Many factors affect the expression of gene silencing. Reports indicate that it can occur at either the transcriptional or posttranscriptional level. Silencing that occurs at the transcriptional level may be caused by direct physical association or pairing of alleles. Also, the ploidy of the plant may sometimes influence gene expression, since reduced transgene expression is observed in triploid, as compared to diploid, Arabidopsis hybrids. Posttranscriptional transgene silencing is affected by both plant developmental stage and environmental factors. The experimental data accumulated to date do not fully explain the mechanism of gene silencing, nor do the suggested models accommodate all experimental data.

To investigate the incidence of gene silencing using different T-DNA configurations under a variety of environmental growth conditions, a team of researchers from the University of Oslo, Norway and University of Skovde, Sweden developed a collection of 111 independent transgenic Arabidopsis thaliana lines. Two vectors, pPCV002 35SGUS and pKOH110 35SGUS, containing gus and nptII reporter genes, were used for the transformation. The 111 lines were divided into six categories according to the number of loci and state of hemizygosity or homozygosity using segregation ratios for kanamycin resistant and kanamycin sensitive phenotypes in T2 seeds. Segregation data were used to determine the extent that nptII-silencing in the different categories was different from the lines screened.

For each transgenic line, eleven four-week-old seedlings were selected and kept at normal growth conditions for three days. Then stress treatment, consisting of 30°C day and night temperatures, was given to seven plants, of which four were also sprayed with insecticide. The other four plants were given a 30°C stress treatment during the day and 4°C during the night. After two weeks, the plants were transferred to original growth conditions (22°C). NptII-silencing was scored by planting dormancy-broken surface-sterilized T3 seeds. Determinations of T-DNA copy number and DNA methylation were made using Southern hybridization.

Sixty-seven and 44 transformants were independently generated for pPCV002 and pKOH110 vectors, respectively. NptII-silencing was scored for each line after germination on medium containing kanamycin. Three silencing phenotypes, I - III, were identified. White or light green cotyledonous plants were designated as Type I; those with white, spotted, and deformed leaves were classified Type II; and Type III had large, green, spotted leaves. Type II and III phenotypes are never found in wild-type seedlings. It was assumed that the white parts of the seedlings were displayed where nptII was silenced, inhibiting normal seed development. Those seedlings displaying nptII-silencing from the stress-treated groups were identified and the overall silencing frequencies of the stress and control groups were calculated for each line. The extra stress treatments, 4°C and insecticide, did not result in additional effects.

NptII-silencing of T3 plants was found in 56% of the 111 transgenic lines. In the majority of lines having a low to medium overall frequency of nptII-silencing (less than 10 - 50%), the frequency was higher in the stressed group than the control group; however, this was not seen for any lines displaying more than 50% silencing frequency. Differences in silencing frequency were not correlated to the T-DNA vector source.

Interestingly, the silencing frequency differed between siblings in all but two lines. In the majority of lines, the frequency of nptII-silencing was significantly different between the stressed group and the control. In several cases, the frequency of silencing in progeny of stress treated plants was higher than for the control group, while in other lines this trend was reversed.

The team investigated whether there was a correlation between homozygosity and silencing in progeny of hemi- and homozygous siblings of the T3 generation and found that silencing could be attributed to homozygosity in only one line. Likewise, a correlation between kanamycin resistance and methylation of the Sac II site in the pnos promoter region, as has been reported in other studies, was found in multi-copy, but not single copy, lines. Additionally, the team found that the frequency of silencing was not correlated with rearranged transgenes.

Stress-induced susceptibility of plants to gene silencing is exhibited at three levels: (i) stress can change the proportion of sibling plants producing silenced progeny; (ii) individual plants may influence the number of seeds in which the transgene is silenced due to a change in the number of cells with silenced transgene(s); and (iii) silencing phenotypes can be altered. All transgene copies in all cells of Type I seedlings are likely to be silenced. However, in seedlings of Type II and III phenotypes, which contain seedlings that have developed beyond the cotyledon stage, transgenes are generally only partially silenced.

Meza et al. report that environmental stress can have either a positive or negative influence on the frequency of silencing. One transgenic line, which exhibited an increase in silencing after stress treatment, displayed both Type I and II phenotypes, while the control seedlings displayed the II and III types. Conversely, in another line in which stress treatment produced a reduction in nptII-silencing, the silenced progeny from the stressed group showed the Type III phenotype, and the control group displayed Types I and II.

The authors suggest that environmental stress produces changes in methylation patterns and/or chromatin conformations. They theorize that those transgenes that integrate into genomic regions which are subject to epigenetic modification during stress treatment are susceptible to environmentally induced silencing. However, in some instances, chromatin configuration may have more of an influence than methylation—as with the Sac II site of the single copy line in which methylation was not detected.

Studies of the sequences flanking the T-DNA in one-locus lines, some of which contain a single copy, may help generate transgenic lines with stable expression patterns. In addition to questions concerning the mechanisms underlying gene silencing, there is interest in isolating genes whose products modify the timing of silencing. There is also a need to clarify the involvement of methylation in silencing and to minimize this effect by targeting transgenes into compatible isochores that include flanking scaffold attachment regions in the constructs.

Sources

1. Meza TJ, et al. 2001. The frequency of silencing in Arabidopsis thaliana varies highly between progeny of siblings and can be influenced by environmental factors. Transgenic Research 10: 53-67.

2. Scheid OM, et al. 1996. A change of ploidy can modify epigenetic silencing. Proceedings of the National Academy of Sciences USA 93: 7114-7119.

3. Matzke MA, Matzke AJM, and Eggleston WB. 1996. Paramutation and transgene silencing: A common response to invasive DNA. Trends in Plant Science 1: 382-388.