Effect of the Conjugation Site on Duplex Stabilization and

Feb 1, 1997 - Department of Chemistry, Merkert Chemistry Center, Boston College, Chestnut Hill, Massachusetts 02167. Received September 26, 1996X...
1 downloads 0 Views 278KB Size
Download PDF
Bioconjugate Chem. 1997, 8, 119−126

119

Oligonucleotides Tethering Hoechst 33258 Derivatives: Effect of the Conjugation Site on Duplex Stabilization and Fluorescence Properties Kristin Wiederholt, Sharanabasava B. Rajur, and Larry W. McLaughlin* Department of Chemistry, Merkert Chemistry Center, Boston College, Chestnut Hill, Massachusetts 02167. Received September 26, 1996X

A series of DNA conjugates have been prepared in which two different derivatives of Hoechst 33258 have been tethered to a sequence containing a 5′-GAATTC-3′ target site. The two derivatives differ only in the length of the tether between the DNA and the Hoechst fluorophore. By using a DNA backbone labeling protocol, one in which the Hoechst dye is tethered to an internucleotide phosphoramidate residue, it was possible to easily vary the site of attachment with respect to the A-T rich binding site. When tethered outside the GAATTC sequence, little if any helix stabilization results upon hybridization of the conjugate to its complementary sequence. As the site of conjugation is moved to one end of the target sequence and finally within the AATT sequence, more effective helix stabilization results. When tethered between the two A residues, or between the A and T residue, a ∆Tm of at least +20 °C is observed. Upon hybridization and formation of the B-form DNA, binding by the tethered Hoechst dye results, and the bound dye becomes brightly fluorescent. Upon a simple titration of the single-stranded conjugate with the complementary target sequence the quantum yield enhancement for hybridization only appears to be 5-7-fold at best. These fluorescence effects, generally less dramatic than those observed with other sequences, result from an increase in quantum yield for the single-stranded conjugate relative to the free Hoechst 33258. Heating the single-stranded conjugate reduces the inherent fluorescence of the single-stranded conjugate to a level comparable with that of the free Hoechst dye. In experiments monitoring absorbance vs temperature, a cooperative transition is observed for the single-stranded conjugate. Both the high quantum yield observed for the singlestranded conjugate and the observed thermally induced transition suggest that the single-stranded conjugate can dimerize (at the GAATTC site), mediated by the groove-binding fluorophore.

INTRODUCTION

The Hoechst dyes are a class of extended heterocycles based upon the bisbenzimidazole ring system that are capable of binding in the minor groove of B-form DNA at A-T-rich sequences. The flexible nature of this extended heterocyclic ring system permits the dye to adopt a conformation that is isohelical with the minor groove and thus optimize binding to double-stranded DNA (1). Crystallographic analyses (2-5) have confirmed the minor groove nature of these dye-DNA complexes and suggested that the preference for binding in A-T-rich sequences is the ability of the benzimidazole rings to penetrate deeply within the groove structure, where they can make hydrogen bonding contacts with specific thymine O2-carbonyls and adenine N3-nitrogens (2-5). Electrostatic effects (6) and van der Waals contacts (7) also play an important role in complex formation. The presence of G residues in the DNA duplex results in one or more N2-amino groups protruding into the minor groove, and the presence of this functional group prevents penetration by the dye into the groove structure (8, 9). The sequence preferences for complex formation by Hoechst 33258 require a minimum of four contiguous A-T (T-A) base pairs (8, 10-14) with the double-stranded sequences AAAA‚TTTT and AATT‚AATT being the preferred binding sites. However, the exact nature of the binding may differ depending upon the target sequence; different sets of crystallographic analyses have indicated * Author to whom correspondence should be addressed [telephone (617) 552-3622; fax (617) 552-2705; e-mail [email protected]]. X Abstract published in Advance ACS Abstracts, February 1, 1997.

S1043-1802(96)00086-9 CCC: $14.00

that the position of the Hoechst 33258 fluorophore in the target sequence (GAATTC)2 can be across the ATTC sequence (2) or across the AATT sequence (3, 5), and similar differences are observed for the target (CAAATTTG)2 (15, 16). In addition to their groove-binding characteristics, the Hoechst dyes have fluorescent properties with a relatively large Stokes shift (17). Although only moderately fluorescent in aqueous solution, upon binding to doublestranded DNA they exhibit a dramatically enhanced quantum yield (18). These effects appear to be related to the tight binding within the minor groove, where the excited state is protected from the aqueous solution and from processes involving nonradiative decay of the excited state (17). The fluorescence characteristics of bisbenzimidazoles have been used to automate DNA content assays (19-21), to determine cell numbers (22, 23), and to sort chromosomes (24). The quantum yield effects for these dyes are sensitive enough to permit the detection of one target cell per million in mixed cell populations using appropriate instrumentation (25). The development of DNA conjugates that rely upon the well-characterized Watson-Crick hydrogen bonding interactions for sequence targeting, and also tether a ligand capable of reporting upon the hybridization event, should be important in the development of DNA-based diagnostics, such as in situ hybridization probes. In the current work, the ligand is a groove-binding agent that requires formation of the B-form DNA minor groove to initiate binding. In this respect, hybridization by the conjugate is required to trigger the binding event by the tethered ligand. The resultant fluorescent signal reports on the success of this event. We have previously described (26) © 1997 American Chemical Society

