Transcriptional interference in transgenic plants

Gene. 109 (1991) 239-242 © 1991 Elsevier Science Publishers B.V. All rights reserved. 0378-1119/91/$03.50

239

GENE 06212

Transcriptional interference in transgenic plants (Agrobacterium vector system; callus; position effect; poly(A) signal; recombinant DNA; terminator)

!. lngelbrecht, P. Breyne, K. Vancompernolle, A. Jacobs, M. Van Montagu and A. Depicker Laboratorium veer Genetica, Universiteit Gent. B-9000 Gent (Belgium)

Received by G.N. Gussin: 10 September 1990 Revised/Accepted:23 August/9 September1991 Receivedat publishers: 30 September 1991

SUMMARY When a promotedess marker gene is transformed into the plant g-aome using the Agrobacterium vector system, on average 30~o of the T-DNA inserts produce gene fusions. This suggests that the T-DNA is preferentially integrated into transcribed regions. Here, we propose that this transcriptional activity is responsible for some of the variation in expression frequently observed among independent transformants. Using hybrid gene constructions, we show that transcriptional readthrough into a downstream gene with opposite orientation substantially reduces expression of this gene both in transient expression and in transgenic plants. Furthermore, a poly(A) signal/terminator can block readthrough ~nd restore the expression of the gene. Finally, enzymatic analysis ofcalli suggests that less variation in neomycin phosphotransferase II synthesis is observed when the gene is separated from plant DNA by promoter and terminator elements.

INTRODUCTION A major problem in the quantitative analysis of cis- and trans-acting sequences in transgenic plants is the variation in expression frequently observed among independent transformants (Odell et al., 1987; Dean et al., 1988). Factors thought to be responsible for this clonal variation include: (i) chromatin structure at the insertion site; (ii) the

Correspondence to: Prof. M. Van Montagu, Laboratoriumvoor Genetica, Universiteit Gent, K.L. Ledeganckstraat 35, B-9000Gent (Belgium) Tel. (32-91)645170; Fax (32-91)645349; BITNET [email protected];Telex 11995gengen b.

Abbreviations: bp, base pair(s); hpt, gene encoding Hy phosphotransferase (see Fig. !); Hy, hygromycin;kb, kilobase(s) or 1000bp; LB, lett border; nos, gene encoding nopaline synthase; NPTII, neomycin phosphotransferase II; nptll, gene encoding NPTII; nt, nucleotide(s); ocs, gene encoding octopine synthase;P, promoter; P35s, cauliflowermosaic virus 35S promoter; PK, plant kinase; R, resistance/resistant; RB, right border; T-DNA, transferred DNA.

copy number of the inserted gene; (ill) de nero methylation of the integrated gene (for a review, see Selker, 1990); and (iv) dominant effects exerted by neighboring plant sequences. In this report, we investigated whether transcriptional activity at the site of integration may also be responsible for some of the observed position effects. Andr~ et al. (1986) demonstrated that a promoterless nptll reporter gene, linked to the right T-DNA border can become fused to transcribed plant DNA sequences upon integration, allowing the isolation of kanamycin-resistant clones. Using similar constructions, Koncz et al. (1989) and Herman et ai. (1990) determined that both in ,4rabidopsis thaliana and in Nicotiana tabacum as many as 30% of the T-DNA inserts induce active gene fusions. This high frequency of activation suggests that T-DNA is preferentially integrated into potentially transcriptionally active loci. Here, we describe three hybrid gene constructions which mimic a transcriptional position effect. We demonstrate that transcriptional readthrough into a downstream gene

240 with opposite orientation virtually eliminates expression of this gene in transgenic plants and, furthermore, that a poly(A) signal/terminator alleviates this inhibition. Finally, analysis of transgenic calli containing an nptH gene at the T-DNA border with and without regulatory elements (the P35s and the 3'-nos element) interposed between the reporter gene and the border shows that more variation is observed in the absence of the regulatory elements.

pGv so3

LB -I]'"

3'nos ''

hpt

Pnos

RB

, •. , - ' ~

j--C

Pnos---'fipt II

pGV100

I

3'ocs I

~,d,s

Pnos"

I

npt II

,/rz

pGVI08

3'ocs .

