RNA adducts with chlorophyll and chlorophyllin: stability and structural features

Porphyrins and their metal derivatives are strong nucleic acids binders. Some of these compounds have been used for radiation sensitization therapy of cancer and are targeted to interact with cellular DNA. Chlorophyll (Chl) binds DNA via guanine N-7 atom (major groove) and the backbone phosphate group (Neault and Tajmir-Riahi. Biophys. J. 76, 2177, 1999), whereas chlorophyllin (Chln) intercalates into A-T and G-C regions (Neault and Tajmir-Riahi. J. Phys. Chem. B. 102, 1610, 1998). This study was designed to examine the interaction of RNA with chlorophyll a and chlorophyllin in aqueous solution at physiological pH with pigment/RNA(phosphate) ratios (r) of 1/80 to 1/2. Fourier transform infrared (FTIR) and UV-visible difference spectroscopic methods were used to characterize the nature of pigment-RNA interaction and to establish correlation between spectral changes and the pigment binding mode, binding constant, RNA secondary structure and structural variations of pigment-RNA complexes in aqueous solution. Spectroscopic results showed that Chl and Chln bind RNA through G-C and A-U bases and the backbone phosphate group with overall binding constants of KChl = 1.95 x 10(5) M(-1) and KChln = 1.61 x 10(5) M(-1). The larger K value obtained for Chl-RNA complexes is attributed to the formation of more stable five or six-coordinate Mg cation in the RNA adducts, while the four-coordination Cu(II) in Chln can be more stable than that of the five or six-coordinated copper ion in the Chln-RNA complexes. Aggregation of pigment-RNA complexes occurs at high metalloporphyrin concentrations. No biopolymer secondary structural changes were observed upon pigment interaction and RNA remains in the A-family structure in these pigment complexes.     

Interaction of antioxidant flavonoids with tRNA: intercalation or external binding and comparison with flavonoid-DNA adducts

Antioxidants are essential to good health. Flavonoids are powerful antioxidants, and prevent DNA damage. The antioxidative protections are related to their binding modes to a DNA duplex and complexation with free radicals in vivo. Recently we reported the interaction of flavonoids with DNA in vitro (Kanakis et al., J. Biomol. Struct. Dyn. 22, 719-724, 2005), where polyphenol different binding modes were discussed. The aim of this study was to examine the interaction of transfer RNA with quercetin (que), kaempferol (kae), and delphinidin (del) in aqueous solution at physiological conditions and to make a comparison with the corresponding pigment-DNA adducts. Constant tRNA concentration (6.25 mM) and various drug/RNA(phosphate) molar ratios of 1/48 to 1/8 were used. FTIR and UV-visible difference spectroscopic methods have been applied to determine the drug binding mode, the binding constants, and the effects of drug complexation on the stability and conformation of tRNA duplex. Both intercalative and external binding modes were observed. Structural analysis showed que, kae, and a del intercalate tRNA duplex with minor external binding to the major or minor groove and the backbone phosphate group with overall binding constants K (que) = 4.80 x 10(4) M(1), K (kae) = 4.65 x 10(4) M(1), and K (del) = 9.47 x 10(4) M(1). The stability of adduct formation is in the order of del > que > kae. A comparison with flavonoids-DNA adducts showed both intercalation and external bindings with the stability order K (que) = 7.25 x 10(4) M(1), K (kae) = 3.60 x 10(4) M(1), and K (del) = 1.66 x 10(4) M(1). Low flavonoid concentration induces helical stabilization, whereas high pigment content causes helix opening. A partial Bto A-DNA transition occurs at high drug concentration, while tRNA remains in the A-family structure.     

