dict.md logo
Advertisement:
Advertisement:

Synthesis and application of charge-modified dye-labeled dideoxynucleoside-5′-triphosphates to ‘direct-load’ DNA sequencing

A novel series of charge-modified, dye-labeled 2′,3′-dideoxynucleoside-triphosphate terminators were synthesized and evaluated as reagents for DNA sequencing. These terminators possess an advantage over existing reagents in that no purification is required to remove unreacted nucleotide or associated breakdown products prior to electrophoretic separation of the sequencing fragments. This obviates the need for a time consuming post-reaction work up, allowing direct loading of DNA sequencing reaction mixtures onto a slab gel. Thermo Sequenase™ II DNA polymerase poorly incorporates the charge-modified terminators compared with regular dye-labeled terminators. However, extending the linker arm between dye and nucleotide and using a mutant form of a related DNA polymerase can in part mitigate the decrease in substrate efficiency. We also present evidence that these charge-modified terminators can relieve gel compression artefacts when used with dGTP in sequencing reactions.

Sanger dideoxy sequencing (1) is the most commonly used method for DNA sequencing, particularly in a high throughput laboratory. The preferred detection method is fluorescence, in which four spectrally distinguishable fluorophores are attached to either the sequencing primer or the dideoxynucleoside-triphosphates (ddNTPs). This allows detection of the dye-labeled Sanger fragments in a single separation (24). Since the development of DNA polymerases that are less discriminating against ddNTPs, dye-primer sequencing has been superseded by dye-terminator methods (58). The primer method is more laborious, requiring four separate reactions and four dye-labeled primers for each template. Another problem associated with dye-primer sequencing is the formation of ‘false stops’ where termination with a deoxynucleotide rather than dideoxynucleotide occurs. For dye-primer sequencing the ‘false stopproduct is fluorescent, but in a dye-terminator reaction the deoxy-product is non-fluorescent and hence undetected.

One of the major bottlenecks associated with dye-terminator sequencing is the requirement for post-reaction purification. During thermal cycling, fluorescent breakdown products from the dye-labeled ddNTPs are formed, both these and unreacted terminator, co-migrate with the labeled Sanger fragments during electrophoresis. Part of the sequence information is obscured unless the unreacted nucleotides and breakdown products are removed. Preferred purification methods include gel filtration to remove unreacted nucleotides, and selective precipitation of sequencing products from fluorescent by-products using an aqueous alcohol mixture (3,1316). Extreme care must be taken during precipitation to ensure selective removal of unreacted nucleotides. If conditions are not optimized, it is possible to remove shorter Sanger fragments whilst trying to ensure complete removal of unreacted nucleotide and associated breakdown products. The purification process is laborious, expensive and time consuming. If this step could either be eliminated, or at least simplified, then a bottleneck in dye-terminator sequencing would be by-passed making the process more efficient and robust.

In order to investigate methods for simplifying or eliminating post-sequencing reaction processing, we have synthesized a series of novel dye-labeled ddNTPs containing a negatively charged linker arm between the nucleobase and fluorescent label. The charged moiety does not degrade during cycle sequencing reaction conditions, so unreacted dye-labeled nucleoside triphosphate and the corresponding by-products migrate faster than the smallest dye-labeled Sanger fragments during electrophoresis and do not interfere with base calling. Therefore, it is possible to carry out electrophoresis of the sequencing reaction products without purification, a method we have termed ‘direct-load’ DNA sequencing.

