\./ 1990 Oxford University Press
Nucleic Acids Research, Vol. 18, No. 21 6413
A method for determining the relative effect of ligands A-T and G-C base pairs in DNA: applications to metal ions, protons and two amino acids
on
S.G.Haroutiunian, E.B.Dalian, V.M.Aslanian, D.Yu.Lando1 and A.A.Akhrem1 Department of Molecular Physics and Biophysics, Yerevan State University, USSR, Yerevan 375049, Mravian 1 and 1lnstitute of Bioorganic Chemistry of Academy of Science, Belorussian SSR, Minsk 220600, Zjodinskaya, 5, USSR Received December 6, 1989, Revised and Accepted October 3, 1990
ABSTRACT A new method is described for the study of specific interactions of low-molecular ligands with the base pairs of DNA. This method is based on the comparative analysis of melting temperature changes in DNAs of different GC-content in the presence of low molecular weight ligands. In this paper, the method is applied to Mn2 +, Ni2 +, Co2+ ions, deprotonation, amino acids, B-alanine and y-aminobutyric acid (ey-ABA). Differences in Tm are affected not only by the changes of relative stability of AT- and GC-pairs, but also by other factors. A theoretical analysis of the sequence specificity of low-molecular ligands on the base pairs in DNA molecules characterized by a high degree of sequence heterogeneity is also presented.
INTRODUCTION A widely used method for studying the specific interaction of ligands on AT- and GC-base-pairs is based on the analysis of UV-spectrophotomeric differential melting curves (DMC) (1-3). A substance which uniformly stabilizes or destabilizes base pairs (in the absence of ligand redistributions during the melting process) will not modify the form of the DMC of DNA (4, 5). The DMA profile will shift to lower or higher temperatures without changes in the width of the helix-coil transition (AT). However, if sequence specific interaction of ligands (metal ions, for example) with particular DNA pairs is involved, AT will also change (6-10). However, it has been shown previously (9-11) that ligands can have both specific and nonspecific interactions with DNA molecules. Certain metal ions (9, 10) or other agents (11) can induce changes in the structure of DNA undergoing a thermal transition (for example, intra- and intermolecular aggregate formation). As a result, the character of the melting process is drastically altered: instead of a simple helix-coil transition, a transition between helix and an aggregated form of DNA can result (10). Such phenomena can lead to additional narrowing or broadening of the melting curve, hence to erroneous conclusions concerning the AT- or GC-specificity of ligand binding.
We believe that conclusions concerning specific effects of a ligand on DNA can be inferred only by analysis of the changes in melting temperatures in a series of DNAs of different GCcontent, in the presence of the ligand in question. If sequence specific effects are convoluted with nonspecific effects, the nonspecific factors can be distinguished by the following procedure: 1) from the melting curves of a series of different G-C content, the dependence of Tm on GC-content of DNA for different concentrations of ligands is determined; 2) from this dependence, TGc-TAT differences are calculated; 3) knowing Toc-TAT and using the correlation between AT and TcC-TAT for DNAs with a high degree of sequence heterogeneity (12), that part of the change in AT which is due exclusively to sequence may be evaluated; 4) subtracting this value from the total change of AT the contribution of nonspecific factors to the change of this parameter is determined. In this paper, we apply this method to study the influence of Mn2+, Ni2+, Co2+ ions, deprotonation, and the amino acids 3alanine and aminobutyric acid (^y-ABA) on the character of the melting process and relative stability of DNAs AT- and GC-pairs.
EXPERIMENTAL DNA from calf thymus (GC = 42%), E. coli (GC = 52%), M. luteus (GC = 72%), and Cl. perfringens (GC = 28%) were obtained from Sigma. DNA from Molluscus modiolus sp. (GC = 32%) was prepared at Moscow State University. The molecular weight in each case is approximately 107 daltons. Metals were introduced as the chlorides. The ion concentration was defined by weight and controlled by titration with AgNO3 in the presence of indicator K2MnO4. Ion concentrations used were from 10-6 to 10-3 M. Ion concentrations were normalized to one mol per phosphate. Preparations of t-alanine and y-ABA from Reanal (Budapest, Hungary) were used after their recrystallization. DMC profiles were obtained by numerical differentiation of initial melting curves. To decrease the quantity of contaminating ions solutions were stored in polyethylene vessels; triple distilled water was used throughout. All experiments were carried out in 1 mM NaCl with IxSSC. (ISSC: 0.15 M NaCl + 0.015 M Na citrate at pH 6.8 and ion strength to [Na+] = 0.192).
6414 Nucleic Acids Research, Vol. 18, No. 21 UV-absorbance spectra and melting curves were measured on a UNICAM SP-8000 spectrophotometer equipped with thermostated cells and on the SPECORD UV-VIS. Calorimetric
aT
(b)
(a)
measurements were carried out with a differential adiabatic
microcalorimeter DASM-1M. Circular dichroism spectra were recorded on a ROUSSEL-JOUAN-2 dichrograph.
.
ii
iI2.0
0.15 i
(a)
I
i
C2+
j\4.0
i/ I''I
0
I,
I.
0.5 ~.,if
1
0.09
N
'
.21.0
.e
I,,
I .
iI
0*03 "
A... "
58
70
i..y-I
\
1'~~~~~~ 94
82
It'
64
76
88
58
Fig 1. DMC of calf thymus DNA in the presence of (a) Mn2+, (b) Co2+ and (c) Ni2+ ions: (--) (-*- ) 2.0 M/P and (--- ) 4.0 M/P ions in 1mM NaCI.
