We analysed protein–DNA and protein–protein interactions relevant to the repair of DNA DSBs (double-strand breaks) by NHEJ (non-homologous end-joining). Conformational transitions in mammalian DNA ligases III (LigIII) and IV (LigIV), as well as in PARP-1 [poly(ADP-ribose) polymerase-1], were analysed upon binding to double-stranded DNA by changes in tryptophan emission and FRET (Förster resonance energy transfer) from tryptophan to DNA-conjugated Alexa Fluor® 532. For LigIII, two non-equivalent high- and low-affinity DNA-binding sites are detected interacting sequentially with DNA. PARP-1 displays a single high-affinity DNA-binding site and can displace bound DNA fragments from the low-affinity site of LigIII, consistent with its mediator role in LigIII–DNA interactions. For the LX [LigIV–XRCC4 (X-ray cross-complementation group 4)] complex, a single DNA-binding site is detected. Binding of Ku to DNA was accompanied by conformational changes in the protein and intermolecular FRET from dansyl chromophores of the labelled Ku to the Alexa Fluor® chromophores of Alexa Fluor® 532-conjugated DNA. The average distance of 5.7 nm calculated from FRET data is consistent with a location of Ku at the very end of the DNA molecule. Binding of LX to Ku–DNA complexes is associated with conformational changes in Ku, translocating the protein further towards the DNA ends. The protein–protein and protein–DNA interactions detected and analysed generate a framework for the characterization of molecular interactions fundamental to the function of NHEJ pathways in higher eukaryotes.

INTRODUCTION

The recognition and repair of DNA DSBs (double-strand breaks) is critical for the maintenance of genomic integrity. It is widely believed that higher eukaryotes utilize predominantly NHEJ (non-homologous end-joining) to remove DSBs from their genome [13]. NHEJ consists of at least two genetically and biochemically distinct pathways. The canonical, or classical, NHEJ is a rapid process that utilizes, in addition to DNA-PKcs [catalytic subunit of DNA-PK (DNA-dependent protein kinase)] and the Ku heterodimer, the LX [LigIV (DNA ligase IV)–XRCC (X-ray cross-complementation group) 4 complex] and the Artemis nuclease, referred to in the present paper as D-NHEJ (DNA-PK-dependent NHEJ). The protein factor XLF (XRCC4-like factor)/Cernunnos binds to LX and promotes adenylation of LigIV, but unlike XRCC4, it seems dispensable for LX stability [4,5].

The second DSB repair pathway, referred to in the present paper as B-NHEJ (back-up-NHEJ) [68], is also frequently referred to as MMEJ (microhomology-mediated end-joining) [9] or A-NHEJ (alternative NHEJ) [10]. It reflects a DNA-PK-independent mechanism of end-joining that has been shown to utilize the LigIII (DNA ligase III), XRCC1 and PARP-1 [poly(ADP-ribose) polymerase-1] module that is a key component of DNA SSBs (single-strand breaks) and base damage repair [1114]. Recently, additional factors including PNK (polynucleotide kinase), Fen-1 (flap endonuclease 1) and the MRN complex, comprising Mre11, Rad50 and Nbs1, as well as CtIP [CtBP (C-terminal-binding protein)-interacting protein] have been implicated in this pathway [14]. B-NHEJ substitutes for D-NHEJ in murine lymphoid class switch recombination, aberrant V(D)J recombination and in the repair of IR (ionizing radiation)-induced DSBs [15,16]. B-NHEJ is, in general, slower than D-NHEJ and shows higher propensity for errors, particularly translocations. It may therefore be critically involved in genomic instability and cancer [14].

The Ku heterodimer, in addition to its role in D-NHEJ, is thought to also contribute to pathway choice in human somatic cells by initiating classical NHEJ and actively suppressing HRR (homologous recombination repair) and B-NHEJ [16]. Ku can also modulate the activities of ATM (ataxia telangiectasia mutated) and ATR (ATM- and Rad3-related) kinases [17,18].

PARP-1, in addition to its role in SSB and base damage repair, has also been implicated in the regulation of DSB-repair pathways [8,19]. PARP-1 is activated upon binding to DNA, particularly in regions of SSBs or DSBs, and catalyses the synthesis of negatively charged PARs [poly(ADP-ribose) polymers] on several target proteins including PARP-1 itself and histones. This post-translational modification marks DNA-damage sites, and induces local chromatin relaxation that facilitates the recruitment of repair factors. Additionally, PARs help in the release of PARP-1 from DNA ends, which allows the completion of the repair process. Recent data show that the proper dynamic assembly of the MRN complex at sites of photo-induced DNA damage is dependent on PARP-1, which accumulates very rapidly and early at DSBs [20]. This function is compatible with a role of both factors in B-NHEJ.

In mammals, the last step of D-NHEJ is ligation by LigIV. On the other hand, LigIII is a more universal ligase involved in SSB repair, as well as in DSB repair through B-NHEJ [14]. Both DNA ligases contain DBDs (DNA-binding domains) and conserved CC (catalytic core) domains consisting of NTase (nucleotidyltransferase) and OB (oligonucleotide/oligosaccharide-binding) domains [2124]. Besides DBD and CC domains, DNA ligases also contain additional domains such as the BRCT [BRCA1 (breast cancer early-onset 1) C-terminal] domains of LigIIIα and LigIV, mediating the interactions with XRCC1 and XRCC4 respectively [25,26]. LigIIIβ lacks the C-terminal BRCT domain and does not form a complex with XRCC1 [25,27]. In contrast to LigI (DNA ligase I) and LigIV, LigIII has a unique N-terminal ZnF (zinc-finger) domain that improves the ligation efficiency of the enzyme [2831]. The amino acid sequence and conformation of this ZnF domain is identical with those of the N-terminal ZnF domains of PARP-1, which mediate the binding of the polymerase to DNA breaks [32,33]. It has been suggested that the unique ZnF domain allows organization of LigIII in two DNA-binding modules, the first one containing ZnF and DBD, and the second containing the CC domain [21,24]. The two structural modules interact separately with DNA and bind two DNA molecules without a requirement for XRCC1 or other DNA repair factors [7,10,34].

The above outline suggests intriguing protein–DNA and protein–protein interactions essential for the processing of DSBs by each repair pathway and possibly also for the selection and co-ordination between repair pathways. Such interactions are likely to be associated with conformational changes in the participating proteins that determine their function in the respective repair pathway. We postulated that detection and analysis of such protein conformational changes will be essential for a molecular understanding of DSB processing and applied fluorescence spectroscopy, including FRET (Förster resonance energy transfer), to characterize selected protein–DNA and protein–protein interactions directly relevant to DSB repair by NHEJ. The results point to intriguing conformational transitions of LigIII, LigIV, Ku and PARP-1 as they engage to DSB repair and generate a useful framework for the detailed characterization of their mechanistic involvement in B-NHEJ and D-NHEJ.

EXPERIMENTAL

DNA substrates

The blunt-ended DNA duplexes used in protein–DNA binding experiments were generated by annealing of a 50-base-long single-stranded oligonucleotide (5′-GAACGAAAACATCGGGTACGAGGACGAAGACTGACCACGACATACTAACA-3′) with a complementary strand of the same length that was in its 5′-end either free, or covalently labelled with Alexa Fluor® 532 (Molecular Probes).

Proteins

Human recombinant PARP-1 was from Biomol. LigIV/XRCC4 and LigIIIβ (referred to as LigIII), were purified as described previously [35] (Figure 1A) and stored at −80°C in buffer A [50 mM Tris/HCl (pH 7.9), 1 mM EDTA, 0.02% Tween 20, 5% (v/v) glycerol, 0.3 M NaCl and 1 mM DTT (dithiothreitol)]. Ku was purified as described previously [36] and was stored at −80°C in buffer A (Figure 1A).

