Characterization of fragmented mitochondrial ribosomal RNAs of the colorless green alga Polytomella parva

Abstract

We have identified previously in mitochondrial DNA of the colorless, chlorophycean, green algal taxon, Polytomella parva, potential coding regions for four small subunit (SSU) and eight large subunit (LSU) rRNA fragments. In this study with P.parva, we isolated RNA from a mitochondrial-enriched preparation, characterized the 12 mitochondrial rRNA transcripts by either northern blot analysis or chemical sequencing and performed secondary structure modeling of the SSU and LSU rRNA sequences. The results show the following features about the mitochondrial SSU and LSU rRNAs of P.parva: (i) they are considerably shorter than their homologs from other green algae, although the main domains typical of conventional rRNAs are conserved; (ii) the rRNA fragmentation pattern is most similar to that of Chlamydomonas reinhardtii among green algae that have been characterized; (iii) three nucleotides are missing from the normally highly conserved GTPase center of the LSU rRNA; and (iv) post-transcriptional modification of the 3′-terminal region of the SSU rRNA is unusual in that it has the ‘eubacterial’ 3-methyluridine (corresponding to m³U at Escherichia coli 16S rRNA position 1498) but lacks the more highly conserved modifications at two adjacent A residues (corresponding to N⁶,N⁶-dimethyladenosine at E.coli 16S rRNA positions 1518 and 1519). This is the first report of the characterization by direct sequencing of fragmented mitochondrial rRNAs from a green alga.

INTRODUCTION

Complete mitochondrial genome sequences from four chlorophycean green algal taxa, i.e. Chlamydomonas reinhardtii (1,2), Chlamydomonas moewusii (= Chlamydomonas eugametos) (3), Chlorogonium elongatum (4) and Scenedesmus obliquus (5,6), revealed the existence of fragmented and scrambled small and large subunit (SSU and LSU, respectively) rRNA-coding regions and the lack of a 5S rRNA gene. The chlorophycean lineage is composed of two main sublineages as supported by both 18S rDNA sequence data and basal body configuration in flagellated cells (7); three of the four taxa discussed above associate with one of these sublineages while S.obliquus associates with the other. The fact that S.obliquus has fewer mitochondrial rRNA gene break points, all of which correspond to variable regions that are also interrupted in the other three taxa, implies that these break points were present in the ancestor of the four taxa. Moreover, the mitochondrial SSU and LSU rRNA break points of C.moewusii and Chlorogonium are in the same variable regions thus indicating a close phylogenetic relationship between these taxa.

The feature of fragmented mitochondrial rRNA-coding regions among chlorophycean algae is distinct compared with green algal taxa outside this group except for Pedinomonas minor whose phylogenetic position among the green algae is uncertain (5 and references therein, 8). However, knowledge of chlorophycean mitochondrial rRNAs is still rather limited. For example, rDNA sequencing, northern blot hybridization and S1 nuclease protection analysis were the main means previously employed to characterize these rRNAs; their direct sequencing has not been reported, principally because the species studied previously, all of which are photosynthetic, typically contain abundant chloroplast rRNA which hampers the isolation of relatively pure mitochondrial rRNA needed for such analysis. Moreover, potential secondary structures of mitochondrial SSU and LSU rRNAs in the Chlorophyceae have been reported only for C.reinhardtii and C.moewusii (1,9,10). Finally, information from additional chlorophycean species is required for understanding the evolutionary pathway of mitochondrial rRNA fragmentation in the chlorophycean lineage.

Polytomella (11) is a green algal genus consisting of colorless, wall-less unicells, which provides advantages in the isolation and characterization of mitochondrial components. For example, Polytomella strain SAG 198.80 has proven useful for the isolation of respiratory proteins (12–14). Similar studies have been difficult with C.reinhardtii, the green alga having the most well studied mitochondrial genetic system, because of contaminating thylakoid components (15,16). In addition, the fact that no plastid rRNAs could be detected in Polytomella parva (= Polytomella agilis), even by means of northern blot analysis (17), implies either the absence or extremely low abundance of plastid ribosomes in this taxon, and thus makes it useful for the study of green algal mitochondrial ribosomes and rRNA. Information obtained about Polytomella mitochondria is particularly applicable to C.reinhardtii, because of the specific evolutionary connection between them among Chlamydomonas-like taxa (18,19).

