Stable high-level transgene expression in ... - Wiley Online Library

The Plant Journal (2004) 39, 440–449

doi: 10.1111/j.1365-313X.2004.02144.x

TECHNICAL ADVANCE

Stable high-level transgene expression in Arabidopsis thaliana using gene silencing mutants and matrix attachment regions Katleen M.J. Butaye1, Inge J.W.M. Goderis1, Piet F.J. Wouters1, Jonathan M.-T.G. Pues1, Stijn L. Delaure´1, Willem F. Broekaert1,†, Ann Depicker2, Bruno P.A. Cammue1,* and Miguel F.C. De Bolle1 1 Centre of Microbial and Plant Genetics, Katholieke Universiteit Leuven, Heverlee, Belgium, and 2 Department of Plant Systems Biology, Universiteit Gent, Gent, Belgium Received 26 January 2004; revised 29 April 2004; accepted 10 May 2004. * For correspondence (fax þ3216321966; e-mail [email protected]). † Present address: CropDesign N.V., Technologiepark 3, B-9052 Gent, Belgium.

Summary Basic and applied research involving transgenic plants often requires consistent high-level expression of transgenes. However, high inter-transformant variability of transgene expression caused by various phenomena, including gene silencing, is frequently observed. Here, we show that stable, high-level transgene expression is obtained using Arabidopsis thaliana post-transcriptional gene silencing (PTGS) sgs2 and sgs3 mutants. In populations of first generation (T1) A. thaliana plants transformed with a b-glucuronidase (GUS) gene (uidA) driven by the 35S cauliflower mosaic virus promoter (p35S), the incidence of highly expressing transformants shifted from 20% in wild type background to 100% in sgs2 and sgs3 backgrounds. Likewise, when sgs2 mutants were transformed with a cyclin-dependent kinase inhibitor 6 gene under control of p35S, all transformants showed a clear phenotype typified by serrated leaves, whereas such phenotype was only observed in about one of five wild type transformants. p35S-driven uidA expression remained high and steady in T2 sgs2 and sgs3 transformants, in marked contrast to the variable expression patterns observed in wild type T2 populations. We further show that T-DNA constructs flanked by matrix attachment regions of the chicken lysozyme gene (chiMARs) cause a boost in GUS activity by fivefold in sgs2 and 12-fold in sgs3 plants, reaching up to 10% of the total soluble proteins, whereas no such boost is observed in the wild type background. MAR-based plant transformation vectors used in a PTGS mutant background might be of high value for efficient high-throughput screening of transgene-based phenotypes as well as for obtaining extremely high transgene expression in plants. Keywords: transgene expression, Arabidopsis thaliana, gene silencing mutants, matrix attachment regions, high-throughput screening, post-transcriptional gene silencing.

Introduction In this post-genomic era, genetic transformation of plants has become a widely used technology that serves multiple purposes in the fields of commerce and research. The use of transgene technology allows the improvement of certain plant traits including disease resistance, stress tolerance, enhanced nutrition and male sterility (reviewed by Lanfranco, 2003). In addition, the use of transgenic plants for the production of various high-value proteins is becoming increasingly important (Giddings et al., 2000). Transgenic plants are 440

also frequently used in fundamental research as a tool to study gene function by overexpressing the target genes (Lloyd, 2003). However, all these applications are hampered by high inter-transformant variation of transgene expression often resulting in a majority of less useful transformants with low-level transgene expression. Various elements are thought to influence this interindividual variation of transgene expression. Promoters, for instance, do not only affect transgene expression levels but ª 2004 Blackwell Publishing Ltd

Boosting transgene expression in Arabidopsis PTGS mutants 441 also the magnitude of expression variability among individual transformants (De Bolle et al., 2003). The widely used 35S promoter of the cauliflower mosaic virus (p35S; Odell et al., 1985) yields a bimodal expression pattern with high expression levels in a limited number of transgenic plants but very low expression levels in the majority of the transformants (De Bolle et al., 2003; Elmayan and Vaucheret, 1996). Other regularly used promoters, such as the derivatives of the promoter of the mannopine synthase gene (pMAS), never reach the high-level expression that is conferred by p35S in some transformants but result in normally distributed expression levels in populations of transformants (De Bolle et al., 2003). Apart from the variability inherent to these regulatory elements, chromosomal position effects caused by the random integration of the transgenic DNA in the plant genome also contribute to the variability of transgene expression. Indeed, transgene integration may occur in regions with higher or lower transcriptional activity. Moreover, surrounding endogenous regulatory sequences may influence transgene expression (reviewed by Meyer, 2000). Targeting the transgenes to a specific integration site in the plant genome might rule out chromosomal position effects, but no efficient techniques for targeted transgene insertion are currently available (Day et al., 2000). In addition, matrix attachment regions (MARs) located in the proximity of transgenes can influence expression levels and variability (reviewed by Allen et al., 2000; Holmes-Davis and Comai, 1998). MARs are non-transcribed regions in eukaryotic genomes that attach to the proteinaceous matrix in the nucleus. It has been speculated that they trigger the formation of chromatin loops and thereby shield genes from chromosomal position effects (reviewed by Bode et al., 2000; Hancock, 2000). Flanking p35S-driven transgenes with the MARs of the chicken lysozyme gene A (chiMAR), for example, resulted in an eightfold reduction of transgene expression variability in tobacco (Mlyna´rova´ et al., 1994, 1995). To date, no such MAR effect has been observed in transgenic Arabidopsis thaliana (De Bolle et al., 2003). Another cause of transgene expression variability in plants is undoubtedly the impediment of transgene expression caused by epigenetic, sequence-specific gene silencing phenomena. Transgene expression can be inhibited at the transcriptional level or the post-transcriptional level, but there is growing evidence that both phenomena are linked (Aufsatz et al., 2002; Matzke et al., 2004). Transcriptional gene silencing (TGS) is associated with methylation of homologous promoter sequences of the transgene (Mette et al., 2000). Probably more impactive on transgene expression in plants is post-transcriptional gene silencing (PTGS) or RNA silencing, assumed to result from sequence homology in transcribed regions (for reviews on RNA silencing see Matzke and Matzke, 2003; Pickford and Cogoni, 2003). In transgenic plants, PTGS is thought to be particularly ª Blackwell Publishing Ltd, The Plant Journal, (2004), 39, 440–449

