The unusual, heterodimeric topoisomerase IB of Leishmania shows functional activity upon reconstitution of the DNA-binding large subunit (LdTOPIL; or L) and the catalytic small subunit (LdTOPIS; or S). In the present study, we generated N- and C-terminal-truncated deletion constructs of either subunit and identified proteins LdTOPIL39–456 (lacking amino acids 1–39 and 457–635) and LdTOPIS210–262 (lacking amino acids 1–210) as the minimal interacting fragments. The interacting region of LdTOPIL lies between residues 40–99 and 435–456, while for LdTOPIS it lies between residues 210–215 and 245–262. The heterodimerization between the two fragments is weak and therefore co-purified fragments showed reduced DNA binding, cleavage and relaxation properties compared with the wild-type enzyme. The minimal fragments could complement their respective wild-type subunits inside parasites when the respective subunits were down-regulated by transfection with conditional antisense constructs. Site-directed mutagenesis studies identify Lys455 of LdTOPIL and Asp261 of LdTOPIS as two residues involved in subunit interaction. Taken together, the present study provides crucial insights into the mechanistic details for understanding the unusual structure and inter-subunit co-operativity of this heterodimeric enzyme.

INTRODUCTION

DNA topoisomerases relieve the torsional strain in DNA that builds up during vital cellular processes. Type IB enzymes nick one DNA strand, form a 3′-phosphotyrosyl link and swivel another strand across the nick by a ‘controlled rotation’ mechanism [1]. The advent of bi-subunit topoisomerase IB in kinetoplastids is a paradigm shift in the type IB family. The DNA-binding large subunit with the ‘VAILCNH’ motif associates with the catalytic small subunit harbouring the consensus ‘SKXXY’ motif to form an active enzyme within the parasite [2,3]. RNAi (RNA interference) of one subunit causes concomitant degradation of the other protein subunit [4]. CPTs (camptothecins) are uncompetitive inhibitors of topoisomerase IB, which trap the enzyme–DNA cleavable complex, stall replication and transcription fork and thereby cause cell death [1].

In monomeric human topoisomerase IB, active complementation of ‘cap’ and ‘catalytic’ domains (linked by a dispensable ‘linker’ domain) and formation of a fully functional enzyme upon deletion of the N-terminal 214 amino acids have been documented [5,6]. A 1:1 molar interaction of two subunits in vitro forms the active LdTOPIL/S (or L/S), which differs from monomeric rat liver topoisomerase I in having reduced affinity for DNA and reduced processivity [7]. Catalytic activity is derived chiefly from five conserved amino acids [8]: Arg314, Lys352, Arg410 and His453 from LdTOPIL (or L) and Tyr222 from LdTOPIS (or S).

The bi-subunit Leishmania topoisomerase IB also harbours stretches of residues in the two subunits, which are dispensable for activity. Our previous studies show that residues 1–39 of L modulate the non-covalent interaction with DNA, while residues 39–99 influence the interaction of the two subunits [9]. In the present study, we characterize the minimal, functionally indispensable core domains of this heterodimer by generating several deletion constructs of each subunit and studying their interaction with each other. Our findings reveal that LdTOPIL39–456 (ΔL) and LdTOPIS210–262 (ΔS) are the minimal interacting fragments capable of carrying out topoisomerization upon reconstitution and could functionally complement their respective wild-type counterparts in vivo. Two residues have also been identified that play a crucial role in heterodimerization of the two subunits. Taken together the present study provides, for the first time, insight into the mechanistic details of functional subunit interaction.

MATERIALS AND METHODS

Construction of recombinant plasmids, overexpression and purification

LdTOPIL and its deletion constructs were cloned in the BamHI/HindIII site of pET28c, while LdTOPIS and its deletion constructs were cloned in the BamHI/EcoRI site of pGEX-5X2. All constructs were transformed into BL21(DE3)pLysS. For reconstituted protein purification, co-transformation was carried out in combinations as stated in Supplementary Table 1 (http://www.BiochemJ.org/bj/409/bj4090481add.htm). Proteins were purified through Ni-NTA (Ni2+-nitrilotriacetate)–agarose [9] or GST (glutathione transferase)–Sepharose 4B column followed by phosphocellulose column (P11 cellulose; Whatman). See Supplementary information I at http://www.BiochemJ.org/bj/409/bj4090481add.htm for more details.

Far-Western analysis

Purified L and its deletion constructs (Figure 1A) were electrophoresed in SDS/12% PAGE gel and transferred on to a PVDF membrane (two sets). The control set was blotted using polyclonal anti-L antibody. The other membrane was rinsed in 1×PBS, washed four times (1 h each) with renaturation buffer [50 mM Tris/HCl, pH 7.5, 50 mM KCl, 5 mM MgCl2, 0.5 mM EDTA, 1 mM DTT (dithiothreitol), 5% (v/v) glycerol and 0.05% Tween 20] containing decreasing concentrations of guanidinium chloride (6, 4, 2 and 1 M) at 4 °C and blocked with 5% (w/v) BSA (2 h) in renaturation buffer [10]. Next, the membrane was incubated overnight at 4 °C with bait protein, S, in renaturation buffer and thereafter blotted using anti-S antibody. Similar experiments were carried out with S and its deletion constructs. One blot was incubated with L as bait and immunoblotted with anti-L antibody. The control set was blotted with anti-S. For all blots, AP (alkaline phosphatase)-conjugated anti-rabbit secondary antibody was used.

