A Pentatricopeptide Repeat Protein Facilitates the trans-Splicing of the Maize Chloroplast rps12 Pre-mRNA

PPR4 Is the Only Predicted Plant Protein with Both a PPR and an RRM Domain

The PlantRBP database (http://plantrbp.uoregon.edu/) provides a tool to search for orthologous proteins in rice and Arabidopsis with specified domain content and predicted intracellular location. A search for proteins that are predicted to be chloroplast targeted and to have both an RRM and a PPR domain returned one Arabidopsis protein, At5g04810, which was assigned to an orthologous group with rice Os04g58780. Orthology between At5g04810 and Os04g58780 is supported by the facts that they are mutual best hits in whole-proteome BLAST comparisons between rice and Arabidopsis, that they cluster in a phylogram that includes the most closely related rice and Arabidopsis proteins (see http://plantrbp.uoregon.edu/), and that their intron number and position are conserved. BLASTN queries of public maize sequence data with the nucleotide sequence of Os04g58780 detected a putative maize ortholog that also encodes a PPR tract and an RRM domain; this sequence detects the initial query, Os0458780, as its top hit in BLASTN searches of the rice genome, supporting their orthology. No evidence for other closely related paralogs exists among available maize sequence data. We named this maize gene ppr4.

A ppr4 cDNA clone obtained from the Z. mays full-length cDNA collection (http://www.genome.arizona.edu/orders/) was sequenced in its entirety. The sequence, which includes the complete open reading frame, has been deposited in GenBank (accession number DQ508419). An alignment of maize PPR4 and its apparent rice and Arabidopsis orthologs is shown in Figure 1 . All three proteins have a predicted chloroplast transit peptide at their N terminus, followed by an RRM domain and 16 tandem PPR motifs (Figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Alignment of PPR4 Orthologs in Maize (Zm), Rice (Os), and Arabidopsis (At).

TargetP (Emanuelsson et al., 2000) and Predotar (Small et al., 2004) algorithms both predict that all three proteins are targeted to the chloroplast. The insertion site of the Mu9 element in ppr4-1 mutants is indicated by an arrowhead. The RRM domain is underlined, with the thicker lines denoting its two core RNP motifs. PPR repeats are indicated below the sequences by double-headed arrows. At PPR4 corresponds to Arabidopsis Genome Initiative locus At5g04810; Os PPR4 corresponds to The Institute for Genomic Research locus Os04g58780.

Recovery of Loss-of-Function ppr4 Alleles

To determine the function of PPR4, we initiated a reverse-genetic approach in both Arabidopsis and maize. We obtained three T-DNA insertions in At5g04810 (At PPR4) from the ABRC (Alonso et al., 2003). The positions of the T-DNA inserts in all three lines were confirmed by PCR analysis and DNA sequencing (see Supplemental Figure 1 online). After self-pollination, seed from each of the three insertion lines was germinated on Murashige and Skoog medium, and >20 of the germinating progeny from each line were analyzed by PCR to detect the T-DNA insertion. Approximately two-thirds of the germinating plants were heterozygous for the expected insert in each line, but no homozygous mutant plants were detected. Furthermore, among ∼250 seeds analyzed from plants heterozygous for each of the three insertions, ∼25% were shriveled and aborted (data not shown). The fact that three independent insertions in At PPR4 are genetically linked to an abortive-seed phenotype strongly suggests that this gene is essential for embryogenesis in Arabidopsis. Disruption of several other genes for chloroplast-targeted PPR proteins has also been reported to cause embryo lethality in Arabidopsis (Lurin et al., 2004; Cushing et al., 2005).

Maize ppr4 mutants were obtained through a PCR-based screen of our collection of Mu transposon–induced nonphotosynthetic mutants (http://chloroplast.uoregon.edu/) (Stern et al., 2004) (see Methods). Two mutant alleles were recovered: the ppr4-1 allele has a Mu9 element in the protein-coding region, near the 3′ end of exon 1 (Figures 1 and 2A ); and the ppr4-2 allele has a Mu1 element inserted in the first intron, 94 bp downstream of the 5′ splice junction (Figure 2A). In contrast with homozygous At PPR4 mutants, the homozygous maize mutants germinated and gave rise to chlorophyll-deficient seedlings: ppr4-1/ppr4-1 seedlings are albino, whereas ppr4-2/ppr4-2 seedlings are slightly pale green (Figure 2B); both are seedling lethal at ∼3 weeks after germination, as is typical of nonphotosynthetic maize mutants. Crosses between ppr4-1/+ and ppr4-2/+ plants yielded mutant plants of an intermediate phenotype (Figure 2B) at the expected 1:3 ratio of mutant-to-normal plants, confirming that the chlorophyll-deficient phenotypes are due to the insertions in the ppr4 gene. Chloroplasts could be isolated from ppr4-1/ppr4-2 plants but not from the albino ppr4-1/ppr4-1 plants, so ppr4-1/ppr4-2 material was used in experiments described below that required chloroplast isolation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Mu-Induced ppr4 Mutants.

(A) Positions of the Mu insertions in ppr4-1 and ppr4-2. The DNA sequence flanking the Mu insertions is shown, with the intron-exon border indicated by a dotted line. Residue numbers with respect to the start codon in the genomic sequence are indicated. A 10-nucleotide duplication flanking the ppr4-2 insertion site is underlined; this sequence is present only once in the progenitor allele. The type of Mu element at each site (i.e., Mu1 or Mu9) was deduced from polymorphisms in the terminal inverted repeat sequences. A schematic representation of the ppr4 gene is shown below. Boxes represent exons, and lines represent introns (drawn to scale). Mu element insertion sites are marked. Arrows illustrate the primers used for RT-PCR analysis of ppr4 mRNA shown in (C).

(B) Phenotypes of ppr4 mutant seedlings. Plants were grown for 2 weeks in soil in a greenhouse.

(C) Decreased abundance of ppr4 mRNA in ppr4 mutants. ppr4 mRNA was assayed by RT-PCR using primers in exons 1 and 2 (primer binding sites are diagrammed in [A]). Conditions were chosen to yield a signal that is roughly proportional to the amount of input RNA. RNA from an hcf7 mutant, whose phenotype resembles that of ppr4-1/ppr4-2 mutants (see Results), was analyzed to address the possibility that ppr4 expression might be regulated in response to the status of plastid development/physiology. The bottom panel shows RT-PCR amplification of tubulin mRNA in the same RNA samples.

