DNA and Protein Requirements for Substrate Conformational Changes Necessary for Human Flap Endonuclease-1-catalyzed Reaction

Abstract

Human flap endonuclease-1 (hFEN1) catalyzes the essential removal of single-stranded flaps arising at DNA junctions during replication and repair processes. hFEN1 biological function must be precisely controlled, and consequently, the protein relies on a combination of protein and substrate conformational changes as a prerequisite for reaction. These include substrate bending at the duplex-duplex junction and transfer of unpaired reacting duplex end into the active site. When present, 5'-flaps are thought to thread under the helical cap, limiting reaction to flaps with free 5'-terminiin vivo Here we monitored DNA bending by FRET and DNA unpairing using 2-aminopurine exciton pair CD to determine the DNA and protein requirements for these substrate conformational changes. Binding of DNA to hFEN1 in a bent conformation occurred independently of 5'-flap accommodation and did not require active site metal ions or the presence of conserved active site residues. More stringent requirements exist for transfer of the substrate to the active site. Placement of the scissile phosphate diester in the active site required the presence of divalent metal ions, a free 5'-flap (if present), a Watson-Crick base pair at the terminus of the reacting duplex, and the intact secondary structure of the enzyme helical cap. Optimal positioning of the scissile phosphate additionally required active site conserved residues Tyr(40), Asp(181), and Arg(100)and a reacting duplex 5'-phosphate. These studies suggest a FEN1 reaction mechanism where junctions are bound and 5'-flaps are threaded (when present), and finally the substrate is transferred onto active site metals initiating cleavage.

Keywords: DNA endonuclease; DNA repair; DNA replication; DNA-protein interaction; circular dichroism (CD); fluorescence resonance energy transfer (FRET); nucleic acid enzymology.

FIGURE 1.

FEN1 DNA bending and double nucleotide unpairing. A, schematic of the FEN1 catalyzed hydrolysis of a double flap DNA yielding single-stranded DNA and double-stranded nicked DNA products. An arrow indicates the site of reaction. Each nucleobase is represented by a different color. B, hFEN1-product complex (Protein Data Bank code 3q8k) showing 100° bent DNA. C, schematic of double nucleotide unpairing proposed to position the scissile phosphodiester bond between the +1 and −1 nt on active site (pink) metal ions (cyan). D, cartoon representation of the active site in the FEN1-product structure (Protein Data Bank code 3q8k) showing the phosphate monoester of the unpaired −1 nt in contact with metal ions (cyan) and helical gateway (base α2-α4) and cap (top of α4 and α5) residues mutated in this study.

FIGURE 2.

FRET data showing DNA bending on complexation with hFEN1 and mutants. A, schematic of double flap (DF, endonucleolytic) and single flap (SF, exonucleolytic) DNA constructs (Table 2) used in FRET studies, donor = fluorescein (blue) and acceptor = TAMRA (red). Nonlabeled (NL), donor only (DOL), acceptor-only (AOL), and donor and acceptor (DAL) versions of these constructs were used. B, variation in energy transfer efficiency of DF(DAL) upon addition of WT hFEN1 measured at pH 7.5 and 37 °C in the presence of Ca²⁺ ions (blue) or EDTA (red) fitted to Equation 4. C, derived (Equation 4) values of K_bend for the DF (double flap) and SF (single flap) substrates (Table 2) with WT and mutated hFEN1s as indicated in Ca²⁺ (purple) and EDTA (pink). MMDF contained a +1 mismatch, 5′-hydroxyl single (3′) flap (HOSF) lacked a 5′-phosphate, and SADF had a 5′-conjugated streptavidin. Standard errors from repeat experiments are shown. D, derived (Equation 4) minimum (E_min) and maximum (E_max) energy transfer in Ca²⁺ (blue) and EDTA (orange) corresponding to the indicated protein with DF (double flap) or SF (single flap) substrates as in C. Duplex DNA was measured for comparison without protein in Ca²⁺-containing buffer. Standard errors from repeat experiments are shown.

FIGURE 3.

hFEN1 and mutant mediated conformational change of 2AP-containing single flap SF₋₁₋₂ monitored by ECCD. All measurements were carried out at 20 °C and pH 7.5. A, divalent metal ion-dependent reduction in 2AP exciton coupling signal occurred when substrate SF₋₁₋₂ was bound to hFEN1, indicative of local substrate conformational change. Unbound SF₋₁₋₂ (black), the corresponding single strand (ssSF₋₁₋₂, dashed line) and SF₋₁₋₂ bound to hFEN1 (blue) all in Ca²⁺-containing buffer. SF₋₁₋₂ bound to hFEN1 in buffer containing 25 mm EDTA (red). B, comparison of molar ellipticity per 2AP residue at 326 nm of SF₋₁₋₂ bound to WT- and mutant hFEN1s in Ca²⁺ (purple) and EDTA (pink) buffers. Standard errors from repeat experiments are shown. C, divalent metal ion-dependent reduction in 2AP exciton coupling signal occurred when substrate HO-SF₋₁₋₂, which lacks a 5′-phosphate, was bound to hFEN1, indicative of local substrate conformational change. Unbound HO-SF₋₁₋₂ (black), the corresponding single strand (ssHO-SF₋₁₋₂, dashed line) and HO-SF₋₁₋₂ bound to hFEN1 (blue) all in Ca²⁺-containing buffer. HO-SF₋₁₋₂ bound to hFEN1 in buffer containing 25 mm EDTA (red). D, comparison of molar ellipticity per 2AP residue at 326 nm of single flap HO-SF₋₁₋₂ free or bound to WT and mutant hFEN1s in Ca²⁺ (purple) and EDTA (pink) buffers. The unbound corresponding single strand is also shown.

