Probing anomalous structural features in polypurine tract-containing RNA-DNA hybrids with neomycin B

Abstract

During (-)-strand DNA synthesis in retroviruses and Saccharomyces cerevisiae LTR retrotransposons, a purine rich region of the RNA template, known as the polypurine tract (PPT), is resistant to RNase H-mediated hydrolysis and subsequently serves as a primer for (+)-strand, DNA-dependent DNA synthesis. Although HIV-1 and Ty3 PPT sequences share no sequence similarity beyond the fact that both include runs of purine ribonucleotides, it has been suggested that these PPTs are processed by their cognate reverse transcriptases (RTs) through a common molecular mechanism. Here, we have used the aminoglycoside neomycin B (NB) to examine which structural features of the Ty3 PPT contribute to specific recognition and processing by its cognate RT. Using high-resolution NMR, direct infusion FTICR mass spectrometry, and isothermal titration calorimetry, we show that NB binds preferentially and selectively adjacent to the Ty3 3' PPT-U3 cleavage junction and in an upstream 5' region where the thumb subdomain of Ty3 RT putatively grips the substrate. Regions highlighted by NB on the Ty3 PPT are similar to those previously identified on the HIV-1 PPT sequence that are implicated as contact points for substrate binding by its RT. Our findings thus support the notion that common structural features of lentiviral and LTR-retrotransposon PPTs facilitate the interaction with their cognate RT.

Figure 1

Comparison of 3′ PPT and flanking RNA sequences from selected lentiviruses and LTR retrotransposons. The PPT RNA sequence is underlined. See ref (4) for a more complete listing of LTR-containing retroelements.

Figure 2

Ty3 PPT-containing hybrid duplex sequences. (a) Wildtype hybrid from the S. cerevisiae LTR retrotransposon Ty3, denoted as PPT_wt. The dotted boxes indicate regions that were mutated in PPT_distal, PPT_prox, and PPT_pol. (b) Mutated sequences. For PPT_distal, the region proposed to be contacted by the Ty3 RT thumb was mutated to an rG·rA polymer to remove the distal NB site. For PPT_prox, the PPT U3 cleavage junction was mutated to an rG·rA polymer to remove the proximal neomycin binding site. For PPT_pol, both NB sites were removed to create an rG·rA 20-mer. Solid boxes indicate regions that were mutated in each sample. (c) Altered “strandedness”: PPT_RNA, both strands of RNA; PPT_DNA, both strands of DNA; PPT_swp, sequences of the RNA and DNA strands swapped. Sequences are numbered from 5′ to 3′, and biological numbering relative to the scissile −1g/+1a phosphodiester bond is indicated. Nucleotides in upper- and lowercase denote DNA and RNA, respectively. The monoisotopic mass observed by nanospray-FTICR analysis for each duplex construct is reported together with the respective mass calculated from the sequence.

Figure 3

Tandem mass spectrometry of the noncovalent 1:1 PPT_wt–NB complex. (a) Summary of the gas-phase fragmentation products afforded by the DNA and RNA strands of the Ty3 PPT hybrid. The lines mark the phosphodiester bonds cleaved on the respective sequences, which highlight nucleotides prevented from undergoing fragmentation in the presence of bound ligand (17). Stars denote products containing noncovalently bound NB. (b) The tandem mass spectrum displays a typical ion series containing noncovalently bound NB. (c) Sequences of the Ty3 PPT hybrid with the gray boxes highlighting binding sites identified using MS/MS footprinting.

Figure 4

1D water flip-back Watergate ¹H NMR spectra of the imino region of the PPT_wt duplex after titration with NB at 10 °C in 80 mM NaCl and 10 mM NaH₂PO₄/Na₂HPO₄ (pH 7.0). The concentration of the PPT_wt duplex was 225 μM. 1D ¹H NMR spectra are shown in the absence (top trace) and presence of 1.0 equiv (middle trace) and 2.0 equiv (bottom trace) of NB. Imino resonances for T(−1), T(−2), u(−13), g(+1), and g(−12), where chemical shift changes can be tracked, are highlighted in bold. Dotted lines and arrows indicate the shifts in position of the resonances.

Figure 5

Sequences of the (a) PPT_distal and (b) PPT_prox hybrids (gray nucleotides indicate mutated bases) with the gray boxes highlighting binding sites identified using MS/MS footprinting.

Figure 6

1D water flip-back Watergate ¹H NMR spectra of the imino region of PPT_distal and PPT_prox. (a) PPT_distal at 10 °C: control (top trace), 1.0 equiv of NB (middle trace), and 2.0 equiv of NB (bottom trace). (b) PPT_prox at 30 °C: control (top trace), 1.0 equiv of NB (middle trace), and 2.0 equiv of NB (bottom trace). Dotted lines and arrows indicate the shift in position of the resonances.

Figure 7

Nanospray FTICR mass spectra of (a) the initial equimolar mixture of the four PPT substrates (5 μM each), (b) products obtained after a 1-fold per substrate addition of NB, and (c) the assemblies after 10-fold ligand addition (see Materials and Methods): (■) PPT_DNA, (▼) PPT_swp, (▲) PPT_wt, and (◆) PPT_RNA. Stars denote the positions of bound ligand. The high resolution and accuracy afforded by this analytical platform enabled unambiguous identification of the different species in solution according to their unique molecular masses (see also Figure 2).

Figure 8

1D water flip-back Watergate ¹H NMR spectra of the imino region of the PPT_swp and PPT_RNA. (a) PPT_swp at 10 °C: control (top trace), 1.0 equiv of NB (middle trace), and 2.0 equiv of NB (bottom trace). (b) PPT_RNA at 30 °C: control (top trace), 1.0 equiv of NB (top middle trace), 2.0 equiv of NB (bottom middle trace), and 3.0 equiv of NB (bottom trace). Dotted lines and arrows indicate the shifts in position of the resonances.

Figure 9

CD spectra with 5.0 equiv of NB (red trace) and without NB (blue trace) of (a) PPT_RNA, (b) PPT_swp, and (c) PPT_WT hybrid samples.