1
2
70
82
pure DNA in 0.1
94
TOC)
SSC, ( ....) 0.5M/P, (- -) l.OM/P, -
Table 1. Tm and AT of DNAs and TGC-TAT at different concentrations of Mn2+, Co2+ and Ni2+ ions. The ionic strength is 1 mM [Na+] Type of ions
Source of DNA
Co2+
Ni2+
Calf thymus
(XGC = 42%)
Me2+(M/P) Tm(OC) 0
Mn2+
M. luteus
(XGC = 72%)
0.40 0.87 3.50 10.00 0.5 1.0 4.0 10.0 0.5 1.0 4.0 10.0
AT(0C)
Cl. perfringens (XGC = 28%)
Tm(OC)
AT(0C)
Tm(OC)
AT(0C)
78.6
9.1
63.0
18.0
54.0
12.0
88.5 84.7 79.5 75.0 87.4 86.5 80.7 76.4 84.9 79.4 74.3 71.7
2.9 2.0 1.7 1.6 4.3 2.0 1.7 1.7 4.0 2.0 2.0 2.0
75.7 77.0 74.8 72.0 76.2 78.0 78.5 74.6 75.3 76.0 74.6 71.5
9.0 6.6 2.3 2.0 11.1 8.2 3.4 1.8 8.3 6.3 2.0 1.3
71.2 73.7 73.4 70.4 66.3 71.5 75.3 74.0 69.9 72.5 73.4 71.5
4.0 3.3 1.6 0.9 5.1 3.7 2.0 0.6 4.8 3.7
AT is defined by tangent in the transition point of melting curve (AT
= l/(8vlat)
1.6 0.7
TGC-TAT
39.0 25.6 13.6 10.0 48.6 34.0 12.3 5.4
34.0 15.2 1.8 0.0
v = 0.5)
90 70
Fig 2. Dependences of Tm concentrations in M/P).
on
GC-content of DNAs
at
different (a) Mn2+, (b) Co2+ and
(c)
Ni2+
concentrations. (The numbers
on
the lines showed the ions
Nucleic Acids Research, Vol. 18, No. 21 6415
RESULTS AND DISCUSSION
approximately equals 1.7-2.0°C, while for DNA with low GC= 28%), AT decreases more weakly (0.6-0.7°C). This effect can be explained by the fact that the metal ions which selectively are adsorbed on GC-pairs interact with each other. Such an interaction can lead to the increase of AT. In AT-rich DNAs, this interaction should be weaker, since the distance between GC-pairs is greater. Hence, AT for the latter DNAs is lower and corresponds to the values obtained in solutions of alkyl ammonium salts (1, 3). From the above, we conclude that the decrease of AT in DNA observed in the presence of the transition metal ions is not simply due to sequence specific effects of AT- and GC-pairs. In the case of Mn2+, Co2+ and Ni2+, nonspecific factors are observed as well (aggregate formation), as a result of which the value of AT slightly increases. Experiments on DNAs of different GC-content also show that aggregation caused by Mn2+, Co2+, and Ni2+ ions is more pronounced in GC-rich DNAs. This is confirmed by the fact that optical density increases at X = 320 nm with the increase of DNA GC-content. The effect is especially evident in the case of Ni2+ ions (the increase of optical density at X = 320 nm on GC-rich DNA has already been observed at concentrations M/P = 0.5). In the case of the amino acids ,3-alanine and y-ABA, the behavior of AT is fundamentally different. The experimental data, presented in Figure 3 and Table 2, indicate a monotonic decrease in the width of helix-coil transition on increasing the alanine concentration. Moreover, AT is changed more strongly than the content (GC
Figure 1 shows the DMC of calf thymus DNA in the presence of Mn2+, Ni2+ and Co2+ ions. In the presence of these ions a significant change in the shape of the melting curve takes place. In particular, at concentrations of Mn2+ (M/P = 2) a sharp narrowing of helix-coil transition width is seen. A similar effect is observed in the presence of Ni2+ and Co2+ ions. The well resolved peaks in the high temperature regions of DMC that characterize the melting of GC-rich regions of DNA merge and transform into one narrow peak. This observation suggests that at stoichiometric concentration of transition metal ions, a shift in the relative stability of AT- and GC-pairs takes place: ATpairs are stabilized, while GC-pairs are destabilized. This suggests in turn that the influence of metal ions on DNA of different GCcontent should be different. Table 1 and Figure 2 summarize the dependence of the melting temperature of DNAs of different GCcontent at different on the concentration of Mn2+, Ni2+ and Co2+ ions. The differences of TGc-TAT were calculated by means of these dependences. As noted earlier, the DNA of higher GC-content the value of AT is in direct proportion with TGCTAT difference (12). Thus, at a given concentration of metal ions, AT and parameter TGC-TAT should change in the same way, if these changes are conditioned only by influences of ions on GC-pairs; although the results in Table 1 indicate that the AT of DNA in presence of given metal ions changes more weakly than the parameter TGC-TAT. This effect is especially strong for Ni2+ ions, where the parameter TGc-TAT decreases down to zero (0) at concentrations such that Ni2+ M/P = 5; AT of calf thymus DNA decreases to 2°C. Note that for DNA with high GC-content (GC = 72 %), the minimal value of AT
Tm(0C)
parameter
I.:,,
80 0
72
64
0
40
80
TGC-TAT.