triggered by multiple, complexly arranged copies of the introduced DNA (Muskens et al., 2000). However, transgene silencing can also occasionally occur in transgenic plants with a single T-DNA copy (De Wilde et al., 2001; Elmayan and Vaucheret, 1996; Meza et al., 2002). Generally, RNA silencing involves double-stranded RNA (dsRNA) that is processed into short-interfering (si)RNAs of 21–25 nucleotides by an RNAse III-like enzyme (Hamilton and Baulcombe, 1999; Jacobsen et al., 1999; Zamore et al., 2000). Several homologous genes controlling RNA silencing in eukaryotic organisms have been identified through genetic screens of mutants impaired in RNA silencing. In A. thaliana, they include genes encoding a putative RNA-dependent RNA polymerase (RdRp) [SDE1/SGS2 (Dalmay et al., 2000; Mourrain et al., 2000)], a coiled-coil protein of unknown function [SGS3 (Mourrain et al., 2000)], a protein containing PAZ and Piwi domains [AGO1 (Fagard et al., 2000)], an RNA helicase [SDE3 (Dalmay et al., 2001)] and an RNase D exonuclease-like protein [WEX (Glazov et al., 2003)]. These and probably other, yet unknown gene products are believed to constitute an RNA silencing cascade (Vance and Vaucheret, 2001). Here, we describe the use of A. thaliana PTGS mutants as a genetic background for transgene expression analysis. We demonstrate that transformed PTGS mutants confer a significantly higher transgene expression level than wild type transformants and that this high level of transgene expression is stably maintained in consecutive generations. Additionally, we found that transformed PTGS mutants expressing transgenes flanked by chiMARs accumulate extremely high amounts of transgene products. To our knowledge, this is the first report describing the successful use of MARs as a genetic tool in A. thaliana to obtain stable, high-level transgene expression in the majority of the transformants, with heterologous protein levels reaching up to 10% of total soluble proteins in vegetative tissues. We believe that A. thaliana PTGS mutants, in conjunction with chiMAR-based plant transformation vectors, are of high value for efficient high-throughput screening of transgenebased phenotypes as well as for obtaining a high production of heterologous proteins in transgenic plants. Results and discussion p35S-driven reporter gene expression in gene silencing mutants In an effort to both increase transgene expression level and reduce transgene expression variability of individual transformants, the impact of different genetic backgrounds of A. thaliana ecotype Columbia 0 (Col0) was evaluated. More specifically in this study, we used A. thaliana wild type plants and the previously characterized PTGS sgs2 and sgs3 mutants (Mourrain et al., 2000) as recipients for transformation. The plant transformation vector used, named

442 Katleen M.J. Butaye et al.

Figure 1. Schematic representation of T-DNA vectors pp35S-uidA, ppCASuidA, ppOMA1-uidA and pMAR-p35S-uidA, constructed using the modular vector system as described (Goderis et al., 2002). Not to scale. uidA: b-glucuronidase coding region; pat: phosphinothricin acetyltransferase coding region; pNOS: nopaline synthase promoter; p35S: cauliflower mosaic virus 35S promoter; pCAS: cassava vein mosaic virus promoter; pOMA1: hybrid octopine and mannopine synthase promoter; tOCS: octopine synthase terminator; tNOS: nopaline synthase terminator; tg7: terminator of gene 7 of Agrobacterium tumefaciens; chiMAR: chicken lysozyme MAR; RB and LB: right and left T-DNA border, respectively.

pp35S-uidA, contains the b-glucuronidase (GUS) reporter gene (uidA; Jefferson et al., 1986) under control of the 35S promoter of the cauliflower mosaic virus (p35S) and a phosphinotricin resistance selectable marker gene (Figure 1). Transformation of A. thaliana wild type plants with pp35S-uidA resulted in about 80% low GUS expressing

pp35S-uidA GUS activity

100000

a

ppCAS-uidA

WT

d

WT

b

sgs2

e

sgs2

c

sgs3

f

sgs3

10000

primary transformants (100 units), a typical bimodal expression pattern that we have repeatedly observed for p35S-driven expression (Figure 2a; see also De Bolle et al., 2003). Maximum GUS activity was 2783 units, whereas average GUS activity was only 320 units, due to the large number of low-expressors (Figure 2a; Table 1). Using the sgs2 mutant as the recipient for transformation with pp35S-uidA, average GUS activity in primary transformants increased almost eightfold, from 320 units in wild type plants to 2280 units in sgs2 plants, which is nearly as high as the level of the highest expressor in the wild type population. In the sgs2 population, 28% of the transformants had an expression level exceeding the maximum GUS activity in the wild type background (2783 units), with the highest expressor reaching up to 13 310 units (Figure 2b). Even more important, all sgs2 transformants had a GUS activity above 180 units, whereas only 20% of the transformants in the wild type background had a GUS activity above 100 units. Therefore, the increase in average GUS activity in sgs2 at the population level is both due to an increase in activity of high-expressing individuals and a drastic reduction of the incidence of individuals with low expression. The overall variability of transgene expression in the sgs2 population is reduced compared with the

ppOMA1-uidA g

pMAR-p35S-uidA

WT

j

WT

h

sgs2

k

sgs2

i

sgs3

l

sgs3

1000 100 10 1

GUS activity

100000 10000 1000 100 10 1

GUS activity

100000 10000 1000 100 10 1

Figure 2. b-Glucuronidase (GUS) activity expressed in units (nmol 4-methylumbelliferone per min per mg total soluble proteins) in T1 A. thaliana wild type, sgs2 and sgs3 background transformed with pp35S-uidA (a–c), ppCAS-uidA (d–f), ppOMA1-uidA (g–i) and pMAR-p35S-uidA (j–l).