HPLC purification was performed using a Waters 600 pump controller in conjunction with a 2487 dual wavelength detection system. Mass spectra were obtained using a Micromass LCT electrospray-TOF spectrometer. Purification solvents were of HPLC grade and were pre-filtered through a 0.45 µm membrane prior to use. 5-Propargylaminopyrimidine and 7-propargylamino-7-deazepurines were synthesized according to Hobbs and Cocuzza (9). Fluorescein and rhodamine succinimidyl esters were purchased from Molecular Probes (Eugene, OR) and are pure 5 regioisomers unless stated. α-Sulfo-β-alanine was obtained from Research Organics (Cleveland, OH). The number preceding the ddNTP indicates the number of atoms in the linker arm attached at the 5 position of pyrimidines and 7 position of 7-deazapurines. The linker arm is made up of aminocaproic acid units attached to the propargylamine modified base. For example, an 11-ddNTP has one caproic acid unit attached to the propargylamino group, an 18-ddNTP has a linker arm containing two aminocaproic acid units attached to the propargylamino group and a 25-ddNTP contains three aminocaproic acid units attached to the propargylamino group. The yield of various charge-modified terminators made in this work varies from 10 to 35% and the yields are not optimized. All other reagents were obtained from Sigma, Aldrich or Fluka unless stated. UV absorption spectra were recorded in TE buffer (pH 8.5). The abbreviations used throughout the paper are as follows: NHS (N-hydroxysuccinamide), DMF (N,N-dimethylformamide), TEAB (triethylammonium bicarbonate), FAM (5-Carboxyfluorescein), DMAP (4-N,N-dimethylaminopyridine), TSTU (O-(N-succinimidyl)-1,1,.3,3-tetramethyluronium tetrafluoroborate), DSC (N,N′-disuccinimidylcarbonate), DIPEA (N,N-diisopropylethylamine).

5-Carboxamidofluorescein–11-ddCTP (1). To a solution of 11-ddCTP (14.6 mg, 14.3 µmol) in NaHCO3/Na2CO3 buffer (0.1 M, pH 8.5) was added 5-carboxyfluorescein–NHS (10 mg, 1.5 eq.) in DMF (2 ml). The reaction mixture was stirred at room temperature for 1 h and the product purified by silica gel chromatography (iPrOH:NH4OH:H2O, 6:3:1 elution), followed by ion-exchange chromatography [Q Sepharose, 10 mm × 10 cm column, A = 0.1 M TEAB/MeCN 3:2 (v/v), B = 1.0 M TEAB/MeCN 3:2 (v/v), 4 ml/min, 0–100% B over 20 column volumes], followed by C18 HPLC (A = 0.1 M TEAB, B = MeCN, Waters Deltapak C18 column, 50 × 300 mm, flow 100 ml/min, 0–100% B over 60 min). Product containing fractions were evaporated to dryness in vacuo and dissolved in TE buffer at pH 8.5 (1.5 µmol, 10%). TOF MS: 974 (1–), λmax 498 nm.

N-(5-carboxamidofluorescein)-(α-sulfo-β-alanine) (2). α-Sulfo-β-alanine (59 mg, 0.35 mmol) was dissolved in DMSO (2 ml) with gentle heating. To the solution was added N,N-diisopropylethylamine (0.9 ml, 15 eq.) and 5-carboxyfluorescein–NHS (200 mg, 1.2 eq.) as a solid. The reaction mixture was stirred at room temperature for 16 h and the product purified by ion-exchange chromatography [Q Sepharose, 10 mm × 10 cm column, A = 0.1 M TEAB/MeCN 3:2 (v/v), B = 1.0 M TEAB/MeCN 3:2 (v/v), 4 ml/min, 0–100% B over 20 column volumes]. Product containing fractions were evaporated to dryness in vacuo and the residue co-evaporated with MeOH (5 × 10 ml) before drying under high vacuum. 1H NMR (DMSO, 300 MHz); 1.2 (t, J = 9.1 Hz, 27H, CH3CH2N), 3.05 (d, J = 9.1 Hz, 18H, CH3CH2N), 3.8–4.1 (m, 3H, CH2CHSO3), 6.55 (m, 4H, aromatic), 6.70 (s, 2H, aromatic), 7.25 (d, J = 9.0 Hz, 2H, aromatic), 7.35 (d, J = 9.0 Hz, 2H, aromatic), 7.39 (d, J = 6 Hz, 1H, aromatic), 8.27 (d, 1H, J = 6.0 Hz, aromatic), 8.43 (s, 1H, aromatic), 9.34 (m, 1H, NH). TOF MS: 526 (1–), λmax 498 nm.

