Document not found! Please try again

thymidylates] and their complexes - Europe PMC

0 downloads 0 Views 1MB Size Report
HPLC analysis of oligomer hydrolysates. Degradation assays were ..... Therefore terminal deoxynucleotidyl transferase enzyme- o. 3.280-. '-4. CO. 3.240. 0. T'-4.
Nucleic Acids Research, Vol. 18, No. 8 2133

\. 1990 Oxford University Press

Biochemical properties of oligo[( + )-carbocyclicthymidylates] and their complexes Janos Sagi, Attila Szemz6, Judit Szecsi and LaszIo Otvos Central Research Institute for Chemistry, Hungarian Academy of Sciences, PO Box 17, H-1525 Budapest, Hungary Received November 21, 1989; Revised and Accepted February 7, 1990

ABSTRACT We report here spectroscopic and biochemical data of a novel series of sugar-modified oligodeoxynucleotides, the carbocyclic oligothymidylates, c(dT)3.20. In c(dT)n a methylene group has been substituted for the oxygen atom of the deoxyribose ring of the natural thymidylate unit. c(dT)10-20 form helical structures, in contrast with oligothymidylates or poly(dT), based on absorbance versus temperature melting profiles. Secondary structure of c(dT)n, where n > 10 is assumed to be double helix. In addition to this, c(dT)n forms as a stable duplex with complementary poly(dA) as does parent (dT)n. On the other hand, c(dT)n-containing oligo/poly duplex is nearly inactive either as a template or as a primer in various DNA polymerase systems, and c(dT)n inhibits DNA replication as well. c(dT)n can efficiently be extended by terminal transferase and shows an increased nuclease stability compared to (dT)n. Base-pairing ability and nuclease stability of c(dT)n suggest that (+ )-carbocyclic nucleoside-containing oligomers could be new potential antisense oligodeoxynucleotides.

INTRODUCTION Modified oligodeoxynucleotides synthesized to serve as antisense oligomers, a new class of potential chemotherapeutic agents (1-5), should posses the following properties: good solubility in water, chemical stability, resistance against nucleases, taken up well by cells and stable complex-forming ability with complementary natural target sequence (mRNA or viral RNA). Such oligodeoxynucleotides are those which have (i) modified phosphodiester linkage, like phosphorothioate (6-12), and the 'nonionic' analogues like methylphosphonate (8, 9, 12-15), phosphoramidate (12, 16) and phosphotriester (8, 9. 17) or (ii) modified sugar residues, like ca-oligodeoxynucleotides (18-21) and (iii) the derivatized oligomers (modified as above or not) that are covalently linked to different compounds through terminal 3' or 5'-OH groups, like intercalators to promote hybridization (22-25), reactive groups to cause selective damage to nucleic acid (26-29) or poly(L-lysine) to stimulate delivery (30, 31). One of the main points in the effectiveness is the stability against nucleases. This is well accomplished by the backbonemodified oligomers. Except for a-oligomers, these modifications lead to mixtures of stereoisomers (7, 9, 18). Stereochemically

pure oligomers can be synthesized, if one starts with optically pure modified synthetic intermediates. This is the case with the carbocyclic nucleotides whose stereospecific synthesis has recently been achieved (32). Carbocyclic thymidine has the same anomeric configuration as thymidine. The only structural difference is that a methylene group has been substituted for the oxygen atom of the furanose ring in the deoxyribose moiety. There are however conformational alterations as determined from crystal structure (33): the usual C3'-exo/C2'-endo pucker of the five-membered ring of thymidine is shifted to the C1 '-exo form giving rise to changes in orientation of 3' and 5' hydroxy groups. In addition, carbocyclic nucleosides possess a more lipophilic character as a consequence of methylene-oxygen exchange, which is an important factor in cell membrane penetration. Determination of hybridization and stability properties of carbocyclic oligothymidylates is the main subject of the present study.

EXPERIMENTAL Chemicals Synthesis of oligothymidylates (dT)3, (dT)I0, (dT)12 and (dT)20, oligo[(+)-carbocyclic-thymidylates] c(dT)3, c(dT)IO, c(dT)12 and c(dT)20, and of the mixed oligomers 5'-(dT)9[c(dT)]-3', and (dT)5[c(dT)]2(dT)5 has recently been reported (34). Carbocyclic dTTP (+)-C-[5-methyl-3H]dTTP (26 mCi/mmol) was prepared from (+)-carbocyclic thymidine according to refs. 32, 36. [3H]dATP (17 Ci/mmol), [3H]dTTP (46 Ci/mmol) and [3H]dGTP (28 Ci/mmol) were purchased from Amersham Int. plc, [14C]dATP (0.39 Ci/mmol) was from UVVR (Prague, Czechoslovakia). [14C]dTTP (4 mCi/mmol) was prepared according to refs. 35, 36. Poly(dA-dT), poly(dA).poly(dT), poly(A).(dT)15 and p(dT)6 were from Boehringer (Mannheim, FRG), poly(dA) and poly(A) from Pharmacia (Sweden). Enzymes E. coli DNA polymerase I Klenow fragment enzyme (7000 units/mg), AMV reverse transcriptase (117000 units/mg), snake venom phosphodiesterase (1 mg/ml) were bought from Boehringer (Mannheim), FPLC-pure terminal deoxynucleotidyl transferase (TdT; 47000 units/mg) was from Pharmacia and calf thumus DNA polymerase ca (1000 units/mg) from P-L Biochemicals Inc.