ª Blackwell Publishing Ltd, The Plant Journal, (2004), 39, 440–449

Boosting transgene expression in Arabidopsis PTGS mutants 443 Table 1 GUS activity in Arabidopsis thaliana wild type versus sgs2 and sgs3 mutants GUS activity Wild type Plasmid

No.

pp35S-uidA ppCAS-uidA ppOMA1-uidA pMAR-p35S-uidA

sgs2 Mean

CV

c

36 32 34 32

No. b

254 211b 64a 246b

320 217b 177a 186b

sgs3 Mean

CV a

36 32 33 34

No. a

2280 860a 166a 11237a

105 108a 87b 95a

Mean

CV

b

33 32 33 30

122a 146b 73b 108a

830 761a 310a 9994a

GUS activity is expressed in units (nmol 4-methylumbelliferone per min per mg total soluble proteins). No., number of primary transformants analyzed; CV, coefficient of variation (%). Different superscript letters for the mean indicate that they differ significantly (P < 0.05) for the genotypes (wild type, sgs2 and sgs3). Different superscript letters for the CV indicate that the population variance differs significantly (P < 0.05) for the genotypes (wild type, sgs2 and sgs3).

or branches of the PTGS cascade. As drastic reduction in the occurrence of low-level expressing pp35S-uidA transformants is observed for two different PTGS mutants, we can conclude that PTGS is the main cause of reduced transgene expression in a large fraction of wild type background plants transformed with pp35S-uidA. To evaluate the variability of the GUS activity within individual transgenic lines at different developmental stages, leaves from at least 12 randomly selected wild type, sgs2 and sgs3 transformants were collected at five different time intervals starting from week 3 after germination up to the flowering stage in week 8. For all analyzed sgs2 and sgs3 transformants, GUS activity remained high throughout the whole period and even increased at the latest time point (Figure 3). For all low-expressing wild type transformants, expression levels remained low throughout development, but for one of the three high expressors, the GUS activity dropped to a very low level at the flowering stage (Figure 3). These results indicate that high-level transgene expression is developmentally stable in PTGS mutants but not always so in wild type plants. To establish a direct link between the mutated PTGS gene and the observed increase in transgene expression, three randomly selected sgs2 plants transformed with pp35S-uidA

wild type population, as indicated by a coefficient of variation (CV) of 105 and 254% for the sgs2 and wild type population, respectively (Table 1). Moreover, the highest and lowest expressors differ by a factor 75 for the sgs2 background, versus a factor 2400 for the wild type background. Both wild type and sgs2 plants were also transformed with a p35S controlled expression cassette containing the green fluorescent protein marker gene (GFP), frequently used for non-destructive visual scoring of transgene expression in plants. In line with the data obtained for uidA, GFP expression in the wild type background was much lower and more variable between different transformants in comparison with GFP expression in the sgs2 background (Figure S1). A similar but less pronounced effect was observed for sgs3 mutants transformed with pp35S-uidA. Average GUS activity of sgs3 transformants (830 units) increased about 2.5-fold in comparison with the average GUS activity obtained in transformed wild type background (320 units) (Table 1, Figure 2c). SGS2 is an RdRp whereas SGS3 is a coiled-coil protein, which plays a yet unknown role in the RNA silencing mechanism (Mourrain et al., 2000). It is supposed that SGS2 and SGS3 are active in different stages

Col0

sgs2

sgs3

10000

GUS activity

1000 100 10 1

24

31

38

45

56

24

31

38

45

56

24

31

38

45

56

Figure 3. b-Glucuronidase (GUS) activity (as in Figure 2) at different time intervals in T1 A. thaliana 13 wild type, 12 sgs2 and 14 sgs3 plants transformed with pp35S-uidA, as measured at 24, 31, 38, 45 and 56 days after germination.


444 Katleen M.J. Butaye et al. were crossed with non-transformed wild type plants, and the GUS activity of the F1 progeny was assessed. GUS activity originally observed in the sgs2 primary transformants (8016, 3354 and 1054 units, respectively) consistently dropped in the respective F1 progeny plants to a level regularly obtained in wild type transformed plants (1, 16 and 84 units, respectively). As sgs2 is a recessive mutation (Elmayan et al., 1998), the effect of the mutation is evidently abolished in the resulting F1 progeny. Moreover, the reduction of the transgene expression level in a single transgenic line of which the PTGS mutant phenotype is crossed out clearly demonstrates that the earlier observed boost of transgene expression in transformed PTGS mutants is directly linked to the mutation in the SGS2 gene. Stable p35S-driven uidA expression in second-generation (T2) PTGS mutants For each genetic background five single-locus transformants expressing the pp35S-uidA transgene were randomly selected and at least 10 progeny plants for each parental line were further grown for GUS activity and Southern blot analyses. GUS activity in T2 populations of wild type, sgs2 and sgs3 plants is shown in Figure 4. Wild type T2 popula-