5FAM-(α-sulfo-β-alanine-α-sulfo-β-alanine)-OH (3). Compound 2 (50 mg, 94 µmol) was dissolved in DMF (3 ml) and N,N-diisopropylethylamine (0.25 ml, 15 eq.) followed by O-(N-succinimidyl)-N,N,N′,N′-tetramethyluronium tetrafluoroborate (42 mg, 1.5 eq.) were added. The reaction mixture was stirred at room temperature for 1 h and α-sulfo-β-alanine (24 mg, 1.5 eq.) added as a solid. The suspension was stirred at room temperature for 16 h to yield a homogeneous solution. The product was purified by anion-exchange chromatography [Q Sepharose, 10 mm × 10 cm column, A = 0.1 M TEAB/MeCN 3:2 (v/v), B = 1.0 M TEAB/MeCN 3:2 (v/v), 4 ml/min, 0–100% B over 20 column volumes], followed by C18 HPLC (A = 0.1 M TEAB, B = MeCN, Waters Deltapak C18 column, 50 × 300 mm, flow 100 ml/min, 0–100% B over 60 min). 1H NMR (DMSO, 300 MHz); 1.22 (t, J = 9.1 Hz, 36H, CH3CH2N), 3.05 (d, J = 9.1 Hz, 24H, CH3CH2N), 3.8–4.2 (m, 6H, CH2CHSO3), 6.57 (m, 4H, aromatic), 6.72 (s, 2H, aromatic), 7.27 (d, J = 9.0 Hz, 2H, aromatic), 7.36 (d, J = 9.0 Hz, 2H, aromatic), 7.39 (d, J = 6 Hz, 1H, aromatic), 8.29 (d, 1H, J = 6.0 Hz, aromatic), 8.41 (s, 1H, aromatic), 9.40 (m, 1H, NH). TOF MS: 677 (1–), λmax 498 nm.

FAM-α-sulfo-β-alanine–11-ddCTP (4). FAM–α-sulfo-β-alanine–OH (2; 10 mg, 19 µmol) was dissolved in DMF (2 ml) and disuccinimidyl carbonate (19 mg, 4 eq.) added at room temperature. The reaction mixture was cooled to –60°C and DMAP (9 mg, 4 eq.) added in DMF (0.5 ml). The reaction was brought to –30°C and a solution of nucleotide (1 eq.) in NaHCO3/Na2CO3 buffer (0.1 M, pH 8.5, 2 ml) was added (9). The reaction mixture was brought to room temperature for 1 h and the product purified by silica gel chromatography [iPrOH:NH4OH:H2O, 6:3:1 (v/v/v)] followed by ion-exchange chromatography [Q Sepharose, 10 mm × 10 cm column, A = 0.1 M TEAB/MeCN 3:2 (v/v), B = 1.0 M TEAB/MeCN 3:2 (v/v), 4 ml/min, 0–100% B over 20 column volumes] followed by C18 HPLC (A = 0.1 M TEAB, B = MeCN, Waters Deltapak C18 column, 50 × 300 mm, flow 100 ml/min, 0–100% B over 60 min). Yield 21%, TOF MS: 1125 (1–), λmax 498 nm.

FAM-(α-sulfo-β-alanine-α-sulfo-β-alanine)–11-ddCTP (5). Compound 5 was synthesized in the same manner as for 4, using compound 3 as starting material on a 22 µmol scale. Yield 25%, TOF MS: 1276 (1–), 637 (M/2), λmax 498 nm.

FAM-(α-sulfo-β-alanine-α-sulfo-β-alanine)–18-ddCTP (6). Same method as for 4 on a 15 µmol scale. Yield 33%, TOF MS: 1389 (1–), 694 (M/2), λmax 498 nm.

FAM-(α-sulfo-β-alanine-α-sulfo-β-alanine)–25-ddCTP (7). Same method as for 4 on a 15 µmol scale. Yield 20%, TOF MS: 1502 (1–), λmax 498 nm.

Synthesis of charge-modified rhodamines (810). α-Sulfo-β-alanine (51 mg, 0.3 mmol) was dissolved in DMF (5 ml) and N,N-diisopropylethylamine (0.26 ml, 5 eq.) was added followed by the appropriate 5-rhodamine–NHS active ester (1.0 eq.) and the reaction stirred at room temperature for 3 h. To the reaction was added further N,N-diisopropylethylamine (0.26 ml, 5 eq.), then HBTU (150 mg, 1.2 eq.) and the reaction stirred at room temperature for 30 min. To the solution was added α-sulfo-β-alanine (51 mg, 0.3 mmol) and stirring continued for an additional 1 h. At this point the process of N,N-N,N-diisopropylethylamine/HBTU/α-sulfo-β-alanine addition was repeated and the reaction left to stir at room temperature for 16 h. The product was isolated by anion-exchange chromatography [Q Sepharose, 16 mm × 10 cm column, A = 0.1 M TEAB/MeCN 3:2 (v/v), B = 1.0 M TEAB/MeCN 3:2 (v/v), 6 ml/min, 0–100% B over 40 column volumes] followed by C18 HPLC (A = 0.1 M TEAB, B = MeCN, Waters Deltapak C18 column, 50 × 300 mm, flow 100 ml/min, 0–100% B over 60 min). The product was evaporated to dryness in vacuo and the residue co-evaporated with MeOH (3 × 50 ml) to yield the desired products. TOF MS: compound 8, observed 910 (1–) (118 mg, 43%); compound 9, observed 882 (1–) (132 mg, 50%); compound 10, observed 988 (1–), (99 mg, 20%).