The accumulation of ppr4 mRNA in ppr4-1/ppr4-2 mutant leaf tissue was compared with that in normal siblings and in leaves from the mutant hcf7, which has a similar chlorophyll-deficient phenotype and a related chloroplast gene expression defect (see below). Primers were designed to the first and second exons, such that amplification products from spliced mRNA and from genomic DNA could be distinguished. As expected, the Mu insertions cause a substantial reduction in the level of ppr4 mRNA (Figure 2C). Residual ppr4 mRNA in ppr4-1/ppr4-2 plants is consistent with the weak phenotype of the ppr4-2 allele, which has a Mu insertion within an intron that could be spliced out of the pre-mRNA.

PPR4 Resides in the Chloroplast Stroma

Antibody was raised to a recombinant PPR4 fragment and used to probe immunoblots of total leaf extract and subcellular fractions. A protein of the expected size (∼100 kD) was strongly enriched in chloroplast extract, with respect to its concentration in total leaf extract where it was barely detectable (Figure 3A ). Analysis of chloroplast subfractions showed that this protein localizes to the stromal fraction (Figure 3A). The abundance of this protein was strongly reduced in stroma from ppr4-1/ppr4-2 mutant chloroplasts (Figure 3B), confirming that it is PPR4. These results show that PPR4 resides in the chloroplast stroma.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

PPR4 Resides in the Chloroplast Stroma.

Immunoblots were probed with affinity-purified anti-PPR4 antibody. The bottom panels show the same blots stained with Ponceau S prior to probing. RbcL, large subunit of Rubisco.

(A) Immunoblot of leaf and subcellular fractions. The total chloroplast fraction (Cp) and the thylakoid+envelope (thy+env) and stroma+envelope (str+env) subfractions were from the fractionated chloroplast preparation described and verified with marker proteins by Williams and Barkan (2003) and are derived from the same quantity of starting chloroplasts. The stromal fraction (str) and its dilution (str 0.3x) are from a separate preparation.

(B) Verification that the 100-kD protein detected by the PPR4 antibody is PPR4. An equal quantity of stromal protein from wild-type and ppr4-1/ppr4-2 chloroplasts (1x) or dilutions thereof (0.3x and 0.1x) were analyzed. The ppr4 mutant material is from a complementation cross between a null and weak allele and retains some ppr4 expression. Antisera from two different rabbits yielded similar results (data not shown).

To determine whether PPR4 is stably bound to other macromolecules in vivo, chloroplast stroma was fractionated in sucrose gradients, and the distribution of PPR4 in the gradient was determined by immunoblotting (Figure 4 , top panel). PPR4 was found primarily in fractions near the top of the gradient; this peak may represent monomeric PPR4. A small fraction of PPR4 was found in high molecular weight complexes that are larger than ribulose-1,5-bis-phosphate carboxylase/oxygenase (Rubisco) (marked with a bar in Figure 4), and some PPR4 was recovered in pelleted material at the bottom of the gradient (last lane). Treatment of stroma with RNase A prior to sedimentation reduced the recovery of PPR4 in the high molecular weight fractions but did not shift its position in the upper fractions (Figure 4, bottom panel). These results indicate that PPR4 is found predominantly in small complexes or as a monomer and that a subpopulation is associated with high molecular weight complexes that include RNA.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Sucrose Gradient Analysis of PPR4-Containing Particles in Stromal Extract.

Stromal extract was treated with RNase A (RNase-treated) or incubated under the same conditions in the absence of RNase (mock-treated) and sedimented through sucrose gradients. An equal volume of each gradient fraction was analyzed on immunoblots probed with PPR4 antibody. Images of the blots stained with Ponceau S are shown below to illustrate the similar fractionation of Rubisco and other abundant proteins in the two gradients. The last lane in each panel contains material that pelleted in the gradients. The position of ribosomes was determined by the pattern of Ponceau S–stained bands and by the appearance of rRNAs in these fractions (data not shown). A high molecular weight, RNase-sensitive peak of PPR4 is indicated with a solid bar.

PPR4 Is Required for the Accumulation of Plastid Ribosomes

The ppr4-1 allele conditions an ivory leaf pigmentation similar to that seen in previously described mutants that lack plastid ribosomes (e.g., ppr2, crs2-1, and iojap) (Walbot and Coe, 1979; Jenkins et al., 1997; Williams and Barkan, 2003), suggesting that ppr4 mutants might have plastid ribosome defects. In fact, RNA gel blots of leaf RNA from both ivory ppr4-1/ppr4-1 and pale green ppr4-2/ppr4-1 mutants revealed defects in plastid rRNAs (Figure 5A ). Mature plastid 23S and 16S rRNAs were undetectable in ppr4-1/ppr4-1 mutants, as in ppr2, iojap, and crs2-1 mutants (Williams and Barkan, 2003). Mature plastid rRNAs are also reduced in pale green ppr4-2/ppr4-1 mutants, albeit to a lesser extent. Interestingly, these mutants accumulate increased levels of a putative precursor to 16S rRNA (Figure 5A), a phenotype similar to that described previously for maize hcf7 mutants (Barkan, 1993); 16S rRNA from hcf7 is shown in Figure 5A for comparison. Accumulation of 16S rRNA precursors in bacteria is indicative of defective assembly of the small ribosomal subunit (Dahlberg et al., 1978; Srivastava and Schlessinger, 1989; Srivastava and Schlessinger, 1990). These results suggested that PPR4 functions in a process that is needed for the biogenesis of small ribosomal subunits in plastids.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Loss of Plastid rRNAs and Plastid-Encoded Proteins in ppr4 Mutants.

(A) RNA gel blot hybridizations showing plastid rRNA defects in ppr4 mutants. Five micrograms of total leaf RNA was analyzed by hybridization to probes for the plastid 16S or 23S rRNA. The same filters stained with methylene blue are shown below: bands corresponding to cytosolic rRNAs (25S and 18S) and plastid rRNAs (16S and 23S*) are marked. 23S* is a breakdown product of the plastid 23S rRNA. The arrow indicates the 16S rRNA precursor that overaccumulates in ppr4 and hcf7 mutants.

