The cisgenesis concept implies that plants are transformed only with their own genetic materials or genetic materials from closely related species capable of sexual hybridization. Furthermore, foreign sequences such as selection genes and vector-backbone sequences should be absent. We used a barley phytase gene (HvPAPhy_a) expressed during grain filling to evaluate the cisgenesis concept in barley. The marker gene elimination method was used to obtain marker-free plant lines. Here, the gene of interest and the selection gene are flanked by their own T-DNA borders to allow unlinked integration of the two genes. We analysed the transformants for co-transformation efficiency, increased phytase activities in the grain, integration of the kanamycin resistance gene of the vector-backbone and segregation between the HvPAPhy_a insert and the hygromycin resistance gene. The frequencies of the four parameters imply that it should be possible to select 11 potentially cisgenic T1-lines out of the 72 T0-lines obtained, indicating that the generation of cisgenic barley is possible at reasonable frequencies with present methods. We selected two potential cisgenic lines with a single extra copy of the HvPAPhy_a insert for further analysis. Seeds from plants homozygous for the insert showed 2.6- and 2.8-fold increases in phytase activities and the activity levels were stable over the three generations analysed. In one of the selected lines, the flanking sequences from both the left and right T-DNA borders were analysed. These sequences confirmed the absence of truncated vector-backbone sequences linked to the borders. The described line should therefore be classified as cisgenic.
Genetic modification is an important breeding tool in the development of crops with improved quality. However, in Europe in particular, the technology has been met with substantial scepticism among the general public and in consequence thereof also by the growers, the agro industry and the retailers. A number of studies have described the major concerns of the general public (Gaskell and Bauer, 2001; Bauer and Gaskell, 2002; Lassen et al., 2002). One concern about transgenic crops is the mingling of genetic materials between species that cannot hybridize by natural means. This reservation is often linked to ethical considerations and fears for potential health risks and for the spreading of new gene combinations in the environment. Also, it is apparent that the public perception of GM crops is highly dependent on whether the traits introduced are considered useful and of benefit to the public.
With the aim of meeting the concerns about mingling of genetic material between species, two new transformation concepts have been proposed, that is, intragenesis (Nielsen, 2003; Rommens, 2004) and cisgenesis (Schouten et al., 2006). In contrast to transgenesis where genes and DNA sequences can be moved between any species, both concepts imply that the plants are transformed only with their own genetic materials or genetic materials from closely related species capable of sexual hybridization. Furthermore, foreign genes such as selection marker genes and vector-backbone genes should be absent or eliminated from the primary transformants or its progeny. The gene pool exploited by intragenesis and cisgenesis is accordingly identical to the gene pool available for traditional breeding. There are two major differences between the two concepts. Intragenesis allows for the utilization of new gene combinations created by in vitro rearrangement of functional genetic elements such as promoter regions, coding regions with or without introns and terminal regions. In cisgenesis, such rearrangements are not permitted and the ‘cisgene’ is accordingly a complete copy of the endogenous gene including the promoter, introns and the terminator in the normal-sense orientation. Furthermore, when using Agrobacterium-mediated transformation, intragenesis requires T-DNA border sequences isolated from the sexual compatible gene pool (P-borders) while cisgenesis uses the T-DNA borders from Agrobacterium tumefaciens.
The implementation of the intragenesis and cisgenesis concepts may imply that crops genetically modified according to these concepts are more acceptable for the public. In February 2010, a Eurobarometer survey was conducted based on representative samples from 32 European countries on the public opinion of cisgenesis as compared to transgenesis using apple resistant to canker, scab and mildew as case studies (http://ec.europa.eu/public_opinion/archives/ebs/ebs_341_winds_en.pdf; Gaskell et al., 2011). The survey showed that across the European Union, 55% of the interviewed people supported cisgenic apples while only 33% supported transgenic apples. The interviewed people expressed the opinion that they found that cisgenic apples were more natural than transgenic apples, less problematic for the environment and more useful.
The implementation of the intragenesis and cisgenesis concepts also carries the anticipation that crops developed according to these guidelines eventually can be deregulated to a level where a minimum of risk assessment is required (Schouten et al., 2006; Rommens, 2010). At present, the very high cost associated with regulatory approval, in particular in the EU area, is a major impediment to the introduction of genetically modified crops. In consequence, many genetically modified crops and traits are currently not included in breeding programmes although they may be considered useful by the general public.
Recently, the first cisgenic plant was reported, a cisgenic apple plant resistant to scab (Vanblaere et al., 2011). In the present study, the challenges associated with implementing cisgenesis were assessed in a major cereal crop, barley. Besides, we wanted to study whether it was possible to improve a particular quality trait by the insertion of extra gene copies from the species itself. As our candidate gene for cisgenesis, we chose a barley phytase (HvPAPhy_a, Dionisio et al., 2011), preferentially synthesized during seed development and thereby responsible for most of the phytase activity deposited in the mature grain.
