Bioconjugate Chem.

2002, 13, 11551158

1155

TECHNICAL NOTES
Evaluation of New Linkers and Synthetic Methods for Internal Modified Oligonucleotides
Troy A. Walton,* Matthew H. Lyttle, Daren J. Dick, and Ronald M. Cook
Biosearch Technologies, Inc., 81 Digital Drive, Novato, California 94949. Received January 21, 2002;
Revised Manuscript Received May 15, 2002

Several fluorescence resonance energy transfer (FRET) oligonucleotide probes were made with different internal linkages between the DNA and the quencher dye. In one example, a 5-fluorescein -actinbased 26-mer DNA sequence was synthesized bearing an internal Tamra quencher. Two different versions were prepared using either conventional C5 [N-(6-aminohexyl)-3-acrylamido]pyrimidinemodified uridine and solution-phase Tamra active ester coupling or solid-phase addition of a Tamra amidite to a C5 [N-(6-hydroxyhexyl)-3-acrylamido]pyrimidine-modified uridine. The products were compared in functional assays. They performed very similarly both in a fluorescence-based melting point assay as well as in quantitative PCR. Another set of -actin probes were synthesized utilizing N4 [N-2-(ethylene glycol ethyl)-5-methyl]cytidine and solid-phase Tamra amidite addition at positions flanking those of the uridine. These versions gave lower Tms than either uridine-labeled probe and did not work as well in quantitative PCR. A control experiment using oligonucleotides with the same modified residues but without fluorophores attached revealed the same trend as the Tm study of internal Tamra-labeled probes. Experimental details for the synthesis, purification, and testing are presented.

INTRODUCTION

A variety of different FRET oligonucleotide probes require internal modifications, including sunrise (1), scorpion (2), tandem dye reporters or FRET reporters (3), and the invader fluorescent assay (4). The maximum FRET effect has been reported with 4-5 nucleotide bases separating reporter and quencher dyes (5, 6), which suggests that the use of internally linked fluorophores (or quenchers) would provide the best performance for longer sequences. Methods for generating internally labeled synthetic oligonucleotides are limited, with solution-phase coupling of active esters to alkylamine internal modifications being prevalent (3). Traditionally, solidphase fluorescent-tagged DNA synthesis is restricted to fluorophore incorporation at the 3 or 5 terminus. In recent work, we developed a method using an internal uridine nucleotide modified at C5 with a protected alkyl hydroxyl group to synthesize internally labeled probes on a solid-phase support (7). We report here application of this technique to the synthesis of dual-labeled FRET probes and compare the performance of these structures to those synthesized using conventional uridine C5 alkylamine and solution-phase N-hydroxysuccinimide (NHS) active ester coupling methods. Probes were also made with a cheaper internal linkage system, utilizing N4-protected alkylhydroxyl functionalized cytidine. Figure 1 shows the structure of the three internal linkages to Tamra. The different probes were compared by a quantitative PCR (qPCR, 3) assay and melting temperature assays (8, 10). Probes were synthesized with
* Address correspondence to this author. Phone: 415-8838400; fax: 415-883-8488; e-mail: Troy.Walton@colorado.edu.

fluorescein on the 5 end and Tamra internally or with no fluorophore modifications.

MATERIALS AND METHODS

Oligo Synthesis. The probes with sequence 5-ATGCCCTCCCCCATGCCATCCTGCGT-3 were targeted for the human -actin gene (3). All oligos were synthesized on Biosearch 8700 DNA synthesizers, using 3-phosphate controlled pore glass (CPG) (p/n BG5-5000). Modified nucleotide amidites (7) and Tamra and fluorescein amidite (FAM) (9) were synthesized in-house along with the Tamra NHS ester used for solution-phase coupling. The modified bases were incorporated at the boldface underlined positions shown above. Syntheses were done at a 1 mol scale. For structure I (see Figure 1), also T7-NHTamra in Table 1, 5-[N-(6-trifluoroacetylaminohexyl)-3(E)-acrylamido]-5-(4,4-dimethoxytrityl)-3-(2-cyanoethyldiisopropylphosphoramidite)-2-deoxyuridine (7) was used in lieu of the T, 7 bases from the 5 end, and the rest of the bases were added followed by FAM addition (7). Workup and Tamra incorporation were done with solution-phase methods as previously reported (3). Figure 2A shows an anion exchange (AX) HPLC trace and photodiode array (PDA) spectrum of the purified probe. For the probe containing internal modification II, also T7-O-Tamra, automated synthesis was halted after the addition of 5-[N-(6-O-levulinoyl-1-aminohexyl)-3-(E)acrylamido]-5-(4,4-dimethoxytrityl)-3-(2-cyanoethyldiisopropylphosphoramidite)-2-deoxyuridine (7) at T7, and the CPG was treated with the dilute hydrazine/pyridine/ acetic acid (HOAc) cocktail outlined in (7). Tamra amidite (9) was coupled onto the now unmasked hydroxyl group, followed by oxidation and capping. The rest of the bases were then added by automated synthesis, followed by