2134 Nucleic Acids Research, Vol. 18, No. 8 Thermal transition measurements Ultraviolet spectra and thermal melting experiments were peformed on a Specord UV-VIS recording spectrophotometer (GDR) interfaced to an IBM AT compatible PC. Regulation of heating, acquisition of absorption and temperature data and calculation of values characterizing thermal transition curves were peformed by the PC with the help of the program 'ABSORG', Version 1.0 (Chemicro Ltd, Salamon u. 13A, H-1 105 Budapest, Hungary). Heating rate was 0.5°C/min. In a temperature range of 80°C, 500 absorption data were collected, smoothed and normalized to the mean of the first 10 values, as 1.0, at the end of the measurement. Tm was calculated from the 1st derivative of the smoothed absorption versus temperature profile. Concentration of (dT)n or c(dT)0 was 80 ,uM(P), the latter based on E(P) values determined (see Results). Data given in the table are mean values of two to four determinations, melting profiles shown in the figure are selected ones. Concentration of c(dT)20 was varied between 13 and 1500 AiM(P) for the determination of thermodynamic parameters, and buffered sodium ion concentration was between 0.3 and 0.8 M for measuring dTm/d(log[Na+]) for c(dT)20.

Wavelength (nm) 250

c 0

4: L-

.0

Circular dichroism Spectra were recorded in a Jobin-Yvon Dichrographe III spectrometer interfaced to a PC. Enzyme reactions Reaction mixtures for the determination of rate and extent of oligomer-primer elongation of poly(dA) and poly(A) duplexes contained 60 mM potassium phosphate buffer (pH 7.4), 6 mM MgCl2, 1 mM 2-mercaptoethanol (ME), 0.16 mM [14C]dTTP (8.9 dpm/pmol) and 0.65 OD260 units/ml of the 1:1 poly(dA) duplexes and 1 OD260 units/ml of the poly(A) duplexes which were prepared by standard annealing procedure in 0.1 M Kphosphate buffer (pH 7.4). Final volume was 0.05 ml. Reactions were started by addition of 0.1 units of Klenow enzyme or 0.25 units of DNA polymerase at (Table 2), and with poly(A) complexes 1 unit of AMV reverse transcriptase or 0.12 units Klenow enzyme (Table 3). Incubations were carried out at 37 and 25°C, respectively. Samples of 0.01 ml were withdrawn, spotted on GF/C filters, washed, dried and counted for acidinsoluble radioactive material. Terminal transferase enzyme was assayed in volumes of 0.05 ml which contained 0.2 M sodium cacodylate buffer (pH 7.2), 2 mM CoCl2 for pyrimidine substrates or 8 mM MgCl2 for purine triphosphates, 0.2 OD, units of oligomers, 0.6 mM of [3H]dTTP (47 dpm/pmol) or (+)-C-[3H]dTTP (104 dpm/pmol) or [14C]dATP (16.6 dpm/pmol). Ractions were started by addition of 28 units of the enzyme. Incubations were at 37°C. Samples of 0.01 ml were taken at times indicated in Table 6, and analysed for acid-insoluble radioactivity. Inhibition experiments were carried out in buffer system as described above for DNA polymerase reactions except that final volumes were 0.035 ml and double labelling was used: [3H]dTTP (47 dpm/pmol) and [14C]dATP (16.6 dpm/pmol) (not shown), furthermore 0.01 ml samples were taken for analysis at 30 and 60 minutes. Concentration of DNAs and inhibitory oligonucleotides are given in Table 5. Template activity of the 20-mers, in complexes with (dA)IO primer was measured in mixtures containing the buffer described above for Klenow DNA polymerase enzyme and [3H]dATP (44.3 dpm/pmol) as the only triphosphate. Template and primer

Wavenumber x 10-3 (cnF1) Figure 1. Ultraviolet spectra of carbocyclic oligothymidylate c(dT)20 in 0.5 M NaCl, 20 mM Tris.HCI (pH 7.4), 10 mM MgCl2 at 200C (Xmax=270 nm) and 60°C (Xmax=274 nm).

(1: 1 for mol(P)] were preincubated together for 3 min at 900C. Solution was then cooled down to room temperature during 5 hours. Annealed (dT)20.(dA)Io or c(dT)20-(dA)Io [5-5, 5-5 nmol(P)s] [more precisely 0.048 OD. (dA)IO 14.85 nmol/P/J plus 0.043 OD,x (dT)20 14.9 nmol/P/J or 0.38 ODnax c(dT)20 15.07 nmol/P/)] were added to 0.05 ml of mix. Reaction was started by addition of 0.17 units of Klenow DNA polymerase and incubation was at 37°C. Samples of 0.01 ml were withdrawn and analysed for acid-insoluble product as above. HPLC analysis of oligomer hydrolysates Degradation assays were carried out in 0.1 ml mix containing 0.1 M Tris.HCl (pH 8.05), 6 mM MgCl2 and 0.8 units of the oligomer. Hydrolysis was started by addition of 0.17 Ag of snake venom phosphodiesterase. Samples of 10 yl were taken at 0, 5, 10, 20, 30, 60, 120, 180 and 240 minutes of incubation at 37°C, mixed with 20 41 cold EtOH. After keeping samples in freezer for 1 hour, they were centrifuged and analysed by HPLC (Isco Inc., Lincoln, Nebraska). The analyses were carried out by linear gradient elution. Eluent A contained 2.5 % triethylammonium acetate, eluent B 2.5 % triethylammonium acetate and 20% acetonitrile. Rate of enzymatic hydrolysis of the synthetic oligomers was based on measuring amount of monomeric unit as a function of incubation time. Amount was calculated from peak area. Chromatographic peaks were identified with standards.