Synthesis of charge-modified rhodamine nucleotides (1113). All labeling reactions and purifications were carried out using the same methodology as for 4 on a 15 µmol scale. TOF MS: compound 11, observed 1736 (1–) (1 µmol, 7%), λmax 527 nm; compound 12, observed 1617 (1–) (2 µmol, 14%), λmax 552 nm; compound 13, observed 1737 (1–) (1.5 µmol, 10%), λmax 581 nm.

Boc-β-phenylalanine-4-(propargylamido-5-fluorescein)-α-sulfo-β-alanine (15). Boc-β-phenylalanine-4-(propargylamido-5-carboxyfluorescein) (10) (14; 400 mg, 0.47 mmol) was dissolved in DMF (4 ml). DIPEA (0.25 ml, 3 eq.) followed by O-(N-succinimidyl)-N,N,N′,N′-tetramethyluronium tetrafluoroborate (155 mg, 1.1 eq.) in DMF (1 ml) was added. The reaction was stirred at room temperature for 5 min, then a solution of α-sulfo-β-alanine (95 mg, 1.2 eq.) in DMSO (3 ml, dissolved by gentle heating) added at room temperature. The reaction was stirred at room temperature for 16 h and DIPEA (0.3 ml, 3 eq.) added followed by O-(N-succinimidyl)-N,N,N′,N′-tetramethyluronium tetrafluoroborate (155 mg, 1.1 eq.) in DMF (1 ml). After stirring for a further 5 min a solution of α-sulfo-β-alanine (95 mg, 1.2 eq.) in DMSO (3 ml, dissolved by gentle heating) was added at room temperature. The reaction was stirred at room temperature for 16 h then evaporated to near dryness in vacuo. The residue was dissolved in water (50 ml) and the product purified by ion-exchange chromatography [Q Sepharose, 10 × 100 mm, A = 0.1 M TEAB/40% MeCN (v/v), B = 1.0 M TEAB/40% MeCN (v/v), flow = 6 ml/min, 0–100% B over 60 min]. Product eluted in 100% buffer B. TLCSiO2 Rf [iPrOH:NH4OH:H2O 6:3:1 (v/v/v)] = 0.25. 1H NMR (300 MHz, CD3OD); 1.39 [s, 9H, C(CH3)3] 2.27 (m, 2H, PhCH2), 2.80–2.85 (m, 3H, CHCH2), 3.45 (m, 1H, CHCH2), 3.75–4.20 (m, 6H, 2× NHCH2CHSO3), 4.05 (m, 1H, CHCH2), 4.43 (s, 2H, propargyl CH2), 6.58 (s, 4H, FAM H1′, H1″, H2′, H2″), 7.00 (d, 2H, J = 11.0 Hz, Ar-H), 7.22 (d, 2H, J = 11.0 Hz, phenylalanine Ar-H), 7.387 (d, 3H, J = 11.0 Hz, 1× FAM Ar-H, +, 2× phenylalanine Ar-H), 8.04 (d, 1H, J = 7.6 Hz, FAM H6), 8.47 (s, 1H, FAM-H4).

N-(β-phenylalanine-4-(propargylamido-5-fluorescein)-α-sulfo-β-alanine)amido-α-sulfo-β-alanine (16). N-(Boc-β-phenylalanine-4-(propargylamido-5-fluorescein)-α-sulfo-β-alanine)amido-α-sulfo-β-alanine (15; 100 mg, 0.10 mmol) was treated with trifluoroacetic acid (10 ml) for 1 h and the reaction evaporated to dryness in vacuo. The residue was triturated with Et2O (30 ml), and the solid product dried under high vacuum.

