Second-Site Mutagenesis of a Hypomorphic argonaute1 Allele Identifies SUPERKILLER3 as an Endogenous Suppressor of Transgene Posttranscriptional Gene Silencing

Abstract

Second-site mutagenesis was performed on the argonaute1-33 (ago1-33) hypomorphic mutant, which exhibits reduced sense transgene posttranscriptional gene silencing (S-PTGS). Mutations in FIERY1, a positive regulator of the cytoplasmic 5'-to-3' EXORIBONUCLEASE4 (XRN4), and in SUPERKILLER3 (SKI3), a member of the SKI complex that threads RNAs directly to the 3'-to-5' exoribonuclease of the cytoplasmic exosome, compensated AGO1 partial deficiency and restored S-PTGS with 100% efficiency. Moreover, xrn4 and ski3 single mutations provoked the entry of nonsilenced transgenes into S-PTGS and enhanced S-PTGS on partially silenced transgenes, indicating that cytoplasmic 5'-to-3' and 3'-to-5' RNA degradation generally counteract S-PTGS, likely by reducing the amount of transgene aberrant RNAs that are used by the S-PTGS pathway to build up small interfering RNAs that guide transgene RNA cleavage by AGO1. Constructs generating improperly terminated transgene messenger RNAs (mRNAs) were not more sensitive to ski3 or xrn4 than regular constructs, suggesting that improperly terminated transgene mRNAs not only are degraded from both the 3' end but also from the 5' end, likely after decapping. The facts that impairment of either 5'-to-3' or 3'-to-5' RNA degradation is sufficient to provoke the entry of transgene RNA into the S-PTGS pathway, whereas simultaneous impairment of both pathways is necessary to provoke the entry of endogenous mRNA into the S-PTGS pathway, suggest poor RNA quality upon the transcription of transgenes integrated at random genomic locations.

Figure 1.

SKI3 deficiency restores L1 S-PTGS in ago1-33. A, GUS activity was measured in leaves at 50 d after germination (DAG). For each of the indicated genotypes, 96 plants were analyzed. S-PTGS efficiency is expressed as the percentage of GUS-negative plants. B, Low-molecular-weight (LMW) RNA was extracted from 18-d-old seedlings of the indicated genotypes, run on a polyacrylamide gel, and hybridized with GUS and U6 probes. C and D, GUS activity was measured in leaves at 40 or 50 DAG. The numbers of independent transformants (C) and descendants of individual transformants (D) analyzed are indicated. S-PTGS efficiency is expressed as the percentage of GUS-negative plants.

Figure 2.

SKI3 deficiency does not confer visible developmental defects. A, Schematic representation of the SKI3 gene and locations of the ethyl methanesulfonate and transfer DNA (T-DNA) mutations. B, Photographs of plants of the indicated genotypes. ago1-33 and ago1-33 ski3-3 are 25 d old, and wild-type Columbia (Col), ski3-3, and ski3-5 are 15 d old.

Figure 3.

SKI3 localizes to the cytoplasm. A, Arabidopsis roots of two independent L1 ago1-33 ski3-3/pSKI3:SKI3-GFP transformants exhibiting restored L1 S-PTGS imaged on a Leica TCS-SP5. B, Magnified image of transformant 3

Figure 4.

ski3 has a weaker effect than xrn4 on S-PTGS. A, GUS activity was measured in leaves at 20 or 40 DAG. For each of the indicated genotypes, 96 plants were analyzed. S-PTGS efficiency is expressed as the percentage of GUS-negative plants. B, LMW RNA was extracted from leaves of 20-d-old plants of the indicated genotypes, run on a polyacrylamide gel, and hybridized with a GUS probe. Ethidium bromide (EtBr) staining is shown as a loading control. C, GUS activity was measured in leaves at 50 DAG. For each of the indicated genotypes, 96 plants were analyzed. S-PTGS efficiency is expressed as the percentage of GUS-negative plants.

Figure 5.

ski3 has a weaker effect than xrn4 on terminatorless constructs. AGO1 silencing was estimated by visual inspection of transformants of the indicated genotypes. The number of transformants analyzed is indicated. S-PTGS efficiency is expressed as the percentage of AGO1-silenced plants.

Figure 6.

Effects of ski3 on various ago1 alleles. A, Schematic representation of the different domains of the AGO1 protein (from left to right : N-terminal coil [N coil], N domain, domain of unknown function1785 [DUF1785], Piwi-Argonaute-Zwille [PAZ] domain, linker2 [L2], domain of the middle of the primary structure [MID], and C-terminal [PIWI] domain. The nature and position of the amino acid change are indicated for each mutation analyzed. B, L1 silencing was quantified by scoring GUS activity in plants of the indicated genotypes. PTGS efficiency is expressed as the percentage of silenced plants (n = 96).