RESULTS Absorption versus temperature melting profiles Secondary structure of carbocyclic oligothymidylates. Ultraviolet spectra of the 20-mer c(dT)20 at two temperatures are

Nucleic Acids Research, Vol. 18, No. 8 2135 shown in Figure 1. Absorbance versus temperature melting profiles of c(dT)1O020 in 0.5 M NaCl, 20 mM Tris.HCl (pH 7.4), 10 mM MgCl2 solution are shown in Figure 2. c(dT)IO displayed a melting profile of low cooperativity

(AT=13.50C, Table 1), and Tm had to be estimated. On the other hand, melting of c(dT)12 proved to be cooperative with a well defined Tm-point. With c(dT)20 thermal stability was further increased. Thermally induced hyperchromicity also increased, whereas width of transition decreased. These reflect formation of more complete and homogeneous helical structures. The transitions were all completely reversible. On the other hand, no such thermal transition curves were obtained with any of the oligothymidylates (dT)n examined, in accordance with the literature (37). Furthermore, carbocyclic thymidine-containing mixed oligothymidylates possessed no helical structure (Table 1). To decide whether the above helices are single or double-

stranded, Tm was determined as function of concentration of c(dT)20, covering a 100-fold range. Figure 3 shows linear least squares fits of I/Tm versus ln CT values, were CT is total strand concentration calculated by using e(P)=7490 M-1cm-1 for c(dT)20 as determined below. There is a dependence of Tm on oligomer concentration, giving a transition enthalpy AH' = -152.8 kcal/mol(P) and entropy ASO = -0.47 kcallmol(P).deg. Insert shows linear least squares fits of Tm of c(dT)20 versus -log of sodium ion concentration in the range of 0.1 to 0.8 M sodium ions, providing a slope of 14.11. Thermally induced hyperchromicity at 270 nm depended strongly on salt concentration and increased from 7.6% in 0.1 M to 17.2% in 0.8 M [Na+], while AT did not show similar dependence (5.1 + 0.40C). At less than 0.1 M [Na+] determination of Tm was uncertain, mainly because of the small hyperchromic values (less than 5%). Formation of double helices with poly(dA). Poly(dA) has a weak single stranded helical structure in neutral salt solutions (37), with a very large transition width (Table 1). Composing 1:1 mixtures with complementary oligothymidylates results in the formation of double helices with thermal stabilities dependent on chain length of the oligomer part (38). However, not only the oligothymidylates but unnatural oligo-carbocyclic thymidylates formed stable 1:1 duplexes with poly(dA) (Table 1). Molar extinction coefficients for c(dT)n- s were calculated from ultraviolet spectra taken before (at 100C) and after hydrolysis to nucleotides at 37°C by snake venom phosphodiesterase in the above high-salt buffer. Complete decomposition to nucleotides was checked by HPLC. Using E273= 10400 (pH 7) for (+)-carbocyclic thymidine (39), e(P)274 of 7900, 7690 and 7490 M-'cm-1 for c(dT)IO, c(dT)12 and c(dT)20, respectively, were obtained. These values were then used for composing 1:1 mixtures with poly(dA). Thermal stability of poly(dA).c(dT)0 proved to be even higher than that of the poly(dA).(dT)n duplexes (Table 1). Difference in stability increased with the higher chain-length analogue complexes. Insertion of one or two carbocyclic thymidines into an oligothymidylate had a small effect on the stability of the complex. Circular dichroism spectra of poly(dA) (dT)20 and poly(dA).c(dT)20 (Figure 4) are basically not different from each other. Differences in AE may originate from the effect of substitution. Primer activity Poly(dA).(dT)n is known to be a very active primer/template with various DNA-polymerases (40). Oligomer-primer chain .

1.) c

Il

2

0 0.

1.:

.0

/0

.1

L-

0~~~~~~~0

o W E

1.1

I

0

:z

0

20

40

80

60

Temperature (IC) Figure 2. Absorption versus temperature melting profiles of c(dT)n s and oligothymidylates in 0.5 M NaCl, 20 mM Tris.HCl (pH 7.4), 10 mM MgCl2 and c(dT)20, . and (dT)n, solution: c(dT)Io c(dT)12, -

-

-0

0-.

