Plant Molecular Biology

, Volume 83, Issue 3, pp 279–285

TAL effector nucleases induce mutations at a pre-selected location in the genome of primary barley transformants


    • Research Centre Flakkebjerg, Department of Molecular Biology and GeneticsAarhus University
  • Preben Bach Holm
    • Research Centre Flakkebjerg, Department of Molecular Biology and GeneticsAarhus University
  • Colby G. Starker
    • Department of Genetics, Cell Biology and Development and Center for Genome EngineeringUniversity of Minnesota
  • Michelle Christian
    • Department of Genetics, Cell Biology and Development and Center for Genome EngineeringUniversity of Minnesota
  • Daniel F. Voytas
    • Department of Genetics, Cell Biology and Development and Center for Genome EngineeringUniversity of Minnesota
  • Henrik Brinch-Pedersen
    • Research Centre Flakkebjerg, Department of Molecular Biology and GeneticsAarhus University
  • Inger Bæksted Holme
    • Research Centre Flakkebjerg, Department of Molecular Biology and GeneticsAarhus University
Brief Communication

DOI: 10.1007/s11103-013-0078-4

Cite this article as:
Wendt, T., Holm, P.B., Starker, C.G. et al. Plant Mol Biol (2013) 83: 279. doi:10.1007/s11103-013-0078-4


Transcription activator-like effector nucleases (TALENs) enable targeted mutagenesis in a variety of organisms. The primary advantage of TALENs over other sequence-specific nucleases, namely zinc finger nucleases and meganucleases, lies in their ease of assembly, reliability of function, and their broad targeting range. Here we report the assembly of several TALENs for a specific genomic locus in barley. The cleavage activity of individual TALENs was first tested in vivo using a yeast-based, single-strand annealing assay. The most efficient TALEN was then selected for barley transformation. Analysis of the resulting transformants showed that TALEN-induced double strand breaks led to the introduction of short deletions at the target site. Additional analysis revealed that each barley transformant contained a range of different mutations, indicating that mutations occurred independently in different cells.


TAL effector nucleasesTargeted mutagenesisHordeum vulgareCereal transformationTALEN


Rapidly advancing DNA sequencing technologies have led to the availability of sequence information that provides a significant asset for future breeding efforts. Recently, the draft assembly was released of the genome of the commercially important cereal, barley (Hordeum vulgare L.) (The International Barley Genome Sequencing Consortium 2012). However, utilizing this sequence information to further improve breeding material remains a major challenge.

Barley is considered a model plant for species within the Triticeae tribe (i.e. wheat and rye) since it is a well-characterized, self-pollinating diploid species. Further, a number of molecular and genetic techniques are available, including an Agrobacterium-mediated transformation system (Schulte et al. 2009; Tingay et al. 1997). For many of the gene sequences recently identified in barley, with corresponding sequences identified in wheat and rye, a function is still not fully assigned. Gene inactivation is a powerful tool for investigating and validating gene function. In species with high transformation efficiencies, inactivation of specific genes is often accomplished by random insertional mutagenesis using transposons or T-DNAs followed by screening of the mutant populations to identify desired mutations (Jeon et al. 2000). Antisense or RNAi techniques may also be implemented to reduce or eliminate gene expression (Matzke et al. 2001). Alternatively, mutations can be randomly induced by chemical or physical mutagens. However, for the species within the Triticeae tribe, these methods are laborious, due to the complexity and size of the genomes and comparatively modest transformation efficiencies. Furthermore, conventional mutagenesis techniques have the disadvantage in that DNA damage is imposed at random positions in the genome, and a plethora of modifications are generated such as substitutions, insertions, deletions, inversions and translocations (Sikora et al. 2011).

Recently, technologies have been developed to enable precise alteration of gene sequences at specific sites in a genome. These technologies are based on DNA sequence-specific binding proteins with nuclease activity, such as meganucleases, zinc finger nucleases (ZFNs) and transcription activator-like effector nucleases (TALENs) (Epinat et al. 2003; Puchta and Hohn 2010; Bogdanove and Voytas 2011). These nucleases can be designed to (1) selectively bind to specific locations in a genome and (2) create double strand breaks (DSBs) at the binding site. Another technology was recently developed that utilizes short RNAs to guide a nuclease (Cas9) to complementary DNA targets where site-specific DSBs are introduced (Jinek et al. 2012). This system, named Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/Cas is, however, not yet optimized for expression and use in plants. The DSBs that are induced by these technologies are repaired by the cell’s native machinery. The two cellular DNA repair mechanisms are non-homologous end joining (NHEJ) and homologous recombination (HR). The primary pathway, NHEJ, is known to result in mutations in the form of short deletions, insertions and/or substitutions.

