Rice Genetics Newsletter 14 (1997) 23

R. Wu, Convener

The ultimate goal of rice biotechnology research is to produce superior transgenic plants that give stable expression of an introduced gene (transgene). Over the last several years, gene silencing in transgenic plants has emerged as a serious problem. In gene silencing, either the transgene is not expressed at all, or it is expressed in plants during earlier stages of growth but not in later stages, or it is expressed in the first generation (R₀) but not in the next generation (R₁). In general, gene silencing is often associated with integration of multiple copies of the introduced gene. The higher the copy number, the more likely the occurrence of gene silencing (Alien et al. 1993; Matzke and Matzke 1995; Mlynarova et al. 1994; Meyer and Saedler 1996; Spiker and Thompson 1996).

Gene silencing is also known as inactivation of homologous transgene expression. Several different mechanisms are thought to be involved in gene silencing, which include cis inactivation (frequently due to methylation of DNA), trans inactivation, or co suppression (sense suppression) of a transgene or a homologous endogenous gene. Gene silencing may also be related to chromosomal insertion sites of the transgene (position effect), or due to degradation of RNA-RNA duplexes (Matzke and Matzke 1995; Alien et al. 1993; 1996).

One way to minimize the problem of gene silencing and to increase transgene expression is to flank the transgene with a DNA sequence known as a scaffold attachment region (SAR) or matrix attachment region (MAR). Experiments on the effects of MAR on gene expression in transgenic dicot plants have been reviewed (see Spiker and Thompson 1996 and references therein). It has been proposed that when MAR sequences are present on both sides of a transgene, they can form a separate transcriptional active loop domain by binding to the nuclear matrix. Even though results from different groups, using different MAR sequences, are not always consistent, it seems that including flanking MAR sequences in plasmid constructs might be useful. In addition, eliminating repeated sequences from plasmid constructs may alleviate problems related to DNA-DNA pairing and de novo methylation. To minimize RNA turnover induced by excess RNA accumulation, moderate transcription rates (driven by a moderately strong promoter) may be preferable to very high rates (driven by a very strong promoter).

Gene silencing in transgenic rice was reported recently at the General Meetings of the International Program on Rice Biotechnology, September 15-19, 1997, Malacca, Malaysia, which was organized by the Rockefeller Foundation. Kumpatia et al. (1997) reported

24 Rice Genetics Newsletter Vol. 14

that direct DNA delivery procedures resulted in plants that contained multiple copies of the transgene, some of which were rearranged. By following bar gene expression, it was found that inactivation of the transgene was due to methylation. The silenced state was inherited, and bar gene expression was restored by germinating R₁ seeds in the presence of 5-azacytidine.

Kloti et al. (1997) found that gene silencing occurred in the T₀ generation or later. In some cases, homozygosity enhanced the silent phenotype. With the RTBV promoter, cell-specific silencing was observed, which leads to lack of expression in the vascular tissue, but not in the epidermis. In these plants, DNA methylation of the promoter fragments was detected, and silencing was reversed by sexual crossing with nontransgenic plants or by treatment with 5 azacytidine.

Thara et al. (1997) observed that transgene silencing occurred in some T₃ and T₄ generation progeny of homozygous rice plants expressing an introduced chitinase gene.

Viegas et al. (1997) found that among five R₀ plants expressing the reverse tran-scriptase gene (RT) of RTBV, one of them no longer produced RT mRNA in the R₁ generation. Since several copies of the same transgene are present in this R₁ line, the results suggest the possibility of gene co-suppression.

From the above examples, it appears that gene silencing is a problem that can arise from attempting to produce transgenic rice plants that stably express the transgene for many generations. It is likely that gene silencing can be largely prevented if only one or two intact copies of the transgene are integrated into the genome without the presence of rearranged copies. One solution to this problem is to use MAR sequences to flank the gene of interest to be introduced into rice cells by transformation (Jayaprakash et al. 1997).

Kloti, A., J. Wunn, P. Burkhardt, S. Bieri, C. Henrich, P. Lucca and I. Potrykus, 1997. Gene silencing in

Kumpatia, S.P., W. Teng, W.G. Buchholz and T.C. Hall, 1997. Gene silencing and reactivation in transgenic

rice. In Intl. Program on Rice Biotechnology. September 15-19, Malacca. Malaysia, p. 64.

Matzke, M.A. and A.J.M. Matzke, 1995. How and why do plants inactivate homologous transgenes? Plant

Meyer, P. and H. Saedler, 1996. Homology dependent gene silencing in plants. Annu. Rev. Plant Physiol.

Report of Committee on Genetic Engineering 25

Spiker, S. and W.R Thompson, 1996. Nuclear matrix attachment regions and transgene expression in plants.