Fluorescence properties of LigIII, PARP-1, LX and Ku

Figure 1
Fluorescence properties of LigIII, PARP-1, LX and Ku

(A) SDS/PAGE analysis of purified human LigIII, LX and Ku. Molecular masses are indicated in kDa. (B) Tryptophan emission spectra of PARP-1, LigIII, LX, Ku, DNS-aziridine-labelled Ku (Ku-DNS) and N-acetyltryptophanamide (Trp) (in buffer or in n-butanol) excited at 295 nm. Fluorescence spectra are corrected for background fluorescence of the buffer and are recorded in arbitrary units (a.u.). (C) Stern–Volmer plots describing quenching by acrylamide of tryptophan emissions of LigIII and LX. (D) Stern–Volmer plots showing quenching of tryptophan emission of unlabelled Ku and quenching of DNS-aziridine emission of DNS-aziridine-labelled Ku by acrylamide (Ku-DNS). The formulae of tryptophan and of the quencher acrylamide are shown within the graphs. F0 and F are intensities of tryptophan or DNS-aziridine emissions in the absence or presence of acrylamide.

Figure 1
Fluorescence properties of LigIII, PARP-1, LX and Ku

(A) SDS/PAGE analysis of purified human LigIII, LX and Ku. Molecular masses are indicated in kDa. (B) Tryptophan emission spectra of PARP-1, LigIII, LX, Ku, DNS-aziridine-labelled Ku (Ku-DNS) and N-acetyltryptophanamide (Trp) (in buffer or in n-butanol) excited at 295 nm. Fluorescence spectra are corrected for background fluorescence of the buffer and are recorded in arbitrary units (a.u.). (C) Stern–Volmer plots describing quenching by acrylamide of tryptophan emissions of LigIII and LX. (D) Stern–Volmer plots showing quenching of tryptophan emission of unlabelled Ku and quenching of DNS-aziridine emission of DNS-aziridine-labelled Ku by acrylamide (Ku-DNS). The formulae of tryptophan and of the quencher acrylamide are shown within the graphs. F0 and F are intensities of tryptophan or DNS-aziridine emissions in the absence or presence of acrylamide.

Fluorescence labelling of Ku

Ku (2–4 μmol) was covalently labelled with DNS-aziridine (5-dimethylaminonaphthalene-1-sulfonyl aziridine, also known as dansylaziridine), a specific thiol-reactive probe. The protein was incubated for 3 h at 4°C in buffer A with a 100-fold dye excess. Free unbound dye was removed by gel filtration using Sephadex G-25 columns equilibrated with buffer B [50 mM Tris/HCl (pH 7.9), 1 mM EDTA, 0.02% Tween 20, 5% (v/v) glycerol and 0.3 M NaCl]. The activity of DNS-aziridine-labelled Ku was determined by EMSA (electrophoretic mobility-shift assay) performed as described in [8] using the oligonucleotides OA (5′-GGCCGCACGCGTCCACCATGGGGTACAA-3′) and OB1 (5′-TTGTACCCCATGGTGGACGCGTGCGGCC-3′) after annealing and 3′-32P-labelling. Ku–DNA binding was assessed by incubating 0.2 ng of 32P-labelled DNA with DNS-aziridine-labelled Ku in buffer C [10 mM Tris/HCl (pH 7.5), 1 mM EDTA, 0.5% glycerol, 150 mM NaCl and 1 mM DTT] at 25°C for 20 min and loaded on to a 6% native polyacrylamide gel in 0.5× TBE [22.5 mM Tris/borate and 0.5 mM EDTA] buffer. Gels were dried and analysed using a Typhoon 9410 Variable Mode Imager (GE Healthcare).

Fluorescence experiments

All fluorescence measurements were performed at room temperature (21–25°C) in buffer D [10 mM Tris/HCl (pH 7.9), 1 mM EDTA, 0.15 M NaCl and 0.5% glycerol] using a PerkinElmer LS-50B spectrofluorimeter. To minimize inner-filter and self-absorption effects, the absorbance of the samples at the excitation wavelength was kept at less than 0.05 by appropriate dilution. The relative tryptophan emission quantum yield (QTrp) was determined by comparing the integrated tryptophan emission spectra of native unlabelled Ku, LX complex, LigIIIβ or PARP-1 with that of an N-acetyltryptophanamide solution. The latter was used as a standard and was normalized to the same absorbance at the excitation wavelength (295 nm). A value of 0.13 was used for the quantum yield of the standard [37]. The relative emission quantum yield of DNS-aziridine-labelled Ku (QDNS) was calculated by comparing its integrated fluorescence spectra with that of a solution of the thiol adduct of 6-acryloyl-2-dimethylaminonaphthalene (acrylodan) prepared in buffer D. DNS-aziridine and acrylodan excitation was at 365 nm, and the fluorescence emission spectra were normalized to the same absorbance at 365 nm. A value of 0.18 was used for the quantum yield of acrylodan [38]. Quenching of tryptophan and DNS-aziridine emissions was performed with acrylamide as external quencher. Data were analysed using the Stern–Volmer equation [37].

Intramolecular FRET from tryptophan residues (donors) to cysteine-attached DNS-aziridine (acceptors) was studied by the decrease in tryptophan fluorescence of Ku in the presence of DNS-aziridine chromophores. Intermolecular FRET between tryptophan residues (donors) of LigIII and LigIV and DNA-attached Alexa Fluor® 532 (acceptor) was measured after excitation at 295 nm. Binding of Ku to DNA was analysed by FRET from DNS-aziridine chromophores in Ku (donors) to the Alexa Fluor® dye on DNA (acceptor) in a reducing environment, under conditions favouring binding of one Ku molecule to one 50-bp-long DNA molecule [39].

The efficiencies of the intra- and inter-molecular energy transfer processes (E) were calculated from the decrease in the donor quantum yields (QD is QTrp or QDNS) in the presence of the acceptor (DNS-aziridine or Alexa Fluor® 532). The average distances, r, between the donor–acceptor pairs were determined using the formula:

 
formula

where R0 is the Förster radius, or the critical distance allowing FRET with a 50% probability; it was determined from the formula:

 
formula

where JAD is the overlap integral between the decadic molar absorbance of the acceptor and the corrected emission spectrum of the donor on a wave number scale normalized to unity; n is the refractive index of the medium and K2 is the orientation factor, determined by the mutual spatial orientation of the transition dipole moments of the donors and acceptors. As no data on the spatial orientation of the transition dipole moments of the chromophores are available, a random orientation of the donor–acceptor pairs was assumed (K2=0.667). A value of 1.36 was taken for the refractive index n [40].

Protein–DNA-binding experiments

Binding affinities of LigIII, LX and PARP-1 for DNA were studied by analysing changes in the intrinsic tryptophan fluorescence of the proteins (excitation wavelength of 295 nm and emission wavelength of 340 nm). Increasing amounts of 50-bp-long DNA duplexes either unlabelled or labelled with Alexa Fluor® 532 were added to 100 nM purified LigIIIβ, LX or PARP-1, and the samples were equilibrated until a steady emission reading was obtained (usually 3–4 min). All titration experiments were performed at room temperature in buffer D, and data were corrected for background fluorescence and dilution, which did not exceed 2.5% of the volume of the sample. The apparent dissociation constants (Kd) and the maximal fluorescence change (ΔFmax) upon saturation of the DNA-binding sites of LigIII, LX and PARP-1 were calculated by non-linear regression of the experimental results and data were fitted to a one-site binding hyperbola where ΔFobs is the observed change in the fluorescence and L is the concentration of the DNA oligonucleotides:

 
formula

RESULTS

Fluorescence properties of LigIII, LX, PARP-1 and Ku

The emission of tryptophan residues in different proteins varies depending on the local environment of the chromophores [41]. Notably, in the same protein, tryptophan emission can change significantly in response to conformational changes associated with engagement in a particular biological process, as a result of changes in the local environment of the chromophores [42]. A major determinant of the tryptophan emission spectrum is the hydrophobicity of the local environment. Thus, whereas tryptophan dissolved in water emits at 365 nm, this emission shifts to 345 nm when tryptophan is dissolved in a polar organic solvent (Figure 1B). The tryptophan emissions of LigIII, LX, PARP and Ku show fluorescence emission maxima in the range 340–349 nm (Table 1 and Figure 1B) suggesting polar environment in the neighbourhoods of the emitting residues [42].