elongation rate (not shown) and extent of poly(dA).(dT)n and poly(dA).c(dT)n were compared with E. coli Klenow DNA polymerase and calf thymus DNA polymerase et enzymes (Table 2). Initial rates of syntheses were different and this difference remained the same at the saturation part of replication. Poly(dA).(dT)10 (1:1) was very active in thymidine incorporation. Poly(dA). [(dT)gc(dT)] behaved like poly(dA).p(dT)6, i.e. enzyme probably cuts out carbocyclic nucleotide first, together with a few thymidine nucleotides, and this primer is then elongated. Poly(dA).c(dT)12 duplex was found to be nearly inactive for elongation of c(dT)12 primer by dTMP units. Poly(dA).(dT)12 and poly(dA).c(dT)12 were equally stable duplexes in solutions of high ionic strength (Table 1). Both duplexes have about the same thermal stability also in 60 mM potassium phosphate (pH 7.4), 6 mM MgCl2, the solution used for DNA polymerase assay: poly(dA).(dT)12, Tm=34.0C, (H=38.8% and AT=4.8°C) and poly(dA).c(dT)12, Tm=33.80C, (H=34.1% and AT=5.70C). At 37°C however, where DNA polymerase assays are performed both seem to be 80-90% denatured. In order to test whether this denatured state is responsible for the difference, replication experiments were repeated at a lower temperature, 25°C (Table 2) where both complexes are duplexes. At 25°C not only rates (not shown) but also saturation were lower than at 37°C, except for poly(dA).p(dT)6. No incorporation of [14C]dTMP was observed into poly(dA).c(dT)12. With calf thymus DNA polymerase ax chain extension of c(dT)12 was observed, although to a limited extent (Table 2). It was about 10% of that of poly(dA).(dT)10 which showed again an optimum. Examination of primer chain extension was also carried out with an error-prone polymerase enzyme, the AMV reverse transcriptase (40). For this ppurpose duplexes with poly(A) template were used (Table 3). Since Klenow DNA polymerase enzyme is also able to use ribo-templates (40), poly(A) complexes were also assayed with this enzyme, in comparison.

2136 Nucleic Acids Research, Vol. 18, No. 8 Table 1. Data of thermal transitions determiined in 0.5 M NaCI, 20 mM Tris.HCI (pH7.4), 10 mM MgC12 DNA

Tm(OC)

H(nm)' (%)

AT2 (°C)

(dT)I0

-

c(dT)IO

19 (calc)

(dT)12

-

c(dT)12

21.6

(dT)20 c(dT)20

-

40.8

(dT)9c(dT)

-

0 (269) 22.5 (274) 1 (267) 26.0 (274) 1 (270) 29.7 (274) 1 (268)

-

2 (269)

-

30

(dT)5c(dT)2 (dT)5 poly(dA)

50 (calc)

30.5 (259)

poly(dA).(dT)6

15.7 39.9 41.5 46.2 49.4 60.7 68.6 39.0 43.5

34.3 55.2 56.2 45.8 63.6 56.9 58.6 45.1 51.1

poly(dA).(dT)10 poly(dA).c(dT)Io poly(dA).(dT)12 poly(dA).c(dT)12 poly(dA).(dT)20 poly(dA).c(dT)20

poly(dA).[(dT)gc(dT)] poly(dA).[(dT)5c(dT)2(dT)5]

(262) (262) (262) (262) (262) (262)

(262) (262) (262)

-

13.5 -

7.5 -

5.3 -

5.6 7.2 11.9 7.6 12.6 7.5 15.1 6.9 7.5

1H(nm) stands for thermal hyperchromicity percentage measured at wavelength in brackets. 2AT is width of the transition between 25 and 75% of the thermal hyperchromicity. Klenow DNA polymerase proved to be as active with the ribo complex poly(A).(dT)1O as with the deoxy poly(dA).(dT)jO. The 9-1 oligomer could hardly be extended, contrary to the poly(dA) system, and c(dT)12 was inactive as a primer similarly as with AMV reverse transcriptase.

Template activity According to results summarized in Table 4 carbocyclic oligothymidylate did not function as a template as well, if complexed with deoxyoligoadenylate as a primer. This was observed with c(dT)20.(dA)IO template-primer duplex (2:1 for nucleotide ratio), in comparison with (dT)20.(dA)IO (2:1).

o

3.320 T

0.

f

40.

3.280-

'-4

3.240 -I-

T'-4

ml.

AM(P).

Terminal transferase reactions Although a limited elongation of the carbocyclic thymidinecontaining oligomer primer was observed with the calf thymus a-polymerase, in all other systems it proved to be inactive. Therefore terminal deoxynucleotidyl transferase enzyme-

.4

-l.1W -lAW -&M1 -4"3

E-

s

V--

oo

"

o

-4L4W -0.M

GMK

lWE D(+]

3.200

3.160q 1sn-

-12

Inhibition of DNA replication Replication of the homopolymeric duplex poly(dA).poly(dT) by Klenow DNA polymerase was stimulated by oligo(dT)6_12 and slightly decreased by (dT)3 [(42), Table 5]. Carbocyclic analogues, even the 3'-end substituted (dT)1o inhibited replication. Strongest inhibition was caused by c(dT)3. Similarly, c(dT)3 was observed to be the most effective inhibitor of the replication of other template-primers, and also of reverse transcription of poly(A).(dT)15 catalyzed by AMV reverse transcriptase enzyme. Lineweaver-Burk plots for the inibition by (dT)3 and c(dT)3 of poly(dA-dT) replication are shown in Figure 5. KM for poly(dA-dT) of the Klenow enzyme was found to be 178.6 AtM(P). Inhibition by both oligos was found to be competititve. K1 for c(dT)3 was 38.2 ftM(P) whereas for (dT)3 it was 202.8

0/

U

J. *.

CO 0

-11

-10

-9

-8

-7

-8

ln CT Figure 3. Linear least squares fits for the van't Hoff plot of data of Tm and concentration of c(dT)20. Slope is -0.013 x 0-3 and intercept is 3.064x l03. Insert shows the linear least squares fits of Tm of c(dT)20 - (-log [Na+]). Slope is 14.11, intercept at -log[Na+]=0 is 42.8°C.

catalyzed reaction, which requires only a primer and no template (41) was studied. Results are summarized in Table 6. Results with dTTP show that sugar-modified oligonucleotides do function as primers. This means that 3'-OH end is available for elongation by this enzyme. Surprisingly, c(dT)12 seemed to have the highest priming activity for the incorporation of dTMP. With dATP as a substrate, again each oligomer tested functioned as a primer although to very different extent. The 5'-phosphorylated oligomer was far the best primer in accordance with the literature (41). Carbocyclic oligothymidylates also functioned. For the incorporation of the carbocyclic-dTTP, [(+)C-[3H]dTTP} none of the natural oligomers functioned as a primer but c(dT)IO and c(dT)12 showed a small activity. Inhibition of incorporation probably comes from the substrate side and not from the primer.