Phytases are the only phosphatases that can initiate the sequential liberation of phosphate groups from phytic acid. They are accordingly very important for the production of bioavailable phosphate from the stored phytic acid that accounts for around 70% of the total phosphate in the seeds (Brinch-Pedersen et al., 2002). Monogastric animals like pigs and poultry have no phytase activity in their digestive tract, and the phytase level of the mature barley grain is inadequate. In consequence, most of the phytic acid of the barley grain is secreted and spread with the manure to the agricultural soils and eventually to the aquatic environment causing algal growth and eutrophication. Large-scale supplementation with rock phosphate, a non-renewable resource, is generally undertaken to meet optimal growth requirements in the monogastric livestock (Steen, 1998). Furthermore, microbial-derived phytase is commonly added to the feed in areas with intensive livestock production to improve the poor bioavailability of phosphate (Brinch-Pedersen et al., 2002). However, many farmers grow their own cereals for feed and do not add phytase. There is accordingly a great need for improving the phytase potential and thereby the phosphate bioavailability in feed crops like barley.
The natural allelic variation for phytase activity in known barley cultivars and wild barley is limited to maximum phytase activities of around 1300 FTU/kg flour (Dai et al., 2010). Classical breeding for higher phytase activity is therefore difficult. Genetic modification has previously been used to generate plants with increased phytase activity in seeds of wheat, maize, soybean and canola using transgenic approaches with phytase transgenes isolated from microorganisms and promoters originating from other genes (Denbow et al., 1998; Brinch-Pedersen et al., 2000, 2003, 2006; Zhang et al., 2000; Chen et al., 2008). These studies have shown that transformation with a microbial phytase encoding gene is as efficient for ensuring phytic acid degradation as addition of microbial phytase to the feed.
In the present study, we show that genetic modification according to the cisgenesis concept is possible in barley at reasonable frequencies. We further show that through the insertion of a genomic clone for HvPAPhy_a, it is possible to increase the activity of phytase.
Generation of transgenic barley
The sequencing of the HvPAPhy_a genomic clone revealed that the coding sequence of 2266 bp consisted of five exons. In addition to this, we decided to include around 2000 bp of the flanking promoter region and 800 bp of the terminator region (Figure 1b). The genomic PAPhy_a was inserted into pClean-G185 with the promoter region at the left T-DNA border site and the terminator region at the right T-DNA border site (pClean-G185-PAPhy_a). This orientation was chosen because smaller deletions at the left border region might not affect the regulatory properties of the promoter region owing to its large size, while deletions of the short terminator region might be more problematic.
Immature embryos of the barley cultivar Golden Promise were transformed with the Agrobacterium strain AGL0 containing the pClean-G185-PAPhy_a and the pClean-S166 vector (Figure 1a). The vector pClean-S166 contains the hygromycin resistance gene with the NOS promoter and NOS terminator (Thole et al., 2007). We used AGL0 which is recA+ enabling homologous recombination within Agrobacterium instead of the recA AGL1strain (Lazo et al., 1991). It is, however, our experience that AGL0 confer higher transformation efficiencies in barley. Moreover, homologous sequences present in both pGreen and pSoup (Hellens et al., 2000) have been removed from the pClean vector-system, minimizing putative homologous recombination between the two binary vectors (Thole et al., 2007). A total of 1500 immature embryos were infected with Agrobacterium. Seventy-two plants survived the hygromycin selection and were transferred to the greenhouse. As expected, this transformation efficiency of 4.8% is lower than what we usually obtain with vectors containing a hygromycin resistance gene with the 35S-CMV promoter (15%), because the NOS promoter is around 30 times weaker than the 35S-CMV promoter (van der Krol et al., 1988).
PCR analysis of the T0-plants
Four primer pairs were used to analyse the T0-plants transferred to the greenhouse for the presence of the T-DNA inserts and the vector-backbone of pClean-G185. The primers amplified PCR products for the hygromycin resistance gene, the kanamycin resistance gene, the promoter-end region (PE-primer pairs) and the terminator-end region (TE-primer pairs) of the genomic PAPhy_a insert. The forward and reverse primers of the PE- and TE-primer pairs, respectively, included the cloning sites of the sequences between the T-DNA borders and the PAPhy_a insert as these sequences are the only sequences inside the T-DNA borders that are different from the endogenous gene. These primer pairs do not give any information about the number of inserts in each transformant, but they were essential for the initial screening of the material and the later identification of marker-free plants. An example of the PCR analysis in T0-plants is shown in Figure 2. As Agrobacterium can still be present in the regenerated T0-transformants and give rise to PCR false positives, the T0-plants showing all four PCR products from the pClean-S166/pClean-G185-PAPhy_a vector-system could be PCR false positive. Additional PCR analysis was therefore performed on T1-plants from T0-plants showing all four PCR products to confirm the presence of the kanamycin resistance gene of the pClean-G185-PAPhy_a vector-backbone.