(B) Immunoblot analysis of photosynthetic enzyme accumulation in ppr4 mutants. Total leaf proteins of ppr4-1/ppr4-2 mutants were analyzed by probing immunoblots with antisera to representative subunits of photosystem I (PsaD), photosystem II (D1), ATP synthase (AtpA), and the cytochrome b₆f complex (PetD). The same filter was stained with Ponceau S to visualize total proteins (bottom); the large subunit of Rubisco (RbcL) is indicated on the stained blot.

As anticipated for plants with a plastid ribosome deficiency, ppr4-2/ppr4-1 mutants accumulate reduced levels of all photosynthetic enzyme complexes that include plastid-encoded subunits (Rubisco, photosystem II, cytochrome b₆f, photosystem I, and ATP synthase) (Figure 5B). In the more severe ppr4-1/ppr4-1 mutants, these proteins were not detectable (data not shown). Thus, ppr4 mutants have a global defect in plastid gene expression.

PPR4 Is Associated in Vivo with the Group II Intron from the trans-Spliced Plastid rps12 Pre-mRNA

Previously studied PPR proteins influence the processing or translation of specific organellar RNAs. Therefore, it seemed plausible that the plastid ribosome deficiency in ppr4 mutants was due to an underlying defect in the expression of just one or several components of the plastid translation machinery. For example, a primary defect in the maturation of the 16S rRNA or of an mRNA encoding a protein in the small ribosomal subunit would be anticipated to result in the ribosome deficiency and aberrant 16S rRNA processing observed in ppr4 mutants.

To gain insight into the direct target(s) of PPR4, we used the RIP-chip assay (Schmitz-Linneweber et al., 2005) to identify plastid RNAs that are associated with PPR4 in chloroplast extract. PPR4 was immunoprecipitated from stromal extract prepared from wild-type chloroplasts, and RNA purified from the immunoprecipitation pellet and supernatant was labeled with red- and green-fluorescing dyes, respectively. The pellet and supernatant RNAs were competitively hybridized to a chloroplast genome tiling microarray. The ratio of red to green fluorescence for each spot reflects the degree to which the corresponding RNA sequence is enriched in the immunoprecipitation pellet. As a negative control, immunoprecipitations were also performed with stromal extract from ppr4-2/ppr4-1 mutant chloroplasts, which accumulate reduced levels of PPR4 protein (Figure 3B). The null ppr4-1/ppr4-1 mutants could not be used for this purpose because we have been unable to obtain sufficient yields of stroma from albino plants for RIP-chip analysis. PPR4-associated RNAs were identified as RNAs that were significantly more enriched in immunoprecipitations from wild-type stroma than in immunoprecipitations from ppr4 mutant stroma.

Two replicate experiments were performed with wild-type stroma and two with ppr4-2/ppr4-1 mutant stroma. The replicate experiments were performed with antibodies from two different immunized rabbits. Data from the four assays were normalized and used to calculate median enrichment ratios (i.e., ratio of red to green fluorescence) for each DNA fragment among the replicate spots (five per array) and replicate experiments (see Supplemental Tables 1 and 2 and Supplemental Figure 2 online). To visualize sequences that were preferentially enriched from the wild-type extract, the difference in median enrichment ratio for each DNA fragment between the wild-type and mutant experiments was plotted as a function of chromosomal position (Figure 6A ). Two prominent peaks of differential enrichment were observed, both of which mapped to DNA fragments encoding the first intron in the rps12 gene (rps12-int1). rps12-int1 is a group II intron that is transcribed in two pieces from distinct chromosomal loci and then spliced in trans (Koller et al., 1987; Hildebrand et al., 1988) (see map in Figure 6D). Array elements from both loci stood out as red fluorescing spots when probed with immunoprecipitated RNA from wild-type stroma but not when probed with immunoprecipitated RNA from ppr4 mutant stroma (Figure 6B; data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Identification of RNAs Associated with PPR4 in Chloroplast Stroma.

(A) Summary of RIP-chip data. The log₂-transformed enrichment ratios (F635:F532) were normalized between two assays involving wild-type stroma and two control assays with ppr4-1/ppr4-2 mutant stroma. The median normalized values for replicate spots from the mutant data were subtracted from those from wild-type data and plotted according to fragment number. Fragments are numbered according to chromosomal position. The data used to generate this figure are provided in Supplemental Tables 1 and 2 online and have been submitted to MIAME Express (accession number E-MEXP-716). The data from each of the four assays are plotted separately in Supplemental Figure 2 online to illustrate the reproducibility of the results.

(B) Excerpts of merged fluorescent images from representative PPR4 RIP-chip experiments involving wild-type and ppr4-1/ppr4-2 mutant stroma. Fragment names are indicated above, and fragment numbers are indicated below. Each DNA fragment is represented five times on the array in clusters of two and three spots. These excerpts show three-spot clusters. AI745002 is a cytosolic cDNA used as a negative control.

(C) Slot blot hybridizations to verify the coimmunoprecipitation of rps12 RNAs with PPR4. One-third of the RNA recovered from each immunoprecipitation pellet (P) and one-sixth of the RNA recovered from each supernatant (S) were applied to replicate slot blots and hybridized with probes A through D (see [D]) or to atpF intron DNA. An immunoprecipitation with antibody to the atpF splicing factor CRS1 was used as a negative control.

(D) Schematic map of the split rps12 gene and its flanking genes (not drawn to scale). Probes used for slot blot hybridizations are indicated with thick bars and are labeled with letters. DNA fragments on the array are indicated with thin bars and are labeled with their fragment number.