Table 1
Fluorescence characteristics of LigIII, LX, PARP-1 and Ku

QTrp is the relative tryptophan emission quantum yield of the proteins; KQ is the bimolecular quenching constant; λmax is the emission maximum position.

ProteinQTrpKQ (M−1)λmax (nm)
LigIIIβ 0.014 5.2 342 
LX 0.025 12.4 340 
PARP-1 0.120 − 342 
Ku70/80 0.015 6.4 349 
ProteinQTrpKQ (M−1)λmax (nm)
LigIIIβ 0.014 5.2 342 
LX 0.025 12.4 340 
PARP-1 0.120 − 342 
Ku70/80 0.015 6.4 349 

For more information regarding the local environment of emitting tryptophan in the above proteins, we carried out quenching experiments using acrylamide. Quenching experiments allow inferences on the localization of emitting tryptophan residues. Specifically, residues buried in the hydrophobic interior of the protein are not quenched, whereas solvent-exposed residues are quenched to an extent commensurate to their solvent accessibility. The results of LigIII, LX and Ku show linear fluorescence quenching with increasing acrylamide concentration suggesting that the majority of emitting tryptophan residues is exposed to the solvent, albeit to different degrees in the different proteins (Table 1, and Figures 1C and 1D). The Stern–Volmer plots of Figures 1(C) and 1(D) allow the calculation of the bimolecular quenching constant KQ. The KQ values of 5.2 and 6.4 M−1 for LigIII and Ku respectively suggest that tryptophan chromophores are only partially exposed to the solvent, whereas the KQ of 12.4 M−1 calculated for LX suggests direct solvent accessibility (Table 1 and Figure 1C). The QTrp varied between proteins (Table 1), again reflecting environmental differences in the vicinity of the emitting tryptophan residues. The lowest values of 0.014 and 0.015 were calculated for LigIII and Ku. On the other hand, the QTrp of PARP-1 approached 0.12 (Table 1), a value that is close to the quantum yield of free tryptophan in water (QTrp=0.13) [37].

Changes in tryptophan emission of LigIII reflect structural changes upon DNA binding

We enquired whether binding of LigIII to DNA is associated with changes in protein conformation that modify the intrinsic tryptophan fluorescence. The binding of 100 nM purified LigIIIβ to 50-bp-long DNA, which is mediated by its ZnF domain (Figure 2A), was analysed by measuring tryptophan fluorescence intensity at different DNA concentrations. Since these measurements were carried out in the absence of ATP and in the presence of EDTA that chelates Mg2+, an essential cofactor for this enzyme, the DNA–enzyme interaction measured approximates the second step of the ligation reaction, i.e. the formation of the ligase–DNA binary complex that follows the enzyme adenylation step. In the range of concentrations studied, interaction of LigIIIβ with DNA is accompanied by a ~30% reduction in the quantum yield of the tryptophan emission (Figure 2B), suggesting protein conformational changes upon binding to DNA. The biphasic nature of fluorescence-intensity reduction with increasing DNA concentration suggests the presence of two DNA-binding sites in LigIII (Figure 2C). A high-affinity binding site (Kd ~3×10−8 M) reacts at low DNA concentrations (up to 100 nM) and reduces tryptophan fluorescence by approximately 12%. At higher DNA concentrations ~15–17% increased quenching is observed, reflecting the saturation of a second, lower-affinity, LigIII DNA-binding site (Kd ~2×10−7 M). The result is in accordance with the modular structure of LigIII that allows sequential and simultaneous binding of two DNA molecules [21,24].

Changes in tryptophan fluorescence intensity of LigIII upon binding to DNA

Figure 2
Changes in tryptophan fluorescence intensity of LigIII upon binding to DNA

(A) Domain organization of LigIIIα and LigIIIβ. All tryptophan residues present in LigIIIβ are indicated. The highly conserved Lys421 of the active site is also shown (K421). aa, amino acids. (B) Tryptophan emission spectra of LigIIIβ, and of the complexes of LigIIIβ with unlabelled or Alexa Fluor® 532-labelled DNA (DNA-AF532). The appearance of emission at 555 nm indicates FRET from tryptophan residues of LigIII to Alexa Fluor® chromophores of DNA. Fluorescence spectra were corrected for background fluorescence of the buffer, or of the buffer containing unlabelled or Alexa Fluor® 532-labelled DNA, and are recorded in arbitrary units (a.u.). The asterisk (*) indicates that the spectrum is not corrected for the contribution of Alexa Fluor® 532-labelled DNA. (C) The biphasic decrease in tryptophan emission upon binding to DNA indicates the presence of high- (Kd ~3×10−8 M) and low- (Kd ~2×10−7 M) affinity DNA-binding sites in LigIII.

Figure 2
Changes in tryptophan fluorescence intensity of LigIII upon binding to DNA

(A) Domain organization of LigIIIα and LigIIIβ. All tryptophan residues present in LigIIIβ are indicated. The highly conserved Lys421 of the active site is also shown (K421). aa, amino acids. (B) Tryptophan emission spectra of LigIIIβ, and of the complexes of LigIIIβ with unlabelled or Alexa Fluor® 532-labelled DNA (DNA-AF532). The appearance of emission at 555 nm indicates FRET from tryptophan residues of LigIII to Alexa Fluor® chromophores of DNA. Fluorescence spectra were corrected for background fluorescence of the buffer, or of the buffer containing unlabelled or Alexa Fluor® 532-labelled DNA, and are recorded in arbitrary units (a.u.). The asterisk (*) indicates that the spectrum is not corrected for the contribution of Alexa Fluor® 532-labelled DNA. (C) The biphasic decrease in tryptophan emission upon binding to DNA indicates the presence of high- (Kd ~3×10−8 M) and low- (Kd ~2×10−7 M) affinity DNA-binding sites in LigIII.

To characterize LigIII–DNA interactions further, we performed experiments using the same DNA oligonucleotides after labelling with Alexa Fluor® 532. This set-up allows the follow-up of changes in LigIII tryptophan fluorescence in the presence of DNA-associated Alexa Fluor® dye. With labelled DNA, the reduction in tryptophan emission is up to 25% greater than that registered with unlabelled DNA fragments (Figure 2B). Notably, the binding of LigIII to Alexa Fluor® 532-labelled DNA leads to the appearance of Alexa Fluor® emission at 555 nm. Since in these experiments excitation was at 295 nm, a wavelength where Alexa Fluor® excitation is close to zero [43], we infer that the 555 nm emission reflects radiationless FRET upon formation of the LigIII–DNA complex: from LigIII tryptophan residues to the DNA-attached Alexa Fluor® chromophores. The difference in the decrease in the LigIII tryptophan emission in the presence of Alexa Fluor® 532-labelled and unlabelled DNA fragments represents the efficiency of the FRET process and is calculated to be ~25% (Table 2). The efficiency of energy transfer depends upon several parameters, including the extent of spectral overlap in the emission spectrum of the donor (LigIII tryptophan residues) and the absorption spectrum of the acceptor (Alexa Fluor® 532–DNA), as determined by the overlap integral (JAD). Table 2 lists the JAD values calculated from the above experiment, as well as the derived R0 distance (2.0 nm). From these values, an average distance (r) of 2.5 nm is calculated between the donor tryptophan and the acceptor Alexa Fluor® 532 chromophores. Alexa Fluor® emission is detectable here at DNA concentrations above 100 nM, suggesting that the tryptophan residues contributing to FRET are close to the lower-affinity DNA-binding site of LigIII.

Table 2
FRET data

E is the FRET efficiency, JAD is the overlapping integral between the donor emission and acceptor absorption, R0 is the donor–acceptor distance for a 50% probability of energy transfer, and r is the average donor–acceptor distance.