Schematic diagram of recombinant constructs and purification of proteins

Figure 1
Schematic diagram of recombinant constructs and purification of proteins

Schematic representation of wild-type and deletion constructs of LdTOPIL (A, left panel) and LdTOPIS (B, left panel). The right panels show Coomassie Brilliant Blue-stained SDS/12% PAGE of recombinant proteins. Proteins were loaded as shown in the left panel.

Figure 1
Schematic diagram of recombinant constructs and purification of proteins

Schematic representation of wild-type and deletion constructs of LdTOPIL (A, left panel) and LdTOPIS (B, left panel). The right panels show Coomassie Brilliant Blue-stained SDS/12% PAGE of recombinant proteins. Proteins were loaded as shown in the left panel.

Co-immobilization assay

6×His-tagged L and LdTOPIL39–456 (ΔL) were co-expressed separately with GST-tagged S and LdTOPIS210–262 (ΔS) and purified through Ni-NTA–agarose or GST–Sepharose 4B. Elution fractions were run in SDS/12% PAGE (three sets). One set was stained with Coomassie Brilliant Blue, while the other two sets were transferred on to nitrocellulose and blotted separately with anti-L and anti-S antibodies respectively, followed by AP-conjugated anti-rabbit secondary antibody [9].

Plasmid relaxation and equilibrium cleavage assay

Co-expressed empty vector-induced lysates and reconstituted enzymes L/S, L/ΔS, ΔL/S and ΔL/ΔS were purified through Ni-NTA–agarose followed by phosphocellulose column chromatography (final protein concentration was 0.8 mg/ml). Purified proteins and pHOT1 plasmid substrate were mixed in 1:2 molar ratios in a relaxation buffer (25 mM Tris/HCl, pH 7.5, 5% glycerol, 0.5 mM DTT, 50 mM KCl, 10 mM MgCl2, 2.5 mM EDTA and 125 μg/ml BSA) at 37 °C for the indicated time periods. Samples were run in 1% agarose gel and later stained with EtBr (ethidium bromide) [7].

The 25-mer oligonucleotide ML25 (5′-GAAAAAAGACT

graphic
TAGAAAAATTTTTA-3′; where the residues in bold indicate the topoisomerase 1B cleavage site and the arrow indicates the position of the cleavage) was 5′-end-labelled with [γ-32P]ATP and annealed to reverse 25-mer MC25 (5′-TAAAAATTTTTCTAAGTCTTTTTTC-3′). Excess of annealed oligonucleotides in the absence and presence of 100 μM CPT was mixed with L/S (50 nM) or ΔL/ΔS (50–1000 nM) in a cleavage reaction buffer (25 mM Tris/HCl, pH 7.5, 0.5 mM DTT, 10 mM MgCl2, 50 mM KCl, 1 mM EDTA and 150 μg/ml BSA) and assayed at 37 °C. It was electrophoresed in a denaturing 20% PAGE gel containing 7 M urea [9].

Substrate competition and fluorescence polarization assay

The same radiolabelled 25-mer duplex (7.5 nM) was either mixed with L/S or ΔL/ΔS (25 and 50 nM) or together with L/S (50 nM) and increasing amounts of ΔL/ΔS (50–600 nM) in a cleavage buffer (25 mM Tris/HCl, pH 7.5, 0.5 mM DTT, 10 mM MgCl2, 50 mM KCl, 1 mM EDTA and 150 μg/ml BSA) for 30 min at 37 °C to induce substrate competition. Samples were boiled with 1% SDS and run in SDS/15% PAGE [11].

Freshly purified S and ΔS were incubated with FITC and dimethylformamide in 100 mM phosphate buffer (pH 8) [12] and incubated at 16 °C for 30 min (Pierce Biotechnology). Samples were desalted and free FITC was removed by passing through PD10 columns (Amersham Biosciences). Labelled S and ΔS (50 or 100 nM) were mixed with increasing amounts of L (50–500 nM) and ΔL (100–1000 nM) separately. Using an F-3010, Hitachi/Japan Polarization system at λex=495 nm and λem=520 nm, the changes in fluorescence polarization were calculated. Thereafter, the fraction of bound proteins (S and ΔS) was calculated as stated in Supplementary information II (http://www.BiochemJ.org/bj/409/bj4090481add.htm) and plotted against respective protein concentration of L and ΔL.

Conditional antisense knockouts, immunoblotting, immunofluorescence and functional complementation

Antisense constructs of L (antiL) and S (antiS) were prepared (detailed in Supplementary information III at http://www.BiochemJ.org/bj/409/bj4090481add.htm) and transfected into Leishmania tarentolae T7.TR strain (Jena Bioscience). DsRed and SV40 (simian virus 40) T-antigen NLS (nuclear localization signal)-tagged ΔL (RΔLn) and GFP (green fluorescent protein)-tagged ΔS (GΔS) were transfected in parasites harbouring antiL and antiS respectively. The transfectant promastigotes antiL, antiL+Rn, antiL+RΔLn, antiS, antiS+G and antiS+GΔS were used in functional complementation [13,14].

