Hydrophobic contribution to the free energy of complexation of aromatic ligands with DNA

The hydrophobic component of complexation energy of double-stranded DNA with biologically active aromatic compounds was calculated using two semi-empirical methods correlations of hydrophobic energy with changes of heat capacity (DCp) and solvent-accessible surface area (SASA). These surface areas were calculated for free ligands and DNA oligomers, unwound DNA duplexes and DNA-ligand complexes. The changes of polar and non-polar SASAs of molecules upon binding ligands to DNA were found. The hydrophobic contribution at both complexation stages was calculated. It was shown that the calculation of hydrophobic energy by SASA method is more correct than (DCp) method for DNA-binding ligands.

In tro duc tion. Bi o log i cally ac tive com pounds (BAC), which be long to ar o matic mol e cules, play an im por tant role in the reg u la tion of many sig nif i cant pro cesses in the liv ing or gan ism and can be widely used in clin i cal prac tice for the treat ment of dif fer ent dis eases. As an ex am ple we can con sider the group of ar o matic an ti bi ot ics such as daunomycin, actinomycin D, novantron etc, which are widely used now a days as ba sic el e ments for che mo ther apy of cancer [1]. Be sides, some ar o matic com pounds pos sess sig nif i cant chemotherapeutic prop er ties. Ac cord ing to the lit er a ture data the mech a nism of the ar o matic BAC ac tion can be ex plained by their im me di ate binding with cel lu lar DNA or DNA-de pend ent pro teins [2][3][4]. This mech a nism is de ter mined by the pres ence in a ligand struc ture of a plane ar o matic chromo phore which pro vides a pos si bil ity of the mol e cule in ter ca lation be tween the base pairs of nu cle o tide se quence.
It was es tab lished that sta bi li za tion of such complex oc curs via stack ing in ter ac tions be tween the ligand chromo phore and ad ja cent base pairs of the inter ca lat ing cav ity, and from the other hand, due to the intermolecular hy dro gen bond ing [4][5][6]. Stack ing inter ac tions in clude van der Waals [7], hy dro pho bic [8] and elec tro static [9] in ter ac tions. Though the ther mody nam ics of ar o matic lig ands bind ing with DNA has been well stud ied [10,11], a rel a tive con tri bu tion of above men tioned in ter ac tions into a to tal en ergy of complexation is still a sub ject of cur rent dis cus sions. The rea son is a lack of com mon ap proach for the calcu la tion of prin ci pal en ergy con tri bu tions into a free en ergy DG of the complexation re ac tion between ar o -matic lig ands and DNA. Par tic u larly, hy dro pho bic con tri bu tion DGhp, which de pends on re leas ing some wa ter mol e cules, bound to DNA and the a ligand upon in ter ca la tion, is one of the most im por tant com ponents of the en ergy of com plex for ma tion with DNA. How ever, at pres ent there are at least two dif fer ent approaches to es ti mate this con tri bu tion: 1) a com pu tation ap proach us ing mo lec u lar dy nam ics [12] and proba bil is tic meth ods [13]; 2) an em pir i cal ap proach, based on the cor re la tion be tween en ergy of hy dro phobic sol va tion and change in SASA (Sol vent Ac ces sible Sur face Area) [14,15], or with change in DCp at the complexation [16]. The sec ond approache is a rou tine method for es ti ma tion of hy dro pho bic con tribu tion for DNA-bind ing ar o matic lig ands [11,17], mean while there is still no com par a tive anal y sis of the SASA and DCp meth ods. It is note wor thy, that es tima tion of hy dro pho bic con tri bu tion dur ing ligand inter ca la tion into DNA mol e cule ac cord ing by SASA method was pre vi ously done only for an ti bi otic daunomycin and its de riv a tives [17]. In the pres ent work we have cal cu lated hy dro pho bic con tri bu tion DGhp to the to tal en ergy of complexation of dif fer ent ar o matic lig ands ( Fig. 1) such as actinomycinD (AMD), daunomycin (DAU), nogalamycin (NOG), novantrone (NOV), ethidium bro mide (EB) and proflavin (PF) with a model DNA frag ment us ing corre la tion of DGhp with SASA and DCp. We have also re viewed both ap proaches.