, I--,

P35S

[

P35S

iI

"

I I

iI

nptll

3'ocs 3'nos

pGV109

: so p

EXPERIMENTAL AND DISCUSSION

(a) Plasmid constructions and transient expression in tobacco protoplasts To determine the influence of transcriptional activity on the expression of a downstream gene in opposite orientation, three fusion genes were made which mimic a transcriptional position effect. The plasmid, pGEMI00, is a pGEM2 (Promega, Madison, Wl) derivative which contains the control chimeric gene consisting of the nptll coding sequence fused to the Pnos promoter (HerreraEstrella et al., 1983), and the 3'-untranslated region of ocs (De Greve et al., 1981). Plasmid pGEM108 is derived from pGEMI00 by cloning the cauliflower mosaic virus 35S promoter (Odeli et al., 1985) downstream from and in the opposite orientation to the nptll gene. Plasmid pGEM 109 is a derivative ofpGEM 108 containing the 3'-untranslated region of nos (Depicker et al., 1982) downstream from and in the same orientation as the P35s. The 3'-nos element contains the signals necessary for 3'-processing and polyadenylation and will be referred to as a poly(A) signal/ terminator. We subcloned the nptH chimeric genes of pGEM 100, pGEM 108 and pGEM 109 between the T-DNA borders of the plant transformation vector pGVI503 (lngelbrecht et al., 1989) yielding pGVI00, pGVI08 and pGV109, respectively (Fig. 1). Plasmids pGEMI00, pGEMI08 and pGEM109 were introduced into tobacco protoplasts (N. tabacum cv. SRI) via electroporation (Dekeyser et al., 1989). After 48 h, the protoplasts were harvested for NPTII enzymatic analysis by an in situ assay after nondenaturing gel electrophoresis (Van den Broeck etal., 1985). The NPTII enzymatic activity produced from pGEMI00 was approximately threefold higher than that produced from pGEM 108 (data not shown). This inhibitory effect of the antisense P35s could be alleviated by linearizing the plasmid between the 3'-nos element and the P3ss, which formally proves that transcription starting at the P3ss significantly reduces expression of the downstream nptll gene. Similarly, inhibition ofgene expression by the presence of an opposing 3' promoter has recently been reported (Paszty and Lurquin, 1990). Our experiment further showed that the NPTII activity resulting from pGEM 109 was similar to that from

Fig. 1. Schematic representation of the DNA region transferred to the plant genome. Indicated are the chimeric Hy R gene (hpt), which served as an independent selection marker and the nptll chimeric gene with the different regulatory elements downstream. Structures of four plasmids are shown.

pGEM 100, illustrating that the presence of the 3'-nos element counters the inhibitory effects of the P35s, thus restoring the expression of the nptll gene.

(b) Analysis of the chimeric genes in transgenic plants Next, we determined whether the phenomena observed in transient expression experiments also occur in transgenic plants. Therefore, the T-DNA constructs, pGVI00, pGVI08 and pGVI09, were transferred from E. coli MC 1061 to Agrobacterium C58C 1Rif R(pGV2260) (Deblaere et al., 1985) using pRK2013 as a helper plasmid (Figurski and Helinski, 1979). Southern-blot analysis was performed on exconjugants with cointegrate plasmids to confirm the presence and integrity of the chimeric nptll genes. Agrobacteria harboring these T-DNA constructions were cocultivated with tobacco protoplast-derived cells, and calli were regenerated and selected on medium containing 50 #g Hy/ml (Depicker et ai., 1985). For every construct, twelve randomly chosen Hy R caUi were analyzed for NPTII activity. A typical result (Fig. 2) clearly de-

GV108

ali ~ ~46,1

GV109

IIIIDIIIIID ~

~

GV100

IIIiI i

IIiD -PK PK

NPTII Fig, 2. Analysis of the NPTII enzymatic activity in crude cell extract of transformed callus tissue. The protein concentration in the crude cell extract was determined using the BioRad assay and equal amounts of total protein were loaded as described (see section a). Positions of the NPTII protein and the plant kinases (PK) are indicated.

241 monstrates that the NPTII activity expressed from the pGVI08 T-DNA is much lower than that found in calli transformed with the control nptIl gene ofpGV100. In fact the P3.~s in pGV108 virtually eliminates expression of the downstream and oppositely oriented nptll gene. In addition, the NPTII activity expressed by the pGV109 construct is comparable to that found in pGV100-transformed calli, which confirms that the 3'-nos element is capable of alleviating the effect of P35sTranscriptional readthrough originating near sites of T-DNA integration could be one source of variable expression of genes located at the T-DNA-plant DNA junction. It could, therefore, be expected that the poly(A) signal/terminator in pGVI09 reduces the variability in NPTII synthesis arising from T-DNA integration. Therefore, we have quantified more precisely the NPTII enzymatic activity in 21 independently transformed pGV 100 and pGV 109 c alli (Table l). The variability as measured by the standard error of the mean for each experiment of each construct was subjected to Wilcoxon's signed rank test (Sokal and Rohlf, 1981). This analysis indicated that variability is higher for the pGV100 construct at a significance level of 93%. In addition, for the control construct pGV 100 there are several 'outliers' with values considerably higher or lower than the mean value. This can be due to two effects. First, readthrough transcripts coming from the plant DNA can be blocked by the 3'-nos element, in this way reducing the number of plants having an NPTII activity value which is lower than the mean. Alternatively, the P3ss in pGVI09 could also block potential enhancing effects (Wasylyk et al., 1983) exerted by the neighboring plant DNA, thus reducing the number of plants with higher than average NPTII activity. The values obtained for the GV100 calli are on average somewhat higher than those of GV 10q. This could be due to the fact that transcriptional activity from the (strong) P3ss is not completely blocked by the 3'-nos element.