To date, sequences of the mitochondrial protein-coding genes, cox1 and cob, from Polytomella strain SAG 198.80 have been reported (20,21), as has most of the mitochondrial DNA (mtDNA) of P.parva, which is in at least two linear pieces of 13.5 and 3.5 kb (22). The latter study (22) identified four SSU (rns_a through rns_d) and eight LSU (rnl_a through rnl_h) potential rRNA-coding modules, which are scattered over 8 kb of the P.parva 13.5 kb mtDNA component (Fig. 1). These coding modules are scrambled in order and interspersed with each other and with additional coding regions; Polytomella, therefore, shares these features with other chlorophycean green algae. The P.parva mitochondrial rRNA-coding modules, however, are distributed on both DNA strands, a feature not previously reported for chlorophycean mtDNA, although this has been reported for the fragmented rRNA genes in apicomplexan mtDNA (23).

Figure 1.

Map of the 13.5 kb mtDNA of P.parva showing the organization of mitochondrial rRNA genes (22). SSU and LSU rRNA genes are fragmented into four (rns_a through rns_d) and eight (rnl_a through rnl_h) potential modules (open rectangles), respectively. These gene modules are distributed on both strands and interspersed with each other and with protein (black rectangles)- and tRNA (hatched rectangle)-coding regions. Two open arrows indicate transcriptional orientations. Two solid arrows represent terminal inverted repeats. Thin lines at the ends indicate unsequenced parts. The partial sequence of this mtDNA is registered under GenBank accession no. AY062933.

In order to gain more information about the evolution of fragmented rRNA in green algae and to enhance the utility of Polytomella for studies of green algal mitochondrial biogenesis, we have characterized the mitochondrial rRNA transcript fragments of P.parva by either RNA sequencing or northern blot analysis, and performed secondary structure modeling of the SSU and LSU rRNA sequences as deduced from the rDNA regions.

MATERIALS AND METHODS

Strain and culture conditions

Polytomella parva (UTEX L 193) was obtained from the culture collection of the University of Texas at Austin. Cultures were grown at 25°C in the medium of Sheeler et al. (24) with shaking for small cultures (100–250 ml) or mild aeration for larger cultures (5–15 l). Cells were harvested in the logarithmic growth phase (∼2.5 × 10⁶ cells/ml) by centrifugation (2000 g) at 4°C for 15 min.

Preparation of a mitochondrial-enriched fraction

The mitochondrial fraction was prepared as described by Spencer et al. (25) with some modifications. A cell pellet from 6 l of culture was washed once in buffer A (50 mM Tris–HCl, pH 8.0, 300 mM mannitol, 0.1% bovine serum albumin, 1 mM β-mercaptoethanol, 3 mM EDTA). After resuspension in 30 ml of buffer A, the cells were disturbed manually in a 40 ml glass homogenizer (Kontes Glass Co.) until 90% of them were broken. Following centrifugation (1000 g) for 10 min, the supernatant was saved and subjected to another centrifugation (12 000 g) for 20 min to collect crude mitochondria. The resulting pellet was resuspended in 2 ml of buffer A and loaded onto a gradient consisting of 1.15 and 1.55 M sucrose (in buffer A). The gradient was centrifuged (25 000 r.p.m., Beckman SW 41 Ti rotor) for 1 h. The mitochondrial band at the interface of the 1.15 and 1.55 M sucrose layers was removed using an 18 gauge needle with a 90° bend. The fraction obtained was slowly diluted with two volumes of buffer B (50 mM Tris–HCl, pH 8.0, 20 mM EDTA), followed by centrifugation (12 000 g) for 15 min. All steps were carried out at 4°C.

RNA isolation

For the isolation of total cellular RNA, cells were washed once in buffer C (50 mM Tris–HCl, pH 8.0, 50 mM NaCl, 50 mM EDTA) and lysed following resuspension in three volumes of buffer C containing 0.1% SDS. For the isolation of mitochondrial-enriched RNA, a mitochondrial-enriched fraction from 6 l of culture was resuspended in 9 ml of buffer D (10 mM Tris–HCl, pH 8.5, 50 mM KCl, 10 mM MgCl₂), followed by the addition of 1 ml of 20% Triton X-100 (in buffer D). Following centrifugation (10 000 g) for 10 min, SDS was added to the supernatant to a final concentration of 2%. RNA extraction from the cell and mitochondrial lysate followed the method of Rochaix and Malnoë (26). All steps were carried out at 4°C.