Table 1 summarizes the data for the DNA fragments whose enrichment from wild-type stroma ranked in the top 10% among the 248 fragments on the array; fragments in the table are ordered according to the degree to which they are differentially enriched from wild-type versus ppr4 mutant extract. A t test was used to evaluate the null hypothesis that sequences corresponding to each fragment were enriched to the same extent from wild-type and ppr4 mutant stroma. A P value below 3 × 10⁻⁴ indicates significant enrichment, as this cutoff would be anticipated to yield <0.1 false positives among the 248 distinct DNA fragments (each in five replicate spots) on the array. Coenrichment of adjacent fragments further increases confidence that the corresponding RNAs were enriched and was observed in both regions encoding rps12-int1 (fragments 193 and 194 and fragments 128 to 131). Furthermore, all five fragments that include sequences from rps12-int1 (fragments 130, 131, 132, 193, and 194) are among the top eight ranking fragments, and all five show highly significant preferential enrichment from wild-type extract (P values <10⁻⁵) (Table 1). Many of the other high-ranking fragments are from sequences adjacent to rps12-int1 (e.g., 129, 190, 195, and 196) and were presumably enriched due to their presence on the same RNA molecules as rps12-int1. Therefore, the RIP-chip data provide strong evidence that PPR4 interacts with rps12 pre-mRNA in vivo, as exon 1, both fragments of intron 1, and exon 2 showed highly significant PPR4-dependent enrichment during PPR4 immunoprecipitations.

To verify the RIP-chip data and to further pinpoint the RNA sequences with which PPR4 is associated, RNAs that coimmunoprecipitate with PPR4 were analyzed by slot blot hybridization. As a control, parallel assays were performed with antibody to the chloroplast splicing factor CRS1, which is known to be associated specifically with the atpF intron in vivo (Till et al., 2001; Ostheimer et al., 2003). RNAs purified from the immunoprecipitation pellets and supernatants were applied through a slot blot manifold to nylon membrane. Duplicate slot blots were probed with radiolabeled PCR products representing different segments of the rps12 gene (see Figure 6D for probe positions) or the atpF intron. All of the rps12 sequences were strongly enriched by immunoprecipitation with PPR4 antibody but not by immunoprecipitation with CRS1 antibody (Figure 6C). Specificity of these results is further shown by the fact that atpF intron RNA coimmunoprecipitated with CRS1 but not with PPR4. Furthermore, the two rps12-int1 fragments were markedly more enriched than the flanking exons in the PPR4 immunoprecipitation pellets (compare probes B and C to probes A and D), as suggested by the RIP-chip data. These results confirm that PPR4 is associated with rps12 pre-mRNA sequences in chloroplast extract and further point to the trans-spliced intron 1 as harboring one or more site with which PPR4 is associated.

The RIP-chip data did not provide strong evidence for other PPR4 interaction sites (see Table 1) but also did not eliminate the possibility that PPR4 might be associated with additional sequences. For example, the second group II intron in rps12 showed significant PPR4-dependent enrichment, albeit with an enrichment ratio ∼16-fold lower than that for rps12-int1 (see fragment 190 in Table 1); this modest enrichment could result from the fact that intron 2 will be coprecipitated with intron 1 due to its location on the same pre-mRNA molecule, or it could reflect a distinct PPR4 binding site. The data also hinted at the possibility that fragments 139 and 142 (petB), 169 (rpl2), 243/244 (ndhA), and 156 (infA) might be enriched in a PPR4-dependent fashion; these are among the top-ranking 20% of fragments in terms of their enrichment from wild-type extract and may be differentially enriched from wild-type versus ppr4 mutant extract (P values <0.01). However, the phenotypic analyses of ppr4 mutants described below suggest that these results do not reflect meaningful physiological interactions.

PPR4 Is Required for rps12 trans-Splicing

The association between PPR4 and the trans-spliced rps12 intron suggested that PPR4 might function in rps12 trans-splicing. A defect of this nature could account for the plastid ribosome deficiency and aberrant 16S rRNA metabolism observed in ppr4 mutants. We therefore analyzed the accumulation of rps12 transcripts in ppr4 mutants.

Mutants with severe plastid ribosome deficiencies are albino, lack all chloroplast-encoded proteins, and exhibit a variety of stereotypical defects in chloroplast RNA metabolism (Han et al., 1993, 1995; Hess et al., 1993; Jenkins et al., 1997; Williams and Barkan, 2003). As a consequence, it can be difficult to identify the underlying molecular lesion in albino mutants through analysis of plastid proteins and transcript patterns (for an example, see Williams and Barkan, 2003). The severe ppr4-1/ppr4-1 mutants are albino and lack plastid rRNAs (Figures 2B and 5A) and so were anticipated to share the stereotypical plastid transcript profiles with other mutants of this type. To reduce these pleiotropic effects, we analyzed RNA from the ppr4-1/ppr4-2 progeny of complementation crosses, as these accumulate some chlorophyll, plastid rRNAs, and chloroplast-encoded proteins (Figures 2B and 5), albeit to reduced levels. To control for effects resulting from aberrant assembly of 30S ribosomal subunits, RNA from hcf7 mutants was analyzed in parallel; hcf7 mutants resemble ppr4-1/ppr4-2 mutants in having a global plastid translation defect, aberrant 16S rRNA processing, and decreased chlorophyll content (Barkan, 1993).

RNAs from the wild type, ppr4-1/ppr4-2 mutants, and hcf7 mutants were analyzed by RNA gel blot hybridization using a series of probes derived from the rps12 gene (Figure 7 ). The smallest transcript detected with both exon 1 and exon 2 probes (probes A and D, ∼1200 nucleotides) was strongly reduced in ppr4 mutants but not in hcf7 mutants (arrows in Figure 7). This transcript was not detected with the intron probes and thus represents a trans-spliced product of the two precursor RNAs. In addition, a putative unspliced precursor (a transcript of ∼2500 nucleotides that hybridized to probes C and D) accumulated to increased levels in ppr4 mutants (asterisk in Figure 7), although this transcript may be at increased levels in the hcf7 mutant as well. Together, these results suggest that rps12 trans-splicing is disrupted in ppr4 mutants, as the mutants accumulate reduced levels of the spliced 1200-nucleotide RNA and increased levels of a putative unspliced precursor.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

RNA Gel Blot Hybridizations of rps12 Transcripts in ppr4 Mutants.

Top panel: Total leaf RNA (5 μg/lane) was fractionated on a single agarose gel and transferred to a nylon membrane. The membrane was cut into four strips, which were hybridized with the indicated probes. Two different ppr4 mutant seedlings (genotype ppr4-1/ppr4-2) and two different normal siblings (wt) were analyzed to provide replicate results. Transcripts that are missing in ppr4 mutants are marked with an arrow. A transcript that accumulates to higher levels in ppr4 mutants than in hcf7 mutants is marked with an asterisk. Middle panel: rRNAs on the same filters were detected by staining with methylene blue. Bottom panel: Schematic of the rps12 gene and probes used for the RNA gel blots.