These results suggest the following conclusions: (1) the narrowing of AT is mainly due to factors other than sequence specificity in the interaction of amino acids, the action of which causes a change in the relative stability of DNA regions melting at different denaturation degrees; (2) both the specific and non-specific effects of these amino acids increase with their concentration; (3) ,B-alanine has different influences on stability of AT- and GC-pairs only at concentrations above 1.5 M, that is, at concentrations of amino acid where its influence on the DNA is probably due to a change of structure of the solvent (13, 14). Analogous results are obtained for 'y-ABA, with the difference that the specificity is more pronounced in this case. It gives the main contribution in the change of AT (80-90%) and begins to manifest itself at lower concentrations ( IM). In fact, the complete quantitative coincidence of the microcalorimetric curves ([DNA] = 1 mM, Fig. 4a) and DMC ([DNA] = 0.01 mM, Fig. 4b) indicates that neither specific nor nonspecific influences of these amino acids are apparently connected with intermolecular aggregations. Possibly, the additional decrease of AT is connected
XGC(%)
Fig. 3. Dependences of Tm on GC-content of DNAs at different ,B-alanine concentrations: (0) pure DNAs in 0.1 SSC, (A) 1.5 M and (0) 5 M ,B-alanine.
Table 2. Tm and AT of DNAs at different ,B-alanine concentrations in 0.1 SSC Source of DNA Concentration of Alanine (M)
Modoilus sp. (XGC = 32%) AT(0C) Tm(OC)
0
63.5 60.4 60.2 64.4
1.0 1.5 5.0
4.6 4.0 3.6
1.5
Calf thymus
E. coli
M.luteus
(XGc = 42%) AT(0C) Tm(OC)
(XGc = 52%) AT(0C) Tm(OC)
Tm(OC)
AT(0C)
72.9 68.9 68.6 72.3
86.3 83.2 81.0 80.7
4.3 4.0 3.9 2.8
10.7 7.6 6.7 4.0
69.9 66.1 65.7 69.2
AT is defined by tangent in transition point of melting curve
(AT
=
1/(av/8t)v
=
0.5)
4.8 3.8 3.7 2.6
(X GC = 72%)
6416 Nucleic Acids Research, Vol. 18, No. 21 TT
(b)
n
0
i.,
0
t2
0.20
Tm(°C)
~~~I
-7-
90
i; 4.0
F
1.5
0.12
70 /I~~~~~1
10
I-
/
co
0)
0.04
50 .
,
0
59
c
65
77
71
i071 .
T
("C)
30
Fig 4. The microcalorimetric (a) and UV-spectrophotometric differential melting curves (b) of calf thymus DNA at different f-alanine concentrations: (--) pure DNA in 0.1 SSC, (---) 1.5 and (----) 4 M f-alanine.
I/
/
10
q&
UAGI7
4W
r r k7') Uv 'LGC
f
Table 3. The values and changes of parameter TGC-TAT and AT of calf thymus DNA at different pH (at 25°C)
pH 7.3 9.0 9.4 9.8 10.2 10.6 11.0 11.3 11.5
(TGC-TAT)O
ATO
TGC-TAT
AT
(OC)
(OC)
TGC-TAT
40.9 40.9 38.6 34.1 30.8 30.8 38.6 52.4 68.1
8.5 8.5 7.4 6.5 6.4 6.6 8.7 11.4 13.2
-
-
1.00 1.06 1.20 1.33 1.33 1.06 0.78 0.60
1.00 1.13 1.29 1.33 1.30 1.01 0.74 0.64
AT
where (TGC-TAT)O and ATo were taken for pH 7.3.
not only with the specific influence in the region of high concentrations (4 M and higher), but also by the fact that these amino acids form some ordered structure around double-helical DNA (complexes of DNA with zwitterions (11, 14)), which is destroyed cooperatively, when the first melted regions appear. Since such a destruction is energetically unfavorable, the melting begins at higher temperatures. At moderate to high degrees of denaturation, this structure is already destroyed and therefore, for the final part of the transitions, no temperature shift is observed. As a result, a decrease of AT takes place. Next, we consider the deprotonation of DNA bases in alkaline solutions. We have investigated the helix-coil transition under these conditions both theoretically and experimentally (15), and find that the bases are deprotonated in both native, as well as denatured DNA. That is, GC-pairs in the helical form are capable of ionizing at pH 10.6 as well. The data in Table 3 indicates that changes of AT of calf thymus DNA and the TGC-TAT at each value of pH are identical. Consequently, the changes in AT of DNA are conditioned by the specific deprotonation of nitrogen bases. 0% of the From extrapolations to XGC 100% and XGC Tm (XGC) profile for each value of the pH, we can extract the pH dependence of TAT and TGC. The results obtained are given in Fig. 6. With increase of pH, TAT(pH) is continuously decreasing at pH 9, while TGC (pH) reaches saturation around pH 11. Our explanation of this behavior is the following. When the relative content of binding sites in the helix is lower than in the coil, the melting temperature will decrease with increasing pH. Once the quantities of binding sites in the helical and melting =
Fig. 5. Dependences of Tm on GC-content of DNAs at different pH. (The numbers on the lines showed the values of Ph).
=
Table 4. Calculation values TGC-TAT and AT of calf thymus DNA at different relative ligand concentrations (C)? (ATo = 10°C, (TGC-TAT)O = 40°C) C 0 10 25 100 1000
TGC-TAT
AT
(OC)
0C)
(TGC-TAT)O TGC-TAT
ATO
40.0 21.6 5.8 -19.6 -38.8
10.0 5.4 1.5 4.9 9.7
1.00 1.85 6.70 2.00 1.03
1.00 1.85 6.85 2.10 1.03
AT
parts become equal, the melting temperature of polynucleotide will not change above a certain pH. According to the mechanism
have suggested (15), AT-pairs ionize only in the denatured TAT(pH) strongly decreases with the increase of pH. But GC-pairs can ionize in both the native and denatured states of DNA. Hence TGC (pH) reaches saturation. On the basis of these investigations, we can conclude that AT is affected by changes of relative stability of AT- and GC-pairs, as well as by other factors. In the case of metal ions, these factors lead to the increase of AT, while f-alanine and -y-ABA decrease it. During deprotonation of DNA, the change of AT is defined exclusively by changes in relative stability of AT- and GC-pairs of DNA, and AT (YGC-TAT); in this case.