Based on the PCR analysis, all 72 T0-plants obtained in this study contained the hygromycin resistance gene. The PCR analysis revealed that 73.6% of the transformants also contained PAPhy_a insert(s) (Table 1). In 38.9% of the co-transformed plants, we could detect both the PE- and the TE-primer pair products of PAPhy_a, indicating that some of the PAPhy_a inserts in these plants contained intact integration of the PAPhy_a T-DNA. In 33.3% of these plants, we could only detect the TE-primer product, indicating deletions at left T-DNA border of all the PAPhy_a inserts. In one transformant (1.4%), only the PCR product of the PE-primers was detected, indicating a deletion at the right T-DNA border of all the PAPhy_a inserts.
Table 1. PCR analysis of T0-plants for the presence of the PAPhy_a insert and the kanamycin resistance gene of the pClean-G185 vector-backbone
PCR amplification product
T0-plants transferred to the greenhouse
T0-plants setting seeds
PE, promoter-end; TE, terminator-end.
No. of T0-plants PCR-positive for the hygromycin resistance gene
Percentage of T0-plants PCR-positive for the TE- and/or PE-primer pairs of PAPhy_a
Percentage of T0-plants PCR-positive for both the TE- and PE-primer pairs of PAPhy_a
Percentage of these T0-plants also PCR-positive for the kanamycin resistance primer pairs
Percentage of T0-plants PCR-positive for only the TE-primer pairs of PAPhy_a
Percentage of these T0-plants also PCR-positive for the kanamycin resistance primer pairs
Percentage of T0-plants PCR-positive for only PE-primer pairs of PAPhy_a
Percentage of these T0-plants also PCR-positive for the kanamycin resistance primer pairs
The test for vector-backbone integration using the primer pairs for the kanamycin resistance gene was also performed on the 72 T0-plants (Table 1). Here, a vector-backbone frequency of 57.1% was found for plants showing intact integration of the PAPhy_a inserts. Verification of the presence of the kanamycin resistance gene and further analysis of the transformants required T0-plants setting seeds. Fifty-five T0-plants produced enough seeds for further analysis (Table 1). PCR analysis of T1-progenies confirmed the presence of the kanamycin resistance gene in the 12 T0-plants setting seeds and showing all four PCR products from the pClean-S166/pClean-G185-PAPhy_a vector-system. Thus, based on the 23 T0-plants setting seeds and showing intact integration of the PAPhy_a inserts, the vector-backbone integration frequency was 52.2% (Table 1). As expected, none of the plants showing deletion at the left T-DNA border region had vector-backbone integration. The transformant with deletion at the right T-DNA border region also showed vector-backbone integration (Table 1).
Phytase activity in seeds of the transformed To-plants
The phytase activity was determined in the T1-seeds of the 55 seed setting T0-plants. The T0-plants will all be hemizygous for the PAPhy_a inserts, and seed samples from each plant will be segregating for the inserts and therefore contain a mixture of seeds without any inserts and seeds hemizygous and/or homozygous for the inserts. The average PAPhy_a gene dosage in the samples should therefore be the same as in seeds hemizygous for PAPhy_a inserts.
The results were divided into three groups: (i) phytase activities in seeds of plants where the PCR product of the PE- and TE-primer pairs were not detected (ii) phytase activities in seeds of plants only showing the PCR product of the TE-primer pairs and (iii) phytase activities in seeds of plants showing the PCR product of both the PE- and TE-primers pairs (Figure 3).
The phytase activity in seeds of the non-transformed Golden Promise was 1350 FTU/kg flour. However, seeds of the 11 transformants not showing the PCR products of the PE- or TE-primer pairs had phytase activities ranging from 950 to 2280 FTU/kg flour (Figure 3a) with an average activity of 1841 FTU/kg flour and a SD of 442. The reason for this variation is not known, but T0-plants regenerated from in vitro cultures are often weaker than seed-derived plants, and T1-seeds of many of these plants showed slightly poorer grain filling than normally observed in the greenhouse. In wheat, it was shown that phytases of the endogenous TaPAPhy genes are deposited in the aleuronic layer (Dionisio et al., 2011). The phytase level per kg flour will therefore vary according to the degree of grain filling. The phytase level in this group of plants was used as a second control for the phytase level in the other two groups.
The phytase level in seeds of the 20 transformants where only the TE-primer pair product was detected ranged between 1250 and 5500 FTU/kg flour (Figure 3b). The SD of 442 FTU/kg flour observed between the T1-seeds of plants without PAPhy_a inserts accounts for some of this variation. However, the phytase activities of 2400–5500 FTU/kg flour observed in seeds of ten of these plants indicate that these transformants contain one or more PAPhy_a inserts where the deletions at the promoter region did not influence the expression of the insert. The remaining ten T0-plants in this group did not show higher expression levels in the seeds than seeds from T0-plants where the PAPhy_a insert was not detected (Figure 3a). These transformants most likely contain inserts where deletions at the promoter-end region abolish the expression of the insert(s).