tions showed highly variable patterns of GUS activity, deviating strongly from the GUS activity level of the parental plants. However, in sgs2 and sgs3 T2 populations GUS activity generally remained stable and reached about the same GUS activity level as the parental plant. Only one of five sgs2 and sgs3 T2 populations showed a moderately deviating pattern of GUS activity and a reduction of average GUS activity in comparison with the GUS activity of the parental plant. Southern blot analysis of the T2 wild type, sgs2 and sgs3 transformants revealed various integration patterns of the T-DNA including single copies, tandem repeats, inverted repeats and mostly multiple copy inserts, regardless of the nature of the genetic background (data not shown). The occurrence of complex T-DNA integration patterns did not appear to hinder the elevation of transgene expression in PTGS-impaired mutants. From the transgene expression analyses in T2 plants, it is concluded that defective PTGS in mutant backgrounds generally ensures maintenance of high-level transgene expression in the next generation progeny of single-locus parental plants. This conclusion is of great significance for the functional analysis of genes via overexpression in transgenic plants, as gene silencing tends to increase over

GUS activity

100000 10000 1000 100 10 1

WT

GUS activity

100000 10000 1000 100 10 1

sgs2

GUS activity

100000 10000 1000 100 10 1

sgs3

Figure 4. b-Glucuronidase (GUS) activity (as in Figure 2) of 16 transgenic progeny plants derived from five single-locus T1 A. thaliana wild type parental plants versus GUS activity of 16 progeny plants derived from five single-locus T1 sgs2 and sgs3 parental plants, all transformed with pp35S-uidA. GUS activity of the parental plant is represented by a horizontal line.


Boosting transgene expression in Arabidopsis PTGS mutants 445 consecutive generations of transgenic plants (Elmayan and Vaucheret, 1996), thus complicating the interpretation of data. Other promoters driving uidA expression in PTGS mutants To assess whether or not the observed increase of transgene expression level in PTGS mutants is specific for p35S, the influence of other promoter elements on the GUS expression level in sgs2 and sgs3 transformants was evaluated. Wild type A. thaliana plants, sgs2 and sgs3 mutants were transformed with ppCAS-uidA (Figure 1) containing a uidA coding sequence regulated by the pCAS promoter, a strong constitutive promoter derived from the cassava vein mosaic virus (Verdaguer et al., 1996). In wild type T1 transformants, average pCAS-driven GUS activity was relatively high (217 units) and comparable with p35S-driven GUS activity in A. thaliana wild type background (320 units) (Figure 2d; Table 1). On T1 population level, the GUS activity pattern of wild type plants transformed with ppCAS-uidA was variable (CV 211%), similar to p35S derived expression patterns (CV 254%), although not bimodally distributed (Figure 2d; Table 1). In the transformed sgs2 plants, average pCASdriven GUS activity increased fourfold relative to the wild type transformed background, from 217 to 860 units, and none of the transformants had an expression level below 70 units (Figure 2e; Table 1). In the sgs3-transformed background, average GUS activity increased nearly fourfold compared with wild type transformed plants, from 217 to 761 units (Figure 2f; Table 1). These results corroborate the results obtained for p35S-driven transgene expression. Similar experiments were conducted with a uidA expression cassette controlled by a hybrid promoter composed of parts of the mannopine and octopine synthase gene promoters (ppOMA1-uidA; Figure 1). In wild type transformed plants, this promoter confers low expression variability (CV 64%) and a rather moderate average GUS activity level (177 units) (Figure 2g; Table 1; De Bolle et al., 2003). However, in contrast to what was observed for p35S- and pCASdriven expression cassettes, no significant increase of expression level or significant reduction of expression variability was found in the sgs2 and sgs3 plants expressing ppOMA1-uidA (Figure 2g–i; Table 1). Taken together, these data indicate that the use of the strong promoters p35S and pCAS for transformation of wild type A. thaliana can lead to high transgene expression but the expression they confer is also more prone to gene silencing than the expression obtained by the weaker pOMA1 promoter (see also Que et al., 1997). We hypothesize that the latter promoter is not strong enough to trigger PTGS, either by read-through transcription and/or by reaching expression levels above a critical threshold. Strong promoters such as p35S are widely used when high transgene expression is required, with the drawback that ª Blackwell Publishing Ltd, The Plant Journal, (2004), 39, 440–449

large numbers of transformants need to be analyzed to select a few high expressors. As an alternative and when only moderately high transgene expression is needed, one could opt for promoters that result in lower inter-transformant variability such as pOMA1 (De Bolle et al., 2003). Here, it is shown that the combination of low inter-transformant expression variability and high-level expression of transgenes can be achieved using strong promoters (p35S, pCAS) in a PTGS-impaired background. This can be of significant importance for applications like high-throughput phenotyping of transgene-overexpressing plants in functional genomics, as is exemplified in the next paragraph. PTGS mutants as a tool for functional genomics To explore the use of PTGS mutants as a tool for facilitated functional analysis of genes, a p35S-driven cyclindependent kinase inhibitor (KRP6) gene was used for the transformation of A. thaliana wild type and sgs2 plants. Overexpression of KRP6 has been reported to dramatically inhibit cell cycle progression in leaf primordial cells resulting in serrated leaves (Figure 5a,b; De Veylder et al., 2001). In the wild type transformed background, the serrated phenotype was observed in about 20% of the primary transformants compared with 100% in the sgs2 transformed background (Figure 5c,d). These data demonstrate that gene silencing mutants as a genetic background for transformation can be used for highthroughput analysis of phenotypes caused by transgene overexpression. A major limitation of current functional genomics programs based on transgene overexpression is that large populations of primary transformants have to be made and screened in order to detect transgene-associated phenotypic alterations. The use of PTGS mutants would allow a reduction of the populations of primary transformants down to a few or theoretically even one individual, thus substantially reducing costs and increasing throughput of functional analysis. The effect of T-DNA bordered by matrix attachment regions in PTGS mutants Although p35S-driven uidA expression is consistently high in transformed PTGS-impaired mutants, it is still variable. This variability could possibly be caused by gene dosage effects, position effects or other silencing-inducing phenomena. In an attempt to reduce the variability of transgene expression caused by position effects, we studied the effect of chicken lysozyme MARs (chiMARs) flanking the T-DNA using the plant transformation vector pMAR-p35SuidA (Figure 1). These chiMARs have been shown to reduce transgene expression variability in tobacco (Mlyna´rova´ et al., 1994, 1995), whereas, so far, no such effect could be demonstrated in wild type transformed

