Although U24 was shown to have two methyl targets, it has been proposed that one snoRNA generally modifies one site. In this work, we have observed three distinct ways in which a single snoRNA appears to direct multiple ribose methyl modifications. The first and most obvious involves ``double guide'' snoRNAs which contain two different guide sequences, one located at the 5' end of the snoRNA, the other at the 3' end. A double guide snoRNA has been previously observed for human and yeast U24 [Kiss-Laszlo et al., 1996]. To this, we add six new double-guide snoRNAs (snR40, snR41, snR47, snR51, snR60, snR67), three of which are at new loci. snR60 is particularly interesting given that the two nucleotides it modifies are on opposing sides at the base of a 12 bp helix formed in the rRNA secondary structure [Gutell, 1994]. It could be imagined that snR60 acts as a chaperone to bring the two ends of the helix together, expediting rRNA folding. However, this is the only example which shows obvious target site proximity. U24, snR41, and snR67 also guide modification of nucleotides within the same ribosomal subunit, although a spatial relationship as in the case of snR60 is not apparent from rRNA secondary structure predictions [Gutell, 1993,Gutell et al., 1993]. snR40, snR47, and snR51 modify nucleotides on different ribosomal subunits, thus proximity is difficult to estimate.
A second method appears to utilize two different D' boxes with the same complementary region to guide methylation at two different sites. Our disruption of snR48 resulted in loss of methylation at Gm2790 and Gm2788, implying that the first methylation site is measured from the D' box ``AUGU'' and the second by D' box ``GUUA'' (the ``GU'' overlaps between D' boxes). These are the two most atypical D' boxes among all confirmed yeast snoRNAs to date. Even so, none of the canonical traits of methylation guide snoRNAs [Kiss-Laszlo et al., 1996] are violated. A second example of this is predicted to occur in a newly identified S. pombe snoRNA but at a different pair of methylation sites (manuscript in preparation).
For U24, snR13, and U18, an additional adjacent ribose methyl modification may be due to a bulge within the snoRNA-rRNA duplex. Previous disruption of U24 has been observed to result in loss of methylation at Am1447 and unexpectedly at Gm1448 [Kiss-Laszlo et al., 1996]. Our disruption of snR13 results in loss of Am2278 and at Am2279. In both cases, we think that a single nucleotide bulge within the snoRNA could ``slide'' the rRNA target one base pair closer to the reference D' box without disrupting the necessary base pairings (see Figure 4.8). The one nucleotide slide places the adjacent site the canonical modification distance (5 bp) away from the D' box. In addition to the sites modified by snR13 and U24, two other pairs of adjacent sites in the rRNA may be modified in a similar manner, LSU-Am647/Cm648 and LSU-Um2918/Gm2919. U18 is predicted to modify LSU-Am647, but would allow a nucleotide bulge in the snoRNA-rRNA duplex to guide at LSU-Cm648 as well (Figure 4.8). We were not able to assay a U18 disruption mutant so this interaction is still hypothetical. Although we could not verify snR52 assignment to the remaining pair of adjacent methylation sites at LSU-Um2918 and LSU-Gm2919, this snoRNA fits the bulge model as well (Figure 4.8). Kiss-Laszlo et al. (1996) proposed an alternative mechanism, that loss of the adjacent methyl site (Gm1448) for the U24 disruption could be due to involvement of an independent methyltransferase that requires a ribose methyl site for sequential addition of an adjacent methyl site.