FIGURE 4.

ECCD monitored conformational change of +1−1 2AP and 5′-modified −1−2 double flap substrates. All measurements were carried out at 20 °C and pH 7.5. ss, single strand. Standard errors from repeat experiments are shown. A, comparison of molar ellipticity per 2AP residue at 326 nm of 5′-streptavidin blocked (BL) and free and bound to hFEN1 and streptavidin-trapped (TR) complexes in Ca²⁺ (purple) and EDTA (pink) buffers. Blocked complex was formed by adding streptavidin to the substrate before the addition of hFEN1, whereas trapped was formed by adding streptavidin to the preformed hFEN1-Ca²⁺-BDF complex. B, comparison of molar ellipticity per 2AP residue at 326 nm of a doubled flap substrate with a +1 mismatch (MMDF₋₁₋₂) when free and bound to WT-hFEN1 in Ca²⁺ (purple) and EDTA (pink) buffers. The corresponding single strand is also shown. C, divalent metal ion-dependent reduction in 2AP exciton coupling signal occurred when substrate DF₊₁₋₁ was bound to hFEN1, indicative of local substrate conformational change. Unbound DF₊₁₋₁ (black), the corresponding single strand (ssDF₊₁₋₁, dashed line) and DF₊₁₋₁ bound to hFEN1 (blue) all in Ca²⁺-containing buffer. DF₊₁₋₁ bound to hFEN1 in buffer containing 25 mm EDTA (red). D, comparison of molar ellipticity per 2AP residue of double flap DF₊₁₋₁ at 326 nm when free and bound to WT and R100A hFEN1s in Ca²⁺ (purple) and EDTA (pink) buffers. The corresponding single strand is also shown. Standard errors from repeat experiments are shown.

FIGURE 5.

ECCD monitored conformational change of single flap +1−1 2AP containing substrates upon binding hFEN1 and mutants. All measurements were carried out at 20 °C and pH 7.5. ss, single strand. A, divalent metal ion-dependent reduction in 2AP exciton coupling signal occurred when single flap SF₊₁₋₁ was bound to hFEN1, indicative of local substrate conformational change. Unbound SF₊₁₋₁ (black), the corresponding single strand (ssSF₊₁₋₁, dashed line) and SF₊₁₋₁ bound to hFEN1 (blue) all in Ca²⁺-containing buffer. SF₊₁₋₁ bound to hFEN1 in buffer containing 25 mm EDTA (red). B, comparison of molar ellipticity per 2AP residue of SF₊₁₋₁ at 326 nm when free and bound to WT and mutant hFEN1s in Ca²⁺ (purple) and EDTA (pink) buffers. The corresponding single strand is also shown. Standard errors from repeat experiments are shown. C, a small divalent metal ion-dependent reduction in 2AP exciton coupling signal occurred when single flap HO-SF₊₁₋₁ that lacks a 5′-phosphate was bound to hFEN1, indicative of deficiency in bringing about local substrate conformational change. Unbound HO-SF₊₁₋₁ (black), the corresponding single strand (ssHO-SF₊₁₋₁, dashed line) and HO-SF₊₁₋₁ bound to hFEN1 (blue) all in Ca²⁺-containing buffer. HO-SF₊₁₋₁ bound to hFEN1 in buffer containing 25 mm EDTA (red). D, comparison of molar ellipticity per 2AP residue of single flap HO-SF₊₁₋₁ at 326 nm when free and bound to WT and mutant hFEN1s in Ca²⁺ (purple) and EDTA (pink) buffers. The corresponding single strand is also shown. Standard errors from repeat experiments are shown.

FIGURE 6.

Schematic model summarizing the responses of hFEN1-substrate complexes to addition of divalent metal ions based on ECCD results. Part (a), in the presence of divalent ions, unmodified substrates interacting with WT and K93A hFEN1s adopt an orientation of the −1 and −2 nt that is unstacked consistent with unpaired DNA. Also, stacking between the −1 and +1 nt is substantially reduced, suggesting control of their relative positions after unpairing. This observed conformational ordering of nucleobases is presumed to effect optimal contact between the scissile bond and active site metal ions and catalytic residues. Part (b), a divalent metal ion-induced substrate state where there is a gross change in the orientation of the −1 and −2 nt suggestive of local DNA unpairing is adopted by R100A, D181A, and Y40A with unmodified substrates and by all proteins (except L130P) with substrates lacking a 5′-phosphate. In these cases, however, there is evidence that stacking reminiscent of ssDNA remains between the −1 and +1 nt, suggesting an unpaired DNA state that is not optimally positioned for reaction. Part (c), the L130P mutation, modifications of the substrate that prevent accommodation of the 5′-flap under the helical cap (i.e. streptavidin conjugation to terminus of 5′-flap), or a mismatch at the +1 position prevent a DNA conformational change on addition of divalent ions. In these cases, the substrate is assumed to remain base-paired.