General methodology for the attachment of rhodamine dyes to 16 (synthesis of 1718). N-(β-phenylalanine-4-(propargylamido-5-fluorescein)-α-sulfo-β-alanine)amido-α-sulfo-β-alanine (16; 20.0 µmol) was dissolved in NaHCO3/Na2CO3 buffer (0.1 M, pH 8.5, 3 ml) and the desired rhodamine–NHS active ester (1.5 eq.) in DMF (3 ml) added. The reaction mixture was stirred at room temperature for 16 h, then evaporated to dryness in vacuo. To the reaction mixture was added water (50 ml) and the product purified by ion-exchange chromatography [Q Sepharose, 10 × 100 mm, A = 0.1 M TEAB/40% MeCN (v/v), B = 1.0 M TEAB/40%MeCN (v/v), flow = 6 ml/min, 0–100% B over 60 min]. Compounds 17 and 18 eluted at 90% B. UV–VIS spectrum: compound 17, 499 nm (0.79A), 555 nm (0.62A); compound 18, 498 nm (0.83A), 595 (0.60A).

General methodology for attachment of modified dyes to alkynylamino-2′,3′-dideoxynucleoside-triphosphates (19,20). Compound 17 or 18 (16.0 µmol) was dissolved in DMF (1 ml) and disuccinimidyl carbonate (8 eq.) added as a solid at room temperature. The reaction was cooled to –60°C, and DMAP (4 eq.) in DMF (0.5 ml) added. The reaction mixture was warmed to –30°C then a solution of aminoalkynl-ddNTP (0.67 eq., Na2CO3/NaHCO3 pH 8.5) added. After stirring for 1 h at room temperature, the reaction mixture was applied directly to a silica gel column. The product was eluted with iPrOH:NH4OH:H2O [4:5:1 (v/v/v)] and evaporated to dryness in vacuo before subsequent purification by ion-exchange chromatography [Q Sepharose, 10 × 100 mm, A = 0.1 M TEAB/40% MeCN (v/v), B = 1.0 M TEAB/40%MeCN (v/v), flow = 6 ml/min, 0–100% B over 60 min] then C18 reverse-phase HPLC (Waters Deltapak C18, 50 × 300 mm, A = 0.1 M TEAB, B = MeCN, 0–100% B over 60 min at 100 ml/min). Compound 19, λmax 499 nm, 555 nm, C91H94N13O32P3S2 observed 2038 (1–), 1019 (M/2); compound 20, λmax 498 nm, 595 nm, C99H102N13O33P3S2 observed 2158 (1–), 1079 (M/2).

The efficiency of incorporation of a synthetic dye-labeled ddNTP relative to that of the unlabeled nucleotide was determined as follows. A reaction mixture containing reaction buffer, 200 ng M13 DNA, 5 pmol universal primer, a thermally stable DNA polymerase, thermally stable inorganic pyrophosphatase, 200 µM each dATP, dCTP, TTP and 1000 µM dITP was prepared on ice and 18 µl aliquots were taken. A dilution series of the nucleotide in question was prepared in 10 mM TrisHCl, pH 8.0, 0.1 mM EDTA. A 2 µl aliquot of each dilution was added separately to the reaction mixture. The samples were then thermally cycled to generate Sanger fragments and prepared for electrophoresis using a typical cycled DNA sequencing protocol. After electrophoresis, the resulting Sanger fragment electropherogram pattern was examined. The ‘reactivity ratio’ of the ddNTP was defined as the ratio of concentration of deoxynucleoside-triphosphate (dNTP) to ddNTP needed to give the highest possible signal strength in each point of the dilution series for the termination product located 600 nt from the primer. In a series of concentrations of terminator the sample with the greatest signal strength at 600 bases would have the optimal ratio of deoxy- to dideoxynucleotide (13). The reactivity ratio of dNTP:ddNTP using either unlabeled dideoxynucleoside-5′-triphosphates or dye-labeled dideoxynucleoside-5′-triphosphates found in DYEnamic ET (10) DNA sequencing kits is approximately 200:1 using DNA polymerases with F to Y modifications (8).