TABLE I

(c) Conclusions

orientation in transient expression and is much more pronounced in transgenic plants. Possibly, the RNA polymerases transcribing from convergent promoters interfere with one another. Alternatively, the production of ant/sense nptll RNA and subsequent base pairing with the sense nptll RNA could be responsible for the reduced expression of the pGVl08 nptll gone (Cornelissen and Vandewiele, 1989). A poly(A) signal/terminator can substantially reduce readthrough from the upstream promoter, and thereby alleviate this interference. Finally, we have measured the NPTII enzymatic activity in calli containing an nptll gone adjacent to and oriented towards the plant DNA and in calli containing the same reporter gene but with regulatory elements at the plant DNA-T-DNA junction. Statistical analysis does not une-

Recent data suggest that DNA introduced into plant cells using the Agrobacterium vector system is preferentially integrated in actively transcribed genomic regions (Koncz et al., 1989; Herman etal., 1990). We investigated whether this transcriptional activity interferes with the expression of transferred genes and, thus, might be a cause of position effects. It has been demonstrated in mammalian cells, that when two genes are placed in tandem, transcriptional readthrough from the first gene into the second reduces the expression of the second gene (Proudfoot, 1986). Similar observations have also been made in prokaryotic systems (Adhya and Gottesman, 1982). Using chimeric gene constructs we demonstrate that transcriptional activity interferes with the expression of a downstream gene in opposite

Quantification of NPTll activity in transgenic tobacco calli Experiment~

pGVIO0t' cpm {10-3)~ 74 113 62 53 28 25 50 4 69 42

!11

IV

V

pGV109h Mean

66.0

38.0

S.E.

cpm (10- ~)

Mean

S.E.

13.9

44 19 50 50 21

36.0

6.9

! 1.0

9 8 7 16 15

11.0

1.9

38.8

7.1

! 13 26 59 28

25.4

9.7

16 51 56 32 39

40 80 22

47.3

17.1

37 41 31

36.3

2.9

72 73 54

66.3

6.1

54 50 24

42.7

9.4

a The different NPTII assays are numbered I-V. b NPTII enzymatic activity of transgenic calli containing the pGVI00 and pGVI09T-DNA (see Fig. 1). Quantification of NPTII activity was performed by determiningcount per minute(cpm)values in the phosphorylated kanamycin spots as described previously (Ingelbrecht et al., 1989). Due to experimental variables,the absolute cpm values obtained differ significantly in independent NPTII assays; therefore, cpm values of different experiments cannot be compared directly. Also indicated are the mean and the standard error of the mean (S.E.).

242

quivocally confirm that transcriptional interference is the major cause for position effects because the main NPTII value of both samples is different only at 93 ~ and not at 95% confidence limits. Nonetheless, this high value suggests that transcriptional interference contributes to the overall variability and might be responsible for the very low NPTII enzymatic activity observed only among pGV100 calli.

ACKNOWLEDGEMENTS

We want to thank Dr. A. Caplan and Dr. G. Gheysen for critical reading of the manuscript. Special thanks go to Dr. O. Hamerlynck who did the statistical analysis. We also thank Martine De Cock for typing and Karei Spruyt and Vera Vermaercke for figures and photographs. This work was supported by grants from the Services of the Prime Minister (U.I.A.P. No. 120C087). II and PB are indebted to the IWONL for a predoctoral fellowship.