we
state, so that
-
THEORETICAL PART The melting process of a DNA molecule with block distribution of base-pairs in the presence of ligands with specific binding character have been modelled by formulae obtained in references 16 and 17. The influence of ligand binding (16, 17) which are characterized by different constants for adenine, thymine, guanine, cytosine on DNA helix-coil transition is considered. The possibility of interaction between ligands bound on DNA is not taken into account. We can introduce a correction for this effect following reference 17. Assuming that the free energy of ligand binding is i-type region with j-type base is a linear function to the degree of filling up of DNA by ligands, the formulae of (17) have the form:
Nucleic Acids Research, Vol. 18, No. 21 6417 4
Oi,i=oo6j j-W E C i,
(1)
ajl "" Kj, exp[2W,( Li= Cl ,i)/RT,] = I1 Cj,1aij,1 C
4
(2)
-
4
4
1
Tov
_ Tv
R AH
WI(
Cl,i)1-W2(
,
j
C2,1)2 +
+
RTV
l (1 Ci,ii,i) -Rl,i I
(3)
(I -C2 iP2 )R2zi
4
4
L Cl,i+v L C2,i C=CO/M+(l-v) i=l =l '
(4)
where j = 1 corresponds to the melted regions on j = 2 to helical regions; i- corresponds to the four bases, A, T, G or C, respectively, Kji = exp(-kO,j/RT) the binding constant with jbasis at the degree of filling type regions according to the i the up of DNA by ligands which strive to zero; Wj proportionality coefficient; Cj,j the degree of filling up of DNA by ligands bound with nitrogen bases i in j-type regions; CO- the molar concentration of free ligands; R - the gas constant; Tv the absolute temperature at which the degree of helicity of DNA complex with ligands equals v; T., - the same the helix-coil parameter for DNA without ligands; AH transition enthalpy; ajj R'jj - the functions from v, which are calculated by the method suggested in (17); C = D/M; D the molar concentration of ligands; M the molar concentration of DNA base pairs. Solving the system of equations (2) together with equation of material balance (4) we obtain 8 values of Cj,i. Inserting them into the equation (3) Tv is determined. In these calculations it is supposed that the ligands interact with one base pair (there is no interaction with AT-pairs), i.e., Ki 1 = K12 = K21 = K22 = 0. For GC-pairs K13 = K14 = 5 x107, and K24 = K23 = 107; AH = 9 kcal/mol base pairs, the melting temperature and width of the transition of calf thymus DNA without ligands is assumed to be 70°C and 10°C, correspondingly (in 0.1 SSC). From Table 4, it follows that in the presence of specific ligands, the ratio of AT -(TGC-TAT) is carried out strictly at any concentrations of ligands. Hence, the decrease or increase of more than the change of (TGC-TAT) is conditioned by the other nonspecific effects mentioned above. -
-
T
(00)
100
-
-
60
-
-
-
20
7
8
9
10
11
pH
-
ACKNOWLEDGEMENT The authors are grateful to Professor M.D.Frank-Kamenetskii for helpful discussion of the results.
REFERENCES 1. Melchior,W.B. and Von Hippel,P.N. (1973) Proc. Natl. Acad. Sci. USA 70, 293-302. 2. Akhrem,A.A. and Lando,D.Y. (1981) Molek. Biol. 15, 1083-1091 (in
Russian). 3. Voskoboinic,A.D. et al. (1985) Molek. Biol. 9, 783-789 (in Russian).
Fig 6. The pH dependences of TAT and TGC (see text). 4. Geiduschek,E.P. and Herskovits,T.T. (1961) Arch. Biochem. Biophys. 95, 114- 129. 5. Herskovits,T.T., Singer,S.J. and Geiduschek,E.P. (1961) Arch. Biochem.
Biophys. 94, 94-114. 6. Forster,W., Bauer,E., Schuts,A. and Berg,H. (1979) Biopolymers 18, 625-661. 7. Shin,J. (1973) Biopolymers 12, 2459-2475. 8. Luck,G. and Zimmer,C. (1972) Eur. J. Biochem. 29, 528-536. 9. Ott,G.S., Bastia,D. and Bauer,W. (1978) Biochem. Biophys. Acta 58, 216-231. 10. Yurgaitis,A.P. and Lazurkin,Y.S. (1981) Biopolymers 20, 967-975. 11. Aslanian,V.M., Haroutiunian,S.G., Lando,D.Y., Dalian,E.B. and Shpakovskii,A.G. (1988) Biofizika 33, 430-436 (in Russian). 12. Berestetskaya,I.V., Frank-Kamenetskii,M.D. and Lazurkin,Y.S. (1974) Biopolymers 13, 193 -205. 13. Aslanian,V.M. and Haroutiunian,S.G. (1984) Biofizika 29, 564-566 (in Russian). 14. Aslanian,V.M. and Haroutiunian,S.G. (1985) Biofizika 30, 741-745 (in Russian). 15. Akhrem,A.A., Haroutiunian,S.G., Aslanian,V.M., Dalian,E.B., Lando,D.Y. and Shpakovskii,A.G. (1989) Molek. Biol. 23, 518-525 (in Russian). 16. Akhrem,A.A. and Lando,D.Y. (1979) Molek. Biol. 13, 1098-1109 (in Russian). 17. Lando,D.Y. et al. (1980) Molek. Biol. 14, 175-181 (in Russian).
