Correction: Misregulation of AUXIN RESPONSE FACTOR 8 Underlies the Developmental Abnormalities Caused by Three Distinct Viral Silencing Suppressors in Arabidopsis

[This corrects the article DOI: 10.1371/journal.ppat.1002035.].

Errors in Figs 4, 5 and 6 were introduced during preparation for publication. Olivier Voinnet, the corresponding author, assumes full responsibility for these errors. All the erroneous figures have been re-prepared from the original raw material and the seeds of the HcPro, P15 and P19 transgenic lines are available upon reasonable request to the journal.
In Fig 4C, the first track of the blot was incorrectly mounted and used as the control track arf8-6 -/-. This track has now been replaced by the correct arf8-6 -/track, alongside its corresponding rRNA loading control track, into the corrected Fig 4C. The blot-splicing operation is now indicated with a white space.
In Fig 4E, the original blot was flipped horizontally for mounting this panel, tracks were spliced without notification, and incorrect/un-flipped rRNA tracks were used as loading control. Fig 4E has been now corrected by splicing out the appropriate tracks from the original unflipped Northern blot, alongside their corresponding rRNA loading control tracks. The blotsplicing operations are now indicated with white spaces.
In Fig 4I, incorrect tracks were mounted and an inappropriate ethidium bromide staining was used. Fig 4I has now been corrected by splicing out the appropriate tracks from the original Northern blot, and by indicating blot-splicing operations with white spaces. Furthermore, since the corresponding original ethidium bromide staining was of poor quality, a pre-loading control from a preparatory agarose gel was used to correct the rRNA loading controls inserted into the corrected Fig 4I. In Fig 5A, incorrect tracks were mounted and spliced without notification. Fig 5A has now been corrected by splicing out the appropriate tracks from the original Western blots, alongside their corresponding coomassie loading control tracks, and by indicating blot-splicing operations with white spaces. Furthermore, for the AGO1 panel, flowers or seedlings were analyzed, which was not properly indicated. Into the correct Fig 5A, the plant tissues used for these analyses are now clearly specified, both on the panel and in the figure caption, and the "Col-0 seedlings" track has been added, alongside its coomassie loading control track, as was the corresponding control for the ago1-36 mutant track.
In Fig 5B, Col-0 and arf8-6 -/tracks together with their U6 loading controls were inverted during the mounting process. Fig 5B has now been corrected by flipping all the original panels horizontally.  Table shows that refining the analysis with additional filters based on transcripts up-regulated in dcl1-9 (pale grey) and hen1-1 (dark grey) backgrounds singularizes ARF4 and ARF8, respectively direct targets of miR390 and miR167, as strong candidates for the underlying developmental defects seen in VSR transgenics.  accumulation in arf8 homozygous seedlings compared to WT seedlings and AGO1 accumulation in arf8 homozygous flowers compared to WT flowers. Due to the strong developmental defects (dcl1) or developmental arrest (ago1) exhibited by miRNA deficient mutants, the negative controls used in this analysis were from seedlings with the hypomorphic dcl1-9 genotype, which accumulate a truncated, non-functional form of the DCL1 protein, and from seedlings with the null ago1-36 genotype. Load: Coomassie staining provides a control for equal loading of total proteins. hybridizations were presented as if they originated from a single membrane, because the corresponding rRNA loading control for the separate CHS siRNA experiment was omitted. To correct Fig 5E, the original membrane used for the CHS RNAi hybridization was re-probed and the new signal was used, alongside its corresponding U6 loading control, and clearly separated from the other hybridizations.
In Fig 6B, the arf8-4 -/track was stretched towards the left during the mounting process for the panel P6 and the wrong ethidium bromide staining was used as rRNA loading control. To correct Fig 6B, the panel P6 was replaced by a new scan of the original Northern blot and the original membrane was probed for the Actin2 housekeeping mRNA, whose hybridization signal is now used as the loading control into the corrected Fig 6B.  Fig 4G was correct. However, for full transparency, this figure has been re-mounted in order to use the exact same raw material as Fig 4C, since both were spliced out from the same original blot.
The authors confirm that these changes do not alter the results of their study and have provided the relevant underlying data as Supporting Information.  hybridized with a mix of random-labeled PCR products corresponding to the 35S terminator and HcPro, allowing simultaneous detection of the P19 and HcPro transcripts (bottom lane), respectively. Blue rectangles indicate the part of the blot used for mounting these figures; samples were loaded according to the track labels; surrounded numbers correspond to the annotated samples on the original Fig 4B and 4F. (B) Original film (left) and ethidium bromide staining (right) used for mounting Fig 4E. The blot was hybridized with random-labeled PCR products corresponding to the 35S terminator, allowing detection of the P15 transcripts. Samples were loaded according to the track labels; surrounded numbers correspond to the annotated samples on the original Fig 4D. (C) Original scan (left) and inappropriate ethidium bromide staining (right) used for mounting Fig 4I. The blot was hybridized with randomlabeled PCR products corresponding to the 35S terminator, allowing detection of the P19 transcripts. Samples were loaded according to the track labels. (D) Original film (left) and original ethidium bromide staining (right) corresponding to samples presented in Fig 4I. The blot was hybridized with random-labeled PCR products corresponding to the 35S terminator, allowing detection of the P19 transcripts. Blue rectangles indicate the part of the blot with the samples used in Fig 4I. Fig 5B. Fig 5B was mounted from films obtained by sequentially stripping and re-hybridizing a single Northern blot membrane with several distinct miRNA probes. U6 and miR168 were hybridized at the same time. The long film exposure (2h30) was selected for miR168, and the short exposure (1h) was selected for U6, to avoid a saturated loading signal. Col-0 and arf8-6 -/are respectively on tracks #7 and #6. rRNA, not used for the mounting of Fig 5B, provides an additional loading control. (C) Original films used for mounting Fig 5E. Fig 5E was mounted from films obtained by sequentially stripping and rehybridizing two Northern blot membranes with several distinct miRNA/siRNA probes. U6 and miR168 were hybridized at the same time. The long film exposure (4h30) was selected for miR168, and the short exposure (1h30) was selected for U6, to avoid a saturated loading signal.  Fig 6B. (D) Original pre-loading control corresponding to the ethidium bromide stainings of 1 μg of total RNA loaded on a 1% agarose gel to check quality and equal loading prior loading of the high molecular Northern blot used in Fig 6B. (E) Original film (left) and ethidium bromide staining (right) used for mounting Fig  6D. The blot was hybridized with random-labeled PCR products corresponding to HcPro. Samples were loaded according to the tracks labels. P15 was detected from 20 μg of total proteins from seedlings. The P15 antibody was used at a 1/10 000 dilution. P19 was detected from 100 μg of total proteins from seedlings. The P19 antibody was used at a 1/5000 dilution. Hc-Pro was detected from 20 μg of total proteins from seedlings. The Hc-Pro antibody was used at a 1/8000 dilution. The red arrow indicates the position of the P19 signal.