120 Bioconjugate Chem., Vol. 8, No. 2, 1997

the synthesis of two Hoechst 33258 analogues used to prepare the DNA-Hoechst conjugates. In the present paper we examine a specific binding site (GAATTC)2 containing four A-T (T-A) base pairs and determine how placement of the tethered Hoechst analogues affects both helix stability and the fluorescence properties of the conjugate/target hybridization complex. EXPERIMENTAL PROCEDURES

Materials. Oligodeoxynucleotides were synthesized using 2′-deoxynucleoside phosphoramidites on an Applied Biosystems 381A DNA synthesizer. The four common 2′-deoxynucleoside phosphoramidites containing aryl- or isobutyrylamides were purchased from BioGenex (San Ramon, CA). H-phosphonate derivatives were obtained from Glen Research (Sterling, VA). The controlled pore glass support containing the 3′-terminal nucleoside was a product of CPG Inc. (Fairfield, NJ). Cystamine dihydrochloride, adamantanecarbonyl chloride, anhydrous solvents, and ammonium hydroxide were all obtained from the Aldrich Chemical Co. (Milwaukee, WI). Nuclease P1 and snake venom phosphodiesterase were products of Boehringer Mannheim (Indianapolis, IN). Acrylamide and bis(acrylamide) were obtained from ICN Biomedicals (Cleveland, OH). Thin layer chromatography was performed on aluminum-backed precoated silica gel 60 F254 plates purchased from EM Science (Gibbstown, NJ). UV measurements for thermal denaturation studies employed an AVIV 14DS spectrophotometer equipped with digital temperature control. Fluorescence emission spectra were collected on a Shimadzu RF5000U fluorescence spectrophotometer containing a Shimadzu DR-15 microprocessor and graphics display terminal. Methods. Syntheses. The synthesis of the two Hoechst derivatives 3 and 4 has been described elsewhere (26), and the DNA-Hoechst conjugates were prepared using a thiol-containing linker attached to the DNA backbone (27). The site of the cystamine-based linker was determined by the position at which the nucleoside H-phosphonate derivative was incorporated. For conjugation sites i and ii, the H-phosphonate of dG was employed, and for conjugation sites iii and iv the Hphosphonate derivative of dA was used. Oligonucleotide Conjugation. To 1 A260 unit of a single diastereomer of the desired dodecamer in 100 mM Tris‚HCl, pH 8, was added DTT to a concentration of 10 mM, and the mixture was incubated at 50 °C for 1 h. HPLC analysis [50 mM triethylammonium acetate, pH 7.0, with a linear gradient of acetonitrile (7-28% over 30 min)] after this time period indicated the complete absence of the triphenylmethylacetyl-protected sequence (retention time ∼ 28 min) and the presence of a new peak (retention time ∼ 9 min). The bromoacetylated Hoechst derivative (3 or 4) in DMF was added to this reaction mixture to a concentration of 15 mM (40% DMF). The reaction mixture was shaken at ambient temperature for 48 h (a precipitate appeared during this period). An equal amount of formamide was added to the reaction mixtures; they were heated to 90 °C (to dissolve aggregates) and loaded directly onto the gel for isolation by electrophoresis. The conjugated product was purified on a 20% denaturing (7 M urea) polyacrylamide gel. A fluorescent band (365 nm excitation) was present at a position that was retarded slightly from that of the oligomer containing the cystamine-based linker. This band was excised from the gel, crushed, and soaked in 0.3 M sodium acetate, pH 6. The gel was removed and the product desalted using a C18 Sep-Pak column and a water/methanol gradient. The conjugated oligomer was eluted with

Wiederholt et al.

approximately 80% methanol in water. Yields varied, typically 0.3-0.5 A260 units. The product exhibited a UV-vis spectrum having characteristics of both DNA (λmax ) 260 nm) and the Hoechst fluorophore (λmax ) 342 nm). Tm Values. Tm values were obtained for complexes containing a 1:1 mixture of the conjugate and its complementary sequence in 21 mM HEPES, pH 7.5, containing 100 mM NaCl and 20 mM MgCl2 at duplex concentrations in the low micromolar range (1-2 µM). Solutions were heated to 90 °C and then cooled slowly to 0 °C prior to analysis. The samples were then heated in 0.5 or 1.0 °C steps, and absorbance readings were taken after a period of temperature stabilization. Absorbance and temperature readings were plotted using Igor software. Tm values were determined from first -and second-order derivatives, as well as graphically from the absorbance vs temperature plots. Fluorescence Measurements. Fluorescence measurements were made in solutions typically containing ∼1 µM of the DNA-Hoechst conjugate in 21 mM HEPES, pH 7.5, containing 100 mM NaCl and 20 mM MgCl2. All emission measurements were made with the following list of parameters: slit width, Ex/Em ) 10 nm/10 nm, low sensitivity, medium speed. Samples were introduced into a 1.25 mL cell thermally isolated with a water jacket. Temperature was controlled with a recirculating water bath typically at 20 or 90 °C. Fluorescence emissions were measured at 470 nm with an excitation wavelength of 342 nm. The fluorescence enhancement (∆F) values were ratios obtained at 450 nm. This emission wavelength represented the emission maximum for the duplex conjugates and was slightly off the emission maximum for the single-stranded conjugates. RESULTS AND DISCUSSION