Donor–acceptor pairE (%)JAD (cm3·M−1)R0 (nm)r (nm)
Tryptophan (LigIIIβ)–Alexa Fluor® (DNA) 25 3.3×10−14 2.0 2.5 
Tryptophan (LX)–Alexa Fluor® (DNA) 36 3.2×10−14 2.3 2.5 
DNS-aziridine (Ku)–Alexa Fluor® (DNA) 25 2.7×10−13 4.6 5.6 
Tryptophan (Ku)–DNS-aziridine (Ku) 70 4.5×10−15 1.8 1.6 
Donor–acceptor pairE (%)JAD (cm3·M−1)R0 (nm)r (nm)
Tryptophan (LigIIIβ)–Alexa Fluor® (DNA) 25 3.3×10−14 2.0 2.5 
Tryptophan (LX)–Alexa Fluor® (DNA) 36 3.2×10−14 2.3 2.5 
DNS-aziridine (Ku)–Alexa Fluor® (DNA) 25 2.7×10−13 4.6 5.6 
Tryptophan (Ku)–DNS-aziridine (Ku) 70 4.5×10−15 1.8 1.6 

PARP-1 has a single DNA-binding site and can displace DNA-bound LigIII

Biochemical and genetic evidence suggests that LigIII and PARP-1 are interacting partners co-operating in the repair of SSBs and DSBs [44]. We studied changes in PARP-1 conformation upon DNA binding, presumably through its ZnF domains (Figure 3A), by analysing changes in tryptophan emission. Both Alexa Fluor® 532-labelled (results not shown) as well as unlabelled (Figure 3B) DNA decreased PARP-1 tryptophan emission by up to 40%. There was no difference in the effect between Alexa Fluor® 532-labelled and unlabelled DNA. The DNA-mediated decrease in tryptophan emission is compatible with the presence of a single high-affinity DNA-binding site in the PARP-1 molecule (Kd ~2×10−8 M) (Figure 3C). Notably, the affinity of PARP-1 for DNA is similar to that of the higher-affinity DNA-binding site of LigIII.

Interactions of PARP-1 with DNA and LigIII

Figure 3
Interactions of PARP-1 with DNA and LigIII

(A) Domain organization of PARP-1. Tryptophan residues in the amino acid sequence of the enzyme are indicated. aa, amino acids. (B) Tryptophan emission spectra of LigIII, and of LigIII complexes with Alexa Fluor® 532-labelled DNA (DNA-AF532) in the presence or absence of PARP-1. Fluorescence spectra were corrected for background fluorescence of the buffer, or of the buffer containing unlabelled or Alexa Fluor® 532-labelled DNA, and are recorded in arbitrary units (a.u.). (C) Reduction of tryptophan PARP-1 emission as a function of DNA concentration suggests the presence of one DNA-binding site (Kd ~2×10−8 M).

Figure 3
Interactions of PARP-1 with DNA and LigIII

(A) Domain organization of PARP-1. Tryptophan residues in the amino acid sequence of the enzyme are indicated. aa, amino acids. (B) Tryptophan emission spectra of LigIII, and of LigIII complexes with Alexa Fluor® 532-labelled DNA (DNA-AF532) in the presence or absence of PARP-1. Fluorescence spectra were corrected for background fluorescence of the buffer, or of the buffer containing unlabelled or Alexa Fluor® 532-labelled DNA, and are recorded in arbitrary units (a.u.). (C) Reduction of tryptophan PARP-1 emission as a function of DNA concentration suggests the presence of one DNA-binding site (Kd ~2×10−8 M).

Addition of PARP-1 to pre-formed complexes of LigIII with Alexa Fluor® 532-labelled DNA changes markedly the spectra of both tryptophan and Alexa Fluor® emissions, consistent with changes in LigIII–DNA interactions (Figure 3B). Specifically, addition of PARP-1 abolishes the decrease in the tryptophan emission normally occurring upon binding of DNA to LigIII, as well as the associated FRET emission. These changes are consistent with a shift in the DNA–LigIII equilibrium towards dissociation and are intriguing because PARP-1 has a Kd value of ~2×10−8 M for DNA which is similar to the Kd value of ~3×10−8 M for the high-affinity DNA-binding site of LigIII, but 10-fold lower than the Kd value of ~2×10−7 M for the lower-affinity DNA-binding site of LigIII. These data suggest mechanistic interactions between the two proteins when bound to DNA and are consistent with the reported formation of DNA–PARP-1 binary and/or DNA–PARP-1–LigIII tertiary complexes [44].

LX binding to DNA

Binding of unlabelled DNA to 100 nM LX, most likely through the DBD of the ligase (Figure 4A), causes a concentration-dependent reduction in tryptophan emission by up to ~20% (Figures 4B and 4C). The result suggests the presence of a single DNA-binding site with a Kd value of ~1.2×10−7 M, a value close to that of the lower-affinity DNA-binding site of LigIII. Similar experiments carried out with Alexa Fluor® 532-labelled DNA show that Alexa Fluor® 532-labelled DNA induces additional reduction of LX tryptophan fluorescence by 36% (Figure 4B). This reduction is accompanied by Alexa Fluor® emission at 555 nm, again suggesting FRET between the LX tryptophan residues and the Alexa Fluor® 532 chromophore in the DNA. Using the QTrp of 0.025 for LX, as well as the overlap integral (JAD=3.25×10−14 cm3·M−1), a critical distance (R0) of 2.3 nm and an average distance (r) of 2.5 nm are calculated between tryptophan and Alexa Fluor® chromophores (Table 2). Notably, this distance is identical with that calculated between tryptophan and Alexa Fluor® chromophores in the LigIII–DNA complex (Table 2). We infer that, in both ligases, conserved tryptophan residue(s) located near the bound DNA molecule are involved in the observed FRET processes.

Binding of DNA to LX

Figure 4
Binding of DNA to LX

(A) Domain organization of LigIV and XRCC4 constituting the LX complex. Positions of all tryptophan residues in the amino acid sequences of the two proteins are indicated. The catalytic Lys273 of LigIV, which binds the AMP cofactor is also shown (K273). Residues of the amino acid sequence of XRCC4 that are involved in binding of LigIV are underlined. aa, amino acids. (B) Tryptophan emission of LX decreases upon saturation of DNA-binding sites with unlabelled or Alexa Fluor® 532-labelled DNA (DNA-AF532). Emission at 555 nm indicates FRET from tryptophan residues of LX to Alexa Fluor® chromophores of DNA. Fluorescence spectra were corrected for background fluorescence of the buffer, or of the buffer containing unlabelled or Alexa Fluor® 532-labelled DNA, and are recorded in arbitrary units (a.u.). (C) Changes in LX tryptophan emission upon binding to DNA indicate the presence of one DNA-binding site (Kd ~1.2×10−7 M).

Figure 4
Binding of DNA to LX

(A) Domain organization of LigIV and XRCC4 constituting the LX complex. Positions of all tryptophan residues in the amino acid sequences of the two proteins are indicated. The catalytic Lys273 of LigIV, which binds the AMP cofactor is also shown (K273). Residues of the amino acid sequence of XRCC4 that are involved in binding of LigIV are underlined. aa, amino acids. (B) Tryptophan emission of LX decreases upon saturation of DNA-binding sites with unlabelled or Alexa Fluor® 532-labelled DNA (DNA-AF532). Emission at 555 nm indicates FRET from tryptophan residues of LX to Alexa Fluor® chromophores of DNA. Fluorescence spectra were corrected for background fluorescence of the buffer, or of the buffer containing unlabelled or Alexa Fluor® 532-labelled DNA, and are recorded in arbitrary units (a.u.). (C) Changes in LX tryptophan emission upon binding to DNA indicate the presence of one DNA-binding site (Kd ~1.2×10−7 M).

Interactions of Ku with DNA are modified by LX

Previous studies have shown that binding to DNA reduces the tryptophan fluorescence of Ku by 15–20% [39,45]. In the present study, we analysed the effect of LX on Ku–DNA complexes using DNS-aziridine-labelled Ku and unlabelled LX which allowed us to better discriminate the effect of DNA binding on the conformation of Ku and LX. Since the tryptophan emission overlaps with the DNS-aziridine absorption spectrum, and the DNS-aziridine emission overlaps with the Alexa Fluor® 532 absorption, the labelling of Ku with DNS-aziridine creates two appropriate donor–acceptor pairs: one for intramolecular FRET (from tryptophan to DNS-aziridine within Ku) and a second pair of DNS-aziridine–Alexa Fluor® chromophores on the DNA, suitable for intermolecular FRET in Ku–DNA complexes. Additionally, since DNS-aziridine emission is highly responsive to the polarity of the environment [46], it allows the monitoring of Ku–DNA complex formation through changes in the DNS-aziridine fluorescence of Ku.