REFERENCES Adhya, S. and Gottesmat:, M.: Promoter occlusion: transcription through a promoter may inhibit its activity. Cell 29 (1982) 939-944. Andrr, D., Colau, D., Scheil, J., Van Montagu, M. and Hernalsteens, J.-P.: Gene tagging in plants by a T-DNA insertion mutagen that generates APH(3') ll-plant gen¢ fusions. Mol. Gen, Genet. 204 (1986) 512-518. Cornelissen, M. and Vandewiele, M.: Both RNA level and translation efficiency are reduced by anti-sense RNA in transgenic tobacco. Nucleic Acids Res. 17 (1989) 833-843. Dean, C., Jones, J., Favreau, M., Dunsmuir, P. and Bedbrook, J.: Influence of flanking sequences on variability in expression levels of an introduced gene in transgenic tobacco plants. Nucleic Acids Res. 16 (1988) 9267-9283. Deblaere, R., Bytebier, B., De Greve, H., Deboeck, F., Scheil, J., Van Montagu, M. and Leemans, L: Efficient octopine Ti plasmid-derived vectors for Agrobacterium.mediated gene transfer to plants. Nucleic Acids Res. 13 (1985) 4777-4788. De Greve, H., Decraemer, H., Seurinck, J., Van Montagu, M. and Schell, J.: The functional organization of the octopine Agrobacterium tumefaciens plasmid pTiB6S3. Plasmid 6 (1981) 235-248. Dekeyser, R., Claes, B., Marichal, M., Van Montagu, M. and Caplan, A.:

Evaluation of selectable markers for rice transformation. Plant Physiol. 90 (1989) 217-223. Depicker, A., Stachel, S., Dhaese, P., Zambryski, P. and Goodman, H.M.: Nopaline synthase: transcript mapping and DNA sequence. J. Mol. Appl. Genet. 1 (1982) 561-573. Depicker, A., Herman, L., Jacobs, A., Schell, J. and Van Montagu, M.: Frequencies of simultaneous transformation with different T-DNAs and their relevance to the Agrobacterium/plant cell interaction. MoL Gen. Genet. 201 (1985) 477-484. Figurski, D.H. and Helinski, D.R.: Replication of an origin-containing derivative ofplasmid RK2 dependent on a plasmid function provided in trans. Prec. Natl. Acad. Sci. USA 76 (1979) 1648-1652. Herman, L., Jacobs, A., Van Montagu, M. and Depicker, A.: Plant chromosome/marker gene fusion assay to study normal and truncated T-DNA integration events. Mol. Gen. Genet. 224 (1990) 248-256. Herrera-Estrella, L., De Block, M., Messens, E., Hernalsteens, J.-P., Van Montagu, M. and Schell, J.: Chimeric genes as dominant selectable markers in plant cells. EMBO J. 2 (1983) 987-995. Ingelbrecht, I.L.W., Herman, L.M.F., Dekeyser, R.A., Van Montagu, M, and Depicker, A.G.: Different 3' end regions strongly influence the level ofgene expression in plant cells, Plant Cell 1 (1989) 671-680. Koncz, C., Martini, N,, Mayerhofer, R., Koncz-Kalman, Z., KOrber, H., Redei, G.P. and Scheil, J.: High-frequency T-DNA-mediated gene tagging in plants. Prec. Natl. Acad. Sci. USA 86 (1989) 8467-8471. Odell, J.T., Nagy, F. and Chua, N.-H.: Identification of DNA sequences required for activity of the cauliflower mosaic virus 35S promoter. Nature 313 (1985) 810-812. Odell, J.T., Nagy, F. and Chua, N.-H.: Variability in 35S promoter expression between independent transformants. In: Key, J.L and Mclntosh, L. (Eds.), Plant Gene Systems and Their Biology(UCLA Symposia on Molecular and Cellular Biology, New Series Vol. 62). Liss, New York, 1987, pp. 321-329. Paszty, C.J.R. and Lurquin, P.F.: Inhibition of transgene expression in plant protoplasts by the presence in cis of an opposing 3'-promoter. Plant Sci. 72 (1990) 69-79. Proudfoot, N.J.: Transcriptional interference and termination between duplicated 0c-giobingene constructs suggests a novel mechanism for gent regulation. Nature 322 (1986) 562-565. Selker, E.U.: DNA methylation and chromatin structure: a view from below. Trends Biochem. Sci. 15 (1990) 103-107. Sokal, R.R. and Rohif, F.J.: Biometry, 2nd ed. Freeman, New York, 1981. Van den Broeck, G., Timko, M.P., Kausch, A.P., Cashmere, A.R., Van Montagu, M. and Herrera-Estrella, L.: Targeting of a foreign protein to chloroplasts by fusion to the transit peptide ofrib~!ose i,5-bisphosphate carboxylase. Nature 313 (1985) 35~-~6~ Wasylyk, B., Wasylyk, C., Augereau, P. and Chambon, P.: The SV40 72 bp repeat preferentially potentiates transcription starting from proximal natural or substitute promoter elements. Cell 32 (1983) 503-514.