Oligonucleotide-Recognizing Topoisomerase Inhibitors (OTIs): Precision Gene Editors for Neurodegenerative Diseases?

Abstract

Topoisomerases are essential enzymes that recognize and modify the topology of DNA to allow DNA replication and transcription to take place. Topoisomerases are divided into type I topoisomerases, that cleave one DNA strand to modify DNA topology, and type II, that cleave both DNA strands. Topoisomerases normally rapidly religate cleaved-DNA once the topology has been modified. Topoisomerases do not recognize specific DNA sequences, but actively cleave positively supercoiled DNA ahead of transcription bubbles or replication forks, and negative supercoils (or precatenanes) behind, thus allowing the unwinding of the DNA-helix to proceed (during both transcription and replication). Drugs that stabilize DNA-cleavage complexes with topoisomerases produce cytotoxic DNA damage and kill fast-dividing cells; they are widely used in cancer chemotherapy. Oligonucleotide-recognizing topoisomerase inhibitors (OTIs) have given drugs that stabilize DNA-cleavage complexes specificity by linking them to either: (i) DNA duplex recognizing triplex forming oligonucleotide (TFO-OTIs) or DNA duplex recognizing pyrrole-imidazole-polyamides (PIP-OTIs) (ii) or by conventional Watson-Crick base pairing (WC-OTIs). This converts compounds from indiscriminate DNA-damaging drugs to highly specific targeted DNA-cleaving OTIs. Herein we propose simple strategies to enable DNA-duplex strand invasion of WC-OTIs giving strand-invading SI-OTIs. This will make SI-OTIs similar to the guide RNAs of CRISPR/Cas9 nuclease bacterial immune systems. However, an important difference between OTIs and CRISPR/Cas9, is that OTIs do not require the introduction of foreign proteins into cells. Recent successful oligonucleotide therapeutics for neurodegenerative diseases suggest that OTIs can be developed to be highly specific gene editing agents for DNA lesions that cause neurodegenerative diseases.

Keywords: CRISPR/Cas9; camptothecin; etoposide; gene editing; inhibitors; topoisomerases.

Figure 1

Human Top1 and Top2 regulate DNA topology at a transcription bubble. (a) In the post-mitotic cells of the central nervous system-human Top1 and human Top2β are expected to be active in regulating DNA topology near transcription bubbles. The only DNA replication in post-mitotic CNS cells will take place in mitochondria; and careful design may be needed to ensure that damage to replicating mitochondrial DNA does not take place [27]. RNAPII, RNA polymerase II, is shown in contact with Top1 [28]. (b) Human Top1 cleaves a single DNA strand forming a 3’ phosphotyrosine. By convention both cleaved and non-cleaved strands are numbered relative to the single DNA-cleavage site. (c) Human Top2 (Top2α or Top2β) can cleave both DNA strands forming two 5’ phosphotyrosines. By convention both strands are numbered relative to the two 4-base-pair staggered DNA-cleavage sites.

Figure 2

Simplified schematics of twelve base-pairs of DNA in TOP1 DNA-cleavage complexes. (a) Simplified schematic of the central twelve base-pairs of DNA in a Top1 DNA-cleavage complex (based on 2.1Å structure pdb code: 1a31). Interactions (<3.5Å) between the protein and the DNA are represented by arrows (adapted from Figure 4–, panel G in [59]). Tyrosine 723 from Top1 has cleaved the top strand. Top1 can be imagined as a hand holding the double-stranded DNA up-stream of the DNA-cleavage site (red arrows and–red numbered nucleotides) and allowing controlled rotation about the phosphodiester bond between the −1 and +1 nucleotides on the intact strand (two green arrows) to relax the DNA. (b) A simplified schematic of the same DNA sequence in a complex with topotecan (a camptothecin derivative). The figure is based on the 2.1Å structure with topotecan (pdb code 1K4T). Note the +1 G-C base-pair (in 1K4T). The topotecan occupies the ‘same’ space as the +1 nucleotide pair in panel a. ‘The intercalation binding site is created by conformational changes of the phosphodiester bond between the +1 and −1 base pairs of the intact strand’ [45].

Figure 3

Type IIA topoisomerases and schematics of three etoposide crystal structures. (a) Simplified schematic of a reaction carried out by a type IIA topoisomerase. The gate or G-DNA (green cylinder) is cleaved and another DNA duplex, the T (or transport segment;-black) is passed through the cleaved DNA before religation. (b) Schematic of a 2.16Å human hTOP2β^core structure (pdb code: 3qx3) with DNA and two etoposides (I) binding at the two DNA cleavage sites, four base-pairs apart. One subunit is shown in red and blue, the other in grey. The DNA sequence (5′-3′) is the same for both strands; the DNA has been cleaved by tyrosine 821. (c) In human TOP2β is a single subunit and functions as a homodimer; the hTOP2β^core is residues 445-1201. Structural domains in hTOP2β^core are TOPRIM (TOP) domain, GK = Greek key domain, WHD = winged helical domain, TOW = tower domain, EX = exit gate domain. (d) Schematic of a 2.8Å structure of S. aureus gyrase^CORE fusion truncate DNA complex containing two etoposide (I) binding at the two DNA cleavage sites (pdb code: 5cdn), four base-pairs apart. One gyrase^CORE fusion truncate is shown in red and blue, the other in grey. (e) Schematic of a 2.45Å structure of S. aureus gyrase^CORE fusion truncate DNA complex containing one etoposide (pdb code: 5cdp) (I) (f) DNA gyrase consists of two subunits, GyrB and GyrA (domains are indicated). Note in the S.aureus gyrase^CORE fusion truncate the GyrB and GyrA subunits are fused into a single ‘subunit’ (B409-B644 + A2-A491) and the small greek key domain (residues B544-B579) has been deleted.

