Transgene expression is not only variable between independent transgenic lines but can also be completely switched off via the process of gene silencing. Gene expression can be modulated at the transcriptional, mRNA processing and translational levels. Transgene silencing, on the other hand, can occur at the transcriptional level where it is known as transcriptional gene silencing (TGS), or at the post-transcriptional level where is termed post-transcriptional gene silencing (PTGS). Understanding transgene silencing and controlling transgene expression are two vital areas of both pure and applied plant science. In this study, several aspects of epigenetics and regulation of gene expression were investigated in plants. Although extremely high levels of transgene expression are sometimes desired, reproducible and consistent levels (whether low or high) of transgene expression can also be important. Previous work in mammalian cell cultures has demonstrated that the number of upstream Open Reading Frames (uORFs) within the 5¡¯-untranslated region (5¡¯-UTR) is inversely proportional to translation efficiency and levels of protein expression. In order to test if this is true in a stably transformed plant system, the same mammalian 5¡¯-UTRs were incorporated into a GUS expression cassette and transformed into Arabidopsis. Similar changes in translation efficiency were seen, but more severe effects were observed at the level of mRNA stability in the plant system. Inclusion of uORFs in modified Tobacco Mosaic Virus (TMV) §Ù 5¡¯-UTRs also resulted in decreased GUS expression in a transient system. It was concluded that uORFs could be of use in transgenic plant systems to gain a range of protein expression levels. A reported feature of the two-component Ds/sAc transposon tagging system in transgenic tomato, is sAc transposase-mediated silencing of the nos:BAR selectable marker gene within the Ds element (Carroll et al., 1995; Reyes et al., 2004). The mechanism of this form of transgene silencing was further elucidated. The results indicated that histone modification occurs prior to DNA methylation, with the later potentially occurring as a form of maintenance silencing in most transposed Ds lines after the sAc has segregated away (Reyes et al., 2004). A silenced Ds was previously utilised to tag a chromosomal region of the tomato genome that appeared to have the potential to alleviate nos:BAR silencing (Reyes, 2002). The tomato sequences flanking this Ds insertion called UQ14 expression modulating sequences (EMS) were cloned and tested for their ability to enhance artificial viral resistance to Potato virus Y (PVY) conferred by a NIa hairpin in transgenic tobacco. Inclusion of the UQ14 EMS sequences in the transgene construct enhanced the frequency of immunity approximately two-fold over non-EMS controls. This demonstrated that the UQ14 sequences have the potential to be used in at least solanaceaeous plant species to modulate transgene expression. To study the mechanisms of long-distance movement of RNA silencing in Arabidopsis, transgenic lines containing a GF-specific hairpin targeting a full length GFP transgene were generated and used in grafting experiments. Using the GF hairpin silenced line as rootstock and a GFP expressing line as a scion, it was shown that once initiated, silencing occurred in all newly formed leaves but did not spread into pre-established older leaf tissue. No genes tested were found to be required for sending the mobile RNA silencing signal. However, there was a correlation between the level of 21-22 nt small interfering RNA (siRNA) and double-stranded RNA (dsRNA) levels in the rootstock genotype, and the speed at which silencing was transmitted to the scion. In addition to this, DCL3-dependent 24 nt siRNA species were found to play no role in the transmission of the silencing signal from the rootstock. Analysis of the genes required for the reception of silencing revealed that multiple RNA silencing pathways were involved. It was shown for the first time that a chromatin silencing pathway, involving RNA polymerase IVa, RNA-dependent RNA polymerase 2 (RDR2), DICER-like 3 (DCL3) and ARGONAUTE 4 is required for reception of long-distance RNA silencing. This pathway produces 24 nt siRNAs, and resulted in decapped RNA, a known substrate for amplification of dsRNA by RDR6. This dsRNA is then processed by DCL4 into 21 nt siRNAs, or in its absence by DCL2 into 22 nt siRNAs, and these guide the degradation of the target GFP transcript in a manner similar to anti-viral defense.