While TALENs were only recently developed, they have already proven effective in inducing site-specific mutations at high efficiencies in various organisms, including the plant species Arabidopsis, tobacco, Brachypodium and rice (Bedell et al. 2012; Cermak et al. 2011; Dahlem et al. 2012; Li et al. 2012; Shan et al. 2013; Tong et al. 2012; Zhang et al. 2013). Even before the development of TALENs, both meganucleases and ZFNs were shown to successfully induce mutations at pre-selected genomic locations; however, the expertise required for their engineering and the cost involved has prevented their widespread use. The main advantages of TALENs are their specificity, ease of assembly and expansive targeting range. Today, TALENs can be designed for any location in any genome, whereas the targeting range of ZFNs and meganucleases is still constrained.

The potential of the TALEN technology for crop improvement efforts was recently demonstrated by Li et al. (2012). This study showed high frequency targeted mutagenesis in rice. Mutations at the target site (the promoter of a bacterial blight susceptibility gene) led to the recovery of pathogen resistant rice plants. The TALENs were introduced into rice by Agrobacterium-mediated transformation; however, since the T-DNA encoding the TALENs is only rarely linked to the target site, Li et al. (2012) were able to identify progeny with mutations that lacked the TALEN T-DNA. This allowed for the propagation of lines without transgenes.

At present, it takes eight to nine months to generate mature T0 barley plants using conventional Agrobacterium-mediated transformation of immature embryos. Implementing and validating new technologies based on transformation is therefore slow. Here we report current results obtained in primary barley transformants using the TALEN technology. We chose as a target for our TALENs a site in the promoter of a barley phytase gene of the purple acid phosphatase group named HvPAPhy_a (Holme et al. 2012). This gene was previously shown to account for most of the phytase activity in the developing barley grain (Dionisio et al. 2011). Within this promoter we targeted a region that contains a group of regulatory motifs involved in the expression of genes during grain development (Madsen et al. 2013). Ultimately, the goal of this study is to validate the importance of these motifs by investigating whether mutations in these sequences result in altered phytase expression and activity in the developing barley grain. In this study we show that it is possible to identify a high percentage of primary transformants with a variety of deletions at this pre-selected location in the barley genome, indicating the potential of TALENs as a powerful tool for barley research and breeding.

Materials and methods

TALEN assembly and yeast-based cleavage assay

For TALEN design, the genomic sequence of the HvPAPhy_a promotor (Holme et al. 2012) was analyzed using the online software tool TALE-NT1.0 (Doyle et al. 2012; Cermak et al. 2011). Three TALEN recognition sites (Online Resources 1) were chosen based on their proximity to the target region. The TALENs were assembled according to their corresponding repeat variable diresidues (RVDs) (Online Resources 1) using the GoldenGate cloning strategy (Cermak et al. 2011; Christian et al. 2010; Addgene Assembled TALEN arrays were cloned into yeast expression vectors pTAL3/pTAL4. The homologue of the genomic target sequence was cloned into the SpeI/BglII (NEB) site of pCP5 (Zhang et al. 2010). Average cleavage frequency of each TALEN was estimated using a yeast-based assay (Townsend et al. 2009; Gietz et al. 1997). Cleavage activity was normalized against a ZFN control (Zif268) and tested on specific and non-specific targets.

Vector construction for plant transformation

The selected TALEN was assembled in truncated pTAL backbones (NΔ288/CΔ231 and NΔ152/CΔ63) as described by Christian et al. (2012). The TALEN cassettes (N-terminus—TALEN Left/Right—C-terminus) were released by co-digestion with XbaI and EcoRV (NΔ288/CΔ231) or XbaI and ScaI (NΔ152/CΔ63). TALEN array Left was cloned into the compatible XbaI/FspI sites of gateway entry vector pTC22. TALEN array Right was cloned into the compatible NheI/Eco53KI sites of gateway entry vector pTC22 + TALEN array Left. Using Gateway cloning, these entry vectors were cloned into the destination vector pH2GW7 (Karimi et al. 2002). This vector contains a hygromycin resistance gene as selectable marker and a 35S promoter and T35S terminator for TALEN expression (Fig. 1). The plant transformation vector was subsequently transformed into the Agrobacterium strain AGL0 using the freeze/thaw method, and positive colonies were selected on medium with spectinomycin (50 μg/ml), hygromycin (100 μg/ml) and rifampicin (25 μg/ml).
Fig. 1

