 g/l NaCl; pH 7.0) containing the antibiotics spectinomycin (50 mg/ml) and rifamycin (25 mg/ml) and grown overnight (28 °C at 225 rpm). The 5 ml suspension was then transferred to 50 ml LB liquid medium for overnight culture at 28 °C to an OD600 of 1.0. The bacteria were harvested by centrifugation at 5500 rpm for 15 min and resuspended in 1 ml of infiltration buffer (10 mM MgCl2, 10 mM MES-K, pH 5.6) supplemented with 100 µM acetosyringone. This step was repeated at least once and the final OD600 was adjusted to approximately 0.2–0.4. The bacterial suspension was taken up into in a 1-ml syringe without a needle. The infiltration leaf area was gently pressured with a thin pipet tip to facilitate infiltration without damage to the leaves. The infiltration was through the abaxial (lower) surface of the leaves.
Illumina deep sequencing of the two potato cultivars
Plants of cv. Russet Burbank and cv. Shepody, without Agrobacterium infiltration, were sequenced. Total DNA was extracted from leaves, roots and stems/or tubers using the Qiagen kit “DNeasy Plant Maxi Kit (Qiagen Gmbh, D-40724, Hilden, Germany), and the OD 260/280 ratio was measured using a NanoDrop 1000 Spectrophotometer (Thermo Scientific, Wilmington, DE, USA). DNA quantity was estimated using a Qubit Fluorometer (Invitrogen, USA) with PicoGreen. The genomic DNA samples were used to construct TruSeq genomic DNA libraries and sequenced with a 100 bp paired-end Illumina HiSeq 2000 sequencer. The Illumina reads were analyzed using the software Trimmomatic (Lohse et al. 2012) (v 0.25, http://www.usadellab.org/cms/in-dex.php?page=trimmomatic), BWA (Li and Durbin 2009) (v 0.6.2, http://bio-bwa.sourceforge.net/), Samtools (Li et al. 2009) (v0.1.18, http://samtools.sourceforge.net/) and Picard (v 1.82, http://picard.sourceforge.net/). All the high-quality reads (passed the quality check of Trimmomatic) at the genome editing sites were selected and aligned to determine the allele numbers at each site in each control (unmutated) cultivar. The sequences were deposited to NCBI Sequence Read Archive (SRA) (www.ncbi.nlm.nih.gov/sra). More details of the Illumina sequencing and related bioinformatic pipeline can be found in Xiang and Li (2015). The run accessions of the cv. Russet Burbank genomic DNA sequencing are SRR1501275, SRR1501287 and SRR1501288. The run accessions of the cv. Shepody genomic DNA sequencing are SRR1501290, SRR1501291 and SRR1501293.
PCR amplification of the targeted genomic mutation region after agroinfiltration
Primers used for amplifying the target mutation regions
Sequence (5′ to 3′)
Restriction site loss assay for detection of targeted DNA mutation
The PCR products were digested with appropriate restriction enzymes to cut into the spacer (i.e., the TALEN target site) between the two TALE-binding sequences. The digestions were size separated by electrophoresis on 2% agarose gels and viewed under UV light. If a DNA molecule is not mutated at the restriction site, the DNA is cleaved by the enzyme and generates two smaller fragments. If the DNA molecule is mutated at the endonuclease restriction site, the molecule cannot be cleaved and therefore remains in the original PCR fragment size. On the electrophoresis gel, the digestion products of the restriction site loss assay would show three fragments, in which the largest fragment is the mutated one.
The agroinfiltrations were repeated four times on different dates. The restriction resistant band (residual, original size band on agarose gel) in the restriction site loss assay was detected every time. We conducted DNA sequencing of this resistant band to verify whether some DNA in the restriction resistant band was from incomplete digestion and to characterize where the mutations happen in the targeted site region. Since the resistant band was faint, we pulled the DNA of this resistant band from all four experiments, established the DNA library and sequenced 7–8 DNA clones for each gene.
Design of TALENs targeting endogenous potato genes
One TALEN was designed and reconstructed to a pK7FWG2 vector (Fig. 1a) for each potato gene (Table 1; Fig. 1b). Each TALEN was named by adding a T to its gene name, as indicated by the following: T-SBE and T-StvacINV2 (Fig. 1c). The TALENs had 18–20 modules in the TALE of each right TALEN monomer and left TALEN monomer (Fig. 1c). PCR primers (Table 2) were designed for each of the two TALEN-targeted regions. The genome editing results, as confirmed by the results of sequencing, are presented as follows for each of the two targeted loci.
Induced mutation at the targeted site of the SBE1 gene
Induced mutation at the targeted site of the StvacINV2 gene
The results presented here clearly confirm that agroinfiltration with non-viral TAEN vectors can effectively introduce TALENs into potato leaf cells and can allow TALENs to induce mutation at specific targeted sites. It is known that plant defense reactions are different in different plants (Higgins et al. 1998; Bhaskar et al. 2009; Xing and Laroche 2011). Cultivar differences were observed in Agrobacterium-mediated transient expression of binary vector-carried protein genes (Bhaskar et al. 2009). The agroinfiltration was used for CRISPR/Cas9-mediated genome editing in plants (Li et al. 2013; Bortesi and Fischer 2015). In the present study, agroinfiltration effectively delivered the TALENs and induced mutation in both Russet Burbank and Shepody. The A. tumefaciens GV3101 used in this study is a strain widely used in genetic transformation of commercial potato cultivars (Bhaskar et al. 2009; Guo et al. 2010). GV3101 showed a higher efficiency for transient expression than LBA4404 in potato leaf agroinfiltration (Bhaskar et al. 2009). The mutation efficiency is usually difficult to determine in agroinfiltration approaches, because agroinfiltration treats intact plant tissues and not individual cells. However, as indicated by Bhaskar et al. (2009), the indirect efficiency, such as the transient gene expression activity represented by the fluorescent light intensity under microscope in the infiltrated tissues, is known to be affected by the type of vectors, infiltration methods and plant cultivars.