Site-directed mutagenesis

Single mutations were introduced into the Leishmania heterodimeric topoisomerase I at positions Lys436, Asn441 and Lys455 of the large subunit and Lys249, Asn256 and Asp261 of the small subunit. Mutagenesis was performed using the Stratagene (La Jolla, CA, U.S.A.) QuikChange® XL kit following the manufacturer's protocol [15]. To carry out the desired mutations, pET28c/LdTOPIL and pGEX-5X2/LdTOPIS were used as templates. The following sense primers (along with antisense counterparts) with their substitution sites in boldface were used: LdTOPILK436A, 5′-GGTCCACCGCGGACGCGCTGGCCTACTTCAAC-3′; LdTOPILN441A, 5′-GCTGGCCTACTTCGCCAAGGCGAACACC-3′; LdTOPILK455A, 5′-CTGTGCAACCATCAAGCGTCCGTCTCGAAG-3′; LdTOPISK249A, 5′-CCGCAACCATCCAGGCGAAGTTTCCGTGGGCC-3′; LdTOPISN256A, 5′-CCGTGGGCCATGGCCGCCGAGAACTTCG-3′; LdTOPISD261A, 5′-CGAGAACTTCGCTTTTTGAGGATCCC-3′. Mutations were confirmed through DNA sequencing in a PerkinElmer (Norwalk, CT, U.S.A.) ABI Prism™ DNA sequencer.

RESULTS AND DISCUSSION

Recombinant protein constructs and their purification

Bi-subunit topoisomerase IB (L/S) may be the phylogenetic predecessor of the latter reconstituted ‘cap’ and ‘catalytic’ domains of its monomeric counterparts, but it originates as distinct subunits from different genes. The crystal structure of the truncated bi-subunit topoisomerase IB of Leishmania indicates that the two subunits contain separate interacting regions around a pocket region, which encompasses the DNA [8]. Knowledge of a functional inter-subunit co-operativity is obtained by identifying the minimal, functionally indispensable interacting fragments of this heterodimer. We had previously reported two N-terminal deletion constructs of L, i.e. LdTOPIL39–635 and LdTOPIL99–635 where residues 1–39 influenced the non-covalent interaction with DNA and residues 39–99 had some role in mediating subunit interaction [9].

In the present study, we generated several N- and C-terminal deletion constructs of L and S. The construct LdTOPIL1–456 was generated keeping the conserved ‘VAILCNH’ motif, while the construct LdTOPIL1–435 deletes this motif. Since LdTOPIL39–635 has been shown to be active [9], the smallest fragment LdTOPIL39–456 was generated that deletes both the unconserved N- and C-termini of L. The smaller subunit has an unconserved N-terminal extension of 200 amino acids, which bears an unusual stretch of serine residues. The construct LdTOPIS80–262 starts with this serine stretch at the N-terminus. Next, the construct LdTOPIS200–262 truncates the unconserved N-terminus, which harbours the serine stretch. Two further N-terminal truncations were carried out to generate the constructs LdTOPIS210–262 and LdTOPIS215–262. The C-terminal deletion constructs LdTOPIS1–255 and LdTOPIS1–245 were generated keeping the conserved ‘SKXXY’ motif harbouring the active-site tyrosine residue intact. Another deletion construct LdTOPIS210–255 was generated deleting the unconserved N-terminal end as well as part of the small C-terminus (Figures 1A and 1B, left panels). The deletion constructs of S were prepared in the pGEX-5X2 vector so that the 26 kDa, GST tag could stabilize the small protein fragments of S.

All recombinant proteins were transformed into BL21(DE3)pLysS cells and purified using Ni-NTA–agarose (constructs of L) or GST–Sepharose 4B (constructs of S) affinity chromatography followed by phosphocellulose column purification. SDS/PAGE analysis (Figures 1A and 1B, right panels) shows that the proteins were purified to near homogeneity.

LdTOPIL39–456 (ΔL) and LdTOPIS210–262 (ΔS) constitute the minimal interacting fragments

Far-Western analysis was carried out to study the in vitro interaction between the proteins as well as to find the interacting regions of two subunits. Using this method, we have identified the minimal fragment of one subunit that interacts with the other intact subunit [11]. The large subunit and its deletion constructs were electrophoresed in two separate polyacrylamide gels and transferred on to a nitrocellulose membrane. One set was immunoblotted with anti-L antibody (Figure 2A). It shows the exact position of the constructs and therefore served as the control. The other membrane was incubated overnight with S and then probed with anti-S antibody (Figure 2B). The proteins LdTOPIL100–635 and LdTOPIL1–435 (lanes 3 and 5) were absent from the blot, indicating that these two proteins failed to interact with intact S. The presence of bands corresponding to LdTOPIL39–635, LdTOPIL1–456 and LdTOPIL39–456 (lanes 2, 4 and 6) indicates that these proteins interact with S. Deletion further downstream of N-terminal residue 39 and upstream of C-terminal residue 456 results in loss of interaction. Hence LdTOPIL39–456 (ΔL) is the minimal fragment of L that interacts with S and the interacting regions for L could be lying between residues 39–99 and residues 435–456.

Far-Western analysis

Figure 2
Far-Western analysis

Constructs of L and S as indicated in Figure 1 were subjected to SDS/12% PAGE (two sets each) and transferred on to a PVDF membrane and renatured. (A) Immunoblot using anti-L. (B) Incubated with bait protein S and immunoblotted with anti-S. (C) Immunoblot using anti-S. (D) Incubated with bait protein L and immunoblotted with anti-L.

Figure 2
Far-Western analysis

Constructs of L and S as indicated in Figure 1 were subjected to SDS/12% PAGE (two sets each) and transferred on to a PVDF membrane and renatured. (A) Immunoblot using anti-L. (B) Incubated with bait protein S and immunoblotted with anti-S. (C) Immunoblot using anti-S. (D) Incubated with bait protein L and immunoblotted with anti-L.