RNA fractionation

RNA was fractionated in 8% polyacrylamide slab gels (20 × 20 × 0.15 cm) containing 7 M urea by electrophoresis at 350 V for 4 h (27) and was detected by staining with ethidium bromide.

Determination of RNA sequences

RNA isolated from the mitochondrial-enriched fraction and semi-purified mitochondrial ribosomes (28) of P.parva was 3′-end-labeled with [³²P]pCp using RNA ligase (29). The end-labeled mitochondrial rRNAs were fractionated in 6% polyacrylamide gels (33 × 40 × 0.05 cm) at 1700 V and eluted according to Schnare et al. (30). The partial chemical degradation method (29) was used to determine the 3′-terminal sequence of mitochondrial rRNAs.

Northern blot hybridization

Following fractionation by polyacrylamide gel electrophoresis, RNA was blotted onto a nylon membrane (Hybond-N+; Amersham Pharmacia Biotech) by the method of capillary transfer (31) using 0.5 mM NaOH as the transfer solution. The Alkphos Direct Labeling and Detection System kit (Amersham Pharmacia Biotech) was employed for probe labeling and the subsequent hybridization, following the kit instructions. Oligonucleotide probes were labeled at 37°C for 2.5–3.5 h. Hybridization was carried out at 38°C overnight in a hybridization oven. The RNA blots were washed twice in the primary wash buffer at 38°C for 10 min and twice in the secondary wash buffer at room temperature for 5 min. Chemiluminescent detection was achieved by exposing the blots to autoradiography film. Oligonucleotide probes used in the northern blot hybridization are as follows: S₁, 5′ TTA TCT CAT AGT GAA AAG CTA GGC AAA GAC 3′; S₂, 5′ TGC GTA AAA CGA TAG TCC TTT GAG ACT ATT 3′; L₁, 5′ TTA TTC GTC TTT TTG TTC CAT CAC TGT ACT 3′; L2, 5′ ATA TTA AAT CGC TGG CCC ATG CTG CAA AAG 3′; L₄, 5′ ATC TCC TTT TGA ACC TTA ACC TAT CCG TTG 3′; L₆, 5′ CCT ATC GTC GCT TTT GTT ACT AAT GCC AGC 3′; L₈, 5′ AGG ATG CGA TGA TCC AAC ATC GAG GTG 3′.

RESULTS

Identification of mitochondrial rRNA fragments

Total cellular and mitochondrial-enriched RNA of P.parva were subjected to polyacrylamide gel electrophoresis (Fig. 2A). The total cellular RNA sample revealed prominent components typical of the nucleocytoplasm, including 25/28S, 18S, 5.8S and 5S rRNAs, as well as tRNAs. The RNA from the mitochondrial-enriched fraction contained, in addition, several potential mitochondrial rRNA species that had sizes significantly smaller than the 25/28S and 18S rRNAs.

Figure 2.

(A) Electrophoretic profile of P.parva total (T) and mitochondrial-enriched (M) RNA in an 8% polyacrylamide gel. Lines with a dot at one end indicate the positions of northern blot hybridization signals obtained with oligonucleotide probes derived from seven mitochondrial rDNA modules (B). Arrows point to the positions of the remaining five mitochondrial rRNA species whose 3′-termini were sequenced (Fig. 3). Positions of nucleocytoplasmic rRNAs including 25/28S, 18S, 5.8S, 5S rRNA and tRNAs, as well as RNA size markers (Sigma), are also shown. (B) Northern blot hybridization analysis to identify seven P.parva mitochondrial rRNA species. RNA species detected with oligonucleotide probes derived from rDNA modules rnl_a, rnl_b, rnl_d, rnl_f, rnl_h, rns_a and rns_b are designated L₁, L₂, L₄, L₆, L₈, S₁ and S₂, respectively. Locations of RNA size markers (Sigma) are indicated.

