Agrobacterium-mediated transformation of Asparagus officinalis L.: molecular and genetic analysis of transgenic plants

Four long-term embryogenic lines of Asparagus officinalis were co-cultured with the hypervirulent Agrobacterium tumefaciens strain AGL1Gin carrying a uidA gene and an nptII gene. 233 embryogenic lines showing kanamycin resistance and -glucuronidase (GUS) activity were obtained. Transformation frequencies ranged from 0.8 to 12.8 transformants per gram of inoculated somatic embryos, depending on the line. Southern analysis showed that usually 1 to 4 T-DNA copies were integrated. Regenerated plants generally exhibited the same insertion pattern as the corresponding transformed embryogenic line. T₁ progeny were obtained from crosses between 6 transformed plants containing 3 or 4 T-DNA copies and untransformed plants. They were analysed for GUS activity and kanamycin resistance. In three progenies, Mendelian 1:1 segregations were observed, corresponding to one functional locus in the parent transgenic plants. Southern analysis confirmed that T-DNA copies were inserted at the same locus. Non-Mendelian segregations were observed in the other three progenies. T₂ progeny also exhibited non-Mendelian segregations. Southern analysis showed that GUS-negative and kanamycin-sensitive plants did not contain any T-DNA, and therefore inactivation of transgene expression could not be responsible for the abnormal segregations.     

Effect of ploidy and homozygosity on transgene expression in primary tobacco transformants and their androgenetic progenies

Expression of a transgene is rarely analysed in the androgenetic progenies of the transgenic plants. Here, we report differential transgene expression in androgenetic haploid and doubled haploid (DH) tobacco plants as compared to the diploid parental lines, thus demonstrating a gene dosage effect. Using Agrobacterium-mediated transformation, and bacterial reporter genes encoding neomycin phosphotransferase (nptII) and β-glucuronidase (uidA/ GUS), driven respectively by the mas 1′ and mas 2′ promoters, we have generated more than 150 independent transgenic (R₀) Nicotiana tabacum plants containing one or more T-DNA copies. Transgene analyses of these R₀, their selfed R₁ lines and their corresponding haploid progenies showed an obvious position effect (site of T-DNA insertion on chromosome) on uidA expression. However, transgene (GUS) expression levels were not proportional to transgene copy number. More than 150 haploids and doubled haploids, induced by treatment with colchicine, were produced from 20 independent transgenic R₀ plants containing single and multiple copies of the uidA gene. We observed that homozygous DH plants expressed GUS at approximately 2.9-fold the level of the corresponding parental haploid plants. This increase in transgene expression may be attributed mainly to the increase (2-fold) in chromosome number. Based on this observation, we suggest a strong link between chromosome number (ploidy dosage effect) and transgene expression. In particular, we demonstrate the effect on its expression level of converting the transgene from the heterozygous (in R₀ plants) to the homozygous (DH) state: e.g. an increase of 50% was observed in the homozygous DH as compared to the original heterozygous diploid plants. We propose that ploidy coupled with homozygosity can result in a new type of gene activation, creating differences in gene expression patterns. Received: 27 April 1998 / Accepted: 12 August 1998     