10.1021/bc0200125 CCC: $22.00 2002 American Chemical Society Published on Web 08/16/2002

1156 Bioconjugate Chem., Vol. 13, No. 5, 2002

Figure 1. Structures of internal Tamra linkages. I ) conventional amine and Tamra NHS coupling; II ) Tamra amidite functionalized T C5 alkylhydroxyl; III ) Tamra amidite functionalized C N4 alkylhydroxyl.
Table 1. Oligonucleotide Sequences of Probes Used in This Studya atgccctcccccatgccatcctgcgt atgccctcccccatgccatcctgcgt atgccctcccccatgccatcctgcgt atgccctcccccatgccatcctgcgt atgccctcccccatgccatcctgcgt atgccctcccccatgccatcctgcgt atgccctcccccatgccatcctgcgt atgccctcccccatgccatcctgcgt T7-NH-Tamra T7-O-Tamra C6-O-Tamra C8-O-Tamra unmodified T7-OH C6-OH C8-OH

a Modified residues are indicated in boldface type. Oligos with internal Tamra modifications were 5-FAM-modified as well.

FAM addition. Figure 2B shows an AX HPLC trace and PDA spectrum of this purified probe. For the other probes containing modification III at the C positions flanking T7, N4-[2-(ethylene glycol-2-levulinate)ethyl]-5-methyl5-(4,4-dimethoxytrityl)-3-O-(2-cyanoethyldiisopropylphosphoramidite)-2-deoxycytidine (7) was used at the appropriate position to generate either C6-O-Tamra or C8-O-Tamra (Table 1), and the other manipulations were the same as those used to generate II. Figure 2C,D has the comparative analysis of these probes. A similar set of probes was synthesized without fluorescent dyes for use as nonlabeled controls in a Sybr Green I melting temperature assay (10). These are referred to as T7-OH, C6-OH, C8-OH, and unmodified. After synthesis, Tamra deprotection cocktail (9) and 16 h of heating at 60 C were used to deprotect and remove the probes from the CPG. Volatile components were removed under vacuum, and the crude sample was dissolved in 1 mL of water. Probes were purified using a double-preparative HPLC method consisting of an anion exchange (AX) purification followed by a reversed phase (RP) purification. A Dionex DNA Pac

Figure 2. Analytical AX HPLC of internally Tamra-labeled FRET probes. The inset shows the PDA spectrum of each sample. Sequences shown in Table 1.

PA100 9 250 mm column was used for AX, and a Hamilton PRP-1 10 250 mm 7 m particle size column was used for RP. A 5 mL injection loop with a 10 mL/ min flow rate for AX and a 6 mL/min flow rate for RP were used. The buffers for the AX preparation consisted of 0.038 M Tris at pH 8.0 plus 15% acetonitrile (ACN) for buffer A, and buffer B was buffer A plus 1 M NaBr. The purification gradient was 10-70% B over 12 min. The buffers for the RP purification consisted of 0.1 M TEAA (triethylamine acetate) + 5% ACN for A, and B was ACN. The gradient was 0-75% B over 15 min. A Waters 600 controller and pump with a Waters 490 multiple wavelength detector were used for purification. Intensely colored fractions were inspected by analytical HPLC, and those containing pure material with the

Bioconjugate Chem., Vol. 13, No. 5, 2002 1157

proper spectrum were pooled and evaporated. The yield of purified probes made at this scale was 60-120 nmol. PCR and Melting Temperature Conditions. The PCR reactions were carried out on an ABI 7700 using 50 L reaction volumes in Taqman Buffer A (Applied Biosystems). The reactions contained 200 nM primers (Applied Biosystems), 400 nM FRET probe, 100 or 1 ng of sheared human genomic DNA (Clontech), with 4 mM MgCl (Applied Biosystems), 200 M each dNTP (Sigma), 0.3 unit of BstXl (Boehringer Mannheim), and 2.5 units of Platinum Taq (Gibco). Thermal cycling conditions were 48 C for 30 min, then 95 C for 10 min followed by 40 cycles of 95 C (25 s) and 60 C (60 s). The ABI 7700 was set for a single dye (FAM) with Tamra as the quencher. The melting curves were also generated on the 7700 in PCR buffer consisting of 10 mM Tris, pH 8.3, 50 mM KCl, and 3.5 mM MgCl2. The temperature was decreased in 1 C steps from 90 to 50 C using the auto increment function with each step lasting 1 min. Probes were suspended at 200 nm, and compliment was suspended at 1 M. For nonlabeled DNA reactions, Sybr Green I was used at a final 1:40 000 dilution from the stock solution provided by the manufacturer.
RESULTS AND DISCUSSION

Figure 3. Melting temperature curves of Tamra-quenched probes generated as a function of increasing fluorescein fluorescence on an ABI 7700. RFU ) relative fluorescence units.