Similarly, the small subunit and its deletion constructs were electrophoresed in two separate polyacrylamide gels and transferred on to nitrocellulose membrane. One set was immunoblotted with anti-LdTOPIS antibody as shown in Figure 2(C). This served as the control, indicating the positions of S and its deletion constructs. Another membrane was incubated overnight with L and thereafter probed with anti-L antibody as shown in Figure 2(D). The proteins LdTOPIS215–262, LdTOPIS210–255 and LdTOPIS1–245 (lanes 5, 6 and 7) failed to interact with L; hence their respective bands were absent from the blot. The presence of bands corresponding to LdTOPIS80–262, LdTOPIS210–262 and LdTOPIS1–255 (lanes 2, 4 and 8) indicates their interaction with L. A peptide flanking residues 215–235, harbouring the catalytic tyrosine residue (Tyr222), also failed to interact with L (results not shown). Since deletion further downstream of N-terminal residue 210 and C-terminal residue 255 of S results in loss of interaction, LdTOPIS210–262 (ΔS) is the minimal fragment of S that interacts with L. The interacting regions of S therefore lie between residues 210–215 and 245–262. From the above results, it is evident that both L and S harbour two stretches of residues on either side of their respective conserved motifs (‘VAILCNH’ for L and ‘SKXXY’ for S), which together are involved in subunit interaction at a time.

Co-immobilization assay was used to find out whether ΔL and ΔS formed the minimal interacting fragments [10]. The 6×His-tagged constructs L and ΔL and the GST-tagged constructs S and ΔS were separately co-transformed and co-expressed. Purification of His-tagged L and ΔL using Ni-NTA–agarose separately co-eluted GST-tagged S and ΔS (Figure 3A, lanes 1–4), while purification of GST-tagged S and ΔS using GST–Sepharose 4B separately co-eluted His-tagged L and ΔL (Figure 3A, lanes 5–8). Two similar polyacrylamide gels were electrophoresed and the proteins were transferred on to a nitrocellulose membrane and blotted using specific antibodies in order to confirm the identity of the co-eluted proteins. Immunoblotting with anti-L antibody lights up bands of 6×His–L (73 kDa) in lanes 1, 2, 5 and 6 and bands of 6×His–ΔL (48.7 kDa) in lanes 3, 4, 7 and 8 in Figure 3(B). The appearance of the bands 6×His–L in lanes 5 and 6 and 6×His–ΔL in lanes 7 and 8 respectively indicate that both GST–S and GST–ΔS can separately co-elute the wild-type as well as the minimal fragment of L through a GST–Sepharose column. When anti-S antibody was used, bands corresponding to GST–S (56 kDa) appear in lanes 1, 3, 5 and 7 and bands of GST–ΔS (34 kDa) appear in lanes 2, 4, 6 and 8 (Figure 3C). The appearance of bands corresponding to GST–S in lanes 1 and 3 and GST–ΔS in lanes 2 and 4 respectively indicates that both 6×His–L and 6×His–ΔL can separately co-elute the wild-type S as well as its minimal fragment through an Ni-NTA column. Therefore ΔL and ΔS were identified as the minimal interacting fragments of L and S.

Co-immobilization assay

Figure 3
Co-immobilization assay

6×His-tagged L and ΔL co-expressed with GST-tagged constructs of S and ΔS, purified and electrophoresed in SDS/12% PAGE gel (three sets). (A) SDS/PAGE of Ni-NTA–agarose (lanes 1–4) and GST–Sepharose purified proteins (lanes 5–8). Lanes 1 and 2, L co-elutes S and ΔS. Lanes 3 and 4, ΔL co-elutes S and ΔS. Lanes 5 and 6, S and ΔS co-elutes L. Lanes 7 and 8, S and ΔS co-elutes ΔL. (B, C) Similar gels were transferred on to nitrocellulose and immunoblotted with anti-L and anti-S antibodies.

Figure 3
Co-immobilization assay

6×His-tagged L and ΔL co-expressed with GST-tagged constructs of S and ΔS, purified and electrophoresed in SDS/12% PAGE gel (three sets). (A) SDS/PAGE of Ni-NTA–agarose (lanes 1–4) and GST–Sepharose purified proteins (lanes 5–8). Lanes 1 and 2, L co-elutes S and ΔS. Lanes 3 and 4, ΔL co-elutes S and ΔS. Lanes 5 and 6, S and ΔS co-elutes L. Lanes 7 and 8, S and ΔS co-elutes ΔL. (B, C) Similar gels were transferred on to nitrocellulose and immunoblotted with anti-L and anti-S antibodies.

LdTOPIL39–456/LdTOPIS210–262 (ΔL/ΔS) is active but 20-fold less sensitive to CPT

Supplementary Table 1 shows all the co-expressed constructs that were tested for plasmid relaxation activity, where ΔL/ΔS formed the minimal functional heterodimer. The two empty vectors pET28c and pGEX-5X2, transformed into BL21(DE3)pLysS, were induced and the cell extracts were purified as stated in the Methods section. The eluted fraction served as the negative control. The purified subunits were found to be inactive in the relaxation assay, indicating the absence of contaminating bacterial topoisomerases. Time kinetics was carried out using co-expressed empty vectors and reconstituted enzymes (shown in boldface in Supplementary Table 1) mixed in 1:2 molar ratio with pHOT1 DNA substrate. Figure 4(A) shows the assay with the eluted fraction of the empty vector control. Figures 4(B)–4(D) show the relaxation activity of the reconstituted wild-type and deletion constructs of L and S. The presence of a large number of intermediate supercoils and a lesser amount of fully relaxed band formation indicate the distributive nature of these enzymes. L/S completely relaxes the substrate in 1 min (Figure 4B, lane 3). L/ΔS and ΔL/S show that reduced activity and complete relaxation occur in 10 min (Figure 4C, lane 5) and in 15 min (Figure 4D, lane 6) respectively. ΔL/ΔS shows a gradual appearance of intermediate topoisomers; therefore the reconstituted enzyme is capable of DNA relaxation (Figure 4E). Since it did not follow steady-state conditions in the estimated time, its fold reduction in relaxation activity was not determined. The reconstituted construct L/LdTOPIS1–255 was found to be active although it lacks few residues at the C-terminal interacting region of S. The presence of the entire wild-type structure of L along with 255 residues of S probably provides the required structural balance to attain an active conformation. Support for our hypothesis is provided by the fact that truncated constructs LdTOPIL1–456 and LdTOPIL39–456 fail to interact with LdTOPIS1–255 (results not shown). But LdTOPIS1–255 was not the minimal fragment of S that interacts with L and hence it was not used in further studies. Davies et al. [8] have recently reported a 2.27 Å (1 Å=0.1 nm) crystal structure of an active truncated L/S comprising residues 27–456 of L and residues 200–262 of S [8]. Therefore the functional activity exhibited by the minimal fragments ΔL/ΔS indicates their structural conformation close to the reported structure.