Nucleic Acids Research, Vol. 18, No. 8 2137 Stability against nuclease Snake venom phosphodiesterase enzyme was used earlier for the analysis of composition of these sugar-modifed oligonucleotides (34). Here we describe the comparison of cleavage rates of (dT)IO and c(dT)IO in the phosphodiesterase enzyme-catalyzed hydrolysis reactions. Progress of degradation was followed by HPLC analysis of the products as described in Experimental section. Time-course of hydrolysis of both oligomers is shown in Figure 5. The modified oligothymidylate c(dT)IO proved to be ten-times as stable as (dT)Io against phosphodiesterase, based

al

wavelength (nm) and Figure 4. Circular dichroism spectra of poly(dA).(dT)20, poly(dA).c(dT)20, -- -, in 60 mM potassium phosphate buffer (pH 7.4), 6mM MgCl2 solution at 15°C (straight lines) and 60°C (broken lines). ,

of incubation times required for the appearance of 50% of the mononucleotide products. This was 6 min with (dT)IO and 60 min in the case of c(dT)I0.

on the comparison

DISCUSSION

Many publications in the literature deal with synthesis and testing of carbocyclic nucleoside derivatives (for reviews see 43, 44). However, there are only two publications reporting about oligoor polynucleotides containing carbocyclic nucleotide analogues. Ikehara and Fukui reported on copolymerization of 5'-diphosphate of Aristeromycin t(-)-carbocyclic-adenosinej with ADP by E. coli polynucleotide phosphorylase enzyme. The analogue alone was no substrate of the enzyme (45). The other reference is about a carbocyclic 2'-5'-oligo-adenosine analogue (46). No detailed information about structure, stability or duplex forming ability is available. The small number of papers published is understandable in the light of the fact that chemical synthesis of the oligomers is not without difficulties (34), and that the 5'-triphosphates of carbocyclic nucleosides are very poor substrates of polymerizing enzymes, like E. coli DNA polymerase I (47, 48). Only terminal transferase enzyme could use (+)carbocyclic [3H]dTTP as a substrate but only to a limited extent with oligo(dC) or poly(dC) primers, or in copolymerization with [14C]dATP using p(dT)6 as a primer (our unpublished results). Recent synthesis of oligo[(+)-carbocyclic thymidylates] (34) now makes possible the evaluation of physical and biochemical properties of carbocyclic oligonucleotides. The very first examination of properties of c(dT)n provided unexpected results: c(dT)n, unlike (dT)n, possesses a helical structure in solution of appropriate ionic strength (Figure 2, Table 1). A single-stranded helical structure can be postulated for c(dT)IO: no definite inflection point could be observed and shape of the profile refers to a low cooperativity melting. This is however far not as low as in the case of poly(dA) which is known to form a single-stranded helix in neutral salt solutions (37). Melting profiles of c(dT)12 and c(dT)20 show cooperative melting. Both hyperchromic and AT values of c(dT)20 approach those of the poly(dA).p(dT)6 1:1 duplex, and Tm of both c(dT)ns is even higher than that of poly(dA).p(dT)6 (Table 1). Based on characteristics of the melting profiles of c(dT)12 and especially of c(dT)20, formation of double helices c(dT)12.c(dT)12 and c(dT)20.c(dT)20, respectively in 0.5 M NaCl, 20 mM Tris.HCl (pH 7.4), 10 mM MgCl2 can be

Table 2. Extent of elongation of the oligomer-primer of poly(dA) duplexes in DNA polymerase systems

Incorporation of [14C]dTMP

Template-primer (1:1)

(0.65 units/ml)

poly(dA).poly(dT) poly(dA).(dT)10 poly(dA).[(dT)5c(dT)2(dT)5] poly(dA)-(dT)12

poly(dA).[(dT)gc(dT)] poly(dA).p(dT)6 poly(dA).c(dT)12 poly(dA).(dT)3 poly(dA).c(dT)3 poly(dA)

370C pmol 372 338 254 236 134 124 5 0 0 0

Klenow 250C pmol (120 min)

% 110 100 75 70 40 37

1.5 0 0 0

a-polymerase 370C %

pmol -

-

45.1

100

-

100 -

-

-

-

132 58 129 0

44 98 0

% (30 min)

-

-

10.4 4.7 4.0 4.4

23 10 9 10

-

-

-

-

-

-

-

-

-

-

-

-

2138 Nucleic Acids Research, Vol. 18, No. 8 Table 3. Extent of primer-chain elongation of poly(A) complexes

Incorporation of [14C]dTMP/120 minutes AMV reverse transc. % pmol

Template-primer (1:1) (1 unit/ml)

Klenow DNA pol. % pmol

poly(A).(dT),o

732 318 154 15 4.5 1.0 0

poly(A).[(dT)5c(dT)2(dT)5] poly(A).(dT)12 poly(A).[(dT)9c(dT)] poly(A).c(dT)12 poly(A).p(dT)6 poly(A)