We have previously described the preparation of a series of DNA-Hoechst conjugates and characterized their thermal denaturation and fluorescence properties for a sequence target containing six contiguous A-T1 base pairs (26). In the present work we examine complex formation between the conjugate and a target sequence containing only a four base-pair binding site for the Hoechst derivative and then determine how the site of conjugation impacts both helix stabilization and the fluorescence properties of the conjugated complexes. A four base-pair target site is required for binding by Hoechst 33258, and both AAAA‚TTTT and (AATT)2 have been shown to be the preferred four base-pair binding sites (8, 10-14). For this study we chose the binding site (GAATTC)2 in which the palindromic four base-pair A-Trich target site is flanked by two G-C base pairs. This target site has been used previously in crystal structure analyses (2, 3, 5) and identified by DNA footprinting studies using Hoechst 33258 (8, 10-14). Synthesis of the Oligonucleotide Conjugates. The parent Hoechst 33258 cannot be covalently tethered to the oligonucleotides of interest without the introduction of appropriate functionality. We have described the synthesis of two Hoechst analogues (3 and 4, Scheme 1) (26) in which a bromoacetamide tether is incorporated into the terminal phenyl moiety of the parent fluorophore. The two analogues differ only in the nature of the tether. Hoechst derivative 4 contains an additional three atoms (-OCH2CH2-) to lengthen the tether slightly with respect to 3. Both derivatives can be conjugated to an alkyl thiol in aqueous solutions at pH 8. 1

All letter abbreviations refer to the 2′-deoxynucleosides.

DNA-Tethered Hoechst 33258 Derivatives

Bioconjugate Chem., Vol. 8, No. 2, 1997 121

Scheme 1

The two Hoechst analogues were conjugated to the DNA sequences using a postsynthetic approach. To vary the site of conjugation within the DNA sequence, we have employed a backbone labeling procedure (27, 28) in which the dye is tethered to a short thiol linker incorporated into the DNA sequence as a phosphoramidate internucleotide linkage. The internucleotide phosphoramidate is prepared by oxidation of an H-phosphonate linkage, immediately after its introduction to the DNA sequence, using pyridine/CCl4 and the N-(triphenylmethylacetyl)cystamine (27). Oxidation of the H-phosphonate generates two isomeric phosphoramidate conjugates, the Rp and Sp diastereomers, tethering the N-triphenylacetamide derivative of cystamine (1, Scheme 1). Placement of the bulky triphenylacetamide on the linker attached to the stereochemical center of the phosphoramidates facilitated the separation of the Rp and Sp isomers during isolation by HPLC in most cases. A typical chromatogram is illustrated in Figure 1. After elution of the failed sequences, both those representing failed couplings and those representing failed oxidation of the

intermediate H-phosphonate, the two phosphoramidate diastereomers (a and b) tethering the triphenylacetamide-protected linker are eluted. The early eluting isomer “a” has been tentatively assigned as the Rp diastereomer (26), while the later eluting isomer “b” has been tentatively assigned as the Sp diastereomer. The yield of dodecamers tethering the cystamine-based linker is roughly half that expected for the native sequence. We believe this loss in yield results from hydrolysis of the putative phosphorochloridate (29, 30), the expected intermediate in the oxidation reaction. Owing to the reactivity of this intermediate with water, effective oxidation in the presence of the cystamine-based linker occurs only under strictly anhydrous conditions, and trace amounts of water will drastically reduce expected yieldssand generally result in increased amounts of failed sequences (note the early eluting peaks in Figure 1). Equal amounts of both diastereomers were obtained in all cases, and only in one case were the diastereomeric sequences unresolvable. When the cystamine-based tether was placed at the internucleotide linkage between

122 Bioconjugate Chem., Vol. 8, No. 2, 1997

Wiederholt et al.