The phytase activities in seeds of the 23 transformants showing the PCR products of both PAPhy_a primer pairs ranged from 1800 to 7500 FTU/kg flour (Figure 3c). Here, seeds of 21 out of 23 transformants showed phytase activities of 2400 FTU/kg flour or more. The higher phytase activity levels in this group are probably due to a higher number of intact PAPhy_a inserts integrated into these T0-plants.
The phytase activities in the seeds of the transformant only showing the PCR product of the PE-primer pair are not included in Figure 3. The seeds of this plant showed a phytase activity of 1800 FTU/kg flour, indicating that the PAPhy_a insert was not expressed because of damaging deletions at the right T-DNA border.
Identification of progeny with segregation between PAPhy and hygromycin inserts
The segregation study was performed in the progeny of five T0-plants obtained from the first transformation experiment. The five T0-plants showed PCR products of both PAPhy_a primer pairs. The plants were named PAPhy01, PAPhy02, PAPhy03, PAPhy05 and PAPhy07 and showed seed phytase activities ranging from 2500 to 3400 FTU/kg. Three of these T0-plants showed the amplification product of the kanamycin primer pairs revealing that at least one of the PAPhy_a inserts from pClean-G185 in each of these T0-plants had vector-backbone integration. In two of the T0-plants (PAPhy01, PAPhy02), no progeny containing the PAPhy_a insert without the hygromycin resistance insert was identified (Table 2). In the remaining three T0-plants (PAPhy03, PAPhy05 and PAPhy07), progeny were identified that contained PAPhy_a inserts but no hygromycin resistance inserts. PAPhy03 showed 22 progeny with PAPhy_a inserts and three progeny without PAPhy_a inserts and thus did not follow the 3 : 1 segregation for a single insert. PAPhy05 and PAPhy07 both almost exactly followed the 3 : 1 segregation for a single PAPhy_a insert, indicating a single PAPhy_a insert in both plants. PAPhy05 and PAPhy07 also lacked the kanamycin resistance gene of the pClean-G185-PAPhy_a vector-backbone as judged by the PCR analysis. An example of the PCR segregation analysis in the progeny of PAPhy07 is shown in Figure 4.
Table 2. Segregation between the PAPhy_a inserts and the hygromycin resistance inserts in progeny of five T0-plants
Identification of marker-free plants homozygous for the PAPhy_a insert and measurement of the phytase activity
The marker-free T1-plants of PAPhy05 and PAPhy07 were further propagated, and plants homozygous for the PAPhy_a insert were identified. These homozygous plants were tested once more by PCR for the absence of the hygromycin resistance gene, the kanamycin resistance gene and the tetracycline resistance gene (Figure S1). Moreover, Southern blot analysis was performed to confirm the presence of a single PAPhy_a insert in each transformant and to document that the PAPhy_a inserts of the two transformants were not inserted in tandem configurations (Figure S2).
The phytase activities were measured in seeds of the PAPhy05 and PAPhy07 plants with no PAPhy_a insert, with a hemizygous PAPhy_a insert and with a homozygous PAPhy_a insert (Figure 5). The phytase activity in the seeds of the hemizygous and homozygous plants of PAPhy05 and PAPhy07 was almost the same. The 1 : 2 : 1 mixture of the PAPhy_a segregating seeds from the hemizygous plants showed a twofold increase in phytase activity. Seeds from the homozygous plants of PAPhy05 and PAPhy07 showed a 2.6- and 2.8-fold increase in phytase activity, respectively.
Flanking regions of the PAPhy_a insert in the PAPhy07 marker-free line
The marker-free PAPhy05 and PAPhy07 plant lines homozygous for the PAPhy_a insert were further analysed for flanking sequences. Unfortunately, we were only able to isolate the right border sequences of the PAPhy_a insert in PAPhy05 (data not shown). For the PAPhy07 plant line, the nucleotides of the junctions between the T-DNA and the flanking genomic regions are shown in Figure 6a. At the right T-DNA border, the border was nicked between the third and the fourth nucleotide of the 25-nucleotide border sequence leaving three nucleotides of the right border in the plant genome. At the left T-DNA border, the border was nicked between the 16th and 17th nucleotide of the border closest to the insert leaving 16 nucleotides of the left border in the plant genome. Additionally, 14 and 22 synthetic nucleotides of the multiple cloning site in pClean-G185-PAPhy_a were integrated in the genome of PAPhy07 at the left and right T-DNA borders, respectively (Figure 1a). Thus, a total of 36 synthetic and 19 T-DNA border nucleotides out of the 36 synthetic and the 101 T-DNA border nucleotides present in the pClean-G185-PAPhy_a were integrated in the genome of PAPhy07. Importantly, the DNA flanking neither the left nor the right T-DNA border nucleotides showed homology to the DNA of the vector-backbone.