Melting Temperature (Tm) Experiments. Instead of conventional Tm measurements using UV absorbance, we chose a system more relevant to the evaluation of FRET probes. Similar buffers and the same instrument were used to generate the Tm data as for qPCR. When the probe hybridizes to exact complement and forms duplex DNA, the fluorescence increases relative to the single-stranded probe free in solution (3). Because of this, the Tm can be determined by measuring the fluorescence over a temperature range the same way one would use absorbance measurements at 260 nm (8). In the nonlabeled DNA reactions, the intercalating dye Sybr Green I was added to the reaction mix. Sybr Green I increases in fluorescence when it binds to double-stranded DNA. In this melting temperature assay, the change in fluorescence of the Sybr Green I with respect to increasing double-stranded DNA concentration vs temperature reveals melting curves similar to those obtained for the dual dye-labeled oligos (10). The data for the dual-labeled probes show that probes containing internal structures I and II anneal to complement at similar Tms. In contrast, both of the modified cytidine compounds (structure III) give lower Tms than those given by the modified thymine compounds. They also have a much broader Tm range than the thymine-based probes, suggesting less specificity during the hybridization event (see Figure 3). This same trend is apparent in Figure 4, which shows a similar melting temperature experiment. In this experiment, the same probe structures were synthesized without fluorescent dyes, and the amine modified probe was replaced by a totally unmodified oligo. The conditions for the nonlabeled DNA experiment were the same except that SYBR Green I was used as the reporter. In the Sybr Green I experiment, the melting curves for the unmodified DNA and the hydroxyl-modified thymine oligo (T7OH in Figure 4) overlap, while the hydroxyl-modified cytidine compounds show a significantly lower melting temperature. When using either amine- or hydroxyl-modified thymine chemistries, the performance of the dual-labeled probe is virtually unaffected. This same trend has been observed when comparing N4-modified cytidine compounds with hydroxyl functionality and amine function-

Figure 4. Melting temperature curves of unlabeled probes generated as a function of increasing Sybr Green I fluorescence on an ABI 7700. RFU ) relative fluorescence units.

ality (data not shown). We can attribute the difference between the T- and C-labeled probes to the modifications that were used for internal labeling, although other factors might be at play. One plausible explanation is that the thymine residue used has a modification on the methyl group at position C5, while the cytidine residue is modified at N4, and N4 in unmodified cytosine participates in an energetically important G-C base pair. The appendage at position C5 of thymine protrudes out into the major groove of B-DNA and should not affect base pairing. Realistically, we have empirically compared several easily accessible linkages in an important application. Different spacer length and geometry, as well as the presence of a phosphate moiety in some of the linkers, may also be having an effect. Although the linkage differences are more directly compared in the nonlabeled DNA experiment using SYBR Green I, the intercalation itself appears to have some effect on the Tm of these oligos (Figure 4). Quantitative PCR Experiments. There are several different chemistries available for qPCR (11), including the 5f3 exonuclease assay, which was used for this comparison (12). In this assay, a dual-labeled fluorescent oligonucleotide is hybridized to the amplicon between the primers and is cleaved by the 5f3 nucleolytic activity of Taq polymerase. This cleavage event generates fluorescence because the reporter fluorophore is no longer in close enough proximity to the quencher for efficient

1158 Bioconjugate Chem., Vol. 13, No. 5, 2002

Table 2. Data from qPCR Experimentsa 1 ng target probe T7-NH-TMR T7-O-TMR C6-O-TMR C8-O-TMR CT 28.85 29.13 31.89 31.09 std dev 0.14 0.06 0.13 0.08 100 ng target CT 21.88 22.2 24.7 24.0 std dev 0.13 0.02 0.04 0.02

a All reactions were run in triplicate. The average C is shown T with the standard deviation for each triplicate reaction. Reactions were done for two different input amounts of target human genomic DNA (1 and 100 ng).