In planta transformation of Arabidopsis thaliana

Transformants of Arabidopsis thaliana can be generated without using tissue culture techniques by cutting primary and secondary inflorescence shoots at their bases and inoculating the wound sites with Agrobacterium tumefaciens suspensions. After three successive inoculations, treated plants are grown to maturity, harvested and the progeny screened for transformants on a selective medium. We have investigated the reproducibility and the overall efficiency of this simple in planta transformation procedure. In addition, we determined the T-DNA copy number and inheritance in the transformants and examined whether transformed progeny recovered from the same Agrobacterium-treated plant represent one or several independent transformation events. Our results indicate that in planta transformation is very reproducible and yields stably transformed seeds in 7–8 weeks. Since it does not employ tissue culture, the in planta procedure may be particularly valuable for transformation of A. thaliana ecotypes and mutants recalcitrant to in vitro regeneration. The transformation frequency was variable and was not affected by lower growth temperature, shorter photoperiod or transformation vector. The majority of treated plants gave rise to only one transformant, but up to nine siblings were obtained from a single parental plant. Molecular analysis suggested that some of the siblings originated from a single transformed cell, while others were descended from multiple, independently transformed germ-line cells. More than 90% of the transformed progeny exhibited Mendelian segregation patterns of NPTII and GUS reporter genes. Of those, 60% contained one functional insert, 16% had two T-DNA inserts and 15% segregated for T-DNA inserts at more than two unlinked loci. The remaining transformants displayed non-Mendelian segregation ratios with a very high proportion of sensitive plants among the progeny. The small numbers of transformants recovered from individual T₁ plants and the fact that none of the T₂ progeny were homozygous for a specific T-DNA insert suggest that transformation occurs late in floral development.National Research Council of Canada Publication No. 38003     

In planta transformation of Arabidopsis thaliana

Transformants of Arabidopsis thaliana can be generated without using tissue culture techniques by cutting primary and secondary inflorescence shoots at their bases and inoculating the wound sites with Agrobacterium tumefaciens suspensions. After three successive inoculations, treated plants are grown to maturity, harvested and the progeny screened for transformants on a selective medium. We have investigated the reproducibility and the overall efficiency of this simple in planta transformation procedure. In addition, we determined the T-DNA copy number and inheritance in the transformants and examined whether transformed progeny recovered from the same Agrobacterium-treated plant represent one or several independent transformation events. Our results indicate that in planta transformation is very reproducible and yields stably transformed seeds in 7–8 weeks. Since it does not employ tissue culture, the in planta procedure may be particularly valuable for transformation of A. thaliana ecotypes and mutants recalcitrant to in vitro regeneration. The transformation frequency was variable and was not affected by lower growth temperature, shorter photoperiod or transformation vector. The majority of treated plants gave rise to only one transformant, but up to nine siblings were obtained from a single parental plant. Molecular analysis suggested that some of the siblings originated from a single transformed cell, while others were descended from multiple, independently transformed germ-line cells. More than 90% of the transformed progeny exhibited Mendelian segregation patterns of NPTII and GUS reporter genes. Of those, 60% contained one functional insert, 16% had two T-DNA inserts and 15% segregated for T-DNA inserts at more than two unlinked loci. The remaining transformants displayed non-Mendelian segregation ratios with a very high proportion of sensitive plants among the progeny. The small numbers of transformants recovered from individual T₁ plants and the fact that none of the T₂ progeny were homozygous for a specific T-DNA insert suggest that transformation occurs late in floral development.     

A broad exploration of a transgenic population of citrus: stability of gene expression and phenotype

A collection of 70 transgenic citrus plants for the uidA and nptII genes have been maintained under screenhouse conditions over a period of 4–5 years. A detailed scanning of the plants allowed us to detect four phenotypic off-type plants and a large variation of transgene integration and expression patterns among the population. Off-type plants were analysed and characterised as nucellar tetraploids, probably originating from tetraploid starting tissues rather than from somaclonal variation events. Transgene integration and expression analyses revealed that: (1) a significant negative correlation was found between copy number and GUS activity; (2) rearrangements of the T-DNA inserts did not imply low expression levels; and (3) stability of integration and expression of the transgenes was confirmed for all the transformants grown under natural environmental conditions. These combined features validate transformation as a tool for the genetic improvement of citrus. Received: 11 January 1999 / Accepted: 19 January 1999     

The transformation of pea (<Emphasis Type="Italic">Pisum sativum</Emphasis> L.): applicable methods of <Emphasis Type="Italic">Agrobacterium tumefaciens</Emphasis>-mediated gene transfer