SDS/PAGE analysis of DNS-aziridine-labelled Ku shows that predominantly the Ku80 subunit is labelled (Figure 5A). DNS-aziridine-labelled Ku retains biological activity as indicated by the DNA binding seen in EMSAs (Figure 5B), albeit at levels lower than that of the unlabelled control. This may derive from some structural distortions induced in Ku by the substitution of the hydrogen atom of the cysteine thiol group with the bulkier dansylaziridine molecule. Indeed, it has been reported that oxidation of cysteine residues inhibits the DNA-binding activity of Ku [39,45,47,48]. Yet, residual activity is robust for further experiments.

Binding of DNA to DNS-aziridine-labelled Ku

Figure 5
Binding of DNA to DNS-aziridine-labelled Ku

(A) DNS-aziridine-labelled Ku analysed by SDS/PAGE. As visualized under UV, Ku80 shows stronger DNS-aziridine emission than Ku70, suggesting that Ku80 is the subunit predominantly labelled with DNS-aziridine. Coomassie Blue staining shows equal amounts of Ku70 and Ku80 in DNS-aziridine-labelled Ku. (B) EMSA with unlabelled and DNS-aziridine-labelled Ku (Ku-DNS). (C) DNS-aziridine emission spectra of DNS-aziridine-labelled Ku (Ku-DNS) (excitation at 350 nm) and of complexes of DNS-aziridine–Ku with unlabelled or Alexa Fluor® 532-labelled DNA. LX induced changes in spectra of complexes of DNS-aziridine–Ku with Alexa Fluor® 532–DNA (DNA-AF532) reflecting conformational changes in Ku upon LX binding. Fluorescence spectra were corrected for background fluorescence of the buffer, or of the buffer containing unlabelled or Alexa Fluor® 532-labelled DNA and are recorded in arbitrary units (a.u.).

Figure 5
Binding of DNA to DNS-aziridine-labelled Ku

(A) DNS-aziridine-labelled Ku analysed by SDS/PAGE. As visualized under UV, Ku80 shows stronger DNS-aziridine emission than Ku70, suggesting that Ku80 is the subunit predominantly labelled with DNS-aziridine. Coomassie Blue staining shows equal amounts of Ku70 and Ku80 in DNS-aziridine-labelled Ku. (B) EMSA with unlabelled and DNS-aziridine-labelled Ku (Ku-DNS). (C) DNS-aziridine emission spectra of DNS-aziridine-labelled Ku (Ku-DNS) (excitation at 350 nm) and of complexes of DNS-aziridine–Ku with unlabelled or Alexa Fluor® 532-labelled DNA. LX induced changes in spectra of complexes of DNS-aziridine–Ku with Alexa Fluor® 532–DNA (DNA-AF532) reflecting conformational changes in Ku upon LX binding. Fluorescence spectra were corrected for background fluorescence of the buffer, or of the buffer containing unlabelled or Alexa Fluor® 532-labelled DNA and are recorded in arbitrary units (a.u.).

The DNS-aziridine fluorescence of Ku has a maximum at 507 nm (Figure 5C), indicative of polar DNS-aziridine environment [49], and an emission quantum yield of 0.22. The value of the KQ constant (12.4 M−1) calculated from the quenching of DNS-aziridine by acrylamide (Figure 1D) suggests solvent accessibility of the chromophores, which is consistent with surface-localization of the labelled cysteine residues. Additionally, it can be deduced that the DNS-aziridine-labelled cysteine residues are located close to the emitting tryptophan residues of Ku, since we observed a strong quenching of tryptophan emission in DNS-aziridine-labelled Ku approaching 70% (Figure 1B). This is indicative of intramolecular FRET from tryptophan to DNS-aziridine chromophores of Ku. By calculating JAD=4.5×10−15 cm3·M−1, QD=0.015 and 70% efficiency for the FRET process, an R0 of 1.8 nm and an average distance (r) of 1.6 nm between tryptophan and DNS-aziridine chromophores are estimated (Table 2).

Although binding of unlabelled Ku to DNA is associated with a reduction in tryptophan emission (see above), binding of DNS-aziridine-labelled Ku to DNA does not further change the residual tryptophan emission of Ku (results not shown), but induces a 20–25% decrease in DNS-aziridine emission, reflecting conformational changes in the molecule (Figure 5C). Use of Alexa Fluor® 532-labelled DNA causes an additional 25% quenching of DNS-aziridine emission and generates a shoulder at 555 nm in the DNS-aziridine fluorescence spectrum of Ku. This implies a dipole–dipole coupling between DNS-aziridine and Alexa Fluor® chromophores. For the DNS-aziridine chromophores of Ku and the Alexa Fluor® 532 dye on the DNA, we calculated values of 4.7 (R0) and 5.6 nm (r), when taking into account a 25% FRET efficiency (Table 2).

Binding of LX to preformed complexes of DNS-aziridine–Ku with Alexa Fluor® 532–DNA is accompanied by an additional ~10% quenching of DNS-aziridine emission (Figure 5C) and a slight increase in the Alexa Fluor® emission shoulder at 555 nm in the DNS-aziridine fluorescence spectrum. These changes reflect conformational changes in the Ku heterodimer upon binding of LX. Conformational changes of Ku may mediate its relocation at the DNA ends, as indicated by the shorter distance of 5.2 nm (r) which was calculated between the DNS-aziridine and Alexa Fluor® chromophores in the tertiary complexes.

DISCUSSION

Fluorescence spectroscopy offers an attractive alternative for the study of dynamic conformational transitions associated with the engagement of proteins in distinct functional steps within a metabolic pathway. In the present study, we took advantage of this technology to examine conformational transitions associated with DNA binding and thus engagements in DNA repair for Ku, LigIII, LigIV and PARP-1.

Our results with LigIII document two non-equivalent DNA-binding sites, operating sequentially with high and low affinity respectively. Taking into account the domain organization of LigIII [21,24], we suggest that the high-affinity DNA-binding site comprises the ZnF motif and the DBD, whereas the low-affinity DNA-binding site comprises the CC domain. Consistent with this postulate, the N-terminal ZnF domains of PARP-1, which are structurally related to the ZnF domain of LigIII [29], show high DNA-binding affinity, similar to that of the ZnF–DBD module of LigIII. Notably, in the LX complex, LigIV, which lacks ZnF domains, displays low DNA-binding affinity (Kd ~1.2×10−7 M), similar to that of the CC module of LigIII (Kd ~2×10−7 M). These observations are compatible with known functional properties of DNA ligases III and IV and for LigIII support the ‘jackknife model’ that elegantly explains the specialized functions of the enzyme [21]. It should be noted that tryptophan residues of XRCC4 probably contribute to the overall intrinsic fluorescence of LX. XRCC4 contains three tryptophan residues (Figure 4A, PDB code 3SR2) which are not a part of its DNA-binding site, as shown recently by Hammel et al. [50]. Notably, the same authors report that XRCC4 displays an order of magnitude lower affinity to DNA (Kd ~10−6 M) than the LX complex. Thus, if some reduction of tryptophan emission of XRCC4 upon DNA binding contributes to the reduction of the tryptophan emission of the LX complex, the effect should be observed at 10-fold higher DNA concentrations than those applied in the present study.