The success in this study with non-viral vectors suggests that the agroinfiltration approach can be used to replace the protoplast approach at least for testing TALEN constructs. Further research about how to increase deletions or amino acid changes in non-viral agroinfiltration approaches may make the targeted genome editing more effective to meet research needs. The remaining step is the regeneration from leaf tissues and screening for mutants. Plant regeneration from leaf explants is expected to be easier than from protoplasts. It is known that the extent of somaclonal variation is very large from protoplast culture (Li et al. 1988a) and much smaller in non-protoplast-culture somatic embryogenesis (Nassar et al. 2011). It is important to avoid the use of a protoplast culture approach in genome editing-based molecular breeding if random mutations are unwanted, even though somaclonal variation is also useful for breeding. An agroinfection approach was used to deliver RNA viral vectors of ZFNs to tobacco plants (Marton et al. 2010) and DNA viral vectors of TALENs (Baltes et al. 2014); an angroinfiltration method was used to deliver CRISPR/cas9 to Nicotiana benthamiana (Li et al. 2013; Nekrasov et al. 2013). Compared with the protoplast culture approach, agroinfiltration is expected to save time, be more cost-effective and less laborious, and increase the chance of obtaining genetically improved plants with mutation only at the target site in the genome with less somaclonal variation.
Both INV2 and SBE TALENs successfully induced mutations in this study. The targeted genes included key genes affecting the degree of starch branching (SBE1) and potato cold sweetening (two acid invertase genes). The Effector Nucleotide Targeter 2.0” program (Doyle et al. 2012) can be used to design TALENs from nearly every nucleotide or every few nucleotides of the input DNA sequences. Therefore, the TALEN approach is powerful and can likely target nearly any gene in potato if needed. Potato is an important food and vegetable crop and the targeted mutation provides a very promising technology for non-GMO-based molecular breeding. Potato cultivars are vegetatively propagated clones and therefore the progeny segregation approach cannot be used to eliminate off-target mutations. Since it is known that TALENs have less off-target mutation than some conventional versions of CRISPRs (De Lange et al. 2014; Fichtner et al. 2014; Wang et al. 2015b), there is likely an advantage to using the TALEN technology in terms of avoiding off-target mutations in further improving elite potato cultivars such as Russet Burbank and Shepody as well as advanced clones in breeding programs.
This work shows several novel aspects in TALEN-induced genome editing in plants. This study (1) confirmed that Agrobacterium-mediated transient expression can effectively induce targeted mutation in plants without using a viral vector and (2) used the second-generation deep sequencing (Illumina) to characterize the zygosity of the targeted genome sites. The development of agrobacterium-mediated/infiltration TALEN approach can stimulate research and facilitate improvement of commercial plant cultivars and advanced breeding lines through TALEN-based genome editing without changing the general genetic makeup of valuable cultivars and advanced lines.
We thank Muhammad Haroon, Agriculture and Agri-Food Canada, for his general support in the laboratory and his technical support in plant and DNA preparation. The research funding was from Agriculture and Agri-Food Canada A-base, the Natural Sciences and Engineering Research Council of Canada (NSERC), and the deep sequencing was supported by the New Brunswick Agricultural Innovation Program.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
- Christian M, Qi Y, Zhang Y, Voytas DF (2013) Targeted mutagenesis of Arabidopsis thaliana using engineered TAL effector nucleases. G3 Genes Genomes, Genet 3:1697–1705Google Scholar
- Clasen BM, Stoddard TJ, Luo S, Demorest ZL, Li J, Cedrone F, Tibebu R, Davison S, Ray EE, Daulhac A, Coffman A, Yabandith A, Retterath A, Haun W, Baltes NJ, Mathis L, Voytas DF, Zhang F (2015) Improving cold storage and processing traits in potato through targeted gene knockout. Plant Biotechnol J 14:169–176CrossRefPubMedGoogle Scholar
- Li X-Q, Prat D, De Paepe R, Pernes J (1988b) Variability induced in Nicotiana sylvestris by two successive cycles of protoplast culture. In: IRRIaA Sinica (ed) Genetic manipulation in crops. Cassell Tycooly, London, pp 131–133Google Scholar
- Zhang F, Maeder ML, Unger-Wallaced E, Hoshaw JP, Reyon D, Christian M, Li X, Pierick CJ, Dobbs D, Peterson T, Joung JK, Voytas DF (2010) High frequency targeted mutagenesis in Arabidopsis thaliana using zinc finger nucleases. Proc Natl Acad Sci USA 107:12028–12033CrossRefPubMedPubMedCentralGoogle Scholar