Three methods of transformation of pea (Pisum sativum ssp. sativum L. var. medullare) were tested. The most efficient Agrobacterium tumefaciens-mediated T-DNA transfer was obtained using embryonic segments from mature pea seeds as initial explants. The transformation procedure was based on the transfer of the T-DNA region with the reporter gene uidA and selection gene bar. The expression of β-glucuronidase (GUS) in the regenerated shoots was tested using the histochemical method and the shoots were selected on a medium containing phosphinothricin (PPT). The shoots of putative transformants were rooted and transferred to non-sterile conditions. Transient expression of the uidA gene in the tissues after co-cultivation and in the course of short-term shoot cultivation (confirmed by histochemical analysis of GUS and by RT-PCR of mRNA) was achieved; however, we have not yet succeeded in proving stable incorporation of the transgene in the analysed plants.     

A tobacco matrix attachment region reduces the loss of transgene expression in the progeny of transgenic tobacco plants

The RB7 matrix attachment region (MAR), when flanking a uidA (GUS) reporter gene, has been previously shown to increase uidA gene expression by 60-fold in stably transformed tobacco suspension cell lines. We have now used the same co-transformation procedure to determine the effect of flanking MARs on uidA gene expression in tobacco plants. The neomycin phosphotransferase selection gene and uidA reporter gene on separate plasmids were co-transformed into seedlings by microprojectile bombardment. In primary transgenic plants, the average uidA expression in plants with MARs was twofold greater than in control plants without MARs, but there was no effect on variation of expression. GUS activity was not proportional to the number of integrated uidA transgenes over the entire range of copy numbers. However, in the lower part of the copy number range, MAR lines show a tendency for expression to increase with copy number. Transgene expression in backcross progenies of the MAR-containing lines averaged threefold higher than in control progenies. MARs also reduced the loss of transgene expression in the BC₁ generation. Sixty-three per cent of the 21 MAR-containing primary transformants, but only 20% of the 14 control primary transformants, produced backcross progenies in which no loss of transgene expression was observed. These observations are discussed in the context of homology-dependent gene silencing.     

Stability and inheritance of endosperm-specific expression of two transgenes in progeny from crossing independently transformed barley plants

To study stability and inheritance of two different transgenes in barley, we crossed a homozygous T₈ plant, having uidA (or gus) driven by the barley endosperm-specific B₁-hordein promoter (localized in the near centromeric region of chromosome 7H) with a second homozygous T₄ plant, having sgfp(S65T) driven by the barley endosperm-specific D-hordein promoter (localized on the subtelomeric region of chromosome 2H). Both lines stably expressed the two transgenes in the generations prior to the cross. Three independently crossed F₁ progeny were analyzed by PCR for both uidA and sgfp(S65T) in each plant and functional expression of GUS and GFP in F₂ seeds followed a 3:1 Mendelian segregation ratio and transgenes were localized by FISH to the same location as in the parental plants. FISH was used to screen F₂ plants for homozygosity of both transgenes; four homozygous plants were identified from the two crossed lines tested. FISH results showing presence of transgenes were consistent with segregation ratios of expression of both transgenes, indicating that the two transgenes were expressed without transgene silencing in homozygous progeny advanced to the F₃ and F₄ generations. Thus, even after crossing independently transformed, homozygous parental plants containing a single, stably expressed transgene, progeny were obtained that continued to express multiple transgenes through generation advance. Such stability of transgenes, following outcrossing, is an important attribute for trait modification and for gene flow studies.     