446 Katleen M.J. Butaye et al.

a

b

c

d

Figure 5. Phenotype of A. thaliana wild type plants and sgs2 mutants overexpressing the cyclin-dependent kinase inhibitor gene (KRP6). (a) Non-transgenic plant. (b) Transgenic plant overexpressing KRP6. (c) Randomly chosen transformed wild type plants overexpressing KRP6. Arrows indicate plants showing the KRP6 overexpression phenotype. (d) Randomly chosen transformed sgs2 mutants overexpressing KRP6.

A. thaliana (De Bolle et al., 2003). In line with previous findings, the presence of chiMARs did not significantly affect either the level or the variability of transgene expression in the wild type background (Figure 2j; Table 1). In the sgs2 background, transgene expression variability was also not influenced by chiMARs (CV 105% versus 95%), but surprisingly, chiMARs caused an additional fivefold boost in the average GUS expression level, from 2280 to 11 237 units (Figure 2k; Table 1). All sgs2 plants transformed with pMAR-p35S-uidA had an expression level higher than 517 units. In a similar experiment with the sgs3 mutant, a 12-fold increase of the average GUS activity attributable to chiMARs was observed (Figure 2l; Table 1). Overall, the average GUS activity in sgs2 and sgs3 plants

Figure 6. SDS-PAGE analysis of total protein extracts (2 lg per lane) from sgs2 plants transformed with pMAR-p35S-uidA (lanes 1 and 2); total protein extracts (2 lg per lane) from a non-transgenic plant (lane 3); 500 ng bovine serum albumin (lane 4); partially purified b-glucuronidase (lane 5). The position of GUS is indicated by the arrow to the right. The position of molecular weight reference proteins is indicated by arrows to the left.

transformed with pMAR-p35S-uidA was 30–40-fold higher relative to expression measured in wild type plants transformed with pp35S-uidA (Table 1). These data are the first to point to a significant role of MARs in the optimization of transgene expression in A. thaliana, when used in conjunction with a PTGS-impaired background. We have analyzed the T-DNA integration pattern of pMARp35S-uidA in wild type and sgs2 single-locus transformants by Southern blot analysis. As for pp35S-uidA, most of the plants transformed with pMAR-p35S-uidA had complex integration patterns, irrespective of whether they had a wild type or sgs2 background (data not shown). Some of the sgs2 transformants containing chiMARflanked transgenes reached extremely high GUS activity levels, up to 41 000 units. Coomassie blue staining of an SDS-PAGE gel revealed a clear band in the total leaf extracts of extremely high GUS accumulating sgs2 plants, which was absent in total leaf extracts of non-transformed sgs2 control plants (Figure 6). This band was situated at the same relative position in the gel as the GUS standard. By densitometric comparison of the intensities of this band with known amounts of bovine serum albumin (BSA), we estimate that GUS accumulated to roughly 10% of the total soluble proteins in these transformed sgs2 plants. To date, no MAR effect has been reported in genetically transformed A. thaliana wild type plants (De Bolle et al., 2003; this study). As such, our data do not support the hypothesis that MARs prevent silencing (Allen et al., 2000; Mlyna´rova´ et al., 2003). However, in a transformed PTGSimpaired background, chiMARs boosted the average transgene expression levels five- to 12-fold. Apparently, chiMARs do no act as suppressors of expression variability in A. thaliana but rather as enhancers of expression, yet the enhancer effect is only observed when the PTGS mechanism is suppressed. Our observation that expression levels are boosted by chiMARs in PTGS mutants but not in wild type plants provide a strong argument in favor of the hypothesis that PTGS mechanisms are only triggered when transgene expression surpasses a certain threshold (Depicker and ª Blackwell Publishing Ltd, The Plant Journal, (2004), 39, 440–449