Figure 1. HPLC isolation of the two phosphoramidate diastereomers (a and b) for the dodecameric sequence 5′ d[GpApGpGpAp(NHCH 2 CH 2 S-SCH 2 CH 2 NHCOCPh 3 )ApTpTpCpGpApG]. HPLC conditions are described under Experimental Procedures. Table 1. DNA-Hoechst Conjugates: Helix Stabilization and Fluorescence Characteristics

conjugate native duplex native duplex

Hoechst site of deriv conjgn

isomera

33258

Tmb (°C) ∆Fc ∆F*d 51 60

2.5e

5 5 6 7 7 8 8

3 3 3 3 3 3 3

i i ii iii iii iv iv

(Rp) (Sp) (Rp) + (Sp) (Rp) (Sp) (Rp) (Sp)

53 60 63 75 61 73 56

1.3 1.1 2.1 2.6 3.8 4.1 2.7

16 12 9 15 13 16 14

9 9 10 11 11 12 12

4 4 4 4 4 4 4

i i ii iii iii iv iv

(Rp) (Sp) (Rp) + (Sp) (Rp) (Sp) (Rp) (Sp)

57 65 57 73 67 72 67

1.1 1.4 2.7 2.0 1.8 6.7 4.7

18 19 20 27 30 35 37

a Assignment of specific diastereomers is based upon arguments made elsewhere (26). The use of parentheses is to indicate that the assignments are tentative and are not yet corroborated by structural data. b Tm values are the average of at least two determinations and have an estimated error range of (1 °C. c ∆F represents the ratio of the emission at 450 nm for the doublestranded conjugate relative to the single-stranded conjugate after titration by the complementary target. d ∆F* represents the ratio of the emission at 470 nm for the double-stranded conjugate relative to the single-stranded conjugate at 90 °C, after titration by the complementary target. e This value was obtained by comparing a solution of 1 µM Hoechst 33258 and 1µM singlestranded GAATTC-containing sequence and then titrating this solution with the complementary target single-stranded DNA. The alternate experiment, comparing the fluorescence of 1 µM Hoechst 33258 and then titrating with excess DNA duplex, results in a ∆F ) 26 (see text).

the G and A residues (site ii, Table 1), the diastereomeric sequences coeluted. For the conjugates 6 and 10 (Table 1), we could only prepare the requisite DNA sequence tethering the cystamine-based linker as a diastereomeric mixture.

Figure 2. HPLC analyses of dodecameric sequences tethering (a) the triphenylacetamide derivative of cystamine (1), (b) the unmasked thiol (2), and (c) a Hoechst 33258 conjugate (11) (see Scheme 1). A peak with the same retention time as that in (c) was also present when the detector was set at 345 nm and a sample of 11 analyzed under the same HPLC conditions. HPLC conditions: 50 mM triethylammonium acetate, pH 7.0, with a linear gradient of acetonitrile (7-28% over 30 min). Retention times for the various dodecameric conjugates did not vary significantly with sequence or with the Hoechst derivative.

The conjugates were formed by cleavage of the disulfide bond to unmask the reactive thiol linker (2) and then treatment with either of the Hoechst analogues 3 or 4 (Scheme 1). After isolation by polyacrylamide gel electrophoresis, conjugates of the type 7 or 11 were obtained as both the Rp and the Sp diastereomers (as noted above, 6 and 10 were obtained as isomeric mixtures). We were unable to isolate the conjugated oligonucleotides by HPLC, largely due to the problems of aggregation of the excess fluorophore in the reaction mixture. The Hoechst derivatives are known to aggregate in aqueous solutions at concentrations above 30 µM, and the conjugations reactions, while performed in aqueous/DMF mixtures, required dye concentrations of roughly 15 mM. However, after completion of the reaction, the aggregates could be dissolved by the addition of excess DMF and then the product isolated by loading the entire reaction mixture onto a denaturing polyacrylamide gel. After electrophoresis, and excision and isolation of the fluorescent band from the gel, the conjugates (stored at concentrations of ∼10 µM) could be analyzed by HPLC. HPLC analysis indicated that the initially prepared triphenylacetamide-protected sequence (Figure 2a), the unmasked thiol (Figure 2b), and the Hoechst conjugate (Figure 2c) could all be effectively resolved. Each conjugate exhibited a UV-vis absorbance spectrum characteristic of the absorption maximum for DNA (260 nm) as well as that for the Hoechst derivative (345 nm) (Figure 3). Tm Characteristics of the Oligonucleotide Conjugates. The native dodecameric duplex exhibited a Tm value of 51 °C (Table 1). In the presence of 1 equiv of Hoechst 33258, the Tm was raised by 9 °C. We cannot determine from this assay whether a single complex is present since the fluorophore is known to bind to DNA by more than one mode. Nevertheless, the increased Tm suggests some minor groove binding, presumably at the A-T-rich site, with attendant helix stabilization. We then examined a series of conjugates prepared from the Hoechst derivative 3. With the fluorophore tethered between the two G residues (5), but outside the (AATT)2 binding site (conjugation site i), moderate helix stabilization was present for one isomer (∆Tm ) +9 °C), but the second diastereomer was less effective than the free fluorophore in stabilizing the dodecameric duplex. Contrary to previous observations, it is the Sp diastereomer that affords the more effective helix stabilizing characteristics. When tethered between the G and A residues at the end of the A-T binding site (conjugate ii), only the diastereomeric mixture was available, but this mixture

DNA-Tethered Hoechst 33258 Derivatives

Figure 3. (a) Structure and UV-vis spectrum of the DNAHoechst conjugate (9) formed by the reaction of 4 with the dodecamer 5′ d(GpApGpGp(NHCH2CH2SH)ApApTpTpCpTpApG) tethering an unmasked thiol.