Expression of two heterologous promoters, Agrobacterium rhizogenes rolC and cauliflower mosaic virus 35S, in the stem of transgenic hybrid aspen plants during the annual cycle of growth and dormancy

We monitored, for the first time, the activity of two model heterologous promoters, the Agrobacterium rhizogenes rolC and the cauliflower mosaic virus (CaMV) 35S, throughout the annual cycle of growth and dormancy in a perennial species, hybrid aspen. Each promoter was fused to the uidA -glucuronidase (GUS) reporter gene and the constructs were introduced into the hybrid aspen genome by Agrobacterium-mediated transformation. Both wildtype and transgenic plants were cultivated under different regimes of photoperiod and temperature to induce passage through one growth-dormancy-reactivation cycle, and at intervals GUS staining was assessed in stem sections. In rolC::uidA transformants, GUS activity in rapidly growing current-year shoots was not only tissue-specific, being localized to the phloem, but also cell-specific at the shoot base, where it was present only in the companion cells. However, during the onset of dormancy induced by short photoperiod, GUS activity shifted laterally from the phloem to include the cortex and pith. After subsequent exposure to chilling temperatures to induce the transition between the dormancy stages of rest and quiescence, GUS activity almost disappeared from all stem tissues, but regained its original phloem specificity and intensity after the shoots were reactivated by exposing them to long photoperiod and high temperatures. In contrast, GUS activity in the stem of 35S::uidA transformants was strong in all tissues except for the vascular cambium and xylem, and did not vary in intensity during the growth-dormancy-reactivation cycle. The lateral shift and increased intensity of GUS activity in the stem of rolC::uidA transformants during dormancy induction was shown to be associated with the accumulation of starch, and to be mimicked by incubating stem sections in sucrose, as well as glucose and fructose, but not sorbitol, prior to the GUS assay. Our results demonstrate that the activities of the rolC and 35S promoters varied in very different, unpredictable ways during the annual cycle of growth and dormancy in a perennial species, and indicate that the spatial and temporal variation in rolC promoter activity that we observed in the stem of transgenic hybrid aspen plants is attributable to cellular and seasonal changes in sucrose content.     

Evaluation of a cotyledonary node regeneration system forAgrobacterium-mediated transformation of pea (Pisum sativum L.)

Summary A rapid regeneration system was used for studies ofAgrobacterium-mediated transformation inPisum sativum L. Cotyledonary node explants were inoculated withAgrobacterium tumefaciens strains containing binary vectors carrying genes for nopaline synthase (NOS),β-glucuronidase (GUS), and neomycin phosphotransferase (NPTII) and placed on selection medium containing either 75 or 150 mg/liter kanamycin. A GUS encoding gene (uidA) containing an intron was used to monitor gene expression from 6 to 21 days postinoculation. GUS activity could be observed 6 days after inoculation in the area of the explant in which regeneration-occurred. Regenerating tissue containing transformed cells was observed in explants on selection medium 21 days postinoculation. Using this system, a single transgenic plant was obtained. Progeny of this plant, which contained two T-DNA inserts, demonstrated segregation for the inserts and for expression of the NOS gene in the selfed R₁ progeny. NPTII activity was observed in the R₂ generation, indicating inheritance and expression of the foreign DNA over at least two generations. Attempts to repeat this procedure were unsuccessful.     

Transformation of microspore-derived embryos of winter oilseed rape (Brassica napus L.) by usingAgrobacterium tumefaciens

Haploid microspore-derived embryos (MDEs) constitute a unique material for the introduction of new traits into winter oilseed rape (Brassica napus). MDEs have been transformed by usingAgrobacterium tumefaciens strains EHA105 and LBA4404, both carrying the binary vector pKGIB containing theuidA gene encoding β-glucuronidase (GUS) and thebar gene as a marker of resistance to phosphinotricin. Transformed embryos expressed GUS and regenerated plants that were resistant to herbicide Basta, as confirmed by a leaf-painting test. Progeny plants of the transformant T-39 were all transgenic, as they inherited T-DNA from their doubled haploid parental plant. Southern-blot analysis confirmed the integration and transmission of T-DNA into T₁ plants. Transformation of MDEs facilitates the obtaining of winter oilseed rape homozygous for the introduced genes.     

