In cultured human (HeLa) cells, oxidized RNA localizes to cytoplasmic ORBs during arsenite-induced oxidative stress. Cells were stained for 8-oxoG (green) and DNA was co-stained with DAPI (blue) and analysed by performing immunofluorescence microscopy. (A) Foci of 8-oxoG were not seen in 95% of cells prior to arsenite treatment. In 40% of arsenite-treated cells, the 8-oxoG immunofluorescence signal was seen in cytoplasmic foci (white arrows) that are distinct from stress granules (seen as foci of the red G3BP1 immunofluorescence signal). ORBs were not seen in arsenite-treated cells which were treated with RNase after fixation. (B) ORBs are also distinct from processing bodies (seen as foci of the red RAK immunofluorescence signal). (C) ORBs are not mitochondria (seen as the red TOM20 immunofluorescence signal). 250 cells were analysed for each experiment. Scale bars: 20 µm.

RBCL affects the level of oxidized RNA, but not oxidized DNA or protein. (A) Levels of 8-oxoG in RNA from the wild-type strain, ΔrbcL, ΔRBCS, the rescued ΔRBCS strain, and the double mutant for rbcL and the RBCS locus are shown under non-stress conditions (dark grey bars) and from cells exposed to 2.0 mM H₂O₂ (for the wild-type strain, ΔrbcL, ΔRBCS) (light grey bars). The bar height represents the percentage of 8-oxoG ECL signal compared with that of the ΔrbcL strain before stress (white bar; 100%); n≥5. (B) The relative levels of 8-oxoG in DNA from the wild-type strain, ΔrbcL and ΔRBCS are presented as described for RNA in A; n≥3. (C) Levels of carbonylated amino acid residues in protein from the wild-type strain, ΔrbcL, and ΔRBCS are presented as described for RNA in A; n=3. Results were analysed and are presented as described in the Materials and Methods. Error bars indicate the s.e.m. *P≤0.05 compared with ΔrbcL, as determined by one-sample Student's t-tests.

A Rubisco-independent pool of RBCL. (A) The relative levels of RBCL in crude lysates of wild-type, ΔRBCS and ΔrbcL cells were determined by SDS-PAGE and immunoblot analysis. Dilutions of the wild-type cell lysate were supplemented with lysate from ΔrbcL to maintain a constant amount of total protein in each lane. The band attributed to RBCL was not a cross-reacting protein as it was not detected in ΔrbcL. The signal from a protein of the 30S subunit of the chloroplast ribosome (30S) was used as a loading control. Black bars indicate adjoined non-adjacent lanes from a single exposure. (B) An illustration of the fractionation scheme used in this study. (C,D) RBCL fractionation during differential centrifugation and solubilisation with Triton X-100 was revealed by immunoblot analyses for (C) the wild-type strain and (D) ΔRBCS. Proteins analysed were RBCL, RBCS, a soluble protein of the chloroplast stroma (HSP70B), an integral thylakoid membrane protein (CP43) and ribosomal proteins of chloroplast ribosome subunits (30S and 50S). Where indicated, fractions were prepared from cells that had been exposed to 2.0 mM H₂O₂. (E) ³⁵S-pulse-labelled (newly synthesized) RBCL in the insoluble (P16) and soluble (S16) fractions from the wild-type strain and ΔRBCS was revealed by SDS-PAGE and phosphoimaging. As a negative control, fractions from ΔrbcL were shown to lack the ³⁵S-pulse-labelled band assigned to RBCL. An unidentified ³⁵S-pulse-labelled protein was detected in both fractions of all three strains (asterisk). Black bars indicate adjoined non-adjacent lanes from a single exposure.