rPPR5 and rCRP1 bind preferentially to single-stranded RNA relative to single-stranded DNA or double-stranded nucleic acids. Gel mobility shift assays were performed with the indicated concentrations of rCRP1 (A) or rPPR5 (B) and 31-mer oligonucleotides in single-stranded DNA (ssDNA), double-stranded DNA (dsDNA), single-stranded RNA (ssRNA), or double-stranded RNA (dsRNA) form. The petA-2 and petA-1 nucleic acid sequences are from two different regions of the petA 5′ untranslated region. For assays with single-stranded RNA or DNA, oligonucleotides were heated and then snap-cooled on ice immediately prior to protein addition. Double-stranded oligonucleotides were generated by solution hybridization with the complementary oligonucleotides and were used either directly in binding assays (left and middle panels), or after gel purification of the double-stranded form (right panels).

Fine-mapping the PPR5 binding site. (A) Identification of a 51-nt segment of the trnG-intron to which rPPR5 binds with specificity. The overlapping ∼50-mer RNAs diagrammed beneath the gene map were used in GMS assays. Each RNA included an additional three G residues at its 5′ end, introduced by the promoter in the transcription template. (B) Mutational analyses to map the RNA sequences that contribute to sequence-specific RNA recognition by PPR5. The sequence of each RNA used in binding reactions is annotated with the position of the predicted EBS1 and α′ group II intron motifs (see Fig. 7A). Each RNA except RNA E included two additional G residues at its 5′ end, introduced by the promoter in the transcription template. The three point mutations in RNA K are underlined.

Identification of a high-affinity rPPR5 binding site within the trnG-UCC intron. Gel mobility shift assays used body-labeled RNAs with the indicated concentrations of rPPR5. To minimize the formation of RNA structure, RNAs were heated in the absence of salts and snap-cooled immediately prior to the addition of salts and protein. (A) Preferential binding of rPPR5 to a segment of the trnG-UCC group II intron. The unspliced trnG-UCC RNA is diagrammed at the top, with exons denoted as rectangles, and the six group II intron domains indicated. The positions of the six RNAs assayed for binding are diagrammed below. The fraction of the input RNA bound to protein was quantified with a phosphorimager, and is shown below. (B) Competition assays demonstrating a sequence-specific interaction between trnG-4 RNA and rPPR5. Binding reactions included radioactive trnG-4 RNA together with a 10- or 100-fold molar excess of the indicated nonradioactive RNAs. [rPPR5] was held constant at 400 nM.

Purification of rCRP1 and rPPR5. (A) Domain organization of PPR5 and CRP1. PPR5 consists of 10 tandem PPR motifs preceded by ∼35 amino acids at the N terminus and followed by 55 amino acids at the C terminus. CRP1 consists of 14 tandem PPR motifs preceded by a region rich in hydrophobic residues of ∼80 amino acids. Both proteins are synthesized with an N-terminal chloroplast transit peptide that is not represented in the recombinant proteins. (B) Elution of rCRP1 and rPPR5 from a Superdex 200 gel filtration column. An equal proportion of contiguous column fractions spanning the elution peaks was analyzed by SDS-PAGE and staining with Coomassie brilliant blue. Elution peaks of BSA (67 kDa) and maltose binding protein (42 kDa) chromatographed under the same conditions are shown.

Biophysical assays of rPPR5. (A) CD spectrum of rPPR5 at 0.12 mg/mL. (B) Change in CD signal at 222 nm as rPPR5 was increased in temperature from 0°C to 70°C in the absence (black line) or presence (gray line) of an equimolar amount of the nonspecific 31-mer RNA oligonucleotide described in Materials and Methods. (C) Sedimentation velocity profile of rPPR5 during analytical ultracentrifugation. (Left) Absorbance at 229 nm plotted as a function of position from the axis of rotation at time intervals of 5 min. Black lines represent nonlinear best fits to a continuous sedimentation distribution model. Gray lines represent experimental data and are shown for every fourth scan. (Right) Sedimentation coefficient, c(s), profile for rPPR5 converted into the differential molar mass distribution, c(M) using SEDFIT. Uncleaved MBP–PPR5 contaminated this preparation (∼5% of the total) and can account for the small peak at ∼98 kDa. The residuals were randomly distributed, with 85% of the residuals having a value <10% of the signal and a maximum spread of 20% (data not shown).

Models of the PPR5 binding site. (A) Predicted secondary structure of the trnG-iii 51-mer RNA to which PPR5 binds with high specificity. The structure was calculated by Mfold (Zuker 2003) at http://frontend.bioinfo.rpi.edu/applications/mfold/, using default parameters, and has a calculated ΔG = −15.3 kcal/mol. The positions of the EBS1 and α′ motifs are marked; EBS1 was assigned based on the trnG-UCC intron model at the Comparative RNA website (http://www.rna.ccbb.utexas.edu/); α′ comes from a model generated by C. Wang and S. Zimmerly (pers. comm.). The 3′ end of the truncated transcript whose accumulation is suppressed by PPR5 maps within several nucleotides of residue 30, on the 3′ side of the diagrammed hairpin. (B) Models for the interaction between PPR5 and its RNA ligand. The models assume that the RNA in contact with PPR5 is single stranded, as suggested by the data in Figure 3. Each of the 10 PPR motifs in PPR5 is represented by a cylinder corresponding to predicted “helix A,” which has been proposed to present the substrate binding face of PPR tracts (for review, see Delannoy et al. 2007). These models diagram a PPR5 monomer because rPPR5 is monomeric under the conditions used in the binding assays; however, cooperative binding of multiple monomers remains a possibility. The models invoke either a single contiguous binding site spanning at least 40 nt (bottom) or a discontinuous binding site that includes sequences near both ends of the 51-mer (top). A discontinuous binding site is best supported by current data.

Mapping the 3′ end of a truncated trnG-UCC transcript whose accumulation is influenced by PPR5. (A) Total leaf RNA (5 μg) from seedlings of the indicated genotypes was analyzed by RNA gel blot hybridization. The probes, diagrammed below a map of the trnG-UCC locus, were the same DNAs used to generate RNAs for the binding assays in Figure 5A. Estimated transcript sizes, indicated to the right, were based on their mobility with respect to RNA size markers. The ∼440 nt transcript was shown elsewhere to include the 5′ exon and to lack the 3′ exon (Beick et al. 2008). The analyzed ppr5 mutant alleles have an incomplete loss of ppr5 function; null mutants are uninformative because their severe plastid ribosome deficiency causes pleiotropic effects that mask defects caused directly by the absence of PPR5. (B) Total leaf RNA (2 μg) from seedlings of the indicated genotypes was analyzed with a ribonuclease protection assay. The diagram shows the position of the 293-nt body-labeled probe with respect to the PPR5 binding site; residue 1 represents the 3′ end of the probe. A reaction using an equivalent amount of yeast tRNA served as a negative control (tRNA). hcf7 mutants exhibit a plastid ribosome deficiency of similar magnitude to that in the ppr5 mutant alleles analyzed and were included to control for secondary effects of this ribosome deficiency. The estimated transcript sizes, indicated to the right, are based upon DNA size standards after calibration for RNA by an adjustment of 11%; this adjustment was based on the mobility of the 293-nt RNA probe with respect to the DNA standards. The protected RNAs of 293 and 272 nt both result from hybridization to intact intron RNA; the 272 fragment results from a sequence polymorphism between the genetic backgrounds used to generate the probe and the leaf RNA (data not shown).