Two approaches were taken for mitochondrial rRNA identification. Northern blot analysis with individual probes derived from seven rRNA-coding modules each identified a single transcript (Fig. 2B). Five of these transcripts corresponded in size to an RNA component detected in the mitochondrial-enriched fraction (Fig. 2A). The other two of these seven transcripts, L₂ and L₄, were not detected by ethidium bromide staining; the former of these transcripts did not stain well because of its small size, and the latter fragment co-migrated with the tRNAs. For the remaining five RNA species, 3′-terminal chemical sequencing data (Fig. 3) established that each of these corresponds to a transcript of one of the rRNA-coding modules (Table 1). We note that the rRNA pairs S₄/L₈ and S₂/L₃ could not be well separated in standard 8% polyacrylamide gels due to their similar sizes within the pairs (Fig. 2A), but could be differentiated using 6% polyacrylamide sequencing gels (data not shown). Due to their similar sizes, rRNA fragments L₁, L₆ and S₁ also co-migrated (Fig. 2).

Figure 3.

Autoradiograms of 20% sequencing gels showing the resolution of limited chemical digests of five 3′-end-labeled P.parva mitochondrial rRNA species. These RNAs were identified as transcripts of rDNA modules rnl_c, rnl_e, rnl_g, rns_c and rns_d, and are designated L₃, L₅, L₇, S₃ and S₄, respectively. In the gel for S₄, the arrows indicate the positions of reduced G-lane and enhanced C-lane cleavage; this sequence is numbered from the 3′-terminus.

Table 1. Summary of features of P.parva mitochondrial rRNA-coding modules and their transcripts^a.

Coding module	rRNA	Minimum lengthb	Maximum lengthc	Actual 3′-terminusd	Length determined by gel electrophoresise	Length and location in structure diagramsf
SSU
rns_a	S1	95 (2485–2391)	116 (2502–2387)g		99	99, 1–99 (2485–2387)
rns_b	S2	167 (10 188–10 354)	188 (10 168–10 355)		198	179, 100–278 (10 177–10 355)
rns_c	S3	362 (7493–7854)	387 (7482–7868)	7854	362	362, 279–640 (7493–7854)
rns_d	S4	339 (9599–9937)	408 (9533–9949)	9937	339	339, 641–979 (9599–9937)
LSU
rnl_a	L1	67 (9466–9532)	102 (9459–9560)		99	100, 1–100 (9461–9560)
rnl_b	L2	48 (7434–7481)	85 (7408–7492)		60	60, 101–160 (7433–7492)
rnl_c	L3	176 (10 429–10 604)	196 (10 429–10 624)	10 624	200	196, 161–356 (10 429–10 624)
rnl_d	L4	71 (10 097–10 167)	120 (10 068–10 187)		73	73, 357–429 (10 095–10 167)
rnl_e	L5	127 (9941–10 067)	159 (9938–10 096)	10 078	141	141, 430–570 (9938–10 078)
rnl_f	L6	92 (10 625–10 716)	185 (10 605–10 789)		99	96, 571–666 (10 625–10 720)
rnl_g	L7	552 (3054–2503)	613 (3098–2486)g	2500	558	558, 667–1224 (3057–2500)
rnl_h	L8	321 (7087–7407)	364 (7070–7433)		332	332, 1225–1556 (7085–7416)

^aLengths are indicated in nucleotides; numbers in parentheses refer to coordinates in the GenBank entry (AY062933).

^bBased on secondary structure modeling.

^cBased on the positions of flanking genes, assuming no overlap; in cases where the gene of interest is flanked by another rRNA gene the coordinates of the neighboring gene were taken from the previous column.

^dGenomic position determined by RNA sequencing.

^eDetected by ethidium bromide staining or northern hybridization.

^fEstimates based on values in previous columns.

^gTranscribed from the opposite strand relative to most of the other rRNA-coding modules.

Potential structures of P.parva mitochondrial SSU and LSU rRNAs

Figures 4 and 5 show the proposed mitochondrial SSU and LSU rRNA secondary structures of P.parva based on the sequences of the rDNA modules and modeled after the proposed secondary structures of their respective Escherichia coli homologs (10). The mitochondrial SSU and LSU rRNA fragments of P.parva have combined lengths of 979 and 1556 nt, respectively. In these structures, theoretically, the four SSU and eight LSU rRNA pieces could be brought together through intermolecular base-pairing to form structures containing three and six domains in the SSU and LSU rRNA, respectively, as in other fragmented chlorophycean counterparts (1,9).