Schematic overview of TALEN plant expression vectors. Two constructs (NΔ288/CΔ231 and NΔ152/CΔ63) with the same TALEN array were prepared and cloned into the destination vector pH2GW7 (Karimi et al. 2002). Both constructs contain the plant selection gene hygromycin (hygR). TAL arrays are separated by a skipping peptide (T2A) for individual expression from the same transcript generated by the constitutive 35S promoter. Cleavage domains are two heterodimeric FokI subunits (Doyon et al. 2011). T35S, the 35S transcriptional terminator

Plant transformation

The spring cultivar Golden Promise was grown in growth cabinets at 15 °C during the day and 10 °C at night with a 16 h light period and a light intensity of 350 μ E m−2 s−1. Immature embryos isolated 12–14 days after pollination were used for Agrobacterium-mediated transformation following the procedure of Holme et al. (2012). Callus was induced on infected immature embryos on hygromycin-containing medium, and plantlets resistant to hygromycin were regenerated.

Molecular analysis

Plant material for molecular analysis was collected from the second leaf of antibiotic resistant primary transformants. Genomic DNA was isolated using The FastDNA™ Kit (MP Biomedicals). Pre-digestion of DNA with DrdI (NEB) was carried out for 2 h at 37 °C in 10 μl total volume (300–500 ng plant DNA, 1X NEB4, 1X BSA, 1U DrdI). PCR on 30–50 ng pre-digested DNA with the forward primer: 5′-taatctggccgaaccttgtta-3′; and reverse primer: 5′-agaatcaatgccttgccatc-3′ was used to amplify the TALEN target site. Phusion High-Fidelity DNA Polymerase (Thermo Scientific) was used for these PCR-reactions according to the manufacturer’s instructions. The amplified PCR products were digested with DrdI for 1 h at 37 °C. Digestion profiles were analyzed on 1 % agarose gels. Undigested bands were gel purified (GenElute™ Gel Extraction Kit, Sigma Aldrich) and cloned using the TOPO Zeroblunt® cloning kit (Invitrogen). Cloned products were sequenced and analyzed by alignment in ApE (A plasmid Editor v2.0.36).

Results and discussion

Initial assessments of TALEN cleavage activity is advantageous when working with plant species like barley, where genetic transformation remains a lengthy process and transformation efficiencies are moderate compared to model plants like tobacco or Arabidopsis. Furthermore, T-DNA cassettes encoding TALENs are very long, which will further decrease transformation efficiencies (Park et al. 2000). In this study three candidate TALENs specific for the same target region in the promoter of the PAPhy_a gene were designed and assembled. Using a rapid, yeast-based assay, we were able to show successful cleavage of the specific target DNA sequence by all three TALENs (Fig. 2). Cleavage of the non-specific target was not observed. TALEN-3 showed fourfold higher cleavage activity relative to the other two TALENs. These differences could be explained by the differences in spacer length (Online Resources 1), which has been reported to effect cleavage efficiency (Christian et al. 2012). Additionally, TALEN-3 contained the highest percentage of strong RVDs (HD and NN) (Online Resources 1), which can improve the binding of TALENs to their target (Streubel et al. 2012). Based on its increased cleavage activity in yeast, TALEN-3 was chosen for barley transformation.
Fig. 2

Yeast-based assay to determine TALEN cleavage frequency and specificity. Each TALEN was tested against the homologue of its specific genomic target site (sp. Target) and against an unspecific binding site (unsp. Target) close to its genomic location. Cleavage frequency was estimated as beta-galactosidase units resulting from the reconstitution of a disrupted lacZ gene. All values were normalized against a Zif268 ZFN control. Error bars depict Standard Deviation (n = 3)

It was recently reported that TALEN cleavage efficiency can be increased when the TAL effector backbone is truncated at the N- and C-termini (Miller et al. 2007; Zhang et al. 2013). We therefore performed experiments with the NΔ288/CΔ231 scaffold (T-DNA size of 11.8 Kb) and the NΔ152/CΔ63 scaffold (T-DNA size of 9.5 Kb) (Fig. 1). A total of 658 immature embryos were infected with the NΔ288/CΔ231 scaffold. For this treatment 89 (13.5 %) antibiotic resistant transformants were recovered. An increase in transformation efficiency was observed when the NΔ152/CΔ63 scaffold was used with the same TAL effector array (i.e. DNA binding specificity of the two TALENs was the same). Here, a total of 473 immature embryos were infected, and from these, 105 (22.2 %) developed antibiotic resistant transformants. The observed increase could be a direct effect of the reduced T-DNA size.