Interaction of arsenic trioxide As2O3 with DNA and RNA

Arsenic salts have been used for centuries to treat a variety of medical conditions ranging from infectious disease to cancer. More recently, trivalent arsenic trioxide was found to exhibit high antitumor activity towards hematological malignancies. Even though much is known about antitumor activity and DNA damage by As2O3, there has been no report on the interaction of arsenic trioxide with isolated DNA or RNA. Therefore, it was of interest to examine the interaction of As2O3 with DNA and RNA in aqueous solution at physiological pH. FTIR and UV-visible difference spectroscopic methods were used to characterize the nature of drug-DNA and drug-RNA interactions and to determine the As binding site, the binding constant, the sequence selectivity, the helix stability, and the biopolymer secondary structure in the As2O3-polynucleotide complexes in vitro. The FTIR spectroscopic studies were conducted with As2O3-polynucleotide (phosphate) ratios of 1/40, 1/20, 1/10, and 1/5, with a final DNA (P) or RNA (P) concentration of 6.25 mmol/l. Spectroscopic results showed As2O3 binds to DNA and RNA at G-C, A-T, and A-U bases, and no interaction with the backbone PO2 group. As2O3-DNA and -RNA adducts showed one type of binding with overall binding constant of K(As2O3-DNA) = 1.24 x 10(5) M(-1) and K(As2O3-RNA) = 2.60 x 10(5) M(-1). The As2O3-polynucleotide complexation is associated with a partial biopolymer aggregation and no major alterations of B-DNA or A-RNA structure.     

Binding characteristics of Hoechst 33258 with calf thymus DNA, poly[d(A-T)], and d(CCGGAATTCCGG): multiple stoichiometries and determination of tight binding with a wide spectrum of site affinities.

Equilibrium binding experiments using fluorescence and absorption techniques have been performed throughout a wide concentration range (1 nM to 30 microM) of the dye Hoechst 33258 and several DNAs. The most stable complexes found with calf thymus DNA, poly[d(A-T)], d(CCGGAATTCCGG), and d(CGCGAATTCGCG) all have dissociation constants in the range (1-3) X 10(-9) M-1. Such complexes on calf thymus DNA occur with a frequency of about 1 binding site per 100 base pairs, and evidence is presented indicating a spectrum of sequence-dependent affinities with dissociation constants extending into the micromolar range. In addition to these sequence-specific binding sites on the DNA, the continuous-variation method of Job reveals distinct stoichiometries of dye-poly[d(A-T)] complexes corresponding to 1, 2, 3, 4, and 6 dyes per 5 A-T base pairs and even up to 1 and 2 (and possibly more) dyes per backbone phosphate. Models are suggested to account for these stoichiometries. With poly[d(G-C)] the stoichiometries are 1-2 dyes per 5 G-C pairs in addition to 1 and 2 dyes per backbone phosphate. Thermodynamic parameters for formation of the tightest binding complex between Hoechst 33258 and poly[d(A-T)] or d-(CCGGAATTCCGG) are determined. Hoechst 33258 binding to calf thymus DNA, chicken erythrocyte DNA, and poly[d(A-T)] exhibits an ionic strength dependence similar to that expected for a singly-charged positive ion. This ionic strength dependence remains unchanged in the presence of 25% ethanol, which decreases the affinity by 2 orders of magnitude. In addition, due to its strong binding, Hoechst 33258 easily displaces several intercalators from DNA.     

DNA interaction with naturally occurring antioxidant flavonoids quercetin, kaempferol, and delphinidin

Flavonoids are strong antioxidants that prevent DNA damage. The anticancer and antiviral activities of these natural products are implicated in their mechanism of actions. However, there has been no information on the interactions of these antioxidants with individual DNA at molecular level. This study was designed to examine the interaction of quercetin (que), kaempferol (kae), and delphinidin (del) with calf-thymus DNA in aqueous solution at physiological conditions, using constant DNA concentration (6.5 mmol) and various drug/DNA(phosphate) ratios of 1/65 to 1. FTIR and UV-Visible difference spectroscopic methods are used to determine the drug binding sites, the binding constants and the effects of drug complexation on the stability and conformation of DNA duplex. Structural analysis showed quercetin, kaempferol, and delphinidin bind weakly to adenine, guanine (major groove), and thymine (minor groove) bases, as well as to the backbone phosphate group with overall binding constants K(que) = 7.25 x 10(4)M(-1), K(kae) = 3.60 x 10(4)M(-1), and K(del) = 1.66 x 10(4)M(-1). The stability of adduct formation is in the order of que>kae>del. Delphinidin with a positive charge induces more stabilizing effect on DNA duplex than quercetin and kaempferol. A partial B to A-DNA transition occurs at high drug concentrations.     