Figure 4.

Potential secondary structure of P.parva mitochondrial SSU rRNA. The structure is constituted by four RNA species, whose 5′- and 3′-termini are indicated. Roman numerals denote the three domains. Thin lines define regions variable among and continuous in rRNAs of P.parva, C.reinhardtii and C.moewusii mitochondria; number of nucleotides in these regions is indicated for P.parva (Pp), C.reinhardtii (Cr) and C.moewusii (Cm). The broken arrows indicate the break points. Small squares, each of which represents a nucleotide, describe structures that are conserved in C.reinhardtii and C.moewusii mitochondria and in E.coli, but altered or absent in the mitochondria of P.parva.

Figure 5.

Potential secondary structure of P.parva mitochondrial LSU rRNA. (A) 5′ half; (B) 3′ half. The structure is constituted by eight RNA species. Roman numerals denote the six domains. The GTPase center of the LSU rRNA is enclosed by thick lines as is the potential structure of the E.coli GTPase center included as an inset for reference. The square in brackets indicates an extra nucleotide in E.coli. Other explanations are as described in Figure 4.

The P.parva mitochondrial rRNAs appear structurally conventional throughout most of the evolutionarily conserved structural and functional cores [see Cannone et al. (10) for structure conservation diagrams], but deletions, relative to the corresponding regions of C.reinhardtii (1) and C.moewusii (9) mitochondrial and E.coli (10) rRNA, occur in three regions not reported previously to be variable. First, the 431–436 region of SSU rRNA lacks 7 nt (Fig. 4). Secondly, two sequences in the stem that results from the interaction of LSU rRNA fragments L₆ and L₇ are eliminated (Fig. 5A). Thirdly, 3 nt are missing from the GTPase center predicted for P.parva mitochondrial LSU rRNA (Fig. 5A); this center is otherwise rather conserved relative to the counterparts from E.coli as well as C.reinhardtii and C.moewusii mitochondria. Although the DNA used for sequencing the region encoding the GTPase center was obtained by PCR (22), the possibility of amplification errors was minimized by a parallel PCR with C.reinhardtii DNA which produced a product having the same sequence found for cloned DNA that had not been PCR amplified (1). In addition, a cDNA of the P.parva mitochondrial GTPase center was obtained by reverse transcriptase–PCR, and its sequence exactly matches the DNA prepared by PCR, thus supporting the absence of the 3 nt in the GTPase center of P.parva LSU rRNA. All DNA sequences were determined for both strands.

Comparisons involving 775 and 1174 nt of mitochondrial SSU and LSU rDNA, respectively, that could be aligned among P.parva, C.reinhardtii and C.moewusii, revealed a nucleotide identity of 61 and 69%, respectively, between P.parva and C.reinhardtii, and an identity of 58 and 68%, respectively, between P.parva and C.moewusii. These results underscore the considerable evolutionary divergence, at the primary sequence level, between the P.parva mitochondrial rDNA sequences and those of C.reinhardtii and C.moewusii.

Features of the 12 P.parva mitochondrial rRNA-coding modules and their transcripts are summarized in Table 1. Approximate locations of the 5′ and 3′ ends of these mitochondrial rRNAs, judged to be within 5–10 nt of the true termini, were estimated on the basis of their sizes as determined by secondary structure modeling, gel electrophoresis and the proximity of neighboring coding regions as determined by DNA sequencing. The 3′ ends of rRNA species L₃, L₅, L₇, S₃ and S₄ were precisely located by direct chemical sequencing. The estimated size of each rRNA fragment (Table 1) is consistent with the size of the corresponding rDNA module identified previously by DNA sequencing, therefore arguing against the presence of additional points of discontinuity within any of the coding modules. The size of S₂ was estimated at 179 nt (Table 1) and could be only 9 nt larger when the positions of flanking genes are considered; it therefore seems likely that its electrophoretic mobility is artifactually slow since it migrates near the 200 nt RNA marker in polyacrylamide gels (Fig. 2). A similar phenomenon has been observed for Euglena gracilis LSU rRNA species 4 (27).