100 43 21 2 0.6 0.1 0

482 446 351 89 1.8

100 93 73 19 0.4

-

-

0

0

Table 4. Templating activity of oligothymidylates

pmol

%

Incorporation of [3H]dAMP/0.01 ml 30 min 1 hour pmol pmol % %

250 0.4

100 0.2

336 4.3

15 min

Template-primer

(dT)20.(dA),O

c(dT)20.(dA)IO

assumed. This assumption was checked by measuring dependence of Tm of c(dT)20 on oligomer and salt concentrations. The former is a direct method to determine whether oligomer exists as a single or double-stranded structure at low temperature (49, 50). There was an increase in Tm on increasing concentration of c(dT)20, and parameters derived from van't Hoff plot of 1/Tm vs ln CT (Figure 3) yielded AH' = -152.8 kcal/mol(P) and ASO = -0.47 kcal/mol(P).deg for the duplex formation of c(dT)20. Slope of Tm vs -log [Na+], i.e. dTm/d(log[Na+]) was as high as 14.11, again characteristic of melting of double helices. In solution, formation of helical structures in deoxyoligonucleotides containing only thymine bases is unusual. The ribo series, like polyuridines and poly(ribothymidine), do form helices and these are self-duplexes (37). Oligothymidylates or poly(dT) do not show temperature melting profiles (37, 51), as it is also shown in Figure 2. What happens in an oligothymidylate upon the oxygen atom - methylene group exchange that can cause or allow the formation of helical structure? One important thing is the change of hydration around the sugar part. Since water is considered to be an integral part of nucleic acids structure, and sugar-ring oxygen atom 04' is intermediate in hydration order (anionic phosphate oxygen atoms are the most hydrated and esterified 03' and 05' backbone atoms are the least hydrated) (52), an altered hydration network can be assumed to contribute to the formation of a helical structure. This is however a less hydrated structure than that of the oligothymidylates, whereas ribo-polythymidylate, which also forms self-duplex is a more hydrophilic structure than the polythymidylate. In addition to the self-duplexes, c(dT)n forms 1:1 double helix with complementary poly(dA) (Table 1). Stability of these structures is as high as those formed between (dT)n and poly(dA). Thus, a structural feature, i.e. stable hybridization with complementary nucleic acid strand, important in the antisense or antimessenger effect is fulfilled also by oligo[(+)-carbocyclic thymidylates]. Stability against a nuclease, venom phosphodiesterase has also been tested (Figure 6). c(dT)IO proved to be ten-times as stable as (dT)IO. This difference is not as high as observed with a few other backbone-modified oligodeoxynucleotides (1, 3, 53), nevertheless it is significant. Although the duplex formed between a carbocyclic

100 1.3

352 2.1

100 0.6

5 hours %

pmol

419 0.4

100 0.1

Table 5. Effect of oligothymidylates and carbocyclic analogues on DNA replication

Incorporation of ['4C]dTMP/30 min, 0.01 mi pmol %

DNA polymerase and

Template/primerl Klenow DNA polymerase

poly(dA).poly(dT) (0.3 units/ml) + (dT)12 + c(dT)12 + + + + +

(dT)1o

(dT)gc(dT) p(dT)6 (dT)3 c(dT)3

poly(dA-dT) (0.33 units/mi) + (dT)3 + c(dT)3 act. calf thymus DNA (0.67 units/mi) + (dT)3 + c(dT)3 AMV reverse transcriptase poly(A).(dT)j5 (1:1) (0.67 units/mi) + (dT)3 + c(dT)3

177 165 51 262 58 204 156 9 155 136 23 143 137 46

100 93 29 148 33 115 88 5 100 88 15 100 96 32

161 153 124

100 95 77

'Template-primers and oligomers were applied in the same units/ml concentrations. With poly(dA).poly(dT) 0.3 units/mni is 0.05 mM(P), with poly(dAdT) 0.33 units/ml is 0.05 mM(P), with calf thymus DNA 0.67 units/mi is 0.034 mg/ml and with poly(A).(dT)15 0.67 units/mi is 0.1 mM(P). With (dT)3 and c(dT)3 0.3 units/mi is 0.034 mM(P).

oligothymidylate and the complementary oligo- or polydeoxynucleotide is as stable as those formed with oligothymidylates, the carbocyclic oligomer is inactive both as a template and a primer (Tables 2-4). At 37°C, in the buffer used for DNA polymerase assays both the natural (38, 54) and the carbocyclic duplexes seem to be 80-90% denatured (page 8). Nevertheless, this state is adequate for the thymidinecontaining duplex to function either as a template or as a primer but inadequate for the carbocyclic oligomer-containing duplex. Stabilization of the initiation complex by the polymerase enzyme has been assumed earlier with natural oligo/polymers (54). Probably, c(dT)n binds strongly to the polymerase enzyme, thus initiation of replication fails. The strong binding may explain the