Nucleic Acids Research, Vol. 18, No. 21 6413
A method for determining the relative effect of ligands A-T and G-C base pairs in DNA: applications to metal ions, protons and two amino acids
on
S.G.Haroutiunian, E.B.Dalian, V.M.Aslanian, D.Yu.Lando1 and A.A.Akhrem1 Department of Molecular Physics and Biophysics, Yerevan State University, USSR, Yerevan 375049, Mravian 1 and 1lnstitute of Bioorganic Chemistry of Academy of Science, Belorussian SSR, Minsk 220600, Zjodinskaya, 5, USSR Received December 6, 1989, Revised and Accepted October 3, 1990
ABSTRACT A new method is described for the study of specific interactions of low-molecular ligands with the base pairs of DNA. This method is based on the comparative analysis of melting temperature changes in DNAs of different GC-content in the presence of low molecular weight ligands. In this paper, the method is applied to Mn2 +, Ni2 +, Co2+ ions, deprotonation, amino acids, B-alanine and y-aminobutyric acid (ey-ABA). Differences in Tm are affected not only by the changes of relative stability of AT- and GC-pairs, but also by other factors. A theoretical analysis of the sequence specificity of low-molecular ligands on the base pairs in DNA molecules characterized by a high degree of sequence heterogeneity is also presented.
INTRODUCTION A widely used method for studying the specific interaction of ligands on AT- and GC-base-pairs is based on the analysis of UV-spectrophotomeric differential melting curves (DMC) (1-3). A substance which uniformly stabilizes or destabilizes base pairs (in the absence of ligand redistributions during the melting process) will not modify the form of the DMC of DNA (4, 5). The DMA profile will shift to lower or higher temperatures without changes in the width of the helix-coil transition (AT). However, if sequence specific interaction of ligands (metal ions, for example) with particular DNA pairs is involved, AT will also change (6-10). However, it has been shown previously (9-11) that ligands can have both specific and nonspecific interactions with DNA molecules. Certain metal ions (9, 10) or other agents (11) can induce changes in the structure of DNA undergoing a thermal transition (for example, intra- and intermolecular aggregate formation). As a result, the character of the melting process is drastically altered: instead of a simple helix-coil transition, a transition between helix and an aggregated form of DNA can result (10). Such phenomena can lead to additional narrowing or broadening of the melting curve, hence to erroneous conclusions concerning the AT- or GC-specificity of ligand binding.
We believe that conclusions concerning specific effects of a ligand on DNA can be inferred only by analysis of the changes in melting temperatures in a series of DNAs of different GCcontent, in the presence of the ligand in question. If sequence specific effects are convoluted with nonspecific effects, the nonspecific factors can be distinguished by the following procedure: 1) from the melting curves of a series of different G-C content, the dependence of Tm on GC-content of DNA for different concentrations of ligands is determined; 2) from this dependence, TGc-TAT differences are calculated; 3) knowing Toc-TAT and using the correlation between AT and TcC-TAT for DNAs with a high degree of sequence heterogeneity (12), that part of the change in AT which is due exclusively to sequence may be evaluated; 4) subtracting this value from the total change of AT the contribution of nonspecific factors to the change of this parameter is determined. In this paper, we apply this method to study the influence of Mn2+, Ni2+, Co2+ ions, deprotonation, and the amino acids 3alanine and aminobutyric acid (^y-ABA) on the character of the melting process and relative stability of DNAs AT- and GC-pairs.
EXPERIMENTAL DNA from calf thymus (GC = 42%), E. coli (GC = 52%), M. luteus (GC = 72%), and Cl. perfringens (GC = 28%) were obtained from Sigma. DNA from Molluscus modiolus sp. (GC = 32%) was prepared at Moscow State University. The molecular weight in each case is approximately 107 daltons. Metals were introduced as the chlorides. The ion concentration was defined by weight and controlled by titration with AgNO3 in the presence of indicator K2MnO4. Ion concentrations used were from 10-6 to 10-3 M. Ion concentrations were normalized to one mol per phosphate. Preparations of t-alanine and y-ABA from Reanal (Budapest, Hungary) were used after their recrystallization. DMC profiles were obtained by numerical differentiation of initial melting curves. To decrease the quantity of contaminating ions solutions were stored in polyethylene vessels; triple distilled water was used throughout. All experiments were carried out in 1 mM NaCl with IxSSC. (ISSC: 0.15 M NaCl + 0.015 M Na citrate at pH 6.8 and ion strength to [Na+] = 0.192).
6414 Nucleic Acids Research, Vol. 18, No. 21 UV-absorbance spectra and melting curves were measured on a UNICAM SP-8000 spectrophotometer equipped with thermostated cells and on the SPECORD UV-VIS. Calorimetric
aT
(b)
(a)
measurements were carried out with a differential adiabatic
microcalorimeter DASM-1M. Circular dichroism spectra were recorded on a ROUSSEL-JOUAN-2 dichrograph.
.
ii
iI2.0
0.15 i
(a)
I
i
C2+
j\4.0
i/ I''I
0
I,
I.
0.5 ~.,if
1
0.09
N
'
.21.0
.e
I,,
I .
iI
0*03 "
A... "
58
70
i..y-I
\
1'~~~~~~ 94
82
It'
64
76
88
58
Fig 1. DMC of calf thymus DNA in the presence of (a) Mn2+, (b) Co2+ and (c) Ni2+ ions: (--) (-*- ) 2.0 M/P and (--- ) 4.0 M/P ions in 1mM NaCI.