Relaxation of pHOT1 DNA by using (A) eluted fraction of empty vector-induced bacterial lysate and enzymes (B) L/S, (C) L/ΔS, (D) ΔL/S and (E) ΔL/ΔS

Figure 4
Relaxation of pHOT1 DNA by using (A) eluted fraction of empty vector-induced bacterial lysate and enzymes (B) L/S, (C) L/ΔS, (D) ΔL/S and (E) ΔL/ΔS

Lane 1, 150 fmol of supercoiled pHOT1 DNA; lanes 2–9, same as lane 1 but incubated with 75 fmol of reconstituted enzymes for indicated time periods.

Figure 4
Relaxation of pHOT1 DNA by using (A) eluted fraction of empty vector-induced bacterial lysate and enzymes (B) L/S, (C) L/ΔS, (D) ΔL/S and (E) ΔL/ΔS

Lane 1, 150 fmol of supercoiled pHOT1 DNA; lanes 2–9, same as lane 1 but incubated with 75 fmol of reconstituted enzymes for indicated time periods.

The sensitivity of an inhibitor of DNA topoisomerase I is generally assessed based on relaxation kinetics in the absence and presence of the drug. Since ΔL/ΔS did not follow steady-state kinetics, its CPT-sensitivity was elucidated using the equilibrium cleavage assay [9]. CPT is known to trap the cleavable complex, and in this assay it releases the cleaved product under denaturing conditions. An equimolar concentration of ΔL/ΔS, compared with L/S, did not form any cleaved product even in the presence of CPT. A 20-fold molar excess of ΔL/ΔS produces 80% of the cleavable complex formed by L/S in the presence of CPT (Figure 5). This indicated that ΔL/ΔS was 20-fold less sensitive to CPT compared with L/S. Since ΔL/ΔS showed reduced CPT-sensitivity, its relative affinity for DNA and interaction between the fragments were studied.

Equilibrium cleavage assay using 5′-end-labelled 25-mer duplex as shown

Figure 5
Equilibrium cleavage assay using 5′-end-labelled 25-mer duplex as shown

Lane 1, labelled 25-mer duplex DNA. Lanes 2 and 3, labelled DNA incubated with 50 nM L/S, in the absence (−) and presence (+) of CPT. Lanes 4–11, same as lane 1, but incubated with indicated amounts of ΔL/ΔS in the absence and presence of CPT alternately.

Figure 5
Equilibrium cleavage assay using 5′-end-labelled 25-mer duplex as shown

Lane 1, labelled 25-mer duplex DNA. Lanes 2 and 3, labelled DNA incubated with 50 nM L/S, in the absence (−) and presence (+) of CPT. Lanes 4–11, same as lane 1, but incubated with indicated amounts of ΔL/ΔS in the absence and presence of CPT alternately.

LdTOPIL39–456/LdTOPIS210–262 (ΔL/ΔS) has reduced affinity for DNA and decreased association between them

A critical amount of a radiolabelled substrate induces competition between L/S and ΔL/ΔS to form the cleavable complex. For the wild-type L/S, phosphotyrosyl linkage is formed between radiolabelled oligonucleotide and Tyr222 of S, while for ΔL/ΔS it is formed by the same oligonucleotide and Tyr222 of ΔS. In an SDS/15% PAGE, the two intact subunits or their fragments are separated but the covalently linked oligonucleotide migrates with S or ΔS and is detected in an autoradiograph indicative of the extent of cleavage. When the concentration of ΔL/ΔS is increased, the amount at which it reduces the cleavable complex of L/S by 50% is the equivalence competitive concentration. Fold excess of ΔL/ΔS at the equivalence concentration required over wild-type L/S gives the KD (equilibrium dissociation constant) value, since it shares a linear relationship [11]. The amount of cleavable complex formed was obtained by densitometry of the autoradiograph in Figure 6. An 8-fold molar excess of ΔL/ΔS was required to reduce substrate binding of L/S by 50%. The KD value of L/S for the oligonucleotide substrate is 3.1×10−7 M [7]; hence the relative KD of ΔL/ΔS for the oligonucleotide substrate is 2.5×10−6 M. The KD values of L/ΔS and ΔL/S obtained by similar experiments were 1.03×10−6 and 1.64×10−6 M respectively. Reconstitution of ‘Δcap’ and ‘topo31’ of human topoisomerase IB occurs at 2:1 or greater molar ratio. It binds DNA with reduced affinity but retains the processive nature. Therefore it differs considerably from reconstituted ΔL/ΔS fragments [6].