The effect of duplications in the T-DNA on the stability of manifestation of heterologous genes in transgenic tobacco plants

The duplication of uidA gene within T-DNA was shown to disturb stability of expression of another marker gene, nptII, in the second generation (T₂) of selfed initial transformants and in F₁ hybrids of the crosses with nontransgenic tobacco. Hybridological analysis of the progeny resulting from various crosses involving T₁ plants demonstrated that the expression of nptII gene was impaired in the hybrids that were hemizygous for the inactivated copy of uidA gene.     

Suppression of transfer of non-T-DNA ‘vector backbone’ sequences by multiple left border repeats in vectors for transformation of higher plants mediated by Agrobacterium tumefaciens

Vectors for transformation of higher plants mediated by Agrobacterium tumefaciens were modified so that one, two or three additional copies of the left border (LB) sequences were inserted close to the original LB of the T-DNA. A gene for -glucuronidase (gusA) was placed outside the T-DNA to monitor the transfer to plants of 'vector backbone' sequences. The expression of GUS in immature embryos of rice that had been co-cultivated with A. tumefaciens carrying these constructs was around one tenth of that with A. tumefaciens carrying an unmodified control vector. Between 88 and 127 of independent transformants were regenerated from rice tissues infected with A. tumefaciens carrying each of these vectors. The GUS expressors among the rice transformed with the modified vectors were much less frequent than ones among the control transformants, and rate of reduction in the ratio of transgenic plants that expressed GUS was higher than 93%. Detection of a fragment across the LB region by the polymerase chain reaction and the gusA gene by Southern hybridization correlated well with GUS expression. These results indicate that transfer of the 'vector backbone' from the control vectors resulted mainly from inefficient termination of formation of the transfer intermediate of the T-DNA and additional LB sequences effectively suppressed such transfer. This approach is simpler than the strategy to place a 'lethal gene' outside the T-DNA and will likely help produce 'clean' transformants efficiently.     

Genetics of homology-dependent gene silencing in Arabidopsis; a role for methylation

Ninety-eight independent transformed (T₁) Arabidopsis plants were generated, containing additional copies of the chalcone synthase (CHS) gene. Three T₂ generation families (A, B and C) were found that showed reduced anthocyanin biosynthesis, consistent with homology- dependent gene silencing of CHS. Clonal sectors of tissue showing CHS silencing were seen in the early generations. Affected individuals in family A showed only slight silencing, in family C such plants were almost completely silenced, and in family B affected individuals were intermediate. Plants homozygous for a single silencing insert were isolated from each family. Plants homozygous or hemizygous for insert A showed variable penetrance and expressivity of silencing. Self-fertilization of plants hemizygous for the B and C-inserts suggested that these CHS-silencing inserts each behave as single Mendelian dominant traits. The CHS mRNA of the C-insert homozygotes was reduced to undetectable levels. Outcrosses of B- and C-insert homozygotes to wild-type plants resulted in F₁ plants that were variegated. This variegation appears to be due to expression of the CHS allele from the wild-type parent, since use of a CHS mutant, tt4, as untransformed parent resulted in uniform green F₁ plants. Southern blots revealed a correlation between DNA methylation and CHS silencing. In addition, derivative plants were generated from C-insert homozygotes that had lost the silencing inserts, and these showed a partial reversion towards wild-type phenotype and methylation of the cellular CHS gene at the TT4 locus. This result suggests that the TT4 copy of CHS became methylated during the C-insert-induced silencing and retained methylation and partial silencing after the silencing T-DNA was lost.     