To more clearly assess splicing of rps12-int1 in ppr4 mutants, the ratio of spliced versus unspliced RNA was quantified with a poisoned primer extension assay. In this assay, a radiolabeled oligonucleotide complementary to the second exon of rps12 was used to prime a reverse transcription reaction in the presence of a dideoxy nucleotide that terminates after different distances on spliced and unspliced RNA templates (see diagram in Figure 8 ). With this assay, all transcripts in which rps12-int1 is spliced are detected as one band, and all those in which this intron is retained are detected as a second band. As suggested by the RNA gel blot hybridizations, this assay showed that the vast majority of rps12 transcripts in ppr4-1/ppr4-2 mutants retain rps12-int1, whereas the ratio of spliced to unspliced RNA is unaltered in hcf7 mutants (Figure 8). These results indicate that PPR4 functions either in the trans-splicing of rps12-int1 or in stabilizing the spliced RNA. A role in trans-splicing rather than mRNA stabilization is supported by the increased accumulation of an unspliced precursor in ppr4 mutants (Figure 7) and by the coimmunoprecipitation data suggesting that PPR4 is associated with intron sequences rather than exon sequences.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

Poisoned Primer Extension Analysis Demonstrating Loss of trans-Spliced rps12 RNA in ppr4 Mutants.

Top panel: Primer extension products were separated on a denaturing polyacrylamide gel. The radiolabeled primer and the extension products from spliced and unspliced RNAs are indicated. As a control, RNA transcribed in vitro from a spliced rps12 cDNA was used as a template (cDNA). The two independent ppr4 mutant samples have the genotype ppr4-1/ppr4-2. Bottom panel: Predicted products of poisoned primer extension reactions. Exon sequences are shaded in gray, and the primer sequence is underlined. Dideoxy CTP included in the extension reaction terminates reverse transcription at the first encountered G residue in the template, yielding 24- and 29-nucleotide (nt) extension products on spliced (S) and unspliced (U) RNA, respectively.

Several previously identified chloroplast group II intron splicing factors are associated with and required for the splicing of multiple introns (Jenkins et al., 1997; Ostheimer et al., 2003). Although the RIP-chip data for PPR4 did not provide strong evidence for its association with introns other than rps12-int1, they suggested the possibility of a weak association with rps12-int2, with the intron-containing petB, rpl2, and ndhA RNAs, and with the intronless infA and rps18 RNAs. To address whether PPR4 influences the metabolism of these RNAs, their processing was assayed by RNA gel blot hybridization and poisoned primer extension (see Supplemental Figure 3 online). Minor changes in the transcript populations from some of these genes were observed, but these changes were similar to those in hcf7 mutants and so could be secondary effects of the ribosome deficiency. Therefore, the RIP-chip data together with the ppr4 mutant phenotype suggest that PPR4 functions specifically in the trans-splicing of the rps12 pre-mRNA and that this single function underlies the plastid ribosome deficiency and 16S rRNA processing defect in ppr4 mutants.

PPR Proteins as Group II Intron Splicing Factors

Group II introns are large ribozymes that are defined by a conserved predicted secondary structure consisting of six helical domains, characteristic interdomain interactions, and small regions of conserved primary sequence (Michel et al., 1989; Michel and Ferat, 1995). Some group II introns, including rps12-int1, have become fragmented during the course of evolution, such that they are transcribed in several pieces. During trans-splicing, the intron fragments assemble via noncovalent interactions and then splice out of the precursor RNA. trans-splicing has been described for several introns of plant mitochondria and algal chloroplasts (reviewed in Barkan, 2004), but in angiosperm chloroplasts, it is restricted to the rps12 gene.

Both cis- and trans-spliced group II introns require protein cofactors to splice efficiently in vivo. These proteins fall into two general classes: conserved maturase proteins that are encoded within some group II introns, and diverse host-encoded proteins that were recruited from the host genomes. Three host-encoded protein complexes have been described that assemble with different subsets of the cis-spliced group II introns in maize chloroplasts and that facilitate their splicing in vivo (Jenkins et al., 1997; Jenkins and Barkan, 2001; Till et al., 2001; Ostheimer et al., 2003). Each of these complexes contains a member of a plant-specific protein family harboring a novel RNA binding domain called the CRM domain. The Arabidopsis PPR protein HCF152 is the only PPR protein previously suggested to function in group II intron splicing, as spliced chloroplast petB mRNA fails to accumulate in hcf152 mutants (Meierhoff et al., 2003). However, excised petB intron accumulates normally in hcf152 mutants, suggesting that HCF152 may function not in splicing but in the stabilization of the spliced petB mRNA.

It seems likely that the complement of proteins involved in group II trans-splicing is more complex than that involved in cis-splicing, as additional proteins may be required to assemble intron fragments. In fact, genetic studies indicate that at least 14 nuclear loci are required for the removal of the two trans-spliced introns in the psaA gene in Chlamydomonas chloroplasts (Goldschmidt-Clermont et al., 1990). Three of these genes have been cloned (Goldschmidt-Clermont et al., 1990; Rivier et al., 2001; Merendino et al., 2006); these proteins are not closely related to proteins in vascular plants, but one of them, Raa1, includes 38 amino acid repeats that resemble the 35–amino acid repeats of the canonical PPR motif (Merendino et al., 2006).

Data presented here show that PPR4 is required for the accumulation of the trans-spliced rps12 intron in maize chloroplasts and that it is associated with the rps12 intron RNA in vivo. These results, together with the fact that other RRM and PPR domains have been shown to bind RNA, strongly suggest that PPR4 facilitates rps12 trans-splicing through direct interaction with the intron RNA. It is noteworthy that a maturase protein is encoded by the trans-spliced rps12 intron in the ancient charophytes Staurastrum and Chaetosphoridium (Turmel et al., 2002, 2005). Charophytes are among the closest algal sister groups of mosses, hornworts, liverworts, and vascular plants, where this maturase open reading frame was lost. It seems plausible that PPR4 might have replaced the ancient maturase proteins in rps12 trans-splicing.