1
2
70
82
pure DNA in 0.1
94
TOC)
SSC, ( ....) 0.5M/P, (- -) l.OM/P, -
Table 1. Tm and AT of DNAs and TGC-TAT at different concentrations of Mn2+, Co2+ and Ni2+ ions. The ionic strength is 1 mM [Na+] Type of ions
Source of DNA
Co2+
Ni2+
Calf thymus
(XGC = 42%)
Me2+(M/P) Tm(OC) 0
Mn2+
M. luteus
(XGC = 72%)
0.40 0.87 3.50 10.00 0.5 1.0 4.0 10.0 0.5 1.0 4.0 10.0
AT(0C)
Cl. perfringens (XGC = 28%)
Tm(OC)
AT(0C)
Tm(OC)
AT(0C)
78.6
9.1
63.0
18.0
54.0
12.0
88.5 84.7 79.5 75.0 87.4 86.5 80.7 76.4 84.9 79.4 74.3 71.7
2.9 2.0 1.7 1.6 4.3 2.0 1.7 1.7 4.0 2.0 2.0 2.0
75.7 77.0 74.8 72.0 76.2 78.0 78.5 74.6 75.3 76.0 74.6 71.5
9.0 6.6 2.3 2.0 11.1 8.2 3.4 1.8 8.3 6.3 2.0 1.3
71.2 73.7 73.4 70.4 66.3 71.5 75.3 74.0 69.9 72.5 73.4 71.5
4.0 3.3 1.6 0.9 5.1 3.7 2.0 0.6 4.8 3.7
AT is defined by tangent in the transition point of melting curve (AT
= l/(8vlat)
1.6 0.7
TGC-TAT
39.0 25.6 13.6 10.0 48.6 34.0 12.3 5.4
34.0 15.2 1.8 0.0
v = 0.5)
90 70
Fig 2. Dependences of Tm concentrations in M/P).
on
GC-content of DNAs
at
different (a) Mn2+, (b) Co2+ and
(c)
Ni2+
concentrations. (The numbers
on
the lines showed the ions
Nucleic Acids Research, Vol. 18, No. 21 6415
RESULTS AND DISCUSSION
approximately equals 1.7-2.0°C, while for DNA with low GC= 28%), AT decreases more weakly (0.6-0.7°C). This effect can be explained by the fact that the metal ions which selectively are adsorbed on GC-pairs interact with each other. Such an interaction can lead to the increase of AT. In AT-rich DNAs, this interaction should be weaker, since the distance between GC-pairs is greater. Hence, AT for the latter DNAs is lower and corresponds to the values obtained in solutions of alkyl ammonium salts (1, 3). From the above, we conclude that the decrease of AT in DNA observed in the presence of the transition metal ions is not simply due to sequence specific effects of AT- and GC-pairs. In the case of Mn2+, Co2+ and Ni2+, nonspecific factors are observed as well (aggregate formation), as a result of which the value of AT slightly increases. Experiments on DNAs of different GC-content also show that aggregation caused by Mn2+, Co2+, and Ni2+ ions is more pronounced in GC-rich DNAs. This is confirmed by the fact that optical density increases at X = 320 nm with the increase of DNA GC-content. The effect is especially evident in the case of Ni2+ ions (the increase of optical density at X = 320 nm on GC-rich DNA has already been observed at concentrations M/P = 0.5). In the case of the amino acids ,3-alanine and y-ABA, the behavior of AT is fundamentally different. The experimental data, presented in Figure 3 and Table 2, indicate a monotonic decrease in the width of helix-coil transition on increasing the alanine concentration. Moreover, AT is changed more strongly than the content (GC
Figure 1 shows the DMC of calf thymus DNA in the presence of Mn2+, Ni2+ and Co2+ ions. In the presence of these ions a significant change in the shape of the melting curve takes place. In particular, at concentrations of Mn2+ (M/P = 2) a sharp narrowing of helix-coil transition width is seen. A similar effect is observed in the presence of Ni2+ and Co2+ ions. The well resolved peaks in the high temperature regions of DMC that characterize the melting of GC-rich regions of DNA merge and transform into one narrow peak. This observation suggests that at stoichiometric concentration of transition metal ions, a shift in the relative stability of AT- and GC-pairs takes place: ATpairs are stabilized, while GC-pairs are destabilized. This suggests in turn that the influence of metal ions on DNA of different GCcontent should be different. Table 1 and Figure 2 summarize the dependence of the melting temperature of DNAs of different GCcontent at different on the concentration of Mn2+, Ni2+ and Co2+ ions. The differences of TGc-TAT were calculated by means of these dependences. As noted earlier, the DNA of higher GC-content the value of AT is in direct proportion with TGCTAT difference (12). Thus, at a given concentration of metal ions, AT and parameter TGC-TAT should change in the same way, if these changes are conditioned only by influences of ions on GC-pairs; although the results in Table 1 indicate that the AT of DNA in presence of given metal ions changes more weakly than the parameter TGC-TAT. This effect is especially strong for Ni2+ ions, where the parameter TGc-TAT decreases down to zero (0) at concentrations such that Ni2+ M/P = 5; AT of calf thymus DNA decreases to 2°C. Note that for DNA with high GC-content (GC = 72 %), the minimal value of AT
Tm(0C)
parameter
I.:,,
80 0
72
64
0
40
80
TGC-TAT.