Ma te ri als and meth ods. The struc tures of ligands and DNA-re cep tor. Three-di men sional structures of the in ves ti gated lig ands, pre sented in Fig.1, were taken from Pro tein Data Bank [18] (PDB IDs 1OFV, 1JO2, 1LOR, 2FUM and 1QVT cor re spondingly). Van der Waals ra di uses of the ligand at oms cor re spond to AMBER 99 force field [19]. It is known that the spec i fic ity of these in ter ca lat ing agents to certain DNA se quences is not very strong, but, ac cord ing to the pre vi ous data [20][21][22][23], all men tioned lig ands show spec i fic ity to CG-and GC-site. Tak ing into consid er ation, that DAU has more af fin ity to the trip let CGA-sites of DNA [20] in stead of CG-dinucleotides, we used self-com ple men tary frag ment d(TCGA) 2 flanked by CG from both sides as a min i mal site for the ligand at tach ment. The self-com ple men tary decamer d(GCGTCGACGC) 2 was used as a model frag ment of DNA. Three-di men sional struc ture of B-DNA du plex was con structed us ing HyperChem 8.0 (Hy per cube Inc., Can ada).
The struc ture of com plexes. Dur ing the lligands inter ca la tion their chromo phores are in serted into the cen tral CpG-site of the du plex. Van der Waals ra di uses and charges of DNA at oms in this work are cor re spondent to AMBER 99 force field [19]. The spa tial structure of the com plexes of DNA oli go mers with BAC were con structed us ing X-PLOR, ver sion 3.1 [24].
The complexation of intercalators with DNA could be con sid ered as two-stage pro cess: the for mation of in ter ca lat ing cav ity in DNA du plex (i.e. tran sition of DNA he lix from B-form into un wound DNA*) and the ligand in ser tion into un wound DNA* [17,25]. The for ma tion of in ter ca la tion site was ac com plished in di vid u ally for each ligand mov ing half of the decamer at oms along the DNA he lix axis on the distance 0.34 nm and turn ing them around axis through an gle DW. The value of the an gle DW was equal to the value cal cu lated dur ing BAC in ter ca la tion into double DNA he lix [26,28]. To make fur ther minimization by po ten tial en ergy in X-PLOR the ini tial struc tures of the ligand-DNA com plexes were set on the ba sis of the data about the in ter ca la tion char ac ter of these mole cule into DNA: it was known that AMD, DAU and EB dur ing complexation pro cess were in serted into CpG-site from the side of the mi nor groove [3,29]; NOV and PF were in serted from the side of ma jor groove [30]; dur ing NOG in ter ca la tion amino sugar ring was lo cal ized in the ma jor groove and sugar res idue (nogalose) was in the mi nor one [29]. Op ti mi zation of the struc tures of the com plexes was done by minimization of their po ten tial en ergy us ing the method of con ju gated gra di ents. The cen tral parts of the spa tial struc tures of DNA decamer com plexes with ar o matic lig ands are shown in Fig.2.
Re sults and dis cus sion. The cal cu la tion of hy dropho bic con tri bu tion us ing cor re la tion of DChp with the SASA change. The hy dro pho bic con tri bu tion was cal cu lated ac cord ing to the ex pres sion (1) [32,33], which is based on the pre vi ously proved lin ear cor rela tion be tween en ergy of hy dro pho bic sol va tion of hy dro car bons and amino ac ids: where g-mi cro scopic sur face ten sion co ef fi cient; DA-the change of SASA at the stages of complexation. Some au thors used the dif fer ent value of g (see dis cussion in [15]), but most of them used g=50 cal/ (mol×A°2) for cal cu la tion of en ergy of pro tein-nu cleic acid binding [33] as well as for en ergy of DNA ni tro gen bases stack ing of [34] and ligand-DNA com plexes [17], consid er ing the dif fer ence of mo lec u lar size of lig ands and sol vent [15]. The to tal SASA of the ligand-DNA complex in cludes Ap, po lar (hy dro philic) and Anp, non-polar (hydrophobic) contributions [11]: where DAnp=Anp(com plex)-{Anp(DNA)+ +Anp(free ligand)}; (3) The to tal SASA and its con tri bu tions were cal culated here us ing GETAREA, ver sion 1.1 [35]. The solvent mol e cule (wa ter) was pre sented as a sphere with ra dius of 0.14 nm (value of van der Waals ra dius for oxy gen atom in wa ter mol e cule [35]). SASA is a sur face area formed by the move ment of the cen tre of probe sphere on van der Waals sur face of ei ther sol ute mol ecules or com plex [36].In turn van der Waals sur face of the mol e cule is an as sem blage of spheres with the centers co in cid ing with the cor re spon dent at oms and radius val ues equal to van der Waals ones for the given at oms. The re sults of the calculations are presented in Table 1.