Figure 4. Absorbance vs temperature plots for the native dodecamer (left-side transition) and the conjugate 12 (Rp diastereomer) complexed to its complementary sequence (righthand transition). The arrow indicates the approximate change in Tm value between the two samples (∆Tm).

provided a 12 °C enhancement in Tmslikely an average value between the Rp and Sp diastereomers. In a previous study (26) the Rp diastereomers were always more effective than the Sp isomers in stabilizing the DNA duplex. This observation suggests that the purified Rp, if available, might exhibit a more significant Tm effect. In fact, this expectation was met with both conjugates 7 and 8. In both of these latter examples, the Rp isomers resulted in Tm values that were 22-24 °C higher than that of the native duplex (Table 1). An example of two typical absorbance vs temperature transitions for a conjugate that effectively stabilizes the DNA duplex is illustrated in Figure 4. Both Sp diastereomers of 7 and 8 were also reasonably effective in stabilizing the DNA dodecamer (∆Tm values of +10 and +5 °C, respectively). We have attributed the differences in helix stabilizing effects for the Rp and the Sp isomers to orientation effects related to the diastereomeric character of the linker. The Rp diastereomer orients the linker more toward the minor groove side of the DNA duplex, while the Sp diastereomer presents the linker more toward the major groove. The more effective helix stabilization by the Rp isomers of conjugates 7 and 8 is again consistent with

Bioconjugate Chem., Vol. 8, No. 2, 1997 123

this relative orientation of the ligands tethered to the DNA backbone (26). The Rp diastereomer 5 was not capable of providing very effective helix stabilization, but this result is more likely to reflect the position of the linker rather than any diastereomeric consequences. It remains unclear why the Sp diastereomer appeared more effective with this sequence, but even these effects are relatively moderate. With the tether placed outside the A-T-rich binding site, it is likely that the linker restricts access by the Hoechst derivative to the A-T-rich binding site. With the tether incorporated into the terminal phenyl ring of the Hoechst fluorophore, one might expect that positioning the linker near the end of the binding site would be most effective. With such a positioning, the Hoechst derivative would be directed along the binding site, ostensibly in its preferred binding mode. In this respect, conjugate 6 would appear to provide the optimal positioning for the linker. However, both sites that place the linker more toward the center of the binding target (iii and iv, Table 1) appear to provide nearly identical, and the most impressive, Tm effects. This result could reflect that there is sufficient linker flexibility to permit similar binding from either tethered position or that more than one binding mode within the (AATT)2 minor groove is possible. Crystal structure analyses indicate that two modes of binding by Hoechst 33258 are possible within the (GAATTC)2 binding site, differing in position by one base pair (2, 3, 5). The DNA conjugates formed from the Hoechst derivative 4 differ from those described above in that the tether consists of an additional three atoms (-CH2CH2O-), which could provide additional flexibility in binding to the target A-T site. Tethering 4 outside the A-T-rich binding site between the two G residues (conjugation site i, Table 1) resulted in slightly more effective helix stabilization that correlated with the results for 3. For the diastereomeric mixture of conjugate 10, which tethered between the G and the A residues (conjugate ii, Table 1), only moderate stabilization was observed. However, for the remaining two conjugation sites (iii and iv, Table 1), both Rp diastereomers (11 and 12) provided in excess of a 20 °C enhancement in Tm values (see Figure 4). The corresponding Sp diastereomers were again not as effective but stabilized the helix somewhat more effectively than did the corresponding conjugates prepared from 3. Although it not clear why, conjugation at site ii with the derivative prepared from 3 resulted in a sequence that was more effective at helix stabilization than the corresponding derivative prepared from 4. This suggests the role of the tether in some complexes is not fully elucidated, particularly with the minimum four base-pair binding site. Nevertheless, the most helix stabilizing conjugates were those that resulted from placing the linker within the targeted binding site. At both sites iii and iv the Rp diastereomers were most effective at helix stabilization, and this observation is also consistent with those regarding the conjugates prepared from 3, as well as those prepared previously for a binding site composed of six contiguous A-T base pairs (26). Fluorescence Properties. The four base-pair sequence (AATT)2 should provide an effective binding site for the fluorophore, but the sequences outside the binding site may impact the nature of the groove, particularly its width, and result in less than optimal protection of the excited state from collisional decay processes involving water (17). At a concentration of 1 µM Hoechst 33258, in the presence of 1 µM single-stranded oligonucleotide, 5′-d(GpApGpGpApApTpTpCpGpApG), titra-