These results suggest the following conclusions: (1) the narrowing of AT is mainly due to factors other than sequence specificity in the interaction of amino acids, the action of which causes a change in the relative stability of DNA regions melting at different denaturation degrees; (2) both the specific and non-specific effects of these amino acids increase with their concentration; (3) ,B-alanine has different influences on stability of AT- and GC-pairs only at concentrations above 1.5 M, that is, at concentrations of amino acid where its influence on the DNA is probably due to a change of structure of the solvent (13, 14). Analogous results are obtained for 'y-ABA, with the difference that the specificity is more pronounced in this case. It gives the main contribution in the change of AT (80-90%) and begins to manifest itself at lower concentrations ( IM). In fact, the complete quantitative coincidence of the microcalorimetric curves ([DNA] = 1 mM, Fig. 4a) and DMC ([DNA] = 0.01 mM, Fig. 4b) indicates that neither specific nor nonspecific influences of these amino acids are apparently connected with intermolecular aggregations. Possibly, the additional decrease of AT is connected
XGC(%)
Fig. 3. Dependences of Tm on GC-content of DNAs at different ,B-alanine concentrations: (0) pure DNAs in 0.1 SSC, (A) 1.5 M and (0) 5 M ,B-alanine.
Table 2. Tm and AT of DNAs at different ,B-alanine concentrations in 0.1 SSC Source of DNA Concentration of Alanine (M)
Modoilus sp. (XGC = 32%) AT(0C) Tm(OC)
0
63.5 60.4 60.2 64.4
1.0 1.5 5.0
4.6 4.0 3.6
1.5
Calf thymus
E. coli
M.luteus
(XGc = 42%) AT(0C) Tm(OC)
(XGc = 52%) AT(0C) Tm(OC)
Tm(OC)
AT(0C)
72.9 68.9 68.6 72.3
86.3 83.2 81.0 80.7
4.3 4.0 3.9 2.8
10.7 7.6 6.7 4.0
69.9 66.1 65.7 69.2
AT is defined by tangent in transition point of melting curve
(AT
=
1/(av/8t)v
=
0.5)
4.8 3.8 3.7 2.6
(X GC = 72%)
6416 Nucleic Acids Research, Vol. 18, No. 21 TT
(b)
n
0
i.,
0
t2
0.20
Tm(°C)
~~~I
-7-
90
i; 4.0
F
1.5
0.12
70 /I~~~~~1
10
I-
/
co
0)
0.04
50 .
,
0
59
c
65
77
71
i071 .
T
("C)
30
Fig 4. The microcalorimetric (a) and UV-spectrophotometric differential melting curves (b) of calf thymus DNA at different f-alanine concentrations: (--) pure DNA in 0.1 SSC, (---) 1.5 and (----) 4 M f-alanine.
I/
/
10
q&
UAGI7
4W
r r k7') Uv 'LGC
f
Table 3. The values and changes of parameter TGC-TAT and AT of calf thymus DNA at different pH (at 25°C)
pH 7.3 9.0 9.4 9.8 10.2 10.6 11.0 11.3 11.5
(TGC-TAT)O
ATO
TGC-TAT
AT
(OC)
(OC)
TGC-TAT
40.9 40.9 38.6 34.1 30.8 30.8 38.6 52.4 68.1
8.5 8.5 7.4 6.5 6.4 6.6 8.7 11.4 13.2
-
-
1.00 1.06 1.20 1.33 1.33 1.06 0.78 0.60
1.00 1.13 1.29 1.33 1.30 1.01 0.74 0.64
AT
where (TGC-TAT)O and ATo were taken for pH 7.3.
not only with the specific influence in the region of high concentrations (4 M and higher), but also by the fact that these amino acids form some ordered structure around double-helical DNA (complexes of DNA with zwitterions (11, 14)), which is destroyed cooperatively, when the first melted regions appear. Since such a destruction is energetically unfavorable, the melting begins at higher temperatures. At moderate to high degrees of denaturation, this structure is already destroyed and therefore, for the final part of the transitions, no temperature shift is observed. As a result, a decrease of AT takes place. Next, we consider the deprotonation of DNA bases in alkaline solutions. We have investigated the helix-coil transition under these conditions both theoretically and experimentally (15), and find that the bases are deprotonated in both native, as well as denatured DNA. That is, GC-pairs in the helical form are capable of ionizing at pH 10.6 as well. The data in Table 3 indicates that changes of AT of calf thymus DNA and the TGC-TAT at each value of pH are identical. Consequently, the changes in AT of DNA are conditioned by the specific deprotonation of nitrogen bases. 0% of the From extrapolations to XGC 100% and XGC Tm (XGC) profile for each value of the pH, we can extract the pH dependence of TAT and TGC. The results obtained are given in Fig. 6. With increase of pH, TAT(pH) is continuously decreasing at pH 9, while TGC (pH) reaches saturation around pH 11. Our explanation of this behavior is the following. When the relative content of binding sites in the helix is lower than in the coil, the melting temperature will decrease with increasing pH. Once the quantities of binding sites in the helical and melting =
Fig. 5. Dependences of Tm on GC-content of DNAs at different pH. (The numbers on the lines showed the values of Ph).
=
Table 4. Calculation values TGC-TAT and AT of calf thymus DNA at different relative ligand concentrations (C)? (ATo = 10°C, (TGC-TAT)O = 40°C) C 0 10 25 100 1000
TGC-TAT
AT
(OC)
0C)
(TGC-TAT)O TGC-TAT
ATO
40.0 21.6 5.8 -19.6 -38.8
10.0 5.4 1.5 4.9 9.7
1.00 1.85 6.70 2.00 1.03
1.00 1.85 6.85 2.10 1.03
AT
parts become equal, the melting temperature of polynucleotide will not change above a certain pH. According to the mechanism
have suggested (15), AT-pairs ionize only in the denatured TAT(pH) strongly decreases with the increase of pH. But GC-pairs can ionize in both the native and denatured states of DNA. Hence TGC (pH) reaches saturation. On the basis of these investigations, we can conclude that AT is affected by changes of relative stability of AT- and GC-pairs, as well as by other factors. In the case of metal ions, these factors lead to the increase of AT, while f-alanine and -y-ABA decrease it. During deprotonation of DNA, the change of AT is defined exclusively by changes in relative stability of AT- and GC-pairs of DNA, and AT (YGC-TAT); in this case.