Results of tryptophan fluorescence changes and the documented intermolecular tryptophan–Alexa Fluor® dipole–dipole coupling to generate FRET indicate intriguing interactions between DNA and LigIII/LigIV that are accompanied by conformational transitions. Since with both DNA ligases, FRET occurs at the same average distance of 2.5 nm, we infer the involvement of conserved tryptophan residues in the effect. From the published crystal structure of LigI [23], LigIII Trp673 and LigIV Trp447 are identified as potential donors in the tryptophan–Alexa Fluor® FRET (Figures 2A and 4A, PDB code 1X9N). These tryptophan residues are highly conserved in all three human DNA ligase families. In the crystal structure of LigI [23], Trp742, which corresponds to LigIII Trp673 and LigIV Trp447, is located near the AMP-binding lysine residue of the active site and is a likely donor in the observed tryptophan–Alexa Fluor® coupling. However, contribution from similarly conserved LigIII Trp816 and LigIV Trp590 in the catalytic modules of the enzymes cannot be ruled out. Moreover, our average distance estimates use the average FRET efficiency calculated from quenching measured for total protein emission. This value sums the collective contributions of all emitting tryptophan residues. However, because tryptophan residues involved in FRET are selectively quenched, we anticipate that FRET efficiency between principal transactors will actually be significantly higher. Assuming a higher, up to 90%, quenching for LigIII Trp673 and LigIV Trp447, average distances of 1.4–1.5 nm between the participating chromophores are calculated, placing these residues very near to the ligase catalytic site.

Despite differences in the overall domain organization, the fluorescence data of the present study suggest similar positioning of DNA ends within the conserved CC domains for LigIII and LigIV. Whereas the DBD of LigIII co-operates with the ZnF domain, the DBD domain of LigIV interacts with its catalytic module [21,51]. Since the DBD domains of LigIII and LigIV lack tryptophan residues [23], we expect that both proteins will utilize the same DNA-engagement mechanism and will alter similarly their CC domains, which contain the most likely candidates for the tryptophan–Alexa Fluor® coupling (Trp673 of LigIII and Trp447 of LigIV). Thus a conserved dynamic architecture is suggested for the active sites of DNA ligases despite heterogeneity in their domain organization.

The high DNA-binding affinity of PARP-1 most certainly derives from the two N-terminal ZnF domains (Figure 3A). Like the homologous ZnF domain of LigIII, these ZnF domains contain two conserved tryptophan residues on identical positions. The first, Trp50, is located, according to NMR structural data [28], close to His52 that co-ordinates metal binding. As a result, its fluorescence may be strongly quenched upon DNA binding, causing the observed effects.

Interactions of LX with DNA and Ku are central in D-NHEJ and were probed in the present study with DNS-aziridine-labelled Ku and Alexa Fluor® 532-conjugated DNA. The DNS-aziridine chromophores quench the intrinsic tryptophan fluorescence of Ku, suggesting efficient intramolecular FRET (average distance of 1.7 nm) between tryptophan and the cysteine residues with conjugated DNS-aziridine. In the structural model of Ku80 (PDB codes 1JEQ and 1JEY), there are two clusters that contain tryptophan and cysteine residues located within a radius of 1.7 nm. The first comprises Trp247, Cys249, Cys339 and Cys418, and the second comprises Trp276, Trp513 and Cys493 [52]. Since binding of Ku to DNA is accompanied by efficient quenching of the DNS-aziridine emission, we infer that the labelled cysteine is located close to the bound DNA in Ku–DNA complexes. Accordingly, Cys249, Cys339 and Cys493 are the most likely candidates for conjugation with DNS-aziridine. These residues are in suitable positions to quench tryptophan emission and to report conformational transitions of the protein upon DNA binding. We infer that Cys493 might be one of the labelled residues because this residue is not directly involved in DNA binding [53] and we also failed to register profound changes in Ku–DNA-binding activity upon DNS-aziridine conjugation.

The FRET signal from the interactions between DNS-aziridine-labelled Ku and Alexa Fluor® 532-labelled-DNA suggests an average distance (r) between DNS-aziridine and Alexa Fluor® 532 of ~5.7 nm. This distance is consistent with a localization of Ku at the very end of the DNA molecule. Considering that, on average, half of the Ku molecules are bound to the unlabelled end of the DNA, the real efficiency of FRET exceeds by a factor of 2 that of 25% which was calculated from quenching of DNS-aziridine by Alexa Fluor® (Table 2). Consequently, the actual distance between DNS-aziridine and Alexa Fluor® 532 should be shorter and close to R0=4.6 nm (Table 2). If we also consider that the free DNA end lies on the Ku70 side [52] and that the Ku80 is predominantly labelled with DNS-aziridine, we can confirm that Ku localizes at the very end of the DNA molecule, as expected from the function of the protein in the initiation, control and regulation of D-NHEJ [16].

Whereas recruitment of Ku to DNA ends does not require other proteins and starts within seconds of DSB induction [54], recruitment of LX to DSBs depends on Ku. Studied by FRET, the interactions between LX and DNA-bound Ku enhance somewhat FRET between DNS-aziridine-labelled Ku and Alexa Fluor® 532-labelled DNA. This suggests a shift in the position of the DNS-aziridine chromophores towards the DNA end, probably driven by conformational changes in Ku upon LX binding.

Abbreviations

     
  • ATM

    ataxia telangiectasia mutated

  •  
  • BRCT

    BRCA1 (breast cancer early-onset 1) C-terminal

  •  
  • CC

    catalytic core

  •  
  • DBD

    DNA-binding domain

  •  
  • DNA-PK

    DNA-dependent protein kinase

  •  
  • DNS-aziridine

    5-dimethylaminonaphthalene-1-sulfonyl aziridine, also known as dansylaziridine

  •  
  • DSB

    double-strand break

  •  
  • DTT

    dithiothreitol

  •  
  • EMSA

    electrophoretic mobility-shift assay

  •  
  • FRET

    Förster resonance energy transfer

  •  
  • KQ

    bimolecular quenching constant

  •  
  • LigI

    DNA ligase I

  •  
  • LigIII

    DNA ligase III

  •  
  • LigIV

    DNA ligase IV

  •  
  • NHEJ

    non-homologous end-joining

  •  
  • B-NHEJ

    back-up-NHEJ

  •  
  • D-NHEJ

    DNA-PK-dependent NHEJ

  •  
  • PAR

    poly(ADP-ribose) polymer

  •  
  • PARP-1

    poly(ADP-ribose) polymerase-1

  •  
  • QDNS

    relative DNS-aziridine–Ku emission quantum yield

  •  
  • QTrp

    relative tryptophan emission quantum yield

  •  
  • SSB

    single-strand break

  •  
  • XRCC

    X-ray cross-complementation group

  •  
  • LX

    LigIV–XRCC4

  •  
  • ZnF

    zinc-finger

AUTHOR CONTRIBUTION

Katia Stankova and Katia Ivanova conducted the experiments and processing the data. Bustanur Rosidi and Aparna Sharma purified the proteins. Emil Mladenov assisted in assay development and preparation of the paper. Rayna Boteva interpreted the data and, together with George Iliakis, designed the study and wrote the paper.

FUNDING

This work was supported by Deutscher Akademischer Austauschdienst (DAAD) [grant number 50512704], the European Space Agency (ESA) AO-08-IBER [grant number BMWWi-50WB0929] and the Bundesministerium für Bildung und Forschung (BMBF) [grant numbers 02NUK001B and 02NUK005C].

References

References
1
Valerie
 
K.
Povirk
 
L. F.
 
Regulation and mechanisms of mammalian double-strand break repair
Oncogene
2003
, vol. 
22
 (pg. 
5792
-
5812
)
2
Jackson
 
S. P.
 
Sensing and repairing DNA double-strand breaks
Carcinogenesis
2002
, vol. 
23
 (pg. 
687
-
696
)
3
Khanna
 
K. K.
Jackson
 
S. P.
 
DNA double-strand breaks: signaling, repair and the cancer connection
Nat. Genet.
2001
, vol. 
27
 (pg. 
247
-
254
)
4
Riballo
 
E.
Woodbine
 
L.
Stiff
 
T.
Walker
 
S. A.
Goodarzi
 
A. A.
Jeggo
 
P. A.
 
XLF-Cernunnos promotes DNA ligase IV–XRCC4 re-adenylation following ligation
Nucleic Acids Res.
2009
, vol. 
37
 (pg. 
482
-
492
)
5
Buck
 
D.
Malivert
 
L.
de Chasseval
 
R.
Barraud
 
A.
Fondaneche
 
M. C.
Sanal
 
O.
Plebani
 
A.
Stephan
 
J. L.
Hufnagel
 
M.
le Deist
 
F.
, et al 
Cernunnos, a novel nonhomologous end-joining factor, is mutated in human immunodeficiency with microcephaly
Cell
2006
, vol. 
124
 (pg. 
287
-
299
)
6
Perrault
 
R.
Wang
 
H.
Wang
 
M.
Rosidi
 
B.
Iliakis
 
G.
 