Boosting transgene expression in Arabidopsis PTGS mutants 447 Van Montagu, 1997; Lechtenberg et al., 2003). Thresholdtriggered PTGS would be responsible for the establishment of a ceiling for maximal expression in wild type plants, which cannot be surpassed even in the presence of expression enhancers such as MARs. In PTGS mutants, however, transgene expression levels are not ceiled, allowing MARs to exhibit their enhancer effect. In conclusion, this study demonstrates that the use of PTGS-impaired A. thaliana mutants holds great promise for obtaining extremely high transgene expression and for performing high-throughput functional analysis of genes. Although these mutants have been reported to be more susceptible to some viral diseases than wild type plants (Mourrain et al., 2000), we have not encountered any problems related to viral or other diseases when growing the plants either in a growth chamber or in a greenhouse. The sgs2 and sgs3 mutants behaved essentially as wild type plants in terms of phenotypic appearance and transformation efficiency. It remains to be demonstrated, however, whether phenotypic defects would become apparent in less contained conditions. The proposed approach is currently limited to the use of A. thaliana. Further investigations are needed to demonstrate whether the results observed here can be translated to other plant species for which PTGS mutants might be identified. Instead of using PTGS-impaired mutants, transgenic plants overexpressing viral gene silencing suppressors including HC-Pro could be implemented. These transgenic plants, however, often suffer from phenotypic abnormalities, including stunting and altered leaf shape (Mallory et al., 2002). High expression of transgenes in A. thaliana sgs2 and sgs3 mutants was not completely unexpected (Elmayan et al., 1998; Mourrain et al., 2000). However, this study shows that this effect is promoter dependent and only conferred by strong promoters. Moreover, we have demonstrated that high-level transgene expression in PTGSimpaired mutants is developmentally stable and maintained in the subsequent generation. Hence, the use of PTGS mutants reduces the requirement to work with large populations of transgenic plants, and avoids flaws in data interpretation caused by erratic transgene expression. In combination with MARs, the expression of transgenes can even be increased further if so desired. Indeed, expression levels in leaves of up to 10% of the total soluble proteins were achieved for a GUS transgene, without any optimization of the coding sequence. Different valuable approaches to use plants as high-expressor systems for heterologous proteins are available to date, including chloroplast transformation (up to 46% of TSP; De Cosa et al., 2001) or the use of Phaseolus vulgaris regulatory sequences in transgenic seeds (up to 36% of TSP; De Jaeger et al., 2002). These systems can be of great value for molecular pharming or other high-yield applicaª Blackwell Publishing Ltd, The Plant Journal, (2004), 39, 440–449

tions. They are, however, limited for other specific purposes, including stable transformation in subsequent progeny and non-seed-specific high transgene expression, respectively. In this study, we show that individuals with very low expression of a transgene and correspondingly weak phenotypic effect occur only very rarely upon transformation of PTGS-impaired mutants. In fact, the current approach virtually eliminates null expressors but does not rule out inter-individual variability entirely. Ultimate alternatives to manage inter-individual transgene variability would be gene targeting (Reiss, 2003) or the use of plant artificial chromosomes (Hall et al., 2004). However, it is clear that the final goal of the routine use of these techniques for basic studies of plant gene functions and for plant biotechnology is still somewhere ahead (Hanin and Paszkowski, 2003).

Experimental procedures Plasmids All plasmids were constructed using a modular vector system (Goderis et al., 2002). Plasmids pp35S-uidA, pMAR-p35S-uidA and ppOMA1-uidA are as previously described (De Bolle et al., 2003). The p35S-uidA expression cassette from pp35S-uidA was replaced by the pCAS-uidA expression unit from pMODUL3198 (Goderis et al., 2002) yielding ppCAS-uidA. The KRP6 gene was derived from a plasmid donated by CropDesign, Belgium, and was inserted in pp35S-uidA replacing the uidA gene.

Plant transformation and propagation The plant transformation vectors were introduced in Agrobacterium tumefaciens GV3101 (pMP90) by electroporation. Arabidopsis thaliana (Columbia 0) wild type, sgs2 and sgs3 plants were grown as described (De Bolle et al., 2003). Plants were transformed using the A. tumefaciens-mediated floral dip transformation method (Clough and Bent, 1998). Primary transformants were selected based on resistance against phosphinotricin (5 mg l)1, Basta; Bayer, Leverkusen, Germany) (Goderis et al., 2002). Transformation efficiencies, expressed as the percentage of phosphinotricin-resistant seedlings to germinated seedlings, were typically between 2 and 3.7%. For each genetic background, A. thaliana populations consisting of at least 30 primary transformants were obtained and further analyzed. After self-pollination, T2 seedlings were grown and sprayed twice with Basta (50 mg l)1) at 10 and 14 days after sowing. Single-locus transformants were selected based on the expected 3:1 segregation of the phosphinotricin selectable marker gene.

Protein assays The activity of the GUS enzyme was measured in a mixture of the fourth, sixth and eighth leaf of 4-week-old transformants, unless stated otherwise. Prior to these experiments, it had been established that GUS activity levels in these individual leaves from one individual plant do not differ in a statistically significant way. GUS activity was measured through a 4-methylumbelliferyl-b-Dglucuronide substrate assay (Jefferson, 1987) and was expressed in units defined as nmol 4-methylumbelliferon per minute and per mg

448 Katleen M.J. Butaye et al. total soluble proteins. Total soluble protein content in leaf extracts was measured as described (De Bolle et al., 2003). Total leaf extracts and GUS standard (Sigma-Aldrich, St Louis, MO, USA) were separated on a 12.5% SDS-PAGE gel and visualized by staining with Coomassie brilliant blue R250.

Statistical evaluation The 3:1 segregation of the phosphinotricin selectable marker gene was statistically evaluated by chi-square analysis. To compare GUS activity levels of different plant populations per genotype, multiple t-tests were performed with a Bonferroni adjustment for multiple comparisons. To compare the variability of GUS activity of different plant populations per genotype, multiple F-tests were performed with a Bonferroni correction. Prior to these analyses, the datasets were normalized by applying the natural logarithm. The CV is a statistical measure of dispersion that measures the deviation of a variable from its mean and was used to reflect the overall variability of each plant population. The CV is given by the standard deviation divided by the mean and expressed in percentage. All tests were performed at significance level 0.05 and with Microsoft Excel Analysis Toolpak.

Acknowledgement The authors thank Dr Herve´ Vaucheret (INRA Versailles) for providing seeds of the sgs-mutants, Dr Vale´rie Frankard (CropDesign, Belgium) for the KRP6-plasmid and Dr Karin Thevissen for critical reading of the manuscript. This work was supported partly by a grant from the Flanders Interuniversity Institute for Biotechnology (PRJ3) and partly by a grant from the Fonds voor Wetenschappelijk Instituut Vlaanderen (G.0118.01). K.M.J. Butaye is indebted to the Instituut voor de Aanmoediging van Innovatie door Wetenschap en Technologie in Vlaanderen for a predoctoral fellowship.