Previously, we showed that the CAF2/CRS2 complex is required for rps12 trans-splicing and that it coimmunoprecipitates with rps12 intron RNA (Ostheimer et al., 2003). Unlike PPR4, however, CAF2/CRS2 also facilitates the cis-splicing of several chloroplast group II introns. Thus, PPR4 may function in a process that is specific for trans-splicing (e.g., assembling the intron fragments or stabilizing their association), with CAF2/CRS2 promoting the subsequent splicing reaction. We have performed coimmunoprecipitation experiments to assess whether CAF2/CRS2 and PPR4 are bound simultaneously to the intron (see Supplemental Figure 4 online; data not shown), but the results have been ambiguous. Because only a small fraction of the PPR4 (Figure 4) and CAF2/CRS2 (Ostheimer et al., 2003) in the stroma are bound to rps12-int1, the failure to detect strong coimmunoprecipitation does not eliminate the possibility that they can be found in the same complex. In any case, both the 5′ and 3′ intron segments were similarly enriched by immunoprecipitation with PPR4 (Figure 6), indicating that PPR4 either associates independently with both fragments and/or remains associated after the intron fragments bind one another.

Function of PPR4 in Ribosome Biogenesis

The defect in rps12 splicing can account for the albino phenotype and plastid ribosome deficiency in ppr4 mutants. rps12 codes for ribosomal protein S12, found in the 30S ribosomal subunit. Escherichia coli S12 is essential for cell viability (http://www.shigen.nig.ac.jp/ecoli/pec/index.jsp). It interacts with the penultimate stem loop of the 16S rRNA (Cukras et al., 2003), positions the small and large subunits of the ribosome during translation elongation, and is involved in the maintenance of translational fidelity. Thus, the lack of mature rps12 mRNA in ppr4 mutants is expected to severely affect plastid ribosome function. This prediction is borne out by the global loss of chloroplast translation products and plastid rRNAs in ppr4 mutants. Plants harboring the weak ppr4-2 allele accumulate increased levels of pre-16S rRNA. We believe this to be a consequence of decreased rps12 expression rather than a direct effect of reduced PPR4 function for the following reasons. First, 16S rRNA was not detectably enriched in the PPR4 RIP-chip assays, suggesting that PPR4 does not interact with pre-16S rRNA or with assembling 30S ribosomes. Furthermore, reduced S12 synthesis is predicted to lead to increased accumulation of pre-16S rRNA, based on studies of ribosome assembly and rRNA processing in E. coli. S12 is added to E. coli ribosomes late in the ribosome assembly pathway (Mizushima and Nomura, 1970; Culver and Noller, 1999), so reduced S12 synthesis is anticipated to result in the accumulation of late intermediates in 30S subunit assembly. Such assembly intermediates harbor incompletely processed rRNAs, as completed ribosome assembly is a prerequisite for the final rRNA processing steps in E. coli (Srivastava and Schlessinger, 1990). Results presented here support a model in which PPR4 is directly involved in S12 synthesis due to its role in rps12 trans-splicing, S12 is necessary for completed 30S subunit assembly, and completed 30S subunit maturation is necessary for the final processing steps of rRNA in chloroplasts, as in bacteria. Previously, we described another mutant, hcf7, with a 16S rRNA processing defect similar to that observed in weak ppr4 mutants (Barkan, 1993). The hcf7 gene product has not been identified, although the results of complementation tests indicate that hcf7 and ppr4 are not allelic (data not shown). In light of the findings presented here, it seems possible that HCF7 is necessary for the expression of a late assembling protein of the small ribosomal subunit in chloroplasts.

T-DNA insertions in the Arabidopsis ppr4 ortholog cause an embryo-lethal phenotype (see Results), as do insertions in the Arabidopsis ortholog of maize ppr2 (At3g06430; see EMB2750 at http://www.seedgenes.org). By contrast, exonic insertions in maize ppr2 and ppr4 condition albino seedlings that lack plastid ribosomes (Williams and Barkan, 2003; this article). In fact, embryo lethality is commonly observed for mutations in Arabidopsis genes encoding chloroplast-targeted PPR proteins (Tzafrir et al., 2004; Cushing et al., 2005). We suspect that embryo lethality in many of these cases is the result of a severe defect in the plastid translation machinery. Strong plastid translation defects seem to impact maize and Arabidopsis very differently, as mutations in maize that condition albino seedlings with severe plastid ribosome deficiencies are commonly observed (e.g., ppr2, ppr4, crs2, iojap, caf1, and caf2) (Walbot and Coe, 1979; Jenkins et al., 1997; Ostheimer et al., 2003; Williams and Barkan, 2003), but analogous Arabidopsis mutants have not been reported. Furthermore, in Arabidopsis, embryo lethality would be anticipated to result from the absence of plastid translation because the Arabidopsis chloroplast genome harbors three open reading frames, accD, ycf1, and ycf2, that are essential for cellular viability (Drescher et al., 2000; Kode et al., 2005). These open reading frames are lacking in the maize chloroplast genome (Maier et al., 1995), and this difference in plastid gene content may account for the seedling viability of maize mutants lacking plastid ribosomes. Therefore, maize is a particularly useful organism in which to study nuclear genes like ppr4, crs2, caf1, and caf2 that are required for the expression of plastid genes encoding components of the translation machinery.

RIP-Chip as a Tool for Facilitating the Genetic Analysis of Chloroplast Gene Expression

RIP-chip data complement genetic analysis of RNA binding proteins by helping to distinguish direct from indirect effects of a mutation. RIP-chip data are also useful to guide analysis of a mutant when the phenotype is uninformative due to genetic redundancy or pleiotropy. For example, identification of specific plastid gene expression defects that underlie a nonphotosynthetic mutant phenotype can be straightforward when a subset of chloroplast-encoded proteins (e.g., a single enzyme complex) accumulate to reduced levels. In such instances, systematic analysis of the expression of genes encoding the missing proteins can identify RNA metabolism or translation defects responsible for the protein losses. However, for mutations like those in ppr4, the global protein deficiencies are not very informative, as they suggest a defect in the expression of one or more of the large number of nuclear and plastid genes that contribute to basal chloroplast gene expression (tRNAs, ribosomal proteins, etc.). It is labor intensive to assess the expression of each candidate gene in search of the underlying molecular lesion. This problem is exacerbated in albino mutants lacking plastid ribosomes because they exhibit stereotypical defects in chloroplast RNA metabolism (Han et al., 1992, 1995; Hess et al., 1993; Jenkins et al., 1997; Williams and Barkan, 2003). These aberrant transcript patterns are caused in part by the loss of the plastid-encoded RNA polymerase, as the nuclear-encoded polymerase activity recognizes a distinct set of promoters (Hajdukiewicz et al., 1997).