Figure 4

50 mer and 30 mer Watson–Crick-OTIs cleave an oncogene target sequence. Adapted from Figure 9 in [43] focusing on Top2α cleavage. (a) The target oncogenic PML-RARA sequence to be cleaved is shown in green on the top line, with the RARA sequence bold and underlined at the 3′ end; the 5′ sequence is from PML. The bottom line shows the 50 mer OTI with an etoposide moiety covalently attached to the T (highlighted in yellow). DNA-cleavage gels showed that the OTI promoted DNA-cleavage at four sites on the target (green) strand with relative intensities (in brackets) 24–25 (17), 23–24 (45), 20–21 (17), 19–20 (5). (b) The same target PML-RARA sequence is shown on the top line, but coloured in cyan at the 3′ end to indicate the position of the major cleavage site. Only two DNA-cleavage sites were observed with the 30 mer OTI, 23–24 (32), 20–21 (9) (red lightning bolts show positions). The red box indicates where the 20 nucleotides are from Xtal structures with etoposide (pdb codes: 3qx3, 5cdn and 5cdp);–based on the major DNA-cleavage site.

Figure 5

A 41 mer comprising a 30 mer etoposide-OTI coupled to a 10 mer AMN-TFO. (a) Depiction of a 41 mer, in which a 30-mer WC-OTI (orange letters with etoposide attached to yellow highlighted T) has been linked (single c in black) to a ten nucleotide TFO (blue letters) with a copper binding artificial metallo-nuclease (AMN) at the 5′ end. The copper binding AMN moiety is covalently attached to a cyan highlighted T. Note the X is a nucleotide designed to recognize a C in a TFO [68]. (b) the OTI recognizes a DNA-duplex (green letters) and the AMN cleaves one strand of the duplex (blue lightning bolt). (c) The rest of the strand-invading OTI (orange letters) can now strand invade and cleave the target oncogene (red lightning bolt–as in Figure 4b).

Figure 6

Deamidation of cytosine to uracil and 5-methylcytosine to thymine. The exocyclic amine of cytosine can be spontaneously deamidated to give uracil and the excocyclic nitrogen of 5-methylcytosine can be spontaneously deamidated to give thymine. Uracil is removed from DNA by uracil-DNA glycosylase while the G:T base-pair resulting from spontaneous deamidation of 5-methylcytosine can be removed by thymine DNA glycosylase (Figure drawn with Marvin-Sketch, from ChemAxon, https://www.chemaxon.com).

Figure 7

Modelling a theoretical correction by three camptothecin based PIP-OTIs of a mutation associated with familial ALS. (a) A model (BDNA, 8 June 2022, from http://www.scfbio-iitd.res.in/software/drugdesign/bdna.jsp) of nucleotides (one strand carbons in black, the other in green) coding amino acids 30–45 (single letter code bottom line) of human SOD1 with a G37R mutation. The single base-pair change causing the glycine to arginine mutation is highlighted in red. (b) Three camptothecin based PIP-OTIs are shown coloured with cyan, magenta and yellow carbons, with the intercalating camptothecin moiety in solid space-fill representation. Three positions (TG) cut on the lower strand, in the presence of Top1 (see also Supplementary Figures S3 and S5) are indicated. In cleavage complexes the T is covalently bonded to Top1 by a 3′ phosphotyrosine bond and the camptothecin moiety intercalates between the TA and GC base-pairs at the DNA-cleavage site. (c) The blue and orange oligonucleotides in b are envisioned to have been removed – and are replaced with an oligonuleotide with a corrected G. The T in the central G-T mismatch should be removed by thymine DNA-deglycoylase–after which the red T in the top strand should be corrected to C.

Figure 8

A theoretical correction, by three camptothecin based PIP-OTIs, of the exon 7 mutation associated with SMA. (a) DNA sequence from human SMN2 gene (NCBI’s RefSeq gene ID: 6607). Experimental evidence suggests that Exon 7 (underlined sequence–bottom line) is skipped because of alternative splicing due to a single base change (AT base-pair highlighted in red). (b) Three positions (TG) positions (5′–3′) on the upper, non-coding strand, to be targeted for cleavage by PIP-OTIs are underlined (note upper strand is drawn 3′–5′). (c) In the presence of Top1 (see also Supplementary Figures S3 and S5) three PIP-OTIs are predicted to cleave the DNA at three positions and remain covalently linked to the T’s. (d) After removal of the PIP-OTIs and covalently attached DNA–a gene editing oligonucleotide with a corrected G is introduced. (e) After the ‘theoretical’ correction of the G-T (in panel d) mismatch to G-C (panel e) exon 7 should be expressed (as in SMN1).