Substrate competition assay using 5′-end-labelled 25-mer duplex (as shown)

Figure 6
Substrate competition assay using 5′-end-labelled 25-mer duplex (as shown)

Lanes 1 and 2, incubated with 25 and 50 nM each of L/S; lanes 3 and 4, incubated with 25 and 50 nM each of ΔL/ΔS. Lanes 5–11, same substrate incubated with 50 nM of L/S mixed with increasing (50–600 nM) amounts of ΔL/ΔS.

Figure 6
Substrate competition assay using 5′-end-labelled 25-mer duplex (as shown)

Lanes 1 and 2, incubated with 25 and 50 nM each of L/S; lanes 3 and 4, incubated with 25 and 50 nM each of ΔL/ΔS. Lanes 5–11, same substrate incubated with 50 nM of L/S mixed with increasing (50–600 nM) amounts of ΔL/ΔS.

Using fluorescence polarization, the association of L and ΔL separately with FITC-labelled S and ΔS was measured at a λex of 495 nm and a λem of 520 nm [12]. When free the smaller FITC-labelled S and ΔS exhibit rapid molecular rotation, thereby depolarizing the polarized beam of light. When these proteins are bound by L or ΔL, their molecular rotation ceases with the extent of binding, thereby causing an increase in fluorescence polarization. The fraction of bound protein (fB) was obtained from the equation stated in Supplementary information II and was plotted against the protein concentrations (F) for S and ΔS separately with L (Figure 7A) and ΔL (Figure 7B). The KD value for the interaction between L and S is 6.1×10−8 M, that between L and ΔS is 1.53×10−7 M, that between ΔL and S is 2.2×10−7 M and that between ΔL and ΔS is 3.9×10−7 M. Therefore the extent of interaction in descending order is L/S>L/ΔS>ΔL/S>ΔL/ΔS where ΔL/ΔS manifests a 6.5-fold reduced association between themselves. Therefore it can be stated that the reduced association between the fragments forms a stringent conformation, which is reflected by its decreased affinity for the DNA substrate and reduced relaxation activity.

Fluorescence polarization assay

Figure 7
Fluorescence polarization assay

(A) The change in polarization value was measured at 518 nm and the fraction bound for 50 nM of S and ΔS mixed with increasing amounts of L (50–500 nM) was plotted. Results shown represent means±S.D. (n=3). (B) Same as (A) but using 100 nM of S and ΔS mixed with increasing amounts (100–1000 nM) of ΔL.

Figure 7
Fluorescence polarization assay

(A) The change in polarization value was measured at 518 nm and the fraction bound for 50 nM of S and ΔS mixed with increasing amounts of L (50–500 nM) was plotted. Results shown represent means±S.D. (n=3). (B) Same as (A) but using 100 nM of S and ΔS mixed with increasing amounts (100–1000 nM) of ΔL.

LdTOPIL39–456 (ΔL) and LdTOPIS210–262 (ΔS) separately complement their wild-type subunits in Leishmania

For complementation studies, wild-type genes were down-regulated using a conditional antisense approach [14]. Figure 8(A) shows the constructs antiL, Rn and RΔLn, which were transfected into the L. tarentolae T7.TR strain to generate transfectant parasites antiL, antiL+Rn and antiL+RΔLn as stated in Supplementary information III. Figure 8(B) shows the nuclear localization of Rn in antiL+Rn parasites and the kinetoplast and nuclear localization of ΔL in antiL+RΔLn parasites. The NLS of SV40 T-antigen (n) facilitates its nuclear localization. The construct antiL was generated from the nucleotide sequence 1725–1908 of wild-type L that was absent from ΔL. Wild-type L was down-regulated in antiL, antiL+Rn and antiL+RΔLn parasites after 24 h of tetracycline induction of antiL construct in these parasites (results not shown). Since one subunit is unstable in the absence of the other subunit [4], a concomitant decrease in wild-type S occurs within 72 h in antiL and antiL+Rn and these parasites die. In antiL+RΔLn, tetracycline induction reduces wild-type L but simultaneously overexpressed RΔLn interacts with wild-type S and stabilizes it and RΔLn/S, similar to the active ΔL/S shown in Figure 4(D), is formed. Hence ΔL functionally complements wild-type L (Figure 8C). Figure 8(D) shows the constructs antiS, G and GΔS, which were transfected into L. tarentolae T7.TR strain to generate transfectant parasites antiS, antiS+G and antiS+GΔS as stated in Supplementary information III. G is localized in the cytoplasm of antiS+G, while GΔS is localized in the nucleus and kinetoplast of antiS+GΔS (Figure 8E). The construct antiS was generated from the nucleotide sequence 240–420 of wild-type S that was absent from ΔS. Wild-type S was down-regulated in antiS, antiS+G and antiS+GΔS parasites after 24 h of tetracycline induction of antiS construct in these parasites (results not shown). Since one subunit is unstable in the absence of the other subunit [4], a concomitant decrease in wild-type L occurs within 72 h in antiS and antiS+G and these parasites die. In antiS+GΔS, tetracycline induction reduces wild-type S, but simultaneously overexpressed GΔS interacts with wild-type L and stabilizes it, and L/GΔS, similar to active L/ΔS shown in Figure 4(C), is formed. Hence ΔS functionally complements the wild-type S (Figure 8F).

