Skip to main content

Structure determination of human and chimpanzee HAR1F RNA by NMR

Final Report Summary - HAR1MC (Structure determination of human and chimpanzee HAR1F RNA by NMR)

Human-accelerated regions (HARs) are parts of human genome with an accelerated nucleotide substitution rate. HAR1 in HAR1F gene encodes for 118-nt non-coding ribonucleic acid (ncRNA) and shows 18 substitutions instead of the expected 0.27 substitutions since our last common ancestor with chimpanzees. HAR1 RNA was found in developing brain, ovary and testes. There are evidences that HAR1 is involved in Huntington's disease by showing a significant lower expression level in its patients.

Different secondary structure models have been offered for the chimpanzee and human HAR1 RNA: cloverleaf-like structures and an extended hairpin. We evaluated secondary structure models of the chimpanzee and human HAR1 RNAs by NMR spectroscopic investigations in order to map the structural changes induced by mutations. We pursued a 'divide-and-conquer' strategy to perform the nuclear magnetic resonance (NMR) assignment by analysing RNA fragments, which mimicked structural regions of the models. 124-nt c124 and h124 RNAs were derived from the full-length chimpanzee and human HAR1 RNAs by addition of three GC base pairs at the 5' end. We further investigated 47-nt c47 and h47 RNAs representing helices C and D, for which the most significant differences between chimpanzee and human have been predicted. 37-nt c37 and h37 RNAs represented helix H1. 54-nt c54 RNA mimicked helices H2 and H3 of model cM2 and was partially overlapping with c37 RNA in helix H1 to simplify the assignment. Additional three RNA fragments were analysed to evaluate model hM2: 39-nt h39 RNA mimicked helix H3, 15-nt h15 RNA represented helix H2 and 32-nt h32 RNA as mimicry for helix H4.

Our NMR data ruled out the chimpanzee and human secondary structure models (cM1 and hM1) initially proposed in 2006 by Pollard et al. On the other hand, an elongated hairpin structure model for the chimpanzee (cM2) published in 2008 by Beniaminov et al. was consistent with our NMR data, since c37 and c47 RNAs folded like their structural parts in the full-length chimpanzee RNA (c124 RNA) and c54 RNA fit except in helix H2. In contrast, several predicted imino proton resonances were not observable in our NMR spectra of the full-length human HAR1 RNA (h124 RNA). Our results were complementary to model hM2 for helices H1 and H3 owning to the NMR spectra of h37 and h39 RNAs, which fit to h124 RNA. We suggest a flexible inner part for h124 RNA, because h15 and h32 RNAs mimicking helices H2 and H4 do not exhibit their imino proton resonance pattern in the spectra of h124 RNA. Surprisingly, these findings show that the secondary structure of the human HAR1 RNA, which is G-rich and should be more stable than the chimpanzee HAR1F RNA, is actually the one that presented more opportunity for structural rearrangements. Pollard et al. already proposed that the structure of the human HAR1 RNA might lack helices C and E in model hM1 due to nucleobases, which were more accessible by DMS.

We further determined high-resolution three-dimensional (3D) structures of helix H1, which is the lower and most conserved part of the chimpanzee and human HAR1 RNAs. We designed and studied 37-nt fragments mimicking helix H1 by NMR in solution. 2D NOESY and 1H-15N HSQC spectra of c37 and h37 RNAs fit well with the spectra of the full-length HAR1 RNAs, c124 and h124, respectively. Moreover, the GAA asymmetric internal loop is the least defined part of both structures and is very interesting structural element, because it has not been characterized yet. The guanine and adenine bases can be either flipped out into the major and minor groove or stacked within the helix. In addition, guanine exhibits the lowest number of NOE contacts in NOESY spectra and does not show any hydrogen-bonding with the neighbour residues.

In conclusion, accelerations in human HAR1 suggest that function and structure of HAR1 in amniotes and human are different. Function of HAR1 RNA still remains unclear, but our results confirm that the human and chimpanzee HAR1 RNAs form different structures, which could consequently lead to their different function. Our analyses reveal that the 18 mutations make the human HAR1 RNA secondary and tertiary structure more dynamic. We propose that the human HAR1 RNA as part of its function is more structurally adaptable, presumably by binding to proteins.