Backup pathways of NHEJ are suppressed by DNA-PK
J. Cell. Biochem.
2004
, vol. 
92
 (pg. 
781
-
794
)
7
Wang
 
H.
Rosidi
 
B.
Perrault
 
R.
Wang
 
M.
Zhang
 
L.
Windhofer
 
F.
Iliakis
 
G.
 
DNA ligase III as a candidate component of backup pathways of nonhomologous end joining
Cancer Res.
2005
, vol. 
65
 (pg. 
4020
-
4030
)
8
Wang
 
M.
Wu
 
W.
Rosidi
 
B.
Zhang
 
L.
Wang
 
H.
Iliakis
 
G.
 
PARP-1 and Ku compete for repair of DNA double strand breaks by distinct NHEJ pathways
Nucleic Acids Res.
2006
, vol. 
34
 (pg. 
6170
-
6182
)
9
McVey
 
M.
Lee
 
S. E.
 
MMEJ repair of double-strand breaks (director's cut): deleted sequences and alternative endings
Trends Genet.
2008
, vol. 
24
 (pg. 
529
-
538
)
10
Audebert
 
M.
Salles
 
B.
Calsou
 
P.
 
Involvement of poly(ADP-ribose) polymerase-1 and XRCC1/DNA ligase III in an alternative route for DNA double-strand breaks rejoining
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
55117
-
55126
)
11
Iliakis
 
G.
 
Backup pathways of NHEJ in cells of higher eukaryotes: cell cycle dependence
Radiother. Oncol.
2009
, vol. 
92
 (pg. 
310
-
315
)
12
Iliakis
 
G.
Wang
 
H.
Perrault
 
A. R.
Boecker
 
W.
Rosidi
 
B.
Windhofer
 
F.
Wu
 
W.
Guan
 
J.
Terzoudi
 
G.
Pantelias
 
G.
 
Mechanisms of DNA double strand break repair and chromosome aberration formation
Cytogenet. Genome Res.
2004
, vol. 
104
 (pg. 
14
-
20
)
13
Iliakis
 
G.
Wu
 
W.
Wang
 
M.
Terzoudi
 
G. I.
Pantelias
 
G. E.
 
Obe
 
G.
Vijayalaxmi
 
 
Backup pathways of nonhomologous end joining may have a dominant role in the formation of chromosome aberrations
Chromosomal Alterations
2007
Berlin
Springer Verlag
(pg. 
67
-
85
)
14
Mladenov
 
E.
Iliakis
 
G.
 
Induction and repair of DNA double strand breaks: the increasing spectrum of non-homologous end joining pathways
Mutat. Res. Fundam. Mol. Mech. Mutagen.
2011
, vol. 
711
 (pg. 
61
-
72
)
15
Corneo
 
B.
Wendland
 
R. L.
Deriano
 
L.
Cui
 
X.
Klein
 
I. A.
Wong
 
S.-Y.
Arnal
 
S.
Holub
 
A. J.
Weller
 
G. R.
Pancake
 
B. A.
, et al 
Rag mutations reveal robust alternative end joining
Nature
2007
, vol. 
449
 (pg. 
483
-
486
)
16
Fattah
 
F.
Lee
 
E. H.
Weisensel
 
N.
Wang
 
Y.
Lichter
 
N.
Hendrickson
 
E. A.
 
Ku regulates the non-homologous end joining pathway choice of DNA double-strand break repair in human somatic cells
PLoS Genet.
2010
, vol. 
6
 pg. 
e1000855
 
17
Abraham
 
R. T.
 
Cell cycle checkpoint signaling through the ATM and ATR kinases
Genes Dev.
2001
, vol. 
15
 (pg. 
2177
-
2196
)
18
Tomimatsu
 
N.
Tahimic
 
C.G.T.
Otsuki
 
A.
Burma
 
S.
Fukuhara
 
A.
Sato
 
K.
Shiota
 
G.
Oshimura
 
M.
Chen
 
D. J.
Kurimasa
 
A.
 
Ku70/80 modulates ATM and ATR signaling pathways in response to DNA double strand breaks
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
10138
-
10145
)
19
Hochegger
 
H.
Dejsuphong
 
D.
Fukushima
 
T.
Morrison
 
C.
Sonoda
 
E.
Schreiber
 
V.
Zhao
 
G. Y.
Saberi
 
A.
Masutani
 
M.
Adachi
 
N.
, et al 
Parp-1 protects homologous recombination from interference by Ku and Ligase IV in vertebrate cells
EMBO J.
2006
, vol. 
25
 (pg. 
1305
-
1314
)
20
Haince
 
J.-F.
McDonald
 
D.
Rodrigue
 
A.
Dery
 
U.
Masson
 
J.-Y.
Hendzel
 
M. J.
Poirier
 
G. G.
 
PARP1-dependent kinetics of recruitment of MRE11 and NBS1 proteins to multiple DNA damage sites
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
1197
-
1208
)
21
Cotner-Gohara
 
E.
Kim
 
I. K.
Tomkinson
 
A. E.
Ellenberger
 
T.
 
Two DNA-binding and nick recognition modules in human DNA ligase III
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
10764
-
10772
)
22
Martin
 
I. V.
MacNeill
 
S. A.
 
ATP-dependent DNA ligases
Genome Biol.
2002
, vol. 
3
 pg. 
REVIEWS3005
 
23
Pascal
 
J. M.
O'Brien
 
P. J.
Tomkinson
 
A. E.
Ellenberger
 
T.
 
Human DNA ligase I completely encircles and partially unwinds nicked DNA
Nature
2004
, vol. 
432
 (pg. 
473
-
478
)
24
Ellenberger
 
T.
Tomkinson
 
A. E.
 
Eukaryotic DNA ligases: structural and functional insights
Annu. Rev. Biochem.
2008
, vol. 
77
 (pg. 
313
-
338
)
25
Nash
 
R. A.
Caldecott
 
K. W.
Barnes
 
D. E.
Lindahl
 
T.
 
XRCC1 protein interacts with one of two distinct forms of DNA ligase III
Biochemistry
1997
, vol. 
36
 (pg. 
5207
-
5211
)
26
Taylor
 
R. M.
Wickstead
 
B.
Cronin
 
S.
Caldecott
 
K. W.
 
Role of a BRCT domain in the interaction of DNA ligase IIIα with the DNA repair protein XRCC1
Curr. Biol.
1998
, vol. 
8
 (pg. 
877
-
880
)
27
Mackey
 
Z. B.
Ramos
 
W.
Levin
 
D. S.
Walter
 
C. A.
McCarrey
 
J. R.
Tomkinson
 
A. E.
 
An alternative splicing event which occurs in mouse pachytene spermatocytes generates a form of DNA ligase III with distinct biochemical properties that may function in meiotic recombination
Mol. Cell. Biol.
1997
, vol. 
17
 (pg. 
989
-
998
)
28
Kulczyk
 
A. W.
Yang
 
J.-C.
Neuhaus
 
D.
 
Solution structure and DNA binding of the zinc-finger domain from DNA ligase IIIα
J. Mol. Biol.
2004
, vol. 
341
 (pg. 
723
-
738
)
29
Mackey
 
Z. B.
Niedergang
 
C.
Menissier-de Murcia
 
J.
Leppard
 
J.
Au
 
K.
Chen
 
J.
de Murcia
 
G.
Tomkinson
 
A. E.
 
DNA ligase III is recruited to DNA strand breaks by a zinc finger motif homologous to that of poly(ADP-ribose) polymerase
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
21679
-
21687
)
30
Taylor
 
A. M.
Whitehouse
 
C. J.
Cappelli
 
E.
Frosina
 
G.
Caldecott
 
K. W.
 
Role of DNA ligase III zinc finger in polynucleotide binding and ligation
Nucleic Acids Res.
1998
, vol. 
26
 (pg. 
4804
-
4810
)
31
Taylor
 
A. M.
Whitehouse
 
C. J.
Caldecott
 
K. W.
 
The DNA ligase III zinc finger stimulates binding to DNA secondary structure and promotes endjoining
Nucleic Acids Res.
2000
, vol. 
28
 (pg. 
3558
-
3563
)
32
Gradwohl
 
G.
Menissier de Murcia
 
J.
Molinete
 
M.
Simonin
 
F.
Koken
 
M.
Hoeijmakers
 
J. H. J.
de Murcia
 
G.
 