Chemical cleavages of rRNA fragment S₄

Sequencing of the 3′-terminus of the P.parva mitochondrial rRNA fragment S₄, which corresponds to the 3′ end region of SSU rRNA from other sources, revealed two sites where chemical cleavage was unusual compared with other sites in this rRNA (Fig. 3). The residue located 45 nt from the 3′ end, where the DNA sequence predicts a U, exhibited an extraordinarily strong C-specific chemical cleavage. This enhanced cleavage is diagnostic for 3-methyluridine (m³U) in the RNA, as was observed, for example, at the homologous position in SSU rRNA from E.coli, wheat mitochondria and E.gracilis chloroplasts (32,33). In addition, the residue located 27 nt from the 3′-terminus, where the DNA sequence predicts a G, was resistant to G-specific chemical cleavage. It is not certain whether this lack of cleavage indicates post-transcriptional modification because there are examples where some unmodified G residues give blanks in the G lane of chemical sequencing gels (29; M.N.Schnare, unpublished data).

Normal chemical cleavage was observed at the two A residues of P.parva (24 and 25 nt from the 3′-terminus of S4 rRNA; Fig. 3) corresponding to E.coli 16S rRNA positions 1518 and 1519. In E.coli and many other eubacterial and eukaryotic SSU rRNAs (32), as discussed later, these sites contain two adjacent N⁶,N⁶-dimethyladenosine residues that are resistant to A-specific chemical cleavage.

DISCUSSION

General structure of P.parva mitochondrial rRNAs

One of the most distinguishing features of the fragmented mitochondrial rRNAs of P.parva is their combined SSU rRNA and LSU rRNA lengths, which are considerably shorter than the fragmented and continuous rRNA counterparts so far characterized from other green algae (Table 2). The observed small size of the P.parva mitochondrial rRNA-coding regions seems to be the result of a trend for smaller SSU and LSU rRNA-coding regions in the lineage leading to P.parva. The existence of such a trend is supported by the lengths of homologous SSU and LSU rRNA variable regions that are unbroken in P.parva, C.reinhardtii and C.moewusii; most of these regions are shorter, sometimes drastically so, in P.parva relative to the other two taxa (Figs 4 and 5).

Table 2. Length^a of green algal mitochondrial rRNAs in nucleotides.

Species	SSU rRNA	LSU rRNA	Accession nos
Mesostigma viride	1558	2843	AF353999
Nephroselmis olivacea	1509	2760	AF110138
Prototheca wickerhamii	1674	3009	U02970
Pedinomonas minor	1178	2110	AF116775
Scenedesmus obliquus	1747	3028	AF204057
Chlorogonium elongatum	1449	1921	Y07814, Y13644
Chlamydomonas moewusii	1240	1916	AF008237
Chlamydomonas reinhardtii	1200	2085b	U03843
Polytomella parva	979	1556	AY062933

^aIn most cases precise termini have not been determined.

^bLength of LSU rRNA does not include L_2b and L_3a (1).

In addition, deletions have been detected in three regions of the P.parva mitochondrial rRNA that are usually conserved in other systems. First, the GTPase center, which corresponds to positions 1051–1108 of E.coli LSU rRNA, is very highly conserved in terms of its 58 nt length and proposed secondary structure. Due to the reduced length of the P.parva GTPase center (55 nt), however, two base pairs, corresponding to pair 1058:1080 and pair 1082:1086 in the secondary structure model for the E.coli GTPase center (Fig. 5A), are not possible. Comparative (34), thermodynamic (35) and crystal structural (36) studies emphasize the fundamental importance of the 1082:1086 base pair. We note, however, that disruption of this base pair also occurs in the mitochondrial rRNA of trypanosomatid protozoa and metazoan animals (10), all of which have highly reduced LSU rRNA sequence lengths. Secondly, length reduction in SSU rRNA fragment S₃ near coordinates 431–436, which corresponds to the 722–733 region of the E.coli homolog, alters the potential secondary structure in this region. Different approaches reveal that this region plays an important role in E.coli protein synthesis [see Zimmermann (37) for a review]. Thirdly, two sequences in the LSU region of interaction between L₆ and L₇, which correspond to positions 1299–1302 and 1626–1641 of E.coli LSU rRNA, are deleted. It has been shown that parts of the missing sequences are in a region involved in ribosomal protein binding (38). Deletions in these three regions may be the result of relaxed functional constraints on the mitochondrial ribosomes of P.parva; it is also possible that proteins may have replaced certain functions of the rRNAs.