The cal cu la tion of hy dro pho bic con tri bu tion us ing cor re la tion of DGhp with the change of heat ca pac ity DCp in complexation re ac tion. As it was pre vi ously shown, the en ergy of hy dro pho bic so lu tion has lin ear cor re la tion with the change of heat ca pac ity DCp in the complexation reaction [37]: The change of heat ca pac ity also lin early cor re lated with the change of non-po lar SASA Anp [37] or (as it was shown in the later work by the same au thors [38]) with the change of po lar and non-po lar SASA of the pro tein-nu cleic ac ids com plexes. Ren J. et al. [11] carried out the cor rec tion of lin ear cor re la tion DGhp with DAnp and DAp on the ba sis of ex per i men tal ca lo ri metric data for the ligand-DNA complexes as the following: We used the ex pres sions (5) and (6) to es ti mate hydro pho bic con tri bu tion in complexation re ac tion of aro matic lig ands with DNA. The cal cu la tions re sults are shown in Table 1.
The com par i son of cal cu la tion meth ods of hy dropho bic con tri bu tion to the en ergy of complexation reac tions be tween intercalators and DNA. The cal cu lation data of DGhp (see Table1) are in agree ment with cor re spon dent en er gies at the stage of ligand in ser tion EB and DAU into DNA cal cu lated by method (5) in [11] and with the val ues of DGhp for DNA un twist ing and DAU in ser tion by the method (5) in [17]. It gave us the ba sis to make a com par i son be tween cal cu la tion meth ods for DGhp de ter mi na tion, which were used by dif fer ent au thors. The most im por tant con clu sion, which can be made af ter DG com par i son (see Ta ble 1) is the fol low ing: the val ues of hy dro pho bic en ergy calcu lated by dis tinct meth ods dif fer by two (EB) or even ten (AMD) times! This es sen tial dif fer ence is prin ci pal and needs to be ex plained.
There are three main as sump tions, which are used as the base for DGhp cal cu la tion as the change of heat ca pac ity (5) [31,[37][38][39]: -ex is tence of lin ear in ter re la tion be tween enthalpy and en tropy of hydrophobic dissolution or between complex formation and the change of heat capacity; -the change of heat ca pac ity is com pletely de termined by hy dro pho bic ef fect; -the change of heat ca pac ity is lin early con nected with the SASA change (see equa tion (6)).