124 Bioconjugate Chem., Vol. 8, No. 2, 1997

Figure 5. Fluroescence emission spectra (λex ) 342) for the single-stranded conjugate 5′ d(GpApGpGpAp(NHCH2CH2SHoechst)ApTpTpCpTpApG) (Rp 11) (a) at 90 °C, (b) at 20 °C, and (c) in the presence of the complementary target at 20 °C after heating and cooling.

tion with the complementary sequence, 5′-‚d(CpTpCpGpApApTpTpCpCpTpC) resulted in a moderate 2.5-fold enhancement in quantum yield. However, the alternate experiment, titration of free Hoechst 33258 by the doublestranded sequence, resulted in a 26-fold enhancement in quantum yield, very similar to what has been observed in other studies (18, 26). The solution containing free Hoechst 33258, and that containing the free fluorophore with one of the single strands of the GAATTC-containing sequence, differed in quantum yield by 10-fold. These differences in quantum yield enhancement suggest that the GAATTC-containing sequence, although not a complete palindrome, is capable of adopting a secondary structure in solution, perhaps mediated by the Hoechst 33258 groove binder, that results in significant enhancement in the emission quantum yield. Tethering the minor groove-binding fluorophore to the oligonucleotide probe results in similar moderate increases in quantum yield upon hybridization. The process can be envisioned to occur in two steps, one a recognition event between the probe and the target sequence. Hybridization creates the minor groove-binding site and triggers the second step, that of ligand binding. Since each sequence tethers but a single ligand, the fluorescence properties of each complex should reflect the presence of a single ligand. For each sequence, the fluorescence of the single-stranded conjugate was measured, and then the sequence was titrated with the complementary target sequence. After the addition of each aliquot, the fluorescence spectrum was again measured, and the result obtained with 1 equiv of the complementary strand is illustrated in Figure 5c. The ratio of the emissions at 470 nm for the initial singlestranded conjugate and the final double-stranded conjugate was used to determine the fluorescence enhancement (∆F) values (Table 1). The most significant fluorescence enhancement occurred with conjugate 12, in which the dye is tethered essentially in the center of the target sequence. One diastereomer results in nearly a 7-fold increase in quantum yield, while the second diastereomer has a value nearly 5-fold larger than that measured for the single-stranded conjugate. These two complexes also exhibit some of the highest Tm values (72 and 67 °C, respectively) observed in this study; both properties suggest that the tethered ligand is most effectively bound to the DNA duplex, and the A-T-rich target site, when tethered in the center of the sequence.

Wiederholt et al.

Although the diastereomers of sequence 11 also exhibit relatively high Tm values, the fluorescence properties are more moderate. These observations suggest that the fluorescence characteristics of such complexes may be more sensitive to the nature of the binding event, and in some cases the ligand may not effectively penetrate the groove structure sufficiently well to avoid some relaxation effects by nonradiative (e.g., collisional) processes. However, it is possible that the single-stranded conjugates can result in the formation of some secondary structure as noted above for the unconjugated sequences in the presence of Hoechst 33258. For example, they can undergo a self-complementary partial duplex assembly using the central palindromic (GAATTC)2 sequence, with additional stabilization by one of the tethered groovebinding agents. Such a complex could also exhibit a significant enhanced fluorescence emission signal such as that observed for the free Hoechst dye in the presence of the unconjugated sequences. Dissociation of this palindromic partial duplex, followed by association with the complementary target sequence during the titration experiments, would result in only moderate apparent quantum yield enhancement effectssexactly what has been observed. Since we could not physically separate the tethered ligand and DNA sequence, instead we heated the single-stranded conjugates in an attempt to destabilize such structures and release the bound fluorophore. Under these conditions we observed a cooperative transition similar to that expected for a helix-to-coil transition. A decrease in the quantum yield of the emission spectrum for the single-stranded conjugate was also observed during heating over the temperature range 10-90 °C. The fluorescence observed for the singlestranded conjugate at 20 °C (Figure 5b) was reduced (Figure 5a) to a level comparable with that of free Hoechst 33258. We then compared the fluorescence enhancement values for the complexes prepared by two procedures: (i) the single-stranded conjugate was titrated with the complementary target at 20 °C, and (ii) after formation of the 1:1 conjugated duplex, the complex was heated and cooled prior to fluorescence measurements. Regardless of the procedure used, the quantum yields observed in the fluorescence emission spectra were essentially the same, suggesting that simple titration by the complementary target sequence disrupts the putative GAATTC dimer and results in the formation of the conjugatetarget duplex (in some cases the emission for the heated/ cooled sample was very slightly reduced, reflecting perhaps minor amounts of degradation). With the presence of the dimerized single-stranded conjugate, and the inherent fluorescence enhancement from this complex, the ∆F values observed for the double-stranded conjugatetarget complex seem, at best, only moderate (Table 1, compare parts b and c of Figure 5). We obtained a second set of fluorescence enhancement values (∆F*, Table 1) in which we compared the fluorescence emission spectra of the double-stranded conjugates (at 20 °C) with the single-stranded conjugates at 90 °C (compare parts a and c of Figure 5). Heating the single-stranded conjugate disrupts the putative dimer and permits a comparison of the fluorescence for the unbound (but tethered) ligand with that of the corresponding conjugated duplex containing the bound ligand. The quantum yield for the single-stranded conjugate typically decreases by roughly an order of magnitude with heating. At higher temperature, nonradiative collisional effects may be slightly increased, thus improving the apparent quantum yield enhancement. For example, when free Hoechst 33258 is titrated with the native, nominally single-stranded