Because of these factors, RIP-chip data are particularly valuable for guiding the phenotypic analysis of mutations that cause a global defect in chloroplast translation. RIP-chip data focus attention on a small set of candidate substrate RNAs, obviating the need to examine in detail the expression of all of the many chloroplast genes that contribute to the gene expression machinery. In the analysis of ppr4 functions, the defect in rps12 mRNA metabolism may ultimately have been discovered in ppr4 mutants without the use of RIP-chip, but the RIP-chip data led much more quickly to the discovery of this defect. Thus, the combined use of RIP-chip and genetic analysis in maize is expected to enhance progress in understanding the functions of the large set of nuclear genes in plants that function in the expression of components of the plastid translation apparatus.

Nucleotide Sequence Analysis of ppr4

The nucleotide sequence of the rice (Oryza sativa) gene Os04g58780 was used to query the maize (Zea mays) sequence at the PlantGDB database (http://www.plantgdb.org/PlantGDB-cgi/blast/PlantGDBblast). The single maize contig identified exhibits high identity to the rice sequence at the amino acid level (83%), had the same intron/exon structure, and detected the rice gene as its top hit when used to query the rice genome. A ppr4 cDNA was obtained from the University of Arizona's collection (accession number DR967823.1), and its nucleotide sequence was determined. The cDNA sequence is deposited in GenBank under accession number DQ508419.

Plant Material

The two ppr4 mutant alleles were identified in a reverse-genetic screen of a collection of ∼2300 Mu-induced nonphotosynthetic maize mutants (http://chloroplast.uoregon.edu/). The PCR-based screen was performed with a ppr4-specific primer (5′-ACTCTAGCCCAACCTTTAACGTGGTA-3′) in conjunction with a Mu terminal inverted repeat primer (5′-AGAGAAGCCAACGCCAWCGCCTCYATTTCGTC-3′) using a method analogous to that described previously for ppr2 (Williams and Barkan, 2003). Allelism between ppr4-1 and ppr4-2 was tested by intercrossing the phenotypically wild-type siblings of mutant plants (which include both +/+ and +/− genotypes). Altogether, 101 independent crosses were performed, 40 of which yielded ∼25% pale green seedlings. These numbers fit the expectation for Mendelian segregation of single, allelic recessive mutations in the two lines that cause a chlorophyll deficiency. The phenotype of the mutant noncomplementing progeny was intermediate between that conditioned by the two parental alleles. Taken together, these results confirmed that the mutant phenotypes segregating in the ppr4-1 and ppr4-2 lines are due to the Mu insertions in ppr4.

Maize hcf7 mutants were used as controls in this work: hcf7 mutants are pale green and show reduced polysome assembly and aberrant metabolism of 16S rRNA (Barkan, 1993). The inbred line B73 (Pioneer HiBred) was used as the source of wild-type tissue for RIP-chip and chloroplast fractionation experiments. Seedlings were grown in soil in a growth chamber under a 16-h-light/8-h-dark cycle at 26°C and harvested between 7 and 10 d after planting.

Nucleic Acid Extraction and Analysis

Seedling leaf DNA for PCR amplification was isolated using plant DNAzol reagent (Invitrogen) according to the manufacturer's protocol. Leaf RNA was isolated using Tri reagent (Molecular Research Center). RNA gel blot hybridizations were performed as described previously with 5 μg of total leaf RNA (Barkan et al., 1994). Hybridization was performed at 65°C in Church hybridization buffer (Sambrook et al., 1989); blots were washed at 65°C in 0.2× SSC and 0.2% SDS. The following DNA fragments were used as probes for chloroplast RNAs: rps12 exon 1, 69281 to 69460; rps12 intron 1 5′, 68793 to 69302; rps12 intron 1 3′, 93161 to 93570; rps12 exon 2, 92876 to 93078; residue numbers from GenBank accession number X86563; rrn16 and rrn23, Bam13 fragment and 3-kb PstI fragment of chloroplast DNA, respectively, described by Barkan (1993).

For poisoned primer extension experiments, 20 pmol of an rps12-exon2 oligonucleotide (5′-GGTTTTTTGGGGTTGATAG-3′) (Figure 8) or rps12-exon3 oligonucleotide (5′-TTGGCTTTTTGGCCCCATATT-3′) (see Supplemental Figure 3B online) was radiolabeled at its 5′ end by incubation with [γ-32P]ATP and T4 polynucleotide kinase (New England Biolabs). Four micrograms of leaf RNA was heated to 95°C and annealed to the oligonucleotide by slow cooling to 45°C in 50 mM Tris-HCl, pH 8.5, 500 mM KCl, and 0.5 mM each of dATP, dGTP, dTTP, and ddCTP (10-μL reaction volume). Primer extension was initiated by adding 30 units of AMV reverse transcriptase (Promega), and the reactions were incubated at 45°C for 30 min. The reactions were stopped by adding 12 μL of 80% formamide and 0.25× TBE (1× TBE is 90 mM Tris-Borate, pH 8.3, and 2 mM EDTA). Ten microliters of each sample was applied to a 12% polyacrylamide gel containing 8 M urea and electrophoresed in 1× TBE. The gel was exposed to a PhosphorImager screen (Molecular Dynamics) and analyzed using ImageQuant software (GE Healthcare).