Structural analysis of DNA-chlorophyll complexes by Fourier transform infrared difference spectroscopy

Porphyrins and metalloporphyrins are strong DNA binders. Some of these compounds have been used for radiation sensitization therapy of cancer and are targeted to interact with cellular DNA. This study was designed to examine the interaction of calf thymus DNA with chlorophyll a (CHL) in aqueous solution at physiological pH with CHL/DNA(phosphate) ratios (r) of 1/160, 1/80, 1/40, 1/20, 1/10, and 1/5. Fourier transform infrared (FTIR) difference spectroscopy was used to characterize the nature of DNA-pigment interactions and to establish correlations between spectral changes and the CHL binding mode, binding constant, sequence selectivity, DNA secondary structure, and structural variations of DNA-CHL complexes in aqueous solution. Spectroscopic results showed that CHL is an external DNA binder with no affinity for DNA intercalation. At low pigment concentration (r = 1/160, 1/80, and 1/40), there are two major binding sites for CHL on DNA duplex: 1) Mg-PO2 and 2) Mg-N7 (guanine) with an overall binding constant of K = 1.13 x 10(4) M-1. The pigment distributions are 60% with the backbone PO2 group and 20% with the G-C base pairs. The chlorophyll interaction is associated with a major reduction of B-DNA structure in favor of A-DNA. At high chlorophyll content (r = 1/10), helix opening occurs, with major spectral alterations of the G-C and A-T bases. At high chlorophyll concentration (1/5), pigment aggregation is observed, which does not favor CHL-DNA complexation.     

Oxidative degradation of cationic metalloporphyrins in the presence of nucleic acids: a way to binding constants?

Mn(III) and Fe(III) complexes of meso-tetrakis(N-methylpyridinium-4-yl)porphyrin (M-TMePyP) and related hybrid molecules ("metalloporphyrin-ellipticine") were activated by potassium monopersulfate in the presence of variable calf thymus (CT) DNA and NaCl concentrations. Monitored by visible spectroscopy (Soret band), fast degradation of the free metalloprophyrin was observed while the DNA-bound form appeared protected. This direct quantitation of free versus bound metalloporphyrin ratios allowed determination of binding constants: Mn- and Fe-TMePyP respectively bind to CT DNA (5 mM phosphate buffer, 0.1 M NaCl, pH 7) with K = 3 X 10(4) and 1.2 X 10(4) M-1. Mn-TMePyP showed a greater affinity for poly[d(A-T)] (K = 1.2 X 10(5) M-1) than for poly[d(G-C)] (K = 0.2 X 10(4) M-1). This method allowed us access to the intrinsic DNA affinity of the metalloporphyrin moiety of the hybrid molecules "metalloporphyrin-ellipticine".     

DNA interaction with human serum albumin studied by affinity capillary electrophoresis and FTIR spectroscopy

The question addressed in this study is how does the protein-DNA complexation affect the structure and dynamics of DNA and protein in aqueous solution. We examined the interaction of calf-thymus DNA with human serum albumin (HSA) in aqueous solution at physiological conditions, using constant DNA concentration of 12.5 mM (phosphate) and various HSA contents 0.25 to 2% or 0.04 to 0.3 mM. Affinity capillary electrophoresis and FTIR spectroscopic methods were used to determine the protein binding mode, the association constant, sequence preference, and the biopolymer secondary structural changes in the HSA-DNA complexes. Spectroscopic evidence showed two types of HSA-DNA complexes with strong binding of K(1) = 4.5 x 10(5) M(-1) and weak binding of K(2) = 6.10 x 10(4) M(-1). The two major binding sites were located on the G-C bases and the backbone PO(2) group. The protein-DNA interaction stabilizes the HSA secondary structure. A minor alteration of B-DNA structure was observed, while no major protein conformational changes occurred.     

DNA interaction with saffron's secondary metabolites safranal, crocetin, and dimethylcrocetin

Saffron comes from the dried red stigmas of the Crocus sativus L. flower. Except for its use in cooking and in traditional medicine, it has numerous applications as an antitoxic, antioxidant, and anticancer agent due to its secondary metabolites and their derivatives (safranal, crocins, crocetin, dimethylcrocetin). However, there has been no information on the interactions of these secondary metabolites with individual DNA at molecular level. This study was designed to examine the interaction of safranal, crocetin (CRT), and dimethylcrocetin (DMCRT) with calf-thymus DNA in aqueous solution at physiological conditions, using constant DNA concentration (6.25 mM) and various drug/DNA(phosphate) molar ratios from 1/48 to 1/2. FTIR and UV-visible difference spectroscopic methods are used to determine the drug binding sites, the binding constants, and the effects of carotenoids and safranal complexation on the stability and conformation of DNA duplex. Both intercalative and external binding modes were observed, with overall binding constants K(safranal) = 1.24 x 10(3) M(-1), K(CRT) = 6.2 x 10(3) M(-1) and K(DMCRT) = 1.85 x 10(5) M(-1) A partial B- to A-DNA transition occurs at high carotenoids and safranal concentrations.     

