Initial estimates of the total number of ribose methyl sites derive from early studies of ribosomal RNA modifications in Saccharomyces carlsbergensis [Klootwijk & Planta, 1973]. These experiments implied 55 distinct 2'-O-ribose methyl groups based on two dimensional gel analysis of 14C-methyl labeled, ribonuclease T1-digested rRNA. In these experiments, 2D gel spots, each containing a T1 fragment, were excised and analyzed for the methylated nucleotide and the surrounding sequence. Using the sequence context information, the locations of 42 ribose methyls were mapped unambiguously to the rRNA sequence. 13 ribose methyls could not be placed precisely due to insufficient sequence information within the T1 fragment [Veldman et al., 1981,Raue et al., 1988].
From the results we described in the previous section, we knew that the S. carlsbergensis methyl site data agreed well with observed concentration-dependent stops in S. cerevisiae rRNA primer extensions. Although one report has described difficulty detecting some ribose methylation sites in vertebrate rRNA by primer extension methods [Yu et al., 1997], our slightly optimized protocol allowed clear detection of all but one of the 42 mapped S. carlsbergensis ribose methyl sites in S. cerevisiae. For SSU-Cm1637, we could only detect a weak, non-concentration dependent band, likely due to two strong stops adjacent to the putative ribose methyl (Figure 4.6). Based on our ability to visualize the known ribose methyl sites by primer extension assay, we believed we would be able visualize unmapped sites as well.
We used three lines of evidence to predict, then experimentally verify the position as well as the snoRNA assignment for each of the unmapped sites. First, we knew between one and five nucleotides of sequence context for each of these sites based on known sequences for T1ribonuclease digest fragments of rRNA that contain the unplaced ribose methyl groups [Klootwijk & Planta, 1973,Veldman et al., 1981]. Second, we checked the existing collection of C/D box snoRNAs for previously unrecognized rRNA complementary regions that could target sites not included in the list of known ribose methyls. Third, we went back to the S. cerevisiae genome search results from our program and extracted all high scoring snoRNAs that could target new rRNA methyl sites.
Using these methods, we identified and verified 12 of the 13 unmapped methyl sites by primer extension. Six new sites were assigned to known C/D box snoRNAs, and the other six were assigned to newly identified snoRNAs (Table 4.2, target sites in boldface). Each of the 12 new methyl sites can be correlated with a T1 digest fragment for one of the 13 unmapped ribose methyls. We could not identify the location of the single unmapped methyl site in small subunit rRNA (T1 fragment GmU). snR190 has also been predicted to target a potential methylation site at LSU-Gm2393 [Kiss-Laszlo et al., 1996]. In our primer extension assay, this site does not give a visible band, nor does its sequence context correspond to an unassigned T1 fragment.
For each of the 12 newly mapped sites, we disrupted the corresponding guide snoRNA (except the intronic snR38 gene), and confirmed loss of the expected methylation site (see Table 4.2, snoRNAs assigned to methyl sites in boldface). None of the verified guide snoRNAs was found to be essential, nor did gene disruption cause noticeably impaired growth.
As in the previous section, several of these snoRNAs guide methylation at more than one site. The snR40 disruption showed loss of the newly mapped LSU-Um896 in addition to the previously predicted SSU-Gm1267 [Kiss-Laszlo et al., 1996]. For snR60, gene disruption showed loss of newly mapped LSU-Gm906, as well as previously mapped LSU-Am815 (see previous section). snR67 disruption showed loss at two newly mapped sites (Table 4.2).
We took advantage of the tandem arrangement of seven snoRNA genes snR72 through snR78 and constructed a septuple deletion mutant of these genes. The septuple snoRNA deletion mutant was still viable with no obvious change in growth rate on rich media. While we did construct a single locus deletion mutant for snR78, we tested the rest of the snoRNAs in the tandem array (snR72-snR77) via the septuple mutant. We therefore cannot prove a 1:1 mapping of the six methyl site losses to snR72-snR77, although the strong rRNA complementarities for each snoRNA support this conclusion.
For each snoRNA disruption mutant, we checked the status of the predicted target methyl site(s) as well as at least two other sites in the neighboring rRNA region. We observed specific methyl site loss only at the predicted target site in all instances but one. In the case of the tandemly grouped snoRNAs, we observed a polarity effect in that snoRNAs downstream from those being disrupted showed either partial or complete loss of methylation at their target sites (unpublished).