DNA-Tethered Hoechst 33258 Derivatives

GAATTC sequence, the quantum yield increases roughly 35-foldsconsistent with expectations for the (GAATTC)2 binding site. However, when this complex is heated at 90 °C, the quantum yield decreases an apparent 46foldspresumably because of the enhanced collisional effects at higher temperatures. Thus, a comparison of the observed emission values for the single-stranded (90 °C) and double-stranded (at 20 °C) conjugates must be done cautiously. Nevertheless, the quantum yield enhancements were now found to be as high as 37-fold, similar to, or greater than, that observed for the free Hoechst 33258 (Table 1). In general, duplexes containing the Hoechst derivative 4 (with the longer linker) tethered near or at the center of the AATT sequence resulted in the highest enhancement values when compared with the heat denatured single-stranded conjugate (27-37-fold). These differences are likely to reflect the greater flexibility afforded the ligand by the longer linker, but even the shorter linker present in complexes prepared from 3 results in quantum yield enhancements from 9- to 16fold. Conclusions. We have shown that the tethered derivative of Hoechst 33258 can be used to target the four base-pair sequence (AATT)2 and provide significant helix stabilization. The most effective position for the linker, employed to tether the fluorophore, is at the center of the targeted binding site. Both types of DNA conjugates, those prepared from either 3 or 4, enhance the Tm values of the duplex by more than 20 °C when present as the tentatively assigned Rp diastereomer. The fluorescence properties after titration of the single-stranded conjugates with the target sequence are less defining and appear at best only moderate. However, fluorescence vs temperature, and absorbance vs temperature, plots indicate that the single-stranded conjugates adopt some secondary structure, most likely dimerization about the palindromic GAATTC core sequence. Heating the singlestranded conjugates reduced their fluorescence properties, consistent with a dimerized complex about the tethered Hoechst dye. Although titration by the complementary target sequence disrupts the secondary structure present in the single-stranded conjugate and permits complex formation, the results of this study suggest that probe sequences containing palindromic sequences rich in A-T base pairs may suffer from dimerization effects mediated by such groove-binding agents. ACKNOWLEDGMENT

This work was supported by a grant from the NIH (GM37065). LITERATURE CITED (1) Goodsell, D., and Dickerson, R. E. (1986) Isohelical analysis of DNA groove-binding drugs. J. Med. Chem. 29, 727-733. (2) Pjura, P. E., Grzeskowiak, K., and Dickerson, R. E. (1987) Binding of Hoechst 33258 to the minor groove of B-DNA. J. Mol. Biol. 197, 257-271. (3) Teng, M.-K., Usman, N., Frederick, C. A., and Wang, A. H.J. (1988) The molecular structure of the complex of Hoechst 33258 and the DNA dodecamer d(CGCGAATTCGCG). Nucleic Acids Res. 16, 2671-2675. (4) de C. T. Carrondo, M. A. A. F., Coll, M., Aymami, J., Wang, A. H.-J., van der Marel, G. A., van Boom, J. H., and Rich, A. (1989) Binding of a Hoechst dye to d(CGCGATATCGCG) and its influence on the conformation of the DNA fragment. Biochemistry 28, 7849-7859. (5) Quintana, J. R., Lipanov, A. A., and Dickerson, R. E. (1991) Low-temperature crystallographic analyses of the binding of