Silver(I) complexes with DNA and RNA studied by Fourier transform infrared spectroscopy and capillary electrophoresis

Ag(I) is a strong nucleic acids binder and forms several complexes with DNA such as types I, II, and III. However, the details of the binding mode of silver(I) in the Ag-polynucleotides remains unknown. Therefore, it was of interest to examine the binding of Ag(I) with calf-thymus DNA and bakers yeast RNA in aqueous solutions at pH 7.1-6.6 with constant concentration of DNA or RNA and various concentrations of Ag(I). Fourier transform infrared spectroscopy and capillary electrophoresis were used to analyze the Ag(I) binding mode, the binding constant, and the polynucleotides' structural changes in the Ag-DNA and Ag-RNA complexes. The spectroscopic results showed that in the type I complex formed with DNA, Ag(I) binds to guanine N7 at low cation concentration (r = 1/80) and adenine N7 site at higher concentrations (r = 1/20 to 1/10), but not to the backbone phosphate group. At r = 1/2, type II complexes formed with DNA in which Ag(I) binds to the G-C and A-T base pairs. On the other hand, Ag(I) binds to the guanine N7 atom but not to the adenine and the backbone phosphate group in the Ag-RNA complexes. Although a minor alteration of the sugar-phosphate geometry was observed, DNA remained in the B-family structure, whereas RNA retained its A conformation. Scatchard analysis following capillary electrophoresis showed two binding sites for the Ag-DNA complexes with K(1) = 8.3 x 10(4) M(-1) for the guanine and K(2) = 1.5 x 10(4) M(-1) for the adenine bases. On the other hand, Ag-RNA adducts showed one binding site with K = 1.5 x 10(5) M(-1) for the guanine bases.     

A Comparative study of Fe(II) and Fe(III) interactions with DNA duplex: major and minor grooves bindings

The involvement of the Fe cations in autoxidation in cells and tissues is well documented. DNA is a major target in such reaction, and can chelate Fe cation in many ways. The present study was designed to examine the interaction of calf-thymus DNA with Fe(II) and Fe(III), in aqueous solution at pH 6.5 with cation/DNA (P) (P = phosphate) molar ratios (r) of 1:160 to 1:2. Capillary electrophoresis and Fourier transform infrared (FTIR) difference spectroscopic methods were used to determine the cation binding site, the binding constant, helix stability and DNA conformation in Fe-DNA complexes. Structural analysis showed that at low cation concentration (r = 1/80 and 1/40), Fe(II) binds DNA through guanine N-7 and the backbone PO(2) group with specific binding constants of K(G) = 5.40 x 10(4) M(1) and K(P) = 2.40 x 10(4) M(1). At higher cation content, Fe(II) bindings to adenine N-7 and thymine O-2 are included. The Fe(III) cation shows stronger interaction with DNA bases and the backbone phosphate group. At low cation concentration (r = 1:80), Fe(III) binds mainly to the backbone phosphate group, while at higher metal ion content, cation binding to both guanine N-7 atom and the backbone phosphate group is prevailing with specific binding constants of K(G) = 1.36 x 10(5) M(-1) and K(P) = 5.50 x 10(4) M(-1). At r = 1:10, Fe(II) binding causes a minor helix destabilization, whereas Fe(III) induces DNA condensation. No major DNA conformational changes occurred upon iron complexation and DNA remains in the B-family structure.     