Nucleic Acids Research, Vol. 18, No. 8 2139

(dT) 3

i=c(dT)3

KMappX10-z

X102

1 -xl

0

1 S

6

4

2

x102

Figure 5. Lineweaver-Burk plots of the data of poly(dA-dT) replication by Klenow DNA polymerase enzyme in the presence of (dT)3 (panel left) and c(dT)3 (panel right). Rate (V) is in pmol [ 4C]dAMP inc./30 min, 10 /d; S is concentration of poly(dA-dT) in uM(P); is concentration of the inhibitor in jtM(P). Inserts show plots of KM(app) vs [i]. Vmax of the replication (i=0) was 964.4 82.2 pmol dAMP incorporation/30 min, 10 1l. Composition of the reaction mix was as described in the Experimental section, except 0.145 mM labeled dATP and 0.2 mM dTTP were used, and 0.4 units of Klenow DNA polymerase were added to 0.035 ml of reaction mix. Table 6. Priming activity of oligomers with terminal transferase Time-course of incorporation of nucleotides

(pmol/0.01 ml) (+)-C-[3H]dTTP

Oligomer

[3H]dTTP hours: (dT)3 c(dT)3 p(dT)6

(dT)I0 c(dT)Io

(dT)9c(dT) c(dT)12

[14C]dATP

1

3

24

1

3

24

1

3

24

3.8 3.0 23.7 21.6 22.2 6.8 29.3

8.9 5.6 43.8 41.7 37.4 9.6 62.9

12.6 3.3 49.0 50.9 51.2

0 0 0 0 0 0.1 0

0 0 0 0 0 0.1 0.3

0 0 0 0 0 0.8 1.1

2.1 0.6 106 1.8 2.7 0.9 0

3.6 1.8 1180 7.8 12.3 1.4 3.3

18.2 3.6 3752 129 102 4.1 9.7

14.7 73.7

inhibitory properties of c(dT)n in DNA replication experiments (Table 5). The trimer c(dT)3 was the most active inhibitor, with a Ki of 38.2 AM(P) with poly(dA-dT) as the template/primer substrate [KM= 178.6 ,uM(P)] of the Klenow DNA polymerase (Figure 5). This provides a further hint for biomedical studies of c(dT)n-s in addition to the antisense effect whose determination with various sequences is under way.

ACKNOWLEDGEMENTS Authors are indebted to Prof. M.Kajtar for CD measurements and Dr. I.Fellegvari for HPLC analysis, and acknowledge the excellent assistance of Ms I. Megyeri, E. Szakonyi, I.Kosztolanyi and Mr B.Hegede. This work was supported by grant from Hungarian Research Fund (OTKA), No. 1015.

2140 Nucleic Acids Research, Vol. 18, No. 8

a._9 0

40

80

Time

of

120

incubation

160

200

(minutes)

Figure 6. Comparison of time-course of hydrolysis by venom phosphodiesterase of an oligothymidylate and its

carbocyclic

0-0O;

analogue: (dT)10,

and c(dT)10,

*-.

REFERENCES 1.

Miller,P.S. and Ts'o,P.O.P. (1988) Annu. Rep. Med. Chem. 23, 295-304.

2.

Toulme,J.-J.

3.

Stein,C.A. and Cohen,J.S. (1988) Cancer Res. 48, 2659-2668.

and

Helene,C.

(1988) Gene 72, 51-58.

TIBS

4.

Loose-Mitchell,D.S. (1988)

5.

Cohen,J.S.

(1989)

Expression,

Macmillan, London.

61

9, 45-47.

Oligodeoxynucleotides:

Antisense Inhibitors of Gene

Potter,B.V.L., Romaniuk,P.J. and Eckstein,F. (1983) 1758-

J1.

Biol. Chem. 258,

1760.

7.

Eckstein,F.

8.

Marcus-Sekura,C .J.,

Zon,G.

and

9.

and (1987)NucnuAcids Quinnan,G.V. Res. 15, 5749 -5763C. Nucl. Zol,G. (1987) J. Prot. Chem. 6, 131-145. Stein,C.A., Subasinghe,C., Shinozuka,K. and Cohen,J.S. (1988) Nucl.

Acids

10.

Res.

(1985) Annu. Rev. Biochem. 54, 367-402.

Woerner,A.M.,

Shinozuka,K.,

16, 3209-3221.

Shibahara,S., Mukai,S., Morisawa,H., Nakashima,H., Kobayashi,S. and Yamamoto,N. (1989) Nucl. Acids Res. 17, 239-252. 12. Agrawal,S., Goodchild,J., Civeira,M., and ZamecniF,P.C. Saain,P.S. (1989) 11.

Nucleos.

and Nucleot. 8, 819-823.

Miller,P.S., Agris,C.H., Blake,K.R., Murakami,A., Spitz,S.A., Reddy,M.P. and Ts'o,P.O.P. (1983) Nucl. Acids Res. 11, 6225-6242 14. Agris,C.H., Blake,K.R., Miller,P.S., Reddy,M.P. and Ts'o,P.O.P. (1986) Biochemistry 25, 6268-6275 15. Kean,J.M., Murakami,A., Blake,K.R., Cushman,C.D. and Miller,P.S. 13.

(1988) Biochemistry 27, 91 13-9121.

16.

Froehler,B.,

Ng,P.

and

Matteucci,M. (1988) Nucl.

Acids

Res.

16,

4831 -4839.

17. Koole,L.H., van Soc. 18.

Genderen,M.H.P.

and Buck,H.M. (1987) J. Am. Chem.

109, 3916-3921.

Nuckeos. and Nucleot. 8,

Imbach,J.-L., Rayner,B. and Morvan,F. (1989) 627 -648.

