Structural basis for suppression of hypernegative DNA supercoiling by E. coli topoisomerase I

Abstract

Escherichia coli topoisomerase I has an essential function in preventing hypernegative supercoiling of DNA. A full length structure of E. coli topoisomerase I reported here shows how the C-terminal domains bind single-stranded DNA (ssDNA) to recognize the accumulation of negative supercoils in duplex DNA. These C-terminal domains of E. coli topoisomerase I are known to interact with RNA polymerase, and two flexible linkers within the C-terminal domains may assist in the movement of the ssDNA for the rapid removal of transcription driven negative supercoils. The structure has also unveiled for the first time how the 4-Cys zinc ribbon domain and zinc ribbon-like domain bind ssDNA with primarily π-stacking interactions. This novel structure, in combination with new biochemical data, provides important insights into the mechanism of genome regulation by type IA topoisomerases that is essential for life, as well as the structures of homologous type IA TOP3α and TOP3β from higher eukaryotes that also have multiple 4-Cys zinc ribbon domains required for their physiological functions.

Figure 1.

Structure of full-length E. coli topoisomerase I (EcTOP1) in complex with single-stranded DNA (ssDNA). (A) Domain arrangement of E. coli topoisomerase I. Between D8 and D9, there is a helical hairpin. (B) A ribbon diagram of full-length EcTOP1 in complex with ssDNA. Full-length EcTOP1 includes four N-terminal domains: D1 (deep salmon), D2 (orange), D3 (cyan) and D4 (green); and five C-terminal domains: D5 (pink), D6 (yellow), D7 (red), D8 (lime) and D9 (grey). The helical hairpin between D8 and D9 is colored in wheat. A ssDNA that binds to the C-terminal domains is colored in blue. Each Zn(II) is represented as a gray sphere. The secondary structures of D2 and part of D4 and D6 are labeled for discussion purposes. A part of the loop (colored in green) between α2 and β6 of D2 includes a charged and conserved sequence of R442KGDEDR, which is highly flexible and was not observed in earlier TOP67 structures. Figures 1B, 2, 4 and 7A are prepared with the program PyMOL (http://www.PyMOL.org).

Figure 2.

Structures of C-terminal domains. (A) A ribbon drawing of the first three C-terminal 4-Cys zinc ribbon domains, D5 (in pink), D6 (in yellow) and D7 (in red). All secondary structures are labeled. D6 has a unique helix (α1) between strands β1 and β2. The Zn(II) ions are drawn as gray spheres. The anomalous difference electron density map (drawn in orange mesh) around each Zn(II) is contoured at 4σ level and calculated at 4.0 Å using diffraction data collected at the Zn absorption peak. (B) Details of D7. The four cysteines that bind Zn(II), the conserved methionine (M718) below the Zn(II)-binding site and the hydrophobic residue (M729 in case of D7) below M718 are drawn in stick format. (C) The β-sheet to β-sheet packing between D5 and D6. The hydrophobic residues across their interface are drawn in stick format. (D) Comparison of the crystal structure and solution structure (PDB: 1YUA, in thin salmon Cα trace format) of D8 and D9 and the linker between them.

Figure 3.

A structure-based sequence alignment of E. coli EcTOP1 C-terminal domains and the sole T. maritima TmTOP1 C-terminal domain. EcTOP1 D5 was used for pairwise structural alignments from all other individual domains using SSM. The secondary structures, β strands, are represented by arrows above appropriate sequences of each domain. The bugles on some of the lines for the β1 strands indicate the presence of one- or two-residue β-bulges on the strands. All cysteines are highlighted in red. The residues that contribute to the cores of 4-Cys zinc ribbon domains or zinc ribbon-like domains (EcTOP1 D8 and D9) are highlighted in blue. The aromatic residues from the β3 strands, which can form π–stacking interactions with nucleotides in the C-terminal domains are highlighted in magenta.