A time-resolved fluorescence study of 4',6'-diamidine-2-phenylindole dihydrochloride binding to polynucleotides

At phosphate/dye (P/D) ratios greater than 30 the quantum yield of 4',6'-diamidine-2-phenylindole dihydrochloride (DAPI)-DNA and DAPI-poly(d(A-T)) complexes was found to be 0.62 and 0.66, respectively. Contrary to earlier reports a fluorescence enhancement of DAPI-poly(d(G-C)) complexes was observed with a quantum yield of 0.22. Time-resolved fluorescence measurements of complexes with a P/D ratio of 150:1 indicate that there were three fluorescent components in DAPI-DNA complexes with lifetimes of 3.86, 1.79 and 0.13 ns. In DAPI-poly(d(A-T)) complexes the lifetimes were 3.91, 1.20 and 0.11 ns. Also, three components with lifetimes of 3.98, 0.87 and 0.12 ns were found in DAPI-poly(d(G-C)) complexes. At low P/D ratios (< 5) another binding form of DAPI was observed which was assigned to the interaction of one or more molecules of DAPI with one previously bound to DNA. It is concluded that DAPI does not exhibit A-T binding specificity and that at high P/D ratios there are two types of binding having similar binding constants.     

Interaction of lac repressor with alternating poly d (A-T) and poly d (G-C). Circular dichroism studies

The binding of lac repressor to poly d(A-T) and poly d(G-C) has been studied using circular dichroism. The results indicate that the binding induces the same conformational change of both polynucleotides and perturbs the same number of nucleic acid bases (28 bases). It is shown that in 0.1 M phosphate buffer the CD measurement can be used to determine the binding constant of lac repressor to poly d(A-T). Competition experiments performed at various salt concentrations show that the stronger interaction of lac repressor for poly d(A-T) than for poly d(G-C) is based on difference in the dissociation rate of the complexes whereas the association rate for both polymers are similar.     

Characterization of ellipticine binding to native calf thymus DNA

The binding of [14C]ellipticine to native calf thymus DNA was studied using equilibrium dialysis. A Scatchard polt revealed the presence of high-and low-affinity binding sites in DNA, the former having a K of 4.0 X 10(7) M(-1) and an n (saturation limiting of binding) of 0.078 (1mol ellipticine/13 mol of DNA nucleotides). The forces involved in stabilizing the high-affinity binding, which has been equated with intercalative binding, were due to a combination of hydrophobic interactions and hydrogen bonding. Difference spectra of ellipticine in the presence of the polydeoxynucleotides, poly d(A-T) or poly d(G-C), showed that there was no base specificity involved in the high-affinity binding. Ellipticine binding to the low-affinity sites, which has been equated with surface binding, was due primarily to the participation of electrostatic interactions of ellipticine with the anionic phosphate groups on the double helical surface of DNA.     

Binding of meso-tetrakis(N-methylpyridiniumyl)porphyrin isomers to DNA: quantitative comparison of the influence of charge distribution and copper(II) derivatization

Factors influencing the binding of tetracationic porphyrin derivatives to DNA have been comprehensively evaluated by equilibrium dialysis, stopped-flow kinetics, etc., for mesotetrakis (4-N-methylpyridiniumyl)porphyrin [TMpyP (4)]. Technical difficulties have previously precluded a comprehensive study of metalloporphyrins. Since electrostatic interactions with the DNA and metal derivatization of the porphyrins have important consequences, we have investigated in greater detail two isomers of TMpyP (4) (meso-tetrakis(3-N-methylpyridiniumyl)porphyrin, [TMpyP (3)] and meso-tetrakis(2-N-methylpyridiniumyl)porphyrin [TMpyP (2)]) in which the position of the charged centers has been varied. A comprehensive study of the Cu(II) derivatives, e.g., CuTMpyP (4), was possible since the difficulties encountered previously with Ni(II) compounds were not a problem with Cu(II) porphyrins [J. A. Strickland, L. G. Marzilli, M. K. Gay, and W. D. Wilson (1988) Biochemistry 27, 8870-8878]. At 25 degrees C, the apparent equilibrium constants [Kobs] decreased with increasing [Na+] for all porphyrins. The Kobs values were comparable for TMpyP (4) and TMpyP (3) binding to either polyd(G-C).polyd(G-C) [poly[d(G-C)2]] or poly[d(A-T)].poly[d(A-T)] [poly[d(A-T)2]]. For the copper(II) porphyrins, the Kobs values were about fivefold greater. The Kobs value for CuTMpyP (2) binding to poly[d(G-C)2] was too small to measure under typical salt conditions; however, Kobs for binding to poly[d(A-T)2] was about two orders of magnitude smaller than those found for CuTMpyP (4) or CuTMpyP (3). Application of the condensation theory for polyelectrolytes suggests about three charge interactions when CuTMpyP (4), CuTMpyP (3), and TMpyP (3) bind to poly[d(G-C)2] or poly[d(A-T)2], a result comparable to that reported for TMpyP (4). At 20 degrees C and 0.115 M [Na+], incorporation of copper decreased the rates of dissociation from poly[d(A-T)2] by a 100-fold compared to those reported for TMpyP (4) but had little effect on the rates of dissociation from poly[d(G-C)2]. Also, movement of the H3CN+ group from the fourth to the third position of the pyridinium ring enhanced the rates of dissociation from poly[d(A-T)2] but decreased the rates of dissociation from poly[d(G-C)2]. From polyelectrolyte theory, the [Na+] dependence of the dissociation rates from poly[d(G-C)2] is consistent with intercalative binding, while that for poly[d(A-T)2] is consistent with an outside binding model. For calf thymus [CT] DNA at 20 degrees C, a greater decrease in the AT than in the GC imino 1H-nmr signal was observed upon addition of CuTMpyP (2), suggesting selective outside binding to the AT regions.(ABSTRACT TRUNCATED AT 400 WORDS)     