Bioconjugate Chem., Vol. 8, No. 2, 1997 125 Hoechst 33258 to the double-helical DNA dodecamer C-G-CG-A-A-T-T-C-G-C-G. Biochemistry 30, 10294-10306. (6) Kissinger, K. L., Drowicki, K., Dabrowiak, J. C., and Lown, J. W. (1987) Molecular recognition between oligopeptides and nucleic acids. Monocationic imidazole lexitropsin that display enhanced GC sequence dependent DNA binding. Biochemistry 26, 5590-5595. (7) Lee, M., Drowicki, K., Hartley, J. A., Pon, R. T., and Lown, J. W. (1988) Molecular recognition between oligopeptides and nucleic acids: Influence of van der Waals contacts in determining the 3′-terminus of DNA sequences read by monocationic lexitropsins. J. Am. Chem. Soc. 110, 3641-3649. (8) Harshman, K. D., and Dervan, P. B. (1985) Molecular recognition of B-DNA by Hoechst 33258. Nucleic Acids Res. 13, 4825-4835. (9) Martin, R. F., and Holmes, N. (1983) Use of an 125I-labeled DNA ligand to probe DNA structure. Nature 302, 452-454. (10) Fox, K. R., and Waring, M. J. (1984) DNA structural variations produced by actinomycin and distamycin as revealed by DNAase I footprinting. Nucleic Acids Res. 12, 92719285. (11) Portugal, J., and Waring, M. J. (1987) Comparison of binding sites in DNA for Berenil, Netropsin and Distamycin as revealed by DNAase I footprinting. Eur. J. Biochem. 167, 281-289. (12) Portugal, J., and Waring, M. J. (1988) Assignment of DNA binding sites for 4′,6-diamidine-2-phenyl indole and bisbenzimidazole (Hoechst 33258). A comparative footprinting study. Biochim. Biophys. Acta 949, 158-168. (13) van Dyke, M. W., Hertzberg, R. P., and Dervan, P. B. (1982) Map of Distamycin, Netropsin and Actinomycin binding sites on heterogeneous DNA: DNA cleavage-inhibition patterns with methidium propyl-EDTA‚Fe(II). Proc. Natl. Acad. Sci. U.S.A. 79, 5470-5474. (14) Abu-Daya, A., Brown, P. M., and Fox, K. R. (1995) DNA sequence preferences of several AT-selective minor groove binding ligands. Nucleic Acids Res. 17, 3385-3392. (15) Spink, N., Brown, D. G., Skelly, J. V., and Neidle, S. (1994) Sequence-dependent effects in drug-DNA interactions: The crystal structure of Hoechst 33258 bound to the d(CGCAAATTTGCG)2 duplex. Nucleic Acids Res. 22, 16071612. (16) Vega, M. C., Garci Saez, I., Aymami, J., Eritja, R., van der Marel, G. A., van Boom, J. H., Rich, A., and Coll, M. (1994) Three dimensional crystal structure of the A-T tract DNA dodecamer d(CGCAAATTTGCG)2 complexed with the minor groove binding drug Hoechst 33258. Eur. J. Biochem. 222, 721-726. (17) Zimmer, C., and Wa¨hnert, U. (1986) Nonintercalating DNA binding ligands: Specificity of the interaction and their use as tools in biophysical, biochemical and biological investigations of genetic material. Prog. Biophys. Mol. Biol. 47, 31112. (18) Lootiens, F. G., Regenfuss, P., Zechel, A., Dumortier, L., and Clegg, R. M. (1990) Binding characteristics of Hoechst 33258 with calf thymus DNA, poly [d(A-T)] and d(CCGGAATTCCGG): Multiple stoichiometries and determination of tight binding with a wide spectrum of site affinities. Biochemistry 29, 9029-9039. (19) Sterzel, W., Bedford, P., and Eisenbrand, G. (1985) Automated determination of DNA using the fluorochrome Hoechst 33258. Anal. Biochem. 147, 462-470. (20) Araki, T., Yamanoto, A., and Yamada, M. (1987) Histochemistry 87, 331-338. (21) Karawajew, L., Rudchenko, S., Wlasik, T., and Trakht, I. (1990) Flow sorting of hybrid hybridomas using the DNA stain Hoechst 33258. J. Immunol. Methods 129, 277-282. (22) Downs, T. R., and Wilfinger, W. W. (1983) Fluorometric quantification of DNA in cells and tissue. Anal. Biochem. 131, 538-547. (23) Adams, C. J., and Storrie, B. (1981) A simple DNAdependent fluorescence enhancement assay for cell number. Histochem. Cytochem. 29, 326-334. (24) Arndt-Jovin, D. J., and Jovin, T. M. (1990) Fluorescence labeling and microscopy of DNA. Cytometry 11, 80-89.

126 Bioconjugate Chem., Vol. 8, No. 2, 1997 (25) Lee, B. R., Haseman, D. B., and Reynolds, C. P. (1989) A digital image microscopy system for rare-event detection using fluorescent probes. Cytometry 10, 256-266. (26) Wiederholt, K., Rajur, S. B., Giuliano, J. J., O’Donnell, M. J., and McLaughlin, L. W. (1996) DNA-tethered Hoechst groove-binding agents: Duplex stabilization and fluorescence properties. J. Am. Chem. Soc. 118, 7055-7062. (27) O’Donnell, M. J., Hebert, N., and McLaughlin, L. W. (1994) The stereospecific introduction of reporter groups to oligodeoxynucleotides by the labeling of individual phosphorus diastereomers. Bioorg. Med. Chem. Lett. 4, 1001-1004.

Wiederholt et al. (28) Fidanza, J., and McLaughlin, L. W. (1992) The use of a thiol tether for the sequence-specific attachment of reporter groups to DNA. J. Org. Chem. 57, 2340-2346. (29) Blackburn, G. M., Cohen, J. S., and Todd, A. R. (1966) Synthesis of dialkyl phosphates from monoalkyl phosphonates. Direct oxidative esterification. J. Chem. Soc., 239-245. (30) Atherton, F. R., Openshaw, H. T., and Todd, A. R. (1945) Dibenzyl chlorophosphonate as a phosphorylating agent. J. Chem. Soc., 660-663.

BC960086E