For RT-PCR analysis of ppr4 mRNA levels in mutant plants, primers were designed that spanned an intron, such that the amplification products from mRNA and genomic DNA could be distinguished. cDNA was generated from the indicated amounts of total leaf RNA by priming reverse transcription with the gene-specific primers PPR4 794REV (5′-CCTCGCTTGGCATAATACAC-3′) or α-tubulin REV (5′-AACACCAAGAATCCCTGCAGCCCAGTGC-3′). Primer was annealed to RNA in a 10-μL reaction containing 2 μL RNA and 1 μL of a 50 μM primer stock in annealing buffer (150 mM NaCl and 20 mM Tris-HCl, pH 8.5) by heating to 95°C for 30 s and cooling at 4°C for 2 min. This mixture was added to 10 μL of 2× Promega AMV buffer (100 mM Tris-HCl, pH 8.3, 100 mM KCl, 20 mM MgCl₂, 1 mM spermidine, and 20 mM DTT) supplemented with 2 mM deoxynucleotide triphosphate and 30 units of RNAsin (Promega). Primer extension was performed with 5 units of AMV reverse transcriptase (Promega) at 48°C for 1 h. PCR amplification was performed with the reverse primers described above, in conjunction with the forward primers PPR4 591FOR (5′-GCGGGCGCGCTGGGTCGAA-3′) or α-tubulin FOR (5′-AGCCCGATGGCACCATGCCCAGTGATACCT-3′). PCR reactions (30 μL) contained 2 μL of cDNA, 1 μM of each primer in 1× PCR buffer (Ex Taq buffer; TaKaRa), 0.8 mM deoxynucleotide triphosphate, 10% DMSO, and 1.25 units of Ex-Taq (TaKaRa) and were amplified under the following conditions: PPR4, 94°C/2 min, followed by 34 cycles of 94°C/30 s, 52°C/30 s, 72°C/30 s, and a final extension of 72°C/5 min; tubulin, 94°C/2 min, followed by 30 cycles of 94°C/30 s, 58°C/30 s, 72°C/30 s, and a final extension of 72°C/5 min. Ten microliters of each reaction were electrophoresed in 1.5% agarose gels. The tubulin cDNA was visualized by staining the gel with ethidium bromide. The ppr4 cDNA was visualized by DNA gel blotting, using a probe generated from a ppr4 cDNA clone using the same primers used for RT-PCR.

Antibody Production

Recombinant PPR4 was generated by expressing amino acids 615 to 886 in the vector pet28b (Novagen) to generate a fusion protein with a C-terminal 6xhistidine tag. This tagged PPR4 fragment was purified on a Ni-NTA agarose column (Qiagen) and used to generate polyclonal antisera in rabbits. The same PPR4 fragment was used to affinity purify the antisera prior to their use for RIP-chip and immunoblotting experiments.

Chloroplast Fractionation and Protein Analyses

Total leaf protein was extracted and analyzed by immunoblotting as described previously (Barkan, 1998). Antisera to OEC23, AtpA, D1, PsaD, and PetD are described by Voelker and Barkan (1995) and McCormac and Barkan (1999). Chloroplast subfractions were those described by Williams and Barkan (2003). Stromal extracts (0.5 mg of stromal protein per experiment) were fractionated by sedimentation through sucrose gradients according to Jenkins and Barkan (2001).

RIP-Chip and Slot Blot Hybridization Analysis of RNAs Bound to PPR4

The maize chloroplast microarray and RIP-chip procedure are described by Schmitz-Linneweber et al. (2005). Briefly, 2 μL of affinity-purified anti-PPR4 antibody was incubated with 100 μL of stromal extract (∼500 μg stromal protein) from the inbred line B73 or from ppr4-1/ppr4-2 mutant (∼100 μg stromal protein). The antibody was collected by incubation with formaldehyde-fixed StaphA cells (IG sorb; Enzyme Centre). RNA was isolated from pellet and supernatant fractions by phenol-chloroform extraction and labeled with Cy3 and Cy5 using the Micromax ASAP RNA labeling kit (Perkin-Elmer Life Sciences). Labeled RNA was purified using Qiaquick spin columns (Qiagen) and hybridized to microarrays covering the entire chloroplast chromosome in overlapping DNA fragments (Array Express accession number A-MEXP-164). Slides were scanned with a Genepix 4000B microarray scanner (Axon Instruments). Data were filtered against elements with low signal-to-noise ratios, and local background was calculated according to default parameters in Genepix Pro 6.0 software. Only spots with a signal-to-background ratio >4 and for which 60% of pixels have a F532 fluorescent signal >2 sd above background were chosen for further analysis. Fragments for which fewer than two spots per array passed these cutoffs were not used for subsequent analyses and appear as gaps when enrichment ratios are plotted according to chromosomal position. Background-subtracted data were used to calculate the median of ratios (pellet RNA F635: supernatant RNA F532). After log₂ transformation, this value is called the enrichment ratio. Normalization was done according to the median log₂F635/F532 value for all above background spots on each array.

RNAs for slot blot hybridizations were prepared in the same way as those used for RIP-chip assays. However, instead of using them to probe an array, they were applied to a nylon membrane with a slot blot manifold and hybridized to specific radiolabeled PCR fragments. The PCR fragments were body-labeled with [³²P]dCTP by the random priming method. Hybridization and washing were performed as described for the slot blot analyses by Schmitz-Linneweber et al. (2005). One-sixth of the RNA recovered from each immunoprecipitation supernatant and one-third of the RNA recovered from each immunoprecipitation pellet were applied to each slot.

Accession Numbers

Sequence data for ppr4 cDNA can be found in the GenBank data library under accession number DQ508419. The PPR4 RIP-chip data have been deposited at MIAME-Express under accession number E-MEXP-716.

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure 1. Positions of T-DNA Insertions in At PPR4 (At5g04810).

• Supplemental Figure 2. Median Log₂-Transformed Enrichment Ratios for Eeplicate Spots in Each of the Four PPR4 RIP-Chip Experiments Plotted against Fragment Number.

• Supplemental Figure 3. Analysis of Transcripts from Putative Secondary Targets of PPR4 in ppr4 Mutants.

• Supplemental Figure 4. Assay for Coimmunoprecipitation of PPR4 with CAF2/CRS2 Complex.

• Supplemental Table 1. Median Log₂ Ratios, Number of Spots Above Background for Each PCR Product on the Array, and sd Values for All Four RIP-Chip Experiments.

• Supplemental Table 2. Values for Combined Replicate RIP-Chip Data Sets.