A typical sequencing reaction contained the following reagents: 50 mM Tris (pH 8.5), 5 mM MgCl2, 200 µM dATP, TTP, dCTP, 1000 µM dITP, 20 U thermally stable DNA polymerase and the corresponding dye-labeled nucleotides. Reaction mixtures were thermally cycled using the following conditions: 95°C for 20 s, 50°C for 30 s, 60°C for 2 min (extension times were 1 min for Thermo Sequenase II DNA polymerase). Reactions were held at 4°C until purification and analysis. For precipitated reactions, 2 µl of 7.5 M ammonium acetate and 30 µl of chilled absolute ethanol were added. After mixing and incubating for 20 min at room temperature, the product DNA was precipitated by centrifugation at room temperature for 15 min using a standard laboratory bench microcentrifuge set on high speed (10 000–16 000 g). Supernatants were removed by aspiration and the DNA pellet rinsed with 200 µl of ice-cold 70% ethanol. The tubes were re-centrifuged for 5 min and supernatants removed. DNA pellets were stripped of liquid under vacuum at room temperature for 2 min without taking them to complete dryness. To the reaction was added 6–8 µl of formamide loading dye (95% formamide, 50 mM EDTA, pH 8.0, 0.2 mg/ml new fuchsin) then mixed by vortexing for 10–15 min at room temperature. The resuspended DNA was denatured at 70–72°C for 3 min and the tubes quenched on ice. Aliquots of 2–5 µl were then loaded in wells of sequencing gels for analysis. For directly loaded sequences, an aliquot of the reaction mixture was removed, treated with formamide (1:1 volume) and 2 µl of the resulting mixture analyzed as above.

The initial goal of this research was to determine the number of negative charges required to alter the electrophoretic mobility of the unreacted nucleotides and corresponding breakdown products to a point where fluorescently labeled by-products from a sequencing reaction migrate ahead of sequence. The choice of the sulfonic acid moiety as a negative charge carrier was ideal for our requirements as the functional group required no chemical protection during synthesis and was found to be stable to cycle sequencing reaction conditions. We began by synthesizing a series of fluorescein-labeled 2′,3′-dideoxycytidine nucleoside triphosphates with various numbers of α-sulfo-β-alanine residues in the linker arm (Scheme 1). Electrophoretic analysis of compounds 1, 4 and 5 indicated that the presence of two sulfonic acid moieties in the linker arm (total charge –3; –2 on the linker arm and –1 on fluorescein dye) gave the desired mobility modification (Fig. 1).

Incorporation efficiency (expressed as ‘reactivity ratio’) of nucleotides 1, 4 and 5 was determined using two enzymes, Thermo Sequenase™ II DNA polymerase and a mutant Taq DNA polymerase with a glutamate to arginine substitution at position 681. An interesting correlation between reactivity ratio and number of charges was observed for both polymerases (Fig. 2). Increased negative charge results in decreased incorporation efficiency. It was also observed that the mutant Taq DNA polymerase (E681R) showed less discrimination against the charge-modified nucleotides than Thermo Sequenase II DNA polymerase. This correlates with decreasing the amount of negative charge present on the enzyme.

A poorly incorporated dye terminator is problematic in DNA sequencing. In order to generate the desired read lengths relatively large quantities of dye-labeled terminator are required. This makes complete removal of the terminator more difficult, leading to the presence of strong ‘dye-blobs’ which can obscure 60 or more sequence bases across multiple lanes on a slab gel. In an attempt to improve the efficiency of the negatively charged nucleotides, a second series of compounds with extended linker arms between the nucleobase and the charged moiety were synthesized (Scheme 2). Single-color sequencing experiments were carried out to determine the nucleotide reactivity ratio and it was demonstrated that increasing the length of the linker arm improved incorporation efficiency (Fig. 3).

The synthesis of the remaining three charge-modified dye-labeled nucleotides was undertaken for a four-color sequencing reaction. As fluorescein is negatively charged at the electrophoretic separation pH (17), with rhodamines considered neutral, three negative charges instead of the original two were attached to the following dyes, 5-carboxy-rhodamine-6G (R6G), 5-carboxy-tetramethylrhodamine (TAMRA) and 5-carboxy-rhodamine-X (ROX) as shown in Scheme 3. Addition of three α-sulfo-β-alanine residues was achieved in high yield in a stepwise manner without purification of intermediates. Also, the ‘failure’ products and N + 1 addition products (formed due to the presence of excess amino acid) were readily removed by anion-exchange chromatography.