Figure 4.

ssDNA binding to EcTOP1 C-terminal domains. (A) The oligo O29-O20 partial duplex used for co-crystallization with EcTOP1. The overhang of the 3′-end of O29 oligo represents a ssDNA segment that is bound to the C-terminal domains of EcTOP1. The nucleotides that were observed in the structure are highlighted in dark blue. The two nucleotides that are highlighted in light blue have disordered bases. Other parts of the oligo O29-O20 are completely disordered. The electron density drawn in gray mesh for the ssDNA segment of the DNA bound to the C-terminal domains is calculated from a weighted 2Fo-Fc map and contoured at 1σ level. (B) The interactions between the 3′-end of ssDNA with the first C-terminal domain D5 as well as the last N-terminal domain D4. Only one coordinating cysteine of the Zn(II), C619, is shown for the purpose of discussion. D6 colored in yellow is not involved in ssDNA binding. (C) The interactions between ssDNA and D7. (D) The interactions between ssDNA and D8 and D9.

Figure 5.

Loss of relaxation activity from C-terminal domain mutations in EcTOP1. The indicated amount of wild-type or mutant enzyme was incubated with the supercoiled plasmid DNA (S) for 30 min in reaction buffer containing either 6 mM or 11 mM MgCl₂.

Figure 6.

C-terminal domain mutations do not inhibit oligonucleotide cleavage and religation. The oligonucleotide substrate (O) was labeled at the 5′-end with ³²P. (A) Cleavage products accumulated by the indicated amount of wild-type and mutant EcTOP1 in the absence of Mg(II). The labeled cleavage products are P1: 5′-GATTATGCAATGCGCT; P2: 5′-GATTATGCAAT; P3: 5′-GATTATGCAATGCGCTTT. (B) Rapid religation of cleavage products from the addition of Mg(II). Following the addition of 2 mM MgCl₂ and 1 M NaCl to the cleavage reactions containing 100 ng of enzyme, the reactions were left on ice for the indicated time before quenching with loading buffer for sequencing gel analysis.

Figure 7.

Assay of complementation of topA temperature sensitive function in E. coli AS17 by wild-type (WT) pLIC-ETOP and its mutant derivatives. Serial dilutions of overnight cultures of AS17 transformants grown at 30°C were spotted on LB plates with kanamycin, followed by incubation at either 30°C or 42°C.

Figure 8.

Models of EcTOP1 in interaction with ssDNAs and its implications for positioning of unwound duplex DNA. (A) Stereo view of a ribbon diagram of full-length EcTOP1 with two ssDNA segments bound at N- and C-terminal domains. The model was created by combining the N-terminal domains/ssDNA structure (PDB code: 3PX7) and the structure reported in this study as described in the text. (B) A single chain and (C) a double chain model of ssDNA chain(s) binding to EcTOP1 for DNA relaxation. In the single chain model, the N- and C-terminal domains interact primarily with the G-strand only. In the double chain model, the N-terminal domains interact primarily with the G-strand while the C-terminal domains interact primarily with the T-strand. D1, drawn in red with a radial gradient does not bind ssDNA. The α1 (alpha1) helix on D6 is modeled as a handle, which presumably can pull or push the hinge between D2 and D4 domains for the regulation of the opening-closing state of N-terminal toroid hole.

Figure 9.

Nuclease footprinting with DNA bubble substrate. (A) Sequence of top (G-strand) and bottom (T-strand) of DNA bubble substrate. The topoisomerase I cleavage site is shown by arrow on the G-strand. Nuclease footprint of EcTOP1-G116S on the bubble substrate is followed by 5′ ³²P end-labeling of either the G-strand or T-strand. (B) DNase I footprinting. Lane 1: Control (C), DNA only; Lane 2: Nuclease only; Lanes 3–7: 10, 20, 40, 60, 100 ng ETOP-G116S followed by nuclease; Lane 8: 100 ng of EcTOP1-G116S only. (C) MNase footprinting. Lane 1: Control (C), DNA only; Lane 2: Nuclease only; Lanes 3–6: 10, 20, 40, 60 ng ETOP-G116S followed by nuclease. Nuclease cleavage products are denoted by * symbol.