19. Lavignon,M., Bertrand,J.-R., Rayner,B., Imbach,J.-L., Malvy,C. and Paoletti,C. (1989) Biochem. Biophys. Res. Commun. 161, 1184-1190. 20. Pauwels,R., Debyser,Z., Balzarini,J., Baba,M., Desmyter,J., Rayner,B., Morvan,F., Imbach,J.-L. and De Clercq,E. (1989) Nucleos. and Nucleot. 8, 995 - 1000. 21.

Cazenave,C.,

22.

Subasinghe,C., Helene,C., Cohen,J.S. and Toulme,J.J. (1989) Nucl. Acids Res. 17, 4255-4273. Toulme,J., Kvisch,H., Loreau,N., Thuong,N.T. and Helene,C. (1986) Proc.

Stein,C.A.,

Natl. Acad. Sci. 23.

USA 83,

Loreau,N.,

Thuong,N.T.,

Neckers,L.M.,

1227-1231.

Zerial,A., Thuong,N.T. and Helene,C. (1987) Nucl. 9909-9919.

Acids

Res.

15,

24. Mori,K., Subasinghe,C., Stein,C.A. and Cohen,J.S. (1989) Nucleos. and Nucleot. 8, 649-657. 25. Durand,M., Maurizot,J.C., Asseline,U., Barbier,C., Thuong,N.T. and Helne.C. (1989) Nucl. Acids Res. 17, 1823-1837. 26. Knorre,D., Vlassov,V. and Zarytova,V. (1985) Biochemie 67, 785-789. 27. Vlassov,V.V., Zarytova,V.F., Kutyavi,I.V. and Mamaev,S.V. (1988) FEBS Lett. 231, 352-354. 28. Lin,S.-B., Blake,K.R., Miller,P.S. and Ts'o,P.O.P. (1989) Biochemistry 28, 1054-1061. 29. Francois,J.-C., Saison-Behmoaras,T., Chassignol,M., Thuong,N.T. and Hel6ne,C. (1989) J. Biol. Chem. 264, 5891-5898. 30. Lemaitre,M., Bayard,B. and Lebleu,B. (1987) Proc. Natl. Acad. Sci. USA 84, 648-652. 31. Lemaitre,M., Bisbal,C., Bayard,B. and Lebleu,B. (1987) Nucleos. and Nucleot. 6, 311-315. 32. Otvos,L., B6res,J., Sagi,Gy., Tomoskozi,I., and Gruber,L. (1987) Tetrahedron Lett. 28, 6381-6384. 33. Kalmdn,A., Koritsanszky,T., B6res,J. and Sagi,Gy. (1989) Nucleos. and Nucleot. accepted. 34. Szemzo,A., Szecsi,J., Sagi,J. and Otvos,L. (1989) Tetrahedron Lett. accepted. 35. Otvos,L., Sagi,J., Kovacs,T. and Walker,R.T. (1987) Nucl. Acids Res. 15, 1763-1777. 36. Szabolcs,A., Kruppa,G., Sigi,J. and Otvos,L. (1978) J. Lab. Comp. Radiopharm. 14, 713-726. 37. Handbook of Biochemistry and Molecular Biology. 3rd Edition, Nucleic Acids - Volume I (Fasman,G.D., Ed.) (1975), p. 581. 38. Cassani,G.R. and Bollum,F.J. (1969) Biochemistry 8, 3928-3936. 39. Shealy,Y.F. and O'Dell,C.A. (1976) J. Heterocycl. Chem. 13, 1041-1047. 40. Kornberg,A. (1980) DNA Replication, Freeman and Co., San Francisco, pp. 223. 41. Bollum,F.J. (1974) In Boyer,P.D. (ed.), 7he Enz;ymes, Academic Press, New York, pp. 145-171. 42. Bardos,T.J. and Ho,Y.K. (1982) An Update on Antitemplates In New Approaches to the Design of Antineoplastic Agents, Bardos,T.J. and Kalman,T.I. (eds.), Elsevier, North Holland, New York, pp. 315-332. 43. Montgomery,J.A. (1986) Acc. Chem. Res. 19, 293-300. 44. Marquez,V.E. and Lim,M.-I. (1986) Med. Res. Rev. 6, 1-40. 45. Ikehara,M. and Fukui,T. (1973) J. Biochem. 73, 945-950. 46. De Clercq,E. (1985) Nucleos. and Nucleot. 4, 3-11. 47. Sagi,J., De Clercq,E., Szemzo,A., Csarnyi,A., Kovacs,T. and Otvos,L. (1987) Biochem. Biophys. Res. Commnun. 147, 1105-1112. 48. Sagi,J., Szecsi,J., Szemzo,A., Sagi,Gy. and Otvos,L. (1987) Nucl. Acids Res. Symp. Ser. 18, 131-135. 49. Marky,L.A. and Breslauer,K.J. (1987) Biopolymers 26, 1601-1620. 50. Breslauer,K.J., Frank,R., Blocker,H. and Marky,L.A. (1986) Proc. Natl. Acad. Sci. USA 83, 3746-3750. 51. Janik,B. (1971) Physicochemical Characteristics of Oligonucleotides and Polynucleotides, IFI/Plenum, New York. 52. Westhof,E. (1988) Ann. Rev. Biophys. Biophys. Chem. 17, 125-144. 53. Quartin,R.S., Brakel,C.L. and Wetmur,J.G. (1989) Nucl. Acids Res. 17, 7253 -7262. 54. Chang,L.M.S., Cassani,G.R. and Bollum,F.J. (1972) J. Biol. Chem. 247, 7718-7723.