The choice of dye/nucleotide combination was determined from previous research, which indicated that purines are more readily incorporated than pyrimidines. We decided to attach the dyes (FAM and R6G) with the absorption maximum closest to the argon ion laser excitation of 488 nm to the poorest substrate (pyrimidines), and dyes (TAMRA and ROX) with absorption maximum further away from laser excitation to the better substrate (purines). A four-color sequencing reaction was subsequently set up using optimized concentrations of dye-labeled nucleotides determined from our single-color experiments. The reaction was thermally cycled as described and a 2 µl aliquot of reaction mixture analyzed by electrophoresis. As shown in Figure 1 with single-color experiments, all the unincorporated nucleotides and their corresponding fluorescent breakdown products indeed migrate ahead of the sequence, demonstrating the applicability of this approach to ‘direct-load’ DNA sequencing.

The chemistry developed for the synthesis of charge-modified single dye-labeled nucleotides is also readily applicable to the synthesis of fluorescence resonance energy transfer (FRET)-labeled nucleotides (1012) and a number of dye-pair, charge-modified nucleotides have been synthesized according to the methods developed in this work (Scheme 4). The incorporation efficiency of FRET terminators, 19–20, was found to be the same as for single dye (–3 charge) terminators, 12–13. The enhanced detection sensitivity obtained from energy transfer labels should be particularly useful for direct-load DNA sequencing, as the sample is not concentrated prior to separation. An optimized four-color set of charge-modified, two single-dye terminators (FAM–25-ddCTP and REG–25-ddUTP) and two ET terminators (FAM–TAMRA–18-ddATP and FAM–ROX–18-ddGTP) was used to demonstrate the utility of these terminators in direct-load sequencing. As shown in Figure 4, high quality sequence data free of any dye-blobs was obtained when the sample was directly loaded on to an ABI PRISM™ 377 DNA sequencer. Ethanol precipitation of sequencing reactions using charge-modified terminators and separation on a capillary sequencer (MegaBACE™) yielded satisfactory data. Studies are ongoing to further develop methods for direct injection of sequencing reactions using these terminators on capillary sequencers. We have observed a slight decrease in resolution with directly loaded samples and we are in the process of determining the factors which affect resolution.

An interesting observation arising from our studies demonstrated that the negatively charged dye-labeled nucleotides could possibly be used in conjunction with dGTP in a dye-terminator sequencing reaction. When dGTP is incorporated in a sequencing reaction, highly uniform terminations are observed. However, gel compressions caused by sequence-specific secondary structures formed during the gel electrophoresis when reaction products have dG bases cause some Sanger fragments to migrate faster than similar size fragments which lack the secondary structure (18). This disrupts the order of elution of the Sanger fragments and leads to inaccurate base calling. Hence, replacements for dGTP, such as 7-deaza-dGTP (19) and dITP, are routinely used to alleviate this problem. A series of other dGTP analogs have been synthesized and evaluated as dGTP replacements in our laboratory in an attempt to improve upon the sequencing results obtained using 7-deaza-dGTP and dITP (22). However, these replacements suffer from poor incorporation by DNA polymerases and uneven termination uniformity (13,20,21).

Our studies using charge-modified nucleotides indicated that sequence obtained was free of compression artefacts when dITP was substituted for dGTP. This led us to study dye-terminator reactions using dGTP to determine if compression artefacts were relieved by the presence of extra negative charge in the linker arm. A series of sequencing reactions was performed substituting dGTP (200 µM) for dITP. The samples were run in gels containing only 4 M urea instead of the customary 8 M to create an environment that would encourage the formation of compressions. Figure 5 shows that the control sequence with uncharged ET terminators contains compression artefacts. Sequence obtained with the negatively charged terminator nucleotides shows that the compressions are reduced in magnitude (note that the nucleotides were not properly mobility matched) suggesting that use of these reagents holds promise for dGTP-based dye-terminator DNA sequencing.

We developed a novel set of charge-modified, dye-labeled nucleotides and a DNA polymerase that together increase the convenience and robustness of dye-terminator DNA sequencing. Under optimized conditions it is possible to directly load a sequencing reaction mixture onto a slab gel sequencer and obtain a legible sequence, obviating the need for post-reaction purification. If a post-reaction work-up is required to concentrate the sample then less precision is required as co-precipitated nucleotides and breakdown products will not interfere with sequence determination. Also of interest is the observation that charge-modified nucleotides could possibly be used with dGTP without the formation of compression artefacts. This could lead to dye-terminator sequencing with termination uniformity comparable to dye-primer methods.