Establishing a gene trap system mediated by T-DNA(GUS) in rice

Two plasmids, p13GUS and p13GUS2, were constructed to create a gene trap system containing the promoterless β-glucuronidase (GUS) reporter gene in the T-DNA region. Transformation of these two plasmids into the rice variety Zhonghua 11 (Oryza sativa ssp. japonica cv.), mediated by Agrobacterium tumefaciens, resulted in 942 independent transgenic lines. Histochemical GUS assays revealed that 31 To plants had various patterns of the reporter gene expression, including expression in only one tissue, and simultaneously in two or more tissues. Hygromycin-resistant (hygr) homozygotes were screened and the copy number of the T-DNA inserts was determined in the GUS-positivs transgenic plants. The flanking sequences of the T-DNA were isolated by inverse-polymerase chain reaction and the insert positions on the rice genome of T-DNA were determined by a basic local alignment search tool in the GUS-positive transgenic plants transformed with plasmid p13GUS. Moreover, calii induced from the seeds of the T1 generation of 911 GUS-negative transgenic lines were subjected to stress and hormone treatments. Histochemical GUS assays were carried out on the calli before and after treatment. The results revealed that calli from 21 lines displayed differential GUS expression after treatment. All of these data demonstrated that this trap system is suitable for identifying rice genes, including those that are sensitive to induction.     

Effect of an enhanced CaMV 35S promoter and a fruit-specific promoter on uida gene expression in transgenic tomato plants

Summary Two different promoters, a cauliflower mosaic virus (CaMV) 35S promoter with a 5′-untranslated leader sequence from alfalfa mosaic virus RNA4 (designated as CaMV 35S/AMV) and an E-8 fruit-ripening-specific promoter, were compared to evaluate their effects on expression of the uidA reporter gene in transgenic tomato plants. In order to generate sufficient numbers of transgenic tomato plants, both a reliable regeneration system and an efficient Agrobacterium transformation protocol were developed using 8-d-old cotyledons of tomato (Lycopersicon ecsulentum Mill. cv. Swifty Belle). Two sets of constructs, both derivatives of the binary vector pBI121, were used in transformation of tomato whereby the uidA gene was driven either by the CaMV 35S/AMV or the E-8 fruit-ripening-specific promoter. Southern blot hybridization confirmed the stable integration of the chimeric uidA gene into the tomato genome. Fruit and leaf tissues were collected from T₀ and T₁ plants, and assayed for β-glucuronidase (GUS) enzyme activity. As expected, both vegetative and fruit tissues of transgenic plants carrying the uidA gene under the control of CaMV 35S/AMV showed varying levels of GUS activity, while no expression was observed in vegetative tissues of transgenic plants carrying the uidA gene driven by the E-8 promoter. All fruits from transgenic plants produced with both sets of constructs displayed expression of the uidA gene. However, when this reporter gene was driven by the CaMV 35S/AMV, GUS activity levels were significantly higher than when it was driven by the E-8 fruit-specific promoter. The presence/absence of the uidA gene in T₁ plants segregated in a 3∶1 Mendelian ratio.     

Reduced variation in transgene expression from a binary vector with selectable markers at the right and left T-DNA borders

A new binary vector for Agrobacterium-mediated plant transformation was constructed, in which two selectable markers, for kanamycin and hygromycin resistance, were placed next to the right and left T-DNA borders, respectively, and a CaMV 35S promoter-driven β-glucuronidase (GUS) gene was placed between these markers as a reporter gene (transgene). Using double antibiotic selection, all transgenic tobacco plants carrying at least one intact copy of the T-DNA expressed the transgene, and this population exhibited reduced variability in transgene expression as compared with that obtained from the parent vector pBI121. Absence of the intact transgene was the major reason for transgenic plants with little or no transgene expression. Integration of truncated T-DNAs was also observed among transgenic plants that expressed the transgene and carried multiple T-DNA inserts. The copy number of fully integrated T-DNAs was positively associated with transgene expression levels in R₀ plants and R₁ progeny populations. Variability due to position effect was determined among 17 plants carrying a single T-DNA insert. The coefficient of variability among these plants was only 35.5%, indicating a minor role for position effects in causing transgene variability. The new binary vector reported here can therefore be used to obtain transgenic populations with reduced variability in transgene expression.