We were able to isolate 1056 bp of the right-border-flanking region and 734 bp of the left-border-flanking region (Figure 6b). The sequences were blasted against the NCBI database nucleotide collection and the EST collection. At E-values of 1e−20 or smaller, two regions included all the alignments with Hordeum vulgare in the nucleotide collection. The region from 628 to 989 bp of the 361-bp showed 87% homology to the 375-bp repetitive element named Xumet_AY661558-1. The region from 1129 to 1298 bp of 169-bp showed 76% homology to the 162-bp miniature inverted repeat transposable element named Thalos_AY661558-1. These elements were first identified in the Hv-eI4E sequence (AY661558.1) by Wicker et al. (2005). The regions also included the majority of the alignments with the EST collection accessions of H. vulgare, indicating that both elements are sometimes transcribed in barley (Figure 6b). The last 151 bp of the left-border-flanking sequences showed 100% homology to the 5′ end of an EST of 480 bp (AJ464005.1). This EST has been isolated from barley shoots 2, 3 and 4 days after germination and has no known function.
Primer pairs designed within both the left and the right flanking areas of the PAPhy_a insert were used for PCR analysis of the wheat–barley addition lines to identify the chromosome number and arm in which the PAPhy_a insert had integrated. The results show that the flanking region of both the left and the right sites of the insert are located on the short arm of chromosome 3H (Figure 6c).
The amplified and sequenced pre-insertion target sequence from the non-transgenic parent plant showed that the integration of the PAPhy_a insert had caused a deletion of 22 nucleotides at the insertion site and a deletion of one nucleotide 5 bp from the left border (Figure 6d). A five-nucleotide overlap between the left T-DNA border and the plant DNA was observed. Sequence similarities at the left border have been suggested to facilitate the T-DNA integration (Tinland, 1996).
In the present study, we have explored the possibility of generating genetically modified barley plants according to the cisgenesis concept (Schouten et al., 2006). In cisgenic plants, the ‘gene of interest’ should originate from the same species or a closely related species with which it can intercross. Moreover, marker genes and vector-backbone sequences should be absent or eliminated from the primary transformants or its progeny. As our ‘gene of interest’ we used a barley genomic clone for the HvPAPhy_a gene, a gene that encodes a phytase enzyme synthesized during grain filling and responsible for most of the phytase activity deposited in the mature grain.
The genomic clone proved to be fully functional when inserted into the barley genome. Transformed T0-plants showed up to sixfold increases in grain phytase activity as compared to the wild type. The increase in phytase activity monitored in two marker-free plant lines with a single-copy PAPhy_a insert was found to be stable over the three generations analysed. Seeds of plants homozygous for the single-copy PAPhy_a insert showed 2.6- to 2.8-fold increases in phytase activity, revealing a positive correlation between gene dosage and gene expression.
We did not observe reduction in the phytase activity as compared to the wild type in any of the plants containing extra copies of the endogenous PAPhy_a. Several studies have shown that suppression of gene expression can occur when the number of homologues to an endogenous host gene is increased (van der Krol et al., 1990; Napoli et al., 1990; Vaucheret et al., 1997; De Paoli et al., 2009). This phenomenon termed co-suppression is regulated post-transcriptionally and is considered to be induced by high transcription levels of the gene (Vaucheret et al., 1997; De Paoli et al., 2009). The co-suppression efficiencies do, however, vary greatly from gene to gene (De Paoli et al., 2009). Further studies could clarify whether PAPhy_a is prone to co-suppression. Crossings between different lines with several copies of PAPhy_a and subsequent generation of double haploids combined with quantitative gene expression analysis can allow for the generation of a linear dosage series that may establish the transcript levels required to trigger the co-suppression.
We used the marker gene elimination method to obtain marker-free plant lines. The plants were co-transformed with the genomic clone of PAPhy_a and the hygromycin resistance gene flanked by their own T-DNA borders on separate pClean vectors (pClean-G185 and pClean-S166; Thole et al., 2007). In this way, unlinked integrations are possible, allowing for the segregation of the two genes in the following generations. The co-transformation frequency of 73.6% obtained in the present study is similar to frequencies in studies where marker-free transformants have been obtained using a two-vector/one-Agrobacterium strain approach in rice (Afolabi et al., 2004) and a one-vector/one-strain approach in barley (Matthews et al., 2001). Moreover, the observed frequency of unlinked integration (60%) is similar to the one reported in rice using a one-vector/one-strain approach (Komari et al., 1996) and higher than the 22%–24% obtained using a two-vector/one-strain approach in rice (Afolabi et al., 2004) and a one-vector/one-strain approach in barley (Matthews et al., 2001). However, it should be stressed that the 60% frequency of unlinked integration in our study is based on the segregation analysis of only five lines.