Finally, we could not identify any RNAs corresponding in sequence to the two RNA species, L_2b and L_3a, from C.reinhardtii mitochondria; these abundant RNAs were initially presumed to correspond to non-core LSU rRNA regions based on their cotranscription with and precise cleavage from LSU rRNA fragments L_2a and L_3b (1). However, as sequences homologous to L_2b and L_3a have also not been identified in other LSU rRNAs, including C.moewusii mitochondrial LSU rRNA (9), it now appears likely that these are not components of C.reinhardtii mitochondrial ribosomes.

Mitochondrial rRNA break points

All the interrupted points, three and seven in P.parva mitochondrial SSU and LSU rRNAs, respectively, are located in regions previously identified as variable in sequence and secondary structure in other rRNAs (39). Figure 6 shows the comparison of mitochondrial SSU and LSU rRNA discontinuity patterns among the Chlamydomonas-like chlorophycean algae. The number and position of broken variable regions in LSU rRNA is identical between P.parva and C.reinhardtii (1), while these two organelles share two out of three break points in their SSU rRNAs. The other break point in P.parva SSU rRNA (separating S₁ and S₂ in the structure models) is shared with C.moewusii/Chlorogonium (4,9). Overall, the fragmentation patterns of P.parva mitochondrial rRNAs are more similar to those of C.reinhardtii than to their counterparts from C.moewusii (Fig. 6), consistent with phylogenetic analysis of nuclear 18S rDNA sequences (18,19).

Figure 6.

Comparison of mitochondrial rRNA fragmentation patterns among Chlamydomonas-like algae. The rRNAs are drawn to the scale of the E.coli homologs. The arrow and arrowheads indicate the break points unique to P.parva/C.moewusii/Chlorogonium and P.parva/C.reinhardtii, respectively.

Post-transcriptional modification near the 3′-terminus of P.parva mitochondrial SSU rRNA

Information about the post-transcriptional modification pattern of rRNA is limited because of technical challenges involved in identifying and localizing modifications in large RNAs (40). However, direct chemical sequencing of 3′-end-labeled SSU rRNA has proven useful for evaluating the presence or absence of the ‘eubacteria-specific’ m³U (corresponding to m³U at E.coli 16S rRNA position 1498) and the two adjacent N⁶,N⁶-dimethyladenosine residues (corresponding to E.coli 16S rRNA positions 1518 and 1519) that are found in both eubacterial and eukaryotic cytoplasmic SSU rRNA (32). The m³U and the two m⁶₂A modifications are present in wheat (32) and Acanthamoeba castellanii (41) mitochondrial rRNAs, two systems in which the rRNA sequences/structures have retained striking similarity to their eubacterial homologs (41,42). The more divergent mitochondrial rRNAs from Tetrahymena pyriformis (43) and fungi (44,45) are missing all three of these modifications, while the small-sized SSU rRNAs from animal mitochondria have the two m⁶₂A modifications but do not have the m³U (46–48). The data presented here for P.parva mitochondria provide the first example of any SSU rRNA that has the ‘eubacterial’ m³U without also having at least one of the adjacent m⁶₂A modifications. Note that the E.gracilis chloroplast 16S rRNA has the m³U but only one of the m⁶₂A residues (33).

CONCLUSION

The highly divergent and fragmented nature of chlorophycean mitochondrial rRNAs raises the question of whether they can actually function in mitochondrial protein synthesis. Indirect evidence that these rRNAs are functional comes from the observations that they are associated with ribosome-sized particles in C.moewusii (28), and that a mutation in the GTPase center of the C.reinhardtii LSU rRNA gene is linked to the suppression of a frame shift mutation in the mitochondrial cox1 gene (49). The work presented here and in Fan and Lee (22) provides the necessary background for more direct functional studies of fragmented mitochondrial rRNAs, using P.parva. In addition, novel structural and post-transcriptional modification features of P.parva mitochondrial rRNAs revealed in this study present interesting questions regarding their potential effect on ribosomal function.