For the molecular analysis of the plantlets, we designed a restriction digestion assay to monitor mutagenesis at the target site. There is a DrdI restriction site in the centre of the spacer region of the TALEN where a mutation is most likely to occur (Fig. 3a). Disruption of this restriction site can be utilized as an indicator of mutation events. For diploid species like barley, it can be anticipated that the majority of TALEN-induced mutations will be heterozygous in the T0 generation, as they will only occur in one of the two homologous chromosomes. In order to identify mutation events against the background of non-mutated chromosomes, we used an enrichment step (Fig. 3b), which was also performed by Cermak et al. (2011) to identify mutations in Arabidopsis. First, genomic DNA was digested with DrdI. PCR amplification using the digested DNA as a template was used to generate a PCR product of the target site. This PCR product was again digested and analyzed by agarose gel electrophoresis (Fig. 3b). PCR products that remained undigested indicated the presence of an altered target site. Undigested PCR products were identified in 16 % (NΔ288/CΔ231 scaffold) and 31 % (NΔ152/CΔ63 scaffold) of the tested transformants. Further, the presence of the TALEN transgene in these transformants was confirmed by PCR. Undigested PCR products from the restriction assay were cloned and sequenced. Sequence analysis revealed a variety of short deletions at the target site (Fig. 4a, b). This indicated that both TALEN constructs were expressed, able to bind to the specific target and induce mutations. The increased frequency of transformants with mutations using the NΔ152/CΔ63 scaffold is in accordance with the findings of Zhang et al. (2013) in tobacco protoplasts. Overall, using the NΔ152/CΔ63 scaffold led to an increase in transformation efficiency as well as mutation frequency compared to the NΔ288/CΔ231 scaffold.
Fig. 3

Overview of an enrichment assay to detect mutations at the target site in the HvPAPhy_a promoter. a TALENs were designed so that the spacer region in between the TALEN binding sites has a restriction site (DrdI). Specific primers (arrows) amplify a 1,051 bp fragment, which can be cut into two fragments of 529 and 522 bp with DrdI. b For the enrichment assay, genomic DNA was pre-digested with DrdI (1). The digested DNA was used for a specific PCR reaction that amplifies the target fragment (2) in antibiotic resistant plants (1–6). As controls, an untransformed sample (wt) and a no template control (H2O) were included. The PCR product was then digested with DrdI (3). Fragments of the final digestion were analyzed on 1 % agarose gels for three types of patterns: (i) fragments of 529 and 522 bp = wild type only (B3, lanes 7, 8); (ii) fragments of 1,051, 529 and 522 bp = wildtype and mutated (B3, lanes 10, 11, 12); (iii) a fragment of 1,051 bp = mutated only (B3, lane 9)
Fig. 4

Sequence alignments of the target region (WT) and the corresponding region in genomic DNA of T0 plantlets. Samples represent cloned PCR products of individual T0 plantlets derived from the enrichment assay. Plants were transformed with either a NΔ288/CΔ231 scaffold or b NΔ152/CΔ63 scaffold. c Several independent clones of the same PCR product from one indivdual T0 plant (Line 12; NΔ152/CΔ63 scaffold) were individually sequenced and aligned to the target region. Underlined sequences represent the TALEN binding sites. The number of deletions for each sample is indicated on the right site

To analyze whether one particular mutation of the initially transformed cell was carried through mitotic cell divisions of one individual plant, we independently sequenced several samples from the same transformant. We found that several different mutations were present in a single plant (Fig. 4c). Therefore independent mutations must have occurred in some cells, rather than all cells being derived from a single mutation event. The occurrence of TALEN-induced mosaicism was also reported in zebrafish (Bedell et al. 2012) that were treated with TALEN mRNA. A potential advantage of a variety of mutations in T0 plantlets is that any of these mutations may be transmitted through the gametes and inherited in the next generation. Therefore, a single transgenic plant may give rise to T1 seedlings each carrying different mutations. This would greatly reduce the time and effort needed to generate mutant libraries from primary transformants.

In this study, we were able to show that large TALEN constructs can be expressed in barley. An average of one out of four antibiotic resistant primary transformants showed the presence of mutations, indicating that it is not necessary to generate large numbers of transformants for analysis. The identification of barley plantlets carrying TALEN-induced mutations was increased using the NΔ152/CΔ63 scaffold, a trend that was also reported for tobacco protoplasts (Zhang et al. 2013). Our results complement earlier findings in plants (Arabidopsis, tobacco, Brachypodium and rice) and report similar mutation patterns as seen in other organisms. Therefore these results should encourage the research community to implement this technology in species with relatively low transformation efficiencies such as barley and other Triticeae species.


This project is funded by the Danish Ministry of Food, Agriculture and Fisheries (3304-FVFP-09-B-006) and a grant from the US National Science Foundation (DBI 0923827).

Supplementary material

11103_2013_78_MOESM1_ESM.pdf (89 kb)
Supplementary material 1 (PDF 88 kb)

Copyright information

© Springer Science+Business Media Dordrecht 2013