toolbox for providing these modifications has been expanded, and both C and T are now available with added amine or hydroxyl functions for additional elaboration. The resulting different linker structures do not significantly change the performance of the modified T probes we tested. The solution-phase amine to active ester coupling method requires more synthetic manipulations. Therefore, we prefer the ease and faster throughput of solid-phase synthesis method made possible with the hydroxyl compounds.
LITERATURE CITED
(1) Nazarenko, I. A., Bhatnagar, S. K., and Hohman, R. J. (1997) A closed tube format for amplification and detection on DNA based on energy transfer. Nucleic Acids Res. 25, 2516-2521. (2) Whitcombe, D., Theaker, J., Guy, S., Brown, T., and Little, S. (1999) Detection of PCR products using self-probing amplicons and fluorescence. Nat. Biotechnol. 17, 804-807. (3) Livak, K., Flood, S., Marmaro, J., Giusti, W., and Deetz, K. (1995) Oligonucleotides with fluorescent dyes at opposite ends provide a quenched probe system useful for detecting PCR product and nucleic acid hybridization. PCR Methods Appl. 4, 1-6. (4) Hall, J., Eis, P., Law, S., Reynaldo, J., Prudent, D., Marshall, H., Allawi, A., and Mast, A. (2000) Sensitive detection of DNA polymorphisms by the serial invasive signal amplification reaction. Proc. Natl. Acad. Sci. U.S.A. 97, 8272-8277. (5) Cardullo, R. A., Agarwal, S., Flores, C., Zamecnik, P. C., and Wolf, D. E. (1988) Detection of Nucleic acid hybridization by nonradiative fluorescence resonance energy transfer. Proc. Natl. Acad. Sci. U.S.A. 85, 8790-8794. (6) Morrison, L. (1992) Detection of energy transfer and fluorescence quenching. Nonisotopic DNA Probe Techniques (Kricka, L. J., Ed.) pp 311-352, Academic Press, San Diego. (7) Lyttle, M., Walton, T., Dick, D., Carter, T., Beckman, J., and Cook, R. (2002) New Reagents and Methods for the Synthesis of Internal and 3-Labeled DNA. Bioconjugate Chem. 13, 1146-1154. (8) Tyagi, S., Bratu, D., and Kramer, F. R. (1998) Multicolor molecular beacons for allele discrimination. Nat. Biotechnol. 16, 49-53. (9) Lyttle, M., Carter, T., Dick, D., and Cook, R. (2000) A tetramethyl rhodamine (Tamra) phosphoramidite facilitates solid-phase-supported synthesis of 5-Tamra DNA. J. Org. Chem. 65, 9033-9038. (10) Ririe, K. M., Rasmussen, R. P., and Wittwer, C. T. (1997) Product Differentiation by Analysis of DNA Melting Curves during the Polymerase Chain Reaction. Anal. Biochem. 245, 154-160. (11) Bustin, S. A. (2000) Absolute quantification of mRNA using real-time reverse transcription polymerase chain reaction assays. J. Mol. Endocrinol. 25, 169-193. (12) Holland, P. M., Abramson, R. D., Watson, R., and Gelfand, D. H. (1991) Detection of specific polymerase chain reaction product by utilizing the 5f3 exonuclease activity of Thermus aquaticus DNA polymerase. Proc. Natl. Acad. Sci. 88, 72767280. (13) Heid, C. A., Stevens, J., Livak, K. J., and Williams, P. M. (1996) Real Time Quantitative PCR. Genome Res. 6, 986-994.

energy transfer via the FRET mechanism (13). By monitoring the increase in fluoresence produced by this cleavage event over several cycles of PCR, the amount of original target DNA in the reaction mix can be measured (13). The cycle in which the fluorescence rises above a certain background level is referred to as threshold cycle (CT) and is used to calculate the amount of target DNA. Comparison of the oligos from Table 1 in a quantitative PCR assay is summarized in Table 2. While the performance difference, as expressed with the CT value, is small between the two thymine compounds (dCT ) 0.28 for 1 ng target), the difference between the thymine and cytidine compounds is quite dramatic (dCT ) 2.76 for 1 ng target, T7-O-Tamra compared with C6-O-Tamra). This same trend is also seen in the 100 ng target samples. The difference in the melting temperature between these probes quite possibly accounts for the difference in performance seen here. The modified cytidine probes have a lower Tm (62 C) compared to the modified thymine probes (73 C) and therefore would not hybridize as well at the 60 C extension temperature of the PCR reaction. This difference in hybridization would lead to fewer of the modified cytidine probes being cleaved by the nucleolytic activity of Taq polymerase, thus requiring more cycles of PCR to generate the same fluorescence as the thymine probes. Reprogramming the thermal cycler used for this qPCR experiment with a lower extension temperature would perhaps give improved or equivalent performance, with respect to CT.
CONCLUSION

Several different ways of attaching Tamra to internal positions of a FRET probe targeting the human -actin gene have been compared in practical applications. We have supplied useful data, which should be helpful in the design of more effective internally labeled FRET probes. A logical extension of this work would be the preparation of analogous C5-modified cytidine derivatives with a free amine at position N4; this work is in progress. Our experiments show that when designing an oligonucleotide with an internal label on a modified nucleotide, N4-modified cytidine compounds do not perform the same as C5-modified thymine compounds. The synthetic

BC0200125