Study of DNA interactions with steroidal and nonsteroidal estrogen-platinum (II)-based anticancer drugs

The interactions of the steroidal and nonsteroidal estrogen-platinum (Pt) (II)-based anticancer drugs 16beta-hydroxymethyl-16alpha-[8-(2-pyridin-2-yl-ethylamino)-3,6-dioxaoctyl]-1,3,5(10)-estratrien-3,17betadiol dichloroplatinum (II) (JPM-39), 4-[6-(2'-pyridylethylamino)-butyloxy)-phenyl]-7-methoxy-2,2-dimethyl-3-phenyl-chroman dichloroplatinum (II) (ATG-99), and 1-[(2-aminoethyl)amino]-9,10,10-tris(4-hydroxyphenyl)-9-decene dichloroplatinum (II) (GEB-28) with calf-thymus DNA in vitro using constant DNA concentration and various drug levels were studied. Fourier transform infrared (FTIR) and circular dichroism (CD) were studied with calf-thymus DNA in vitro using constant DNA concentration and various drug levels. FTIR, UV-visible, and CD spectroscopic methods were used to characterize the drug binding mode, the binding constant, and structural variations of DNA in aqueous solution. Spectroscopic evidence showed that the various Pt-based drugs bind indirectly to the major and minor grooves of DNA duplex with some degree of drug-phosphate interaction. The overall binding constants for JPM-39, GEB-28, and ATG-99 are K(JPM-39) = 4.2 (+/-0.75) x 10(3) M(-1), K(GEB-28) = 3.4 (+/-0.65) x 10(3) M(-1), and K(ATG-99) = 2.1 (+/-0.45) x 10(3) M(-1). DNA aggregation occurs at high drug concentration, while DNA remains in the B-family structure.     

Interaction of tRNA with Safranal, Crocetin, and Dimethylcrocetin

Saffron is the red dried stigmas of Crocus sativus L. flowers and used both as a spice and as a drug in traditional therapeutic. The biological activity of saffron in modern medicine is in development. Its numerous applications as an anti-oxidant and anti-cancer agent are due to its secondary metabolites and their derivatives (safranal, crocins, crocetin, dimethylcrocetin). The aim of this study was to examine the interaction of transfer RNA with safranal, crocetin, and dimethylcrocetin in aqueous solution at physiological conditions. Constant tRNA concentration (6.25 mM) and various drug/tRNA (phosphate) molar ratios of 1/48 to 1/8 were used. FT-IR and UV-Visible difference spectroscopic methods have been applied to determine the drug binding mode, the binding constants and the effects of drug complexation on the stability and conformation of tRNA duplex. External binding mode was observed for safranal crocetin and dimethylcrocetin, with overall binding constants K(safranal) = 6.8 (+/- 0.34) x 10(3) M(-1), K(CRT) = 1.4 (+/- 0.31) x 10(4) M(-1), and K(DMCRT) = 3.4 (+/- 0.30) x 10(4) M(-1). Transfer RNA remains in the A-family structure, upon safranal, crocetin and dimethylcrocetin complexation.