As an additional line of evidence to confirm our 22 newly identified snoRNA loci, we verified gene expression by primer extension, and mapped 5' snoRNA ends to nucleotide resolution. Most previously isolated human and yeast guide snoRNAs end four to five nucleotides upstream of the C box, and two to four bases downstream of the D box ([Kiss-Laszlo et al., 1996]; Genbank C/D box snoRNA entries). Our program assumes a 5' end that is four nucleotides upstream of the C box and two nucleotides 3' of the D box. All 5' ends we mapped for new snoRNAs occurred 4-5 nucleotides from the true C box (Figure 4.7). We also assayed the 5' end of snR13 since the Genbank entry showed it extends 21 nucleotides 5' of the C box, the only guide snoRNA exception to the 4-5 bp rule. Our primer extension agreed with the unusually long 5' end given in the snR13 Genbank entry (SCU16692).
5' end mapping allowed us to check that the program had chosen the correct C boxes of new snoRNAs. C' boxes have been observed to occur in the interior of snoRNAs [Kiss-Laszlo et al., 1996,Kiss-Laszlo et al., 1998], and could be mistaken for legitimate C boxes. For 21 of 22 new snoRNAs tested, our program picked a C box that matched the experimentally determined 5' end. One snoRNA, snR63, surprised us by giving a primer extension product over 200 bp in length, far longer than any other methylation guide snoRNA identified. Assuming that the 3' end D box prediction is correct, snR63 appears to be 255 bp in length.
Transcription of the only two apparently redundant snoRNAs, snR59 and snR39, was also verified. Since these snoRNAs are intronic, mapping of the predicted mature 5' ends showed that both appear to be processed from their host gene mRNA transcripts.
Finally, we also tested each snoRNA disruption strain to verify that that we had eliminated snoRNA production completely. Each disruption strain showed complete loss of the snoRNA extension product of the expected size (Figure 4.7).