The second zinc-finger domain of poly(ADP-ribose) polymerase determines specificity for single-stranded breaks in DNA
Proc. Natl. Acad. Sci. U.S.A.
1990
, vol. 
87
 (pg. 
2990
-
2994
)
33
Menissier de Murcia
 
J.
Molinete
 
M.
Gradwohl
 
G.
Simonin
 
F.
de Murcia
 
G.
 
Zinc-binding domain of poly(ADPR-ribose) polymerase participates in the recognition of single strand breaks on DNA
J. Mol. Biol.
1989
, vol. 
210
 (pg. 
229
-
233
)
34
Wu
 
W.
Wang
 
M.
Singh
 
S. K.
Mussfeldt
 
T.
Iliakis
 
G.
 
Repair of radiation induced DNA double strand breaks by backup NHEJ is enhanced in G2
DNA Repair
2008
, vol. 
7
 (pg. 
329
-
338
)
35
Rosidi
 
B.
Wang
 
M.
Wu
 
W.
Sharma
 
A.
Wang
 
H.
Iliakis
 
G.
 
Histone H1 functions as a stimulatory factor in backup pathways of NHEJ
Nucleic Acids Res.
2008
, vol. 
36
 (pg. 
1610
-
1623
)
36
Ono
 
M.
Tucker
 
P. W.
Capra
 
J. D.
 
Production and characterization of recombinant human Ku antigen
Nucleic Acids Res.
1994
, vol. 
19
 (pg. 
3918
-
3924
)
37
Lehrer
 
S. S.
 
Solute perturbation of protein fluorescence: the quenching of the tryptophyl fluorescence of model compounds and of lysozyme by iodide ion
Biochemistry
1971
, vol. 
10
 (pg. 
3254
-
3263
)
38
Prendergast
 
F. G.
Meyer
 
M.
Carlson
 
G. L.
Iida
 
S.
Potter
 
J. D.
 
Synthesis, spectral properties, and use of 6-acryloyl-2-dimethylaminonaphthalene (Acrylodan): a thiol-selective, polarity-sensitive fluorescent probe
J. Biol. Chem.
1983
, vol. 
258
 (pg. 
7541
-
7544
)
39
Andrews
 
B. J.
Lehman
 
J. A.
Turchi
 
J. J.
 
Kinetic analysis of the Ku–DNA binding activity reveals a redox-dependent alteration in protein structure that stimulates dissociation of the Ku–DNA complex
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
13596
-
13603
)
40
Horrocks
 
W.D.W.
Snyder
 
A. P.
 
Measurement of distance between fluorescent amino acid residues and metal ion binding sites: quantitation of energy transfer between tryptophan and terbium(III) or europium(III) in thermolysin
Biochem. Biophys. Res. Commun.
1981
, vol. 
100
 (pg. 
111
-
117
)
41
Vivian
 
J. T.
Callis
 
P. R.
 
Mechanisms of tryptophan fluorescence shifts in proteins
Biophys. J.
2001
, vol. 
80
 (pg. 
2093
-
2109
)
42
Lakowicz
 
J. R.
 
Lakowicz
 
J. R.
 
Principles of fluorescence spectroscopy
Protein Fluorescence
2004
Berlin
Springer Verlag
(pg. 
446
-
485
)
43
Panchuk-Voloshina
 
N.
Haugland
 
R. P.
Bishop-Stewart
 
J.
Bhalgat
 
M. K.
Millard
 
P. J.
Mao
 
F.
Leung
 
W.-Y.
Haugland
 
R. P.
 
Alexa dyes, a series of new fluorescent dyes that yield exceptionally bright, photostable conjugates
J. Histochem. Cytochem.
1999
, vol. 
47
 (pg. 
1179
-
1188
)
44
Leppard
 
J. B.
Dong
 
Z.
Mackey
 
Z. B.
Tomkinson
 
A. E.
 
Physical and functional interaction between DNA ligase IIIα and poly(ADP-ribose) polymerase 1 in DNA single-strand break repair
Mol. Cell. Biol.
2003
, vol. 
23
 (pg. 
5919
-
5927
)
45
Arosio
 
D.
Cui
 
S.
Ortega
 
C.
Chovanec
 
M.
Di Marco
 
S.
Baldini
 
G.
Falaschi
 
A.
Vindigni
 
A.
 
Studies on the mode of Ku interaction with DNA
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
9741
-
9748
)
46
Lin
 
T.-I.
 
Fluorimetric studies of actin labeled with dansyl aziridine
Arch. Biochem. Biophys.
1978
, vol. 
185
 (pg. 
285
-
299
)
47
Ayene
 
I. S.
Stamato
 
T. D.
Mauldin
 
S. K.
Biaglow
 
J. E.
Tuttle
 
S. W.
Jenkins
 
S. F.
Koch
 
C. J.
 
Mutation in the glucose-6-phosphate dehydrogenase gene leads to inactivation of Ku DNA end binding during oxidative stress
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
9929
-
9935
)
48
Song
 
J. Y.
Lim
 
J. W.
Kim
 
H.
Morio
 
T.
Kim
 
K. H.
 
Oxidative stress induces nuclear loss of DNA repair proteins Ku70 and Ku80 and apoptosis in pancreatic acinar AR42J cells
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
36676
-
36687
)
49
Cheung
 
H. C.
Wang
 
C.-K.
Garland
 
F.
 
Fluorescence energy transfer studies of skeletal troponin C proximity between methionine-25 and cysteine-98
Biochemistry
1982
, vol. 
21
 (pg. 
5135
-
5142
)
50
Hammel
 
M.
Rey
 
M.
Yu
 
Y.
Mani
 
R. S.
Classen
 
S.
Liu
 
M.
Pique
 
M. E.
Fang
 
S.
Mahaney
 
B. L.
Weinfeld
 
M.
, et al 
XRCC4 protein interactions with XRCC4-like factor (XLF) create an extended grooved scaffold for DNA ligation and double strand break repair
J. Biol. Chem.
2011
, vol. 
286
 (pg. 
32638
-
32650
)
51
Cotner-Gohara
 
E.
Kim
 
I. K.
Hammel
 
M.
Tainer
 
J. A.
Tomkinson
 
A. E.
Ellenberger
 
T.
 
Human DNA ligase III recognizes DNA ends by dynamic switching between two DNA-bound states
Biochemistry
2010
, vol. 
49
 (pg. 
6165
-
6176
)
52
Walker
 
J. R.
Corpina
 
R. A.
Goldberg
 
J.
 
Structure of the Ku heterodimer bound to DNA and its implications for double-strand break repair
Nature
2001
, vol. 
412
 (pg. 
607
-
614
)
53
Bennett
 
S.
Neher
 
T.
Shatilla
 
A.
Turchi
 
J.
 
Molecular analysis of Ku redox regulation
BMC Mol. Biol.
2009
, vol. 
10
 pg. 
86
 
54
Mari
 
P.-O.
Florea
 
B. I.
Persengiev
 
S. P.
Verkaik
 
N. S.
Bruggenwirth
 
H. T.
Modesti
 
M.
Giglia-Mari
 
G.
Bezstarosti
 
K.
Demmers
 
J. A. A.
Luider
 
T. M.
, et al 
Dynamic assembly of end-joining complexes requires interaction between Ku70/80 and XRCC4
Proc. Natl. Acad. Sci. U.S.A.
2006
, vol. 
103
 (pg. 
18597
-
18602
)

Author notes

1

Present address: University Technology Malaysia, Institute of Bioproduct Development, Johor Bahru, Malaysia.