Structural elements of dynamic RNA strings.

Short transient secondary structures form while RNA is being transcribed, and these become the initial sites for protein anchoring. The insulin gene transcript (IGT) of chain length 1,430 nucleotides can be folded into 92 stem-loops, at an average of one stem-loop per 15.5 nucleotides (range: 9-35). The 25-hydroxyvitamin D3 1-alpha-hydroxylase gene transcript (HDHGT) of 4,825 nucleotides can fold into 274 stem-loops, at one stem-loop per 17.6 nucleotides (range: 9-45). We found no differences in transient secondary structures between the exons of IGT and HDHGT but there were significant differences between the introns. RNA chain shortening by folding ranged from 2.57 to 9.6 fold. Contraction ratios for IGT were 2.79 for minimal contraction and 7.77 for maximal contraction, and for HDHGT 2.57 and 8.80 respectively. The maximal contraction ratios but not the minimal contraction ratios differed significantly between IGT and HDHGT. This implies that initial RNP fibril formation may proceed by shared mechanisms whereas the final degree of compaction can differ in different hnRNPs. Metastable co-transcriptional folding may be necessary for "chaperones"/"match makers" to refold the RNA correctly for splicing and other maturation process. Branch point sequences are not consistent and are not included in the analysis. However, 5' and 3' splice regions have more disordered secondary structures, and 3' and 5' exon regions contain intrinsic snap-back complementarity that can bring 3' and 5' nucleotides together for joining. Upon splicing, the remaining exons undergo no change except for a few stem-loops flanking the splice sites.