Fragment complementation experiments

Figure 8
Fragment complementation experiments

Schematic representation of (A) transfected constructs of L and (B) localization of Rn and RΔLn. (C) Functional complementation by RΔLn. Promastigotes (1×105) were cultured with (+) or without (−) tetracycline (T). Live promastigotes were calculated at 12 h intervals and plotted against time. (D) Transfected constructs of S and (E) localization of G and GΔS. (F) Functional complementation by GΔS as stated in (C).

Figure 8
Fragment complementation experiments

Schematic representation of (A) transfected constructs of L and (B) localization of Rn and RΔLn. (C) Functional complementation by RΔLn. Promastigotes (1×105) were cultured with (+) or without (−) tetracycline (T). Live promastigotes were calculated at 12 h intervals and plotted against time. (D) Transfected constructs of S and (E) localization of G and GΔS. (F) Functional complementation by GΔS as stated in (C).

Residues Lys455 of LdTOPIL and Asp261 of LdTOPIS are involved in heterodimerization

Alanine substitutions of three charged residues were carried out in the C-terminal interacting region of both subunits. For L, the residues were Lys436, Asn441 and Lys455, while for S they were Lys249, Asn256 and Asp261 respectively. The reconstituted mutant enzymes LdTOPILK436A/S, LdTOPILN441A/S, L/LdTOPISK249A and L/LdTOPISN256A did not show any appreciable change in the relaxation pattern compared with the wild-type enzyme L/S. Besides, the reconstitution of each of the two mutants of L and S with each other also showed negligible changes in the plasmid relaxation pattern compared with L/S (results not shown). Time kinetics of relaxation was carried out using wild-type enzyme L/S (Figure 9A) and two mutant constructs. Figure 9(B) shows that reconstitution of L with LdTOPISD261A (SD261A) causes a 15-fold decrease in the relaxation activity compared with L/S. This is evident from the fact that complete relaxation by the mutant enzyme L/SD261A occurs in 15 min compared with 1 min for L/S. Similarly, reconstitution of LdTOPILK455A (LK455A) with S causes a 25-fold decrease in the relaxation activity compared with L/S, since complete relaxation by the mutant enzyme LK455A/S occurs in 25 min compared with 1 min for L/S (Figure 9C). The reconstitution of LK455A and SD261A did not show any relaxation activity (Figure 9D). The reconstituted enzymes formed by one mutant and the other wild-type subunit induce a structural stringency at one end of the subunit harbouring the mutation. But the intact subunit balances the overall structure by forming a partially interacted conformer having reduced relaxation activity. The inhibitory effect on relaxation by mutation in LK455A of the reconstituted enzyme LK455A/S is greater than that of the mutation in SD261A of the reconstituted enzyme L/SD261A. Reconstitution of two mutant subunits, LK455A and SD261A, fails to assume an interacted conformation and hence it does not show any relaxation activity.

Relaxation of pHOT1 DNA using reconstituted wild-type and mutant enzymes

Figure 9
Relaxation of pHOT1 DNA using reconstituted wild-type and mutant enzymes

(A) L/S (B) L/SD261A, (C) LK455A/S and (D) LK455A/SD261A. Lane 1, 150 fmol of supercoiled pHOT1 DNA; lanes 2–9, same as lane 1 but incubated with 75 fmol of reconstituted enzymes for the indicated time periods.

Figure 9
Relaxation of pHOT1 DNA using reconstituted wild-type and mutant enzymes

(A) L/S (B) L/SD261A, (C) LK455A/S and (D) LK455A/SD261A. Lane 1, 150 fmol of supercoiled pHOT1 DNA; lanes 2–9, same as lane 1 but incubated with 75 fmol of reconstituted enzymes for the indicated time periods.

CONCLUSION

Overall, the present study identifies the minimal functionally active fragments of the heterodimeric topoisomerase IB of Leishmania (Table 1). The search for important residues governing subunit interaction is narrowed to the active ΔL/ΔS. Therefore, taking this into account, we have identified residues Lys455 of L and Asp261 of S as two key amino acids modulating subunit interaction. The reduced fragment association between ΔL and ΔS results in a weakly reconstituted enzyme having lower DNA binding affinity, reduced cleavage and relaxation activity. The bi-subunit nature of topoisomerase IB in Leishmania allows functional complementation using truncated constructs. Therefore functional dissection of these individual subunits will help in better understanding the molecular architecture of this unusual heterodimeric type IB topoisomerase of the parasite.

Table 1
Various properties of the reconstituted wild-type and minimal fragments
Enzyme Relaxation activity (fold reduced compared with L/S) DNA-binding affinity (KD)DNA CPT sensitivity (fold reduced compared with L/S) Inter-subunit affinity (KD)protein In vivo complementation in Leishmania 
LdTOPIL/S (L/S) 100-fold 3.1×10–76.1×10–8− 
LdTOPIL/LdTOPIS210–262 (L/ΔS) 82-fold 1.03×10−67-fold 1.53×10−7GΔS complements wild-type S 
LdTOPIL39–456/LdTOPIS (ΔL/S) 65-fold 1.64×10−69-fold 2.2×10−7RΔLn complements wild-type L 
LdTOPIL39–456/LdTOPIS210–262 (ΔL/ΔS) 21-fold 2.5×10−620-fold 3.9×10−7− 
Enzyme Relaxation activity (fold reduced compared with L/S) DNA-binding affinity (KD)DNA CPT sensitivity (fold reduced compared with L/S) Inter-subunit affinity (KD)protein In vivo complementation in Leishmania 
LdTOPIL/S (L/S) 100-fold 3.1×10–76.1×10–8− 
LdTOPIL/LdTOPIS210–262 (L/ΔS) 82-fold 1.03×10−67-fold 1.53×10−7GΔS complements wild-type S 
LdTOPIL39–456/LdTOPIS (ΔL/S) 65-fold 1.64×10−69-fold 2.2×10−7RΔLn complements wild-type L 
LdTOPIL39–456/LdTOPIS210–262 (ΔL/ΔS) 21-fold 2.5×10−620-fold 3.9×10−7− 

We are grateful to Professor S. Roy, Director, Indian Institute of Chemical Biology, for his interest in this work. We thank Professor S. M. Beverley and Professor G. A. M. Cross for the gift of the Leishmania transfection vectors. This work was supported by a grant from the Department of Biotechnology, Government of India (BT/PR6399/BRB/10/434/05), to H. K. M. S. B. D. was supported by a Senior Research Fellowship from the CSIR (Council of Scientific and Industrial Research), Government of India.