The first as sump tion is ex per i men tally ver i fied for the large set of hy dro car bons and some pro teins [37,39]. The sec ond as sump tion in its turn sup poses neg ligi ble con tri bu tion in the change of vi bra tion de grees of free dom, it can be in di rectly proved for the re ac tions with pro tein par tic i pa tion [37]. And the third as sump - tion, as a rule, is a re sult of struc tural ther mo dy namic anal y sis of ex per i men tal data [37,38]. Above men -tioned con di tions are the basis for both expression (6) and the following ratio: Where Th=295 K; Ts=386 K-the val ues of tem pera ture, when the enthalpy and entropic con tri bu tion are cor re spond ingly equal to zero [37,39]. If T=Th we can get (5) from (7). As it fol lows from the ta ble data and (6) the con tri bu tion of po lar com po nent in SASA change is at least three times less as com pared with non-po lar one. So we can ne glect the sec ond summand in (6) and, tak ing into con sid er ation (5), obtain: The com par i son of (8) and (2) al lows to im ply at least two times dif fer ence in DGhp de ter mi na tion by SASA method and DCp method (tak ing into ac count, that dur ing the bind ing of the lig ands with DNA DA>DAnp, as it fol lows from the ta ble). So we can make a con clu sion, that there is a sys tem atic dif fer ence between two con cerned meth ods for es ti ma tion of hy dropho bic con tri bu tion of DNA-in ter ca lat ing lig ands. We think that there are two main rea sons, ac cord ing to which we can prove the in cor rect ness of the us age of DGhp cal cu la tion method based on the heat ca pac ity change (6), (7) for DNA-in ter ca lat ing lig ands. In the first place as it fol lows from (5), the co ef fi cient value 80 in (5) is com pletely de ter mined by the val ues of tran sition tem per a tures Th and Ts in (7). In turn these tem pera ture val ues are ex per i men tally es tab lished for aliphatic hy dro car bons [31,37,39], but it is known that even for el e men tary ar o matic groups the tem per a ture val ues of Th and Ts sig nif i cantly change [31]. There fore, to make DGhp cal cu la tion dur ing the bind ing of ar o matic lig ands with DNA by the method (5) we rec om mend the us age of the cor rected co ef fi cient value at DCp.
Sec ondly, the cal cu la tion method based on heat capac ity change can be cor rect enough only in cases when hy dro pho bic in ter ac tions give the main con tri bu tion into the re ac tion en ergy of sol va tion or com plex for mation (as it takes place for pro teins) [37]. Un der this condi tion the change of heat ca pac ity is ex pected com par atively big ger and cor re la tions (6), (7) are sta tis ti cally sig nif i cant. In the case of com plex for ma tion be tween intercalators and DNA the change of heat ca pac ity is sig nif i cantly less as com pared with pro tein sys tems (see data [11] and [37]) and hy dro pho bic in ter ac tions are not com pletely dom i nated [11,17,21]. Be sides, the con tri bu tion of vi bra tion de grees of free dom in to tal value of heat ca pac ity of the re ac tion is still un clear. In con trast to above men tioned method for cal cu la tion of hy dro pho bic con tri bu tion the SASA method (1) is based on the sin gle as sump tion of lin ear cor re la tion between the en ergy of hy dro pho bic dis so lu tion of hy drocar bons and the SASA value, which is ex per i men tally and the o ret i cally proved [15,34]. The main prob lem of the method is a search for the cor rect value of g co ef ficient cor re spond ing to the real hy dro pho bic con tri bution. Re cently we have "cal i brated" g co ef fi cient [40] and proved the cor rect ness of us ing its stan dard value g=50cal/ (mol×A°2) for ar o matic lig ands. Con se quently, we have car ried out the anal y sis of hy dro pho bic con tribu tion by the SASA method (see DGg in the table).
The anal y sis of hy dro pho bic con tri bu tion to the energy of com plex for ma tion be tween lig ands and DNA. The ob tained val ues of DA and DGhp de scrib ing the bind ing of DNA du plex and DAU (see Ta ble) at the both stages of in ter ca la tion are in ac cor dance with [17]; the small dif fer ence can be ex plained by the fact, that in [17] the au thors used van der Waals ra dius of the at oms cor re spond ing to CHARMM force field. Ac cord ing to the ta ble data hy dro pho bic ad van tage of the ligand bind ing to DNA is in the fol low ing or der: NOG>AMD>DAU>NOV>EB>PF. The ob tained order is in good cor re la tion with the de gree of branch ing of the ligand side chains: from mas sive mol e cules NOG and AMD to com par a tively small PF mol e cule con taining only two hy dro philic amino groups in 3 and 6 po sitions of the chromo phore (see Fig. 1, f). The or der is also in ac cor dance with the de crease of hy dro pho bic con tri bu tion, which was shown in [11]: AMD>DAU>EB. The brunchness of side chains de termines the ef fi ciency of wa ter dis place ment from hydration shells of DNA and ligand dur ing the com plex for ma tion. So, more brunched side groups of intercalators in sert ing into DNA du plex grooves provide big ger hy dro pho bic con tri bu tion. The ef fect of wa ter dis place ment is in di rectly proved by the fact, that total SASA change at the stage of DNA untwisting and ligand insertion is negative (see Table).