Supplementary material The following material is available from http://www. blackwellpublishing.com/products/journals/suppmat/TPJ/TPJ2144/ TPJ2144sm.htm Figure S1. Expression of the green fluorescent protein gene (GFP) in Arabidopsis thaliana wild type plants versus sgs2 mutants. C, non-transgenic plants; WT, transformed wild type plants expressing GFP; sgs2, transformed sgs2 mutants expressing GFP. Plants were randomly chosen.

References Allen, G.C., Spiker, S. and Thompson, W.F. (2000) Use of matrix attachment regions (MARs) to minimize transgene silencing. Plant Mol. Biol. 43, 361–376. Aufsatz, W., Mette, M.F., van der Winden, J., Matzke, A.J.M. and Matzke, M. (2002) RNA-directed DNA methylation in Arabidopsis. Proc. Natl Acad. Sci. USA, 99, 16499–16506. Bode, J., Benham, C., Knopp, A. and Mielke, C. (2000) Transcriptional augmentation: modulation of gene expression by scaffold/ matrix-attached regions (S/MAR elements). Crit. Rev. Eukaryot. Gene Expr. 10, 73–90. Clough, S.J. and Bent, A.F. (1998) Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J. 16, 735–743.

Dalmay, T., Hamilton, A., Rudd, S., Angell, S. and Baulcombe, D.C. (2000) An RNA-dependent RNA polymerase gene in Arabidopsis is required for posttranscriptional gene silencing mediated by a transgene but not by a virus. Cell, 101, 543–553. Dalmay, T., Horsefield, R., Braunstein, T.H. and Baulcombe, D.C. (2001) SDE3 encodes an RNA helicase required for post-transcriptional gene silencing in Arabidopsis. EMBO J. 20, 2069–2078. Day, C.D., Lee, E., Kobayashi, J., Holappa, L.D., Albert, H. and Ow, D.W. (2000) Transgene integration into the same chromosome location can produce alleles that express at a predictable level, or alleles that are differentially silenced. Genes Dev. 14, 2869–2880. De Bolle, M.F.C., Butaye, K.M.J., Coucke, W.J.W., Goderis, I.J.W.M., Wouters, P.F.J., van Boxel, N., Broekaert, W.F. and Cammue, B.P.A. (2003) Analysis of the influence of promoter elements and a matrix attachment region on the inter-individual variation of transgene expression in populations of Arabidopsis thaliana. Plant Sci. 165, 169–179. De Cosa, B., Moar, W., Lee, S.B., Miller, M. and Daniell, H. (2001) Overexpression of the Bt cry2Aa2 operon in chloroplasts leads to formation of insecticidal crystals. Nat. Biotechnol. 19, 71–74. De Jaeger, G., Scheffer, S., Jacobs, A., Zambre, M., Zobell, O., Goossens, A., Depicker, A. and Angenon, G. (2002) Boosting heterologous protein production in transgenic dicotyledonous seeds using Phaseolus vulgaris regulatory sequences. Nat. Biotechnol. 20, 1265–1268. De Veylder, L., Beeckman, T., Beemster, G.T., Krols, L., Terras, F., Landrieu, I., van der Schueren, E., Maes, S., Naudts, M. and Inze´, D. (2001) Functional analysis of cyclin-dependent kinase inhibitors of Arabidopsis. Plant Cell, 13, 1653–1667. Depicker, A. and Van Montagu, M. (1997) Post-transcriptional gene silencing in plants. Curr. Opin. Cell Biol. 9, 373–382. De Wilde, C., Podevin, N., Windels, P. and Depicker, A. (2001) Silencing of antibody genes in plants with single-copy transgene inserts as a result of gene dosage effects. Mol. Genet. Genomics, 265, 647–653. Elmayan, T. and Vaucheret, H. (1996) Expression of single copies of a strongly expressed 35S transgene can be silenced post-transcriptionally. Plant J. 9, 787–797. Elmayan, T., Balzergue, S., Beon, F. et al. (1998) Arabidopsis mutants impaired in cosuppression. Plant Cell, 10, 1747–1758. Fagard, M., Boutet, S., Morel, J.B., Bellini, C. and Vaucheret, H. (2000) AGO1, QDE-2, and RDE-1 are related proteins required for post-transcriptional gene silencing in plants, quelling in fungi, and RNA interference in animals. Proc. Natl Acad. Sci. USA, 97, 11650–11654. Giddings, G., Allison, G., Brooks, D. and Carter, A. (2000) Transgenic plants as factories for biopharmaceuticals. Nat. Biotechnol. 18, 1151–1155. Glazov, E., Phillips, K., Budziszewski, G.J., Meins, F. and Levin, J.Z. (2003) A gene encoding an RNase D exonuclease-like protein is required for post-transcriptional silencing in Arabidopsis. Plant J. 35, 342–349. Goderis, I.J.W.M., De Bolle, M.F.C., Franc¸ois, I.E.J.A., Wouters, P.F.J., Broekaert, W.F. and Cammue, B.P.A. (2002) A set of modular plant transformation vectors allowing flexible insertion of up to six expression units. Plant Mol. Biol. 50, 17–27. Hall, A.E., Keith, K.C., Hall, S.E., Copenhaver, G.P. and Preuss, D. (2004) The rapidly evolving field of plant centromeres. Curr. Opin. Plant Biol. 2, 108–114. Hamilton, A.J. and Baulcombe, D.C. (1999) A species of small antisense RNA in posttranscriptional gene silencing in plants. Science, 286, 950–952.