Abbreviations

     
  • AP

    alkaline phosphatase

  •  
  • CPT

    camptothecin

  •  
  • DTT

    dithiothreitol

  •  
  • GST

    glutathione transferase

  •  
  • LdTOPIL (or L)

    large subunit of topoisomerase IB of Leishmania

  •  
  • LdTOPIS (or S)

    small subunit of topoisomerase IB of Leishmania

  •  
  • ΔL

    LdTOPIL39–456

  •  
  • ΔS

    LdTOPIS210–262

  •  
  • Ni-NTA

    Ni2+-nitrilotriacetate

  •  
  • NLS

    nuclear localization signal

  •  
  • SV40

    simian virus 40

References

References
1
Champoux
J. J.
DNA topoisomerases, structure, function and mechanism
Annu. Rev. Biochem.
2001
, vol. 
70
 (pg. 
369
-
413
)
2
Villa
H.
Otero Marcos
A. R.
Reguera
R. M.
Balana-Fouce
R.
Garcia-Estrada
C.
Perez-Pertejo
Y.
Tekwani
B. L.
Myler
P. J.
Stuart
K. D.
Bjornsti
M. A.
Ordonez
D.
A novel active DNA topoisomerase I in Leishmania donovani
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
3521
-
3526
)
3
Balana Fouce
R.
Redondo
C. M.
Perez-Pertezo
Y.
Diaz-Gonzales
R.
Ruguera
M.
Targeting atypical trypanosomatid DNA topoisomerase I
Drug Discov. Today
2006
, vol. 
11
 (pg. 
733
-
740
)
4
Bakshi
R.
Shapiro
T. A.
RNA interference of Trypanosoma brucei topoisomerase IB, both subunits are essential
Mol. Biochem. Parasitol.
2004
, vol. 
136
 (pg. 
249
-
255
)
5
Ireton
G.
Stewart
L.
Parker
L. H.
Champoux
J. J.
Expression of human topoisomerase I with a partial deletion of the linker region yield monomeric and dimeric enzymes that respond differently to camptothecin
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
25820
-
25830
)
6
Yang
Z.
Champoux
J. J.
Reconstitution of enzymatic activity by the association of the cap and catalytic domains of human topoisomerase I
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
30815
-
30823
)
7
Das
B. B.
Sen
N.
Ganguly
A.
Majumder
H. K.
Reconstitution and functional characterization of the unusual bi-subunit type I DNA topoisomerase from Leishmania donovani
FEBS Lett.
2004
, vol. 
565
 (pg. 
81
-
88
)
8
Davies
D. R.
Mushtaq
A.
Interthal
H.
Champoux
J. J.
Hol
W. J.
The structure of the transition state of the heterodimeric topoisomerase I of Leishmania donovani as a vanadate complex with nicked DNA
J. Mol. Biol.
2006
, vol. 
357
 (pg. 
1202
-
1210
)
9
Das
B. B.
Sen
N.
BoseDasgupta
S.
Ganguly
A.
Majumder
H. K.
N-terminal region of the large subunit of Leishmania donovani bi-subunit topoisomerase I is involved in DNA relaxation and interaction with small subunit
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
16335
-
16344
)
10
Wu
L.
Hickson
I. D.
The Bloom's syndrome helicase stimulates the activity of topoisomerase IIIα
Nucleic Acids Res.
2002
, vol. 
30
 (pg. 
4823
-
4829
)
11
Stewart
L.
Ireton
G. C.
Champoux
J. J.
Reconstitution of human topoisomerase I by fragment complementation
J. Mol. Biol.
1997
, vol. 
269
 (pg. 
355
-
372
)
12
Park
S. H.
Raines
R. T.
Fluorescence polarization assay to quantify protein–protein interactions
Methods Mol. Biol.
2004
, vol. 
261
 (pg. 
161
-
166
)
13
Robinson
K. A.
Beverley
S. M.
Improvements in transfection efficiency and tests of RNA interference (RNAi) approaches in the protozoan parasite Leishmania
Mol. Biochem. Parasitol.
2003
, vol. 
128
 (pg. 
217
-
228
)
14
Goswami
S.
Dhar
G.
Mukherjee
S.
Mahata
B.
Chaterjee
S.
Home
P.
Adhya
S.
A bi-functional tRNA import receptor from Leishmania mitochondria
Proc. Natl. Acad. Sci. U.S.A.
2006
, vol. 
103
 (pg. 
8354
-
8359
)
15
Das
B. B.
BoseDasgupta
S.
Ganguly
A.
Mazumder
S.
Roy
A.
Majumder
H. K.
Leishmania donovani bisubunit topoisomerase I gene fusion leads to an active enzyme with conserved type IB enzyme function
FEBS J.
2007
, vol. 
274
 (pg. 
150
-
163
)

Supplementary data