The vector pClean-G185 used for PAPhy_a insertion carries two left T-DNA borders to reduce read-through of the left border resulting in the transfer and integration of vector-backbone sequences (Kuraya et al., 2004). Vector-backbone integration in barley has previously been investigated using a vector with the normal single left border sequence. In that study, 48% of the lines that possessed intact copies of the T-DNA also contained vector-backbone (Lange et al., 2006). This frequency compared to the 52.2% found in the plants showing intact integration of the PAPhy_a inserts does not suggest a reduction in read-through by the two left borders of pClean-G185.
In the present study, we were able to identify two hygromycin- and kanamycin resistance gene-free T1-lines containing a single PAPhy_a insert. Although the kanamycin resistance gene was not present in the two lines, short truncated vector-backbone sequences linked to the left border and sometimes to the right border are frequently reported (Kim et al., 2003; Wu et al., 2006; Petty et al., 2009). Isolation of flanking regions of the single PAPhy_a insert in the two marker-free lines was therefore attempted. However, we were only able to isolate 188 bp of the RB region of the PAPhy_a insert in the marker-free PAPhy05 plant line, which showed insertion in a very AT-rich area (data not shown). The left border region could not be isolated with the primers used in this study, indicating some rearrangements in the left border region. Hence, we do not know whether the plant line contains truncated sequences of vector-backbone linked to the left border. We can therefore not unequivocally classify the plant PAPhy05 as cisgenic.
Flanking regions were isolated from both the left and right T-DNA border sites of the PAPhy_a insert in the marker-free plant line PAPhy07. The flanking sequences did not show any homology to vector-backbone sequences of pClean-G185-PAPhy_a, revealing that there was no read-through into the vector-backbone at neither the left nor the right T-DNA borders in this plant line. The marker-free transformed plant line PAPhy07 can therefore be classified as a cisgenic plant line according to the definitions of Schouten et al. (2006).
It is apparent that it is very difficult with present methods to generate cisgenic plants that are without rearrangements at the T-DNA integration site. A thorough study of T-DNA insertions in Arabidopsis showed rearrangements in the target sequence in 97% of the integration events, the most common being deletions, but duplications of the target site sequences and even major chromosomal rearrangements such as translocations were also observed (Forsbach et al., 2003). The results of our study showed that the integration of the PAPhy_a insert into the PAPhy07 plant line did not induce translocations because the flanking regions of both the right and the left T-DNA borders mapped to the same chromosome, that is, the short arm of chromosome 3H. However, the integration did induce a 22-bp deletion at the target site. Still, the BLASTn searches of the target site–flanking sequences in the NCBI databases did not indicate that the PAPhy_a insert interrupted any coding sequence and the phenotype of the plant was completely normal.
It is also apparent that it is not possible with the method used in the present study to generate plants that are totally devoid of foreign DNA. In PAPhy07, a total of 19 T-DNA border nucleotides and 36 synthetic nucleotides were integrated into genome. The Agrobacterium-derived T-DNA nucleotides could have been avoided by substituting the borders with P-borders isolated from a sexually compatible gene pool as is required for intragenesis. However, the presence of at least a few extra synthetic nucleotides for cloning is difficult to leave out. Moreover, we also cannot exclude the presence of truncated sequences of T-DNA-unlinked vector-backbone or Agrobacterium chromosomal sequences in the PAPhy07 genome as have been reported in some studies (Kononov et al., 1997; Ülker et al., 2008; Petti et al., 2009). However, such T-DNA unlinked sequences could be eliminated by backcrosses to the parent or to breeding lines. Current progress in genome sequencing will enable the detection and thus facilitate the elimination of such foreign sequences.
In summary, we have documented that genetically modified barley lines can be obtained according to the cisgenesis concept. The introduced gene, a phytase, caused a 2.8-fold increase in the phytase activity of the grain, and the increase in activity was stable over three generations. This illustrates that increasing dosage by the insertion of extra copies of an endogenous gene might be a useful tool for crop improvement.
The increase in phytase activity from a single extra PAPhy_a insert amounted to around 2200 FTU/kg flour. Normally, microbial phytase is added to the feed at a concentration of 1500 FTU/kg, which makes about 60% of the phosphate bound in phytate bioavailable (Kerr et al., 2010). This implies that the current approach will be commercially interesting for farmers who do not purchase commercial phytase-supplemented feed but mix their own feed from home-grown cereals. If needed, the number of PAPhy_a inserts can easily be increased by crossing different cisgenic lines followed by double haploid production until a possible threshold for increased phytase activity is reached.