Boosting transgene expression in Arabidopsis PTGS mutants 449 Hancock, R. (2000) A new look at the nuclear matrix. Chromosoma, 109, 219–225. Hanin, M. and Paszkowski, J. (2003) Plant genome modification by homologous recombination. Curr. Opin. Plant Biol. 6, 157–162. Holmes-Davis, R. and Comai, L. (1998) Nuclear matrix attachment regions and plant gene expression. Trends Plant Sci. 3, 91–97. Jacobsen, S.E., Running, M.P. and Meyerowitz, E.M. (1999) Disruption of an RNA helicase/RNAse III gene in Arabidopsis causes unregulated cell division in floral meristems. Development, 126, 5231–5243. Jefferson, R.A. (1987) Assaying chimeric genes in plants: the GUS gene fusion system. Plant Mol. Biol. Rep. 5, 387–405. Jefferson, R.A., Burgess, S.M. and Hirsh, D. (1986) b-Glucuronidase from Escherichia coli as a gene-fusion marker. Proc. Natl Acad. Sci. USA, 83, 8447–8451. Lanfranco, L. (2003) Engineering crops, a deserving venture. Riv. Biol. 96, 31–54. Lechtenberg, B., Schubert, D., Forsbach, A., Gils, M. and Schmidt, R. (2003) Neither inverted repeat T-DNA configurations nor arrangements of tandemly repeated transgenes are sufficient to trigger transgene silencing. Plant J. 34, 507–517. Lloyd, A. (2003) Vector construction for gene overexpression as a tool to elucidate gene function. Methods Mol. Biol. 236, 329– 344. Mallory, A.C., Reinhart, B.J., Bartel, D., Vance, V.B. and Bowman, L.H. (2002) A viral suppressor of RNA silencing differentially regulates the accumulation of short interfering RNAs and microRNAs in tobacco. Proc. Natl Acad. Sci. USA, 99, 15228–15333. Matzke, M.A. and Matzke, A.J. (2003) RNAi extends its reach. Science, 301, 1060–1061. Matzke, M.A., Aufsatz, W., Kanno, T., Daxinger, L., Papp, I., Mette, M.F. and Matzke, A.J. (2004) Genetic analysis of RNAmediated transcriptional gene silencing. Biochim. Biophys. Acta, 1677, 129–141. Mette, M.F., Aufsatz, W., van der Winden, J., Matzke, M.A. and Matzke, A.J. (2000) Transcriptional silencing and promoter methylation triggered by double-stranded RNA. EMBO J. 19, 5194–5201. Meyer, P. (2000) Transcriptional transgene silencing and chromatin components. Plant Mol. Biol. 43, 221–234. Meza, T.J., Stangeland, B., Mercy, I.S. et al. (2002) Analyses of single-copy Arabidopsis T-DNA-transformed lines show that the


presence of vector backbone sequences, short inverted repeats and DNA methylation is not sufficient or necessary for the induction of transgene silencing. Nucl. Acids Res. 30, 4556– 4566. Mlyna´rova´, L., Loonen, A., Heldens, J., Jansen, R.C., Keizer, P., Stiekema, W.J. and Nap, J.P. (1994) Reduced position effect in mature transgenic plants conferred by the chicken lysozyme matrix-associated region. Plant Cell, 6, 417–426. Mlyna´rova´, L., Jansen, R.C., Conner, A.J., Stiekema, W.J. and Nap, J.P. (1995) The MAR-mediated reduction in position effect can be uncoupled from copy number-dependent expression in transgenic plants. Plant Cell, 7, 599–609. Mlyna´rova´, L., Hricova, A., Loonen, A. and Nap, J.P. (2003) The presence of a chromatin boundary appears to shield a transgene in tobacco from RNA silencing. Plant Cell, 15, 2203–2217. Mourrain, P., Beclin, C., Elmayan, T. et al. (2000) Arabidopsis SGS2 and SGS3 genes are required for posttranscriptional gene silencing and natural virus resistance. Cell, 101, 533–542. Muskens, M.W., Vissers, A.P., Mol, J.N. and Kooter, J.M. (2000) Role of inverted DNA repeats in transcriptional and post-transcriptional gene silencing. Plant Mol. Biol. 43, 243–260. Odell, J.T., Nagy, F. and Chua, N.H. (1985) Identification of DNA sequences required for activity of the cauliflower mosaic virus 35S promoter. Nature, 313, 810–812. Pickford, A.S. and Cogoni, C. (2003) RNA-mediated gene silencing. Cell Mol. Life Sci. 60, 871–882. Que, Q., Wang, H.Y., English, J.J. and Jorgensen, R.A. (1997) The frequency and degree of cosuppression by sense chalcone synthase transgenes are dependent on transgene promoter strength and are reduced by premature nonsense codons in the transgene coding sequence. Plant Cell, 9, 1357–1368. Reiss, B. (2003) Homologous recombination and gene targeting in plant cells. Int. Rev. Cytol. 228, 85–139. Vance, V. and Vaucheret, H. (2001) RNA silencing in plants – defense and counterdefense. Science, 292, 2277–2280. Verdaguer, B., de Kochko, A., Beachy, R.N. and Fauquet, C. (1996) Isolation and expression in transgenic tobacco and rice plants of the cassava vein mosaic virus (CVMV) promoter. Plant Mol. Biol. 31, 1129–1139. Zamore, P.D., Tuschl, T., Sharp, P.A. and Bartel, D.P. (2000) RNAi: double-stranded RNA directs the ATP-dependent cleavage of mRNA at 21 to 23 nucleotide intervals. Cell, 101, 25–33.