we
state, so that
-
THEORETICAL PART The melting process of a DNA molecule with block distribution of base-pairs in the presence of ligands with specific binding character have been modelled by formulae obtained in references 16 and 17. The influence of ligand binding (16, 17) which are characterized by different constants for adenine, thymine, guanine, cytosine on DNA helix-coil transition is considered. The possibility of interaction between ligands bound on DNA is not taken into account. We can introduce a correction for this effect following reference 17. Assuming that the free energy of ligand binding is i-type region with j-type base is a linear function to the degree of filling up of DNA by ligands, the formulae of (17) have the form:
Nucleic Acids Research, Vol. 18, No. 21 6417 4
Oi,i=oo6j j-W E C i,
(1)
ajl "" Kj, exp[2W,( Li= Cl ,i)/RT,] = I1 Cj,1aij,1 C
4
(2)
-
4
4
1
Tov
_ Tv
R AH
WI(
Cl,i)1-W2(
,
j
C2,1)2 +
+
RTV
l (1 Ci,ii,i) -Rl,i I
(3)
(I -C2 iP2 )R2zi
4
4
L Cl,i+v L C2,i C=CO/M+(l-v) i=l =l '
(4)
where j = 1 corresponds to the melted regions on j = 2 to helical regions; i- corresponds to the four bases, A, T, G or C, respectively, Kji = exp(-kO,j/RT) the binding constant with jbasis at the degree of filling type regions according to the i the up of DNA by ligands which strive to zero; Wj proportionality coefficient; Cj,j the degree of filling up of DNA by ligands bound with nitrogen bases i in j-type regions; CO- the molar concentration of free ligands; R - the gas constant; Tv the absolute temperature at which the degree of helicity of DNA complex with ligands equals v; T., - the same the helix-coil parameter for DNA without ligands; AH transition enthalpy; ajj R'jj - the functions from v, which are calculated by the method suggested in (17); C = D/M; D the molar concentration of ligands; M the molar concentration of DNA base pairs. Solving the system of equations (2) together with equation of material balance (4) we obtain 8 values of Cj,i. Inserting them into the equation (3) Tv is determined. In these calculations it is supposed that the ligands interact with one base pair (there is no interaction with AT-pairs), i.e., Ki 1 = K12 = K21 = K22 = 0. For GC-pairs K13 = K14 = 5 x107, and K24 = K23 = 107; AH = 9 kcal/mol base pairs, the melting temperature and width of the transition of calf thymus DNA without ligands is assumed to be 70°C and 10°C, correspondingly (in 0.1 SSC). From Table 4, it follows that in the presence of specific ligands, the ratio of AT -(TGC-TAT) is carried out strictly at any concentrations of ligands. Hence, the decrease or increase of more than the change of (TGC-TAT) is conditioned by the other nonspecific effects mentioned above. -
-
T
(00)
100
-
-
60
-
-
-
20
7
8
9
10
11
pH
-
ACKNOWLEDGEMENT The authors are grateful to Professor M.D.Frank-Kamenetskii for helpful discussion of the results.
REFERENCES 1. Melchior,W.B. and Von Hippel,P.N. (1973) Proc. Natl. Acad. Sci. USA 70, 293-302. 2. Akhrem,A.A. and Lando,D.Y. (1981) Molek. Biol. 15, 1083-1091 (in
Russian). 3. Voskoboinic,A.D. et al. (1985) Molek. Biol. 9, 783-789 (in Russian).
Fig 6. The pH dependences of TAT and TGC (see text). 4. Geiduschek,E.P. and Herskovits,T.T. (1961) Arch. Biochem. Biophys. 95, 114- 129. 5. Herskovits,T.T., Singer,S.J. and Geiduschek,E.P. (1961) Arch. Biochem.
Biophys. 94, 94-114. 6. Forster,W., Bauer,E., Schuts,A. and Berg,H. (1979) Biopolymers 18, 625-661. 7. Shin,J. (1973) Biopolymers 12, 2459-2475. 8. Luck,G. and Zimmer,C. (1972) Eur. J. Biochem. 29, 528-536. 9. Ott,G.S., Bastia,D. and Bauer,W. (1978) Biochem. Biophys. Acta 58, 216-231. 10. Yurgaitis,A.P. and Lazurkin,Y.S. (1981) Biopolymers 20, 967-975. 11. Aslanian,V.M., Haroutiunian,S.G., Lando,D.Y., Dalian,E.B. and Shpakovskii,A.G. (1988) Biofizika 33, 430-436 (in Russian). 12. Berestetskaya,I.V., Frank-Kamenetskii,M.D. and Lazurkin,Y.S. (1974) Biopolymers 13, 193 -205. 13. Aslanian,V.M. and Haroutiunian,S.G. (1984) Biofizika 29, 564-566 (in Russian). 14. Aslanian,V.M. and Haroutiunian,S.G. (1985) Biofizika 30, 741-745 (in Russian). 15. Akhrem,A.A., Haroutiunian,S.G., Aslanian,V.M., Dalian,E.B., Lando,D.Y. and Shpakovskii,A.G. (1989) Molek. Biol. 23, 518-525 (in Russian). 16. Akhrem,A.A. and Lando,D.Y. (1979) Molek. Biol. 13, 1098-1109 (in Russian). 17. Lando,D.Y. et al. (1980) Molek. Biol. 14, 175-181 (in Russian).