Although we only selected two potentially cisgenic lines for further analysis in the present study, the methods used to generate cisgenic plants seemed to be quite efficient. In the total material of 72 T0-lines, we obtained 19 plants that were co-transformed with both T-DNAs, expressed the PAPhy_a insert and did not show the PCR product of the kanamycin resistance gene from the vector-backbone of pClean-G185-PAPhy_a. Based on the 60% unlinked integration frequency obtained, we can predict that it should be possible to select 11 potentially cisgenic T1-lines out of the 72 T0-lines obtained.
Identification and isolation of the genomic HvPAPhy_a-gene
A lambda phage genomic barley library was used for the isolation of the genomic HvPAPhy_a clone (Stratagene no. 946104; Stratagene, Cedar Creek, TX). The DNA library was made from the barley winter cultivar Igri and contains 9- to 23-kb inserts of genomic barley DNA cloned into the Lambda FIX®II/XhoI partial fill-in vector kit (Stratagene). The whole library can be represented in 2 × 106 plaque-forming units (PFU), so 40 (12 × 12 cm) plates with 50 000 PFU per plate were used for the initial screening. The probe used was the cDNA of PAPhy_a isolated and characterized by Dionisio et al. (2011). Transfer of the PFU to filters, hybridization, screening and selection of positive plaques was done according to the procedure of Sambrook et al. (1989). The initial PFU selection was followed by two additional selections where the selected PFU were diluted to 1000 and 50 plaques per 9-cm plates, respectively. The probe was labelled with [32P] using Ready-To-Go DNA labelling beads (Amersham Biosciences, Piscataway, NJ). Two PFU-clones hybridizing with the cDNA of PAPhy_a in the third selection were amplified using the lambda DNA isolation kit (Qiagen Denmark, Copenhagen, Denmark). These clones were sequenced and both clones contained the genomic PAPhy_a gene. We sequenced 2708 bp of the promoter region, 2266 bp of the coding region and 878 bp of the terminator region (Figure 1b).
Agrobacterium transformation vectors
The vector-system pClean based on the pGreen/pSoup system was used for transformation (Hellens et al., 2000; Thole et al., 2007). Details of the pClean vector-systems are available on the pClean website (http://www.jic.ac.uk/staff/philippe-vain/pclean/pclean1.htm). The vectors pClean-G185 and pClean-S166 were chosen as vector-pairs (Figure 1a). The pClean-S166 vector contains a hygromycin resistance gene with a NOS promoter and a NOS terminator within its T-DNA and a tetracycline resistance gene in the vector-backbone (Thole et al., 2007). The pClean-G185-vector contains two left T-DNA borders, a multiple cloning site with six different cloning restriction sites within the T-DNA and a vector-backbone with a virGwt gene and a kanamycin resistance gene (Thole et al., 2007). The genomic PAPhy_a gene was amplified by PCR from the selected lambda clone using primers with flanking restriction sites corresponding to restriction sites within the T-DNA of pClean-G185. We chose the NotI and the ApaI as cloning sites. The primer at the promoter site was flanked with the NotI restriction site sequences, and the primer at the terminator site was flanked with the ApaI restriction site sequences. Both flanking restriction sites were followed by additional nucleotides to ensure the cutting of the restriction sites. The sequence of the primer pairs and the PCR conditions are shown in Table S1. The primer pairs amplified a 5208-bp product of the genomic PAPhy_a gene (Figure 1b). The PCR product was subsequently inserted into the NotI/ApaI-cloning site in pClean-G185 using standard procedures. The insertion was checked by sequencing. This vector is named pClean-G185-PAPhy_a throughout this study. The Agrobacterium strain AGL0 was co-transformed with the pClean-G185-PAPhy_a and the pClean-S166 vectors using the freeze/thaw method and selected on medium with 50 mg/L kanamycin and 7.5 mg/L tetracycline according to Thole et al. (2007).
The spring cultivar Golden Promise was grown in growth cabinets at 15 °C at day and 10 °C at night with 16-h light period at a light intensity of 350 μ E/m2/s. Immature embryos isolated 12–14 days after pollination were used for Agrobacterium transformation following the procedure of Bartlett et al. (2008) with the exception that 500 μm acetosyringone was included in the CI medium used for co-cultivation (Hensel et al., 2008), 30 mg/L hygromycin was used instead of 50 mg/L in the transition, regeneration and rooting medium, and 0.1 mg/L benzylaminopurine was added to the regeneration medium. The plants that developed roots in the rooting medium were transferred to the greenhouse.
PCR analysis of transformants
Four primer pairs were used for PCR analysis of the T0-plants: (i) hygromycin resistance gene primers to amplify a 727-bp fragment of the HptII gene (Vain et al., 2003); (ii) TE-primers (Figure 1b) to amplify a 458-bp fragment of the TE region of the PAPhy_a insert including the ApaI-cloning site in the reverse primer; (iii) PE-primers (Figure 1b) to amplify a 747-bp fragment of the promoter-end region of the PAPhy_a insert including the NotI-cloning site in the forward primer; and (iv) kanamycin resistance gene primers to amplify a 694-bp fragment of the nptII-gene. T0-plants setting seeds and showing all four PCR products of the primer pairs were further analysed in the next generation to eliminate PCR false positives caused by the presence of Agrobacterium in T0-plants. Four to ten T1-progeny from each of these T0-plants were PCR-analysed with the PAPhy_a insert and nptII-gene primers.
Segregation analysis between the hygromycin resistance genes and the PAPhy_a inserts was performed on progeny from five T0-plants. T1-progeny were PCR-analysed using the primers described above.
The two selected hygromycin- and kanamycin resistance gene-free plant lines were further propagated to obtain lines homozygous for the PAPhy_a insert. Ten to twenty progeny from each marker-free T1-plants were analysed for the PAPhy_a insert by PCR. One homozygous T1-plant from the PAPhy05 line was identified in the T2-generation. For the PAPhy07-plant line, T3-plants were analysed in the same way, leading to the identification of a homozygous T2-plant. Plant lines homozygous for the PAPhy_a insert were PCR-analysed again with the four different primers described above and a primer pair amplifying a 711-bp fragment of the tetracycline resistance gene located in the vector-backbone of pClean-S166. All primer sequences and PCR conditions used are shown in Table S1.
Analysis of gene expression in transgenic plants by phytase activity assays
The phytase activities were analysed in the crude protein extracts from milled mature seeds according to Brinch-Pedersen et al. (2000) and Engelen et al. (1994). The phytase activity in seeds of all plants was determined in duplicate. Random samples of 20–25 seeds from each transformant were milled and the flour was used for phytase activity analysis.
Genomic DNA gel blot analysis of the marker-free plant lines
Genomic DNA gel blot analysis was performed to confirm the presence of a single non-tandem inserted PAPhy_a insert and the absence of the hygromycin resistance gene in each of the marker-free lines. Genomic DNA (15 μg) from the two marker-free plants homozygous for the PAPhy_a insert and a transgenic control plant containing the hygromycin resistance gene but not the PAPhy_a insert was digested with HindIII and separated on a 1% agarose gel. HindIII cuts the PAPhy_a insert into two fragments. One fragment of 2447 bp included the whole promoter region and a small part of the coding region and one fragment of 2761 bp included the rest of the coding region and the terminator region (Figure 1b). Transfer to Hybond N+ and hybridization were performed as described by Holme et al. (2006). The probe used for the PAPhy_a insert was a 32P-labelled 1358-bp PCR product from the promoter region located within the 2447-bp HindIII fragment (Figure 1b). For primer details, see Table S1. The probe should hybridize to a 2892-bp fragment of the endogenous PAPhy_a because a HindIII site is located 2892 bp upstream of the HindIII site in the coding region (Figure 1b). For tandem insertions of the PAPhy_a insert in one locus, the probe should hybridize to fragments of around 6 kb. The hygromycin resistance gene probe was a 420-bp fragment released by the digestion of pVec8GFP (Murray et al., 2004) by BamHI and PstI labelled by Ready-To-Go DNA labelling beads (Amersham Biosciences).
Identification and analysis of the flanking areas of the PAPhy_a insert in the PAPhy07 marker-free plant line
Flanking regions of the PAPhy_a insert in the PAPhy07 marker-free plant line were amplified using the DNA Walking SpeedUp™Premix Kit (Seegene Germany, Eschborn, Germany). The kit provides a PCR master mix and primers for the unknown flanking sequences. We designed three nested primers in the known sequence for both the promoter and terminator region of the PAPhy_a gene (Table S1). The final PCR products were cloned into TOPO®TA® cloning kit (Invitrogen, Carlsbad, CA) and the cloned products were sequenced. Flanking sequences were blasted against the NCBI database using BLASTn, and homology to published barley sequences was analysed. The wheat–barley chromosome addition lines made by hybridization between Chinese spring wheat and barley cv Betzes (Islam et al., 1981; Islam and Shepherd, 2000) were used to identify the barley chromosome in which the PAPhy_a insert of PAPhy07 had integrated. Primer pairs (Table S1) were designed within the left-border- and right-border-flanking regions of the PAPhy_a insert to amplify a 557-bp PCR product of the left border region and a 395-bp PCR product of the right border region, respectively.
Primers (Table S1) designed in the flanking regions to reach over the site of the PAPhy_a integration in line PAPhy07 were used to amplify the pre-insertion site in the non-transgenic parent. The primer pairs amplified a 1526-bp fragment. The amplified PCR product was cloned into the Zero Blunt® TOPO® PCR Cloning Kit (Invitrogen) and the cloned product was sequenced.
The authors thank Ole Bråd Hansen, Heidi Petersen and Lis Holte for skilful technical assistance. We also thank Vera Thole and Philippe Vain for providing the pClean vectors. The work was funded by a grant from the Danish Ministry of Food, Agriculture and Fisheries.