Synthesis of Releasable Electrophore Tags for Applications in Mass

A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 0.93 g of IV, 0.30 g of piperidine, 1.0 g of potassium carbo...
0 downloads 0 Views 245KB Size
1002

Bioconjugate Chem. 2002, 13, 1002−1012

Synthesis of Releasable Electrophore Tags for Applications in Mass Spectrometry Xin Zhang, Linxiao Xu,† Poguan Wang, Zhixian Wang,‡ and Roger W. Giese* Department of Pharmaceutical Sciences in the Bouve College of Health Sciences, Barnett Institute, and Chemistry Department, Northeastern University, Boston, Massachusetts 02115. Received July 25, 2001; Revised Manuscript Received November 23, 2001

Releasable electrophore mass tags (electrophore tags) are compounds for use as labels in ligand assays such as hybridization assays and immunoassays. In such assays, the electrophore-tagged reagent (e.g., DNA probe or antibody) is quantified at the conclusion of the assay by cleaving a bond in the attached tag so that the electrophore part can be brought into the gas phase (usually thermally) for detection by electron capture mass spectrometry (EC-MS) or a related technique. Interest in these tags is promoted mainly by their potential to provide highly sensitive and multiplexed assays. The high multiplexing arises from the opportunity to measure many such tags simultaneously in the mass spectrometer, where each tag has an electrophore part with a unique mass. In this study five precursors of electrophore mass tags are presented. Each precursor can lead to a large library of electrophore tags in a practical way, since each precursor can be converted to many different electrophore tags by reaction with commonly available phenols that provide a variation in mass. The phenol-reactive part of the tag is either a polyfluorobiphenyl or a benzyl chloride moiety. Representative library compounds are prepared and detected in an inert ester form by gas chromatography electron capture mass spectrometry (GC-EC-MS). Further, one tag is conjugated to DNA, and the resulting product is detected by laser-induced electron capture time-of-flight mass spectrometry on a silver surface. A calculation by the semiempirical method AM1 for an ion formed by one of the electrophores suggests that ring rotation promotes dissociative electron capture. The features of practical synthesis, simple composition, physicochemical stability, high multiplicity, high sensitivity, and potential for high throughput detection make releasable electrophore mass tags attractive for highly multiplexed assays. This includes their use in SNP assays or dideoxy DNA sequencing for detection of mutations in individuals, where the combination of high accuracy and speed is essential.

INTRODUCTION

A variety of substances such as radioisotopes and fluorescent dyes are employed as molecular tags in assays requiring high sensitivity. Examples of such assays are immunoassays, DNA sequencing, and SNP detection. One shortcoming of many current molecular tags is the limited number of them that can be detected simultaneously. For example, it is difficult to quantify more than four fluorescent tags with high sensitivity under ordinary and practical conditions when present as a mixture. This is because of the broadness of the emission bands along with differences, commonly, in their absorbance characteristics. One consequence of this is that automated DNA sequencing by means of capillary electrophoresis with fluorescence detection is done with only one sequence (or less) loaded into each capillary at a time (1). This barrier to high throughput is unfortunate since DNA sequencing is regarded as the gold standard for detecting mutations. Screening for DNA mutations in a population of people can be successful when some false positive and negative results are present, because the true positive and negative results are typically present in a larger abundance (2). However, as the practice of mutation detection turns to scoring individuals as part of routine diagnostics, high accuracy along with high speed will be essential. Increas* To whom correspondence should be addressed. Tel: 617373-3227; fax: 617-373-8720; e-mail: [email protected]. † Current address: Sigma-RBI, Natick, MA. ‡ Current address: Tropix, Inc., Bedford MA.

ing the multiplexing of dideoxy sequencing assays by employing molecular tags with high multiplicity is one way to accomplish this, for example, with releasable electrophore mass tags. The concept and rationale for using releasable electrophore mass tags (electrophore tags) in chemical analysis is as follows. Each tag contains, effectively, three functional groups: electrophore, release, and reactivity. A tag is attached covalently to an analytical reagent of interest, such as a DNA oligomer or antibody, by means of the reactivity group. Once the assay employing the tagged reagent has reached the detection step, the electrophore part is released by a cleavage reaction at the release group and brought into the gas phase for ionization by electron capture (EC) with subsequent detection based on its mass. Since a mass spectrometer can detect many masses simultaneously, there is a potential for high multiplicity (e.g., 400 electrophore tags) in the assay. While mass spectrometry can detect a moderate number of unlabeled DNA molecules at the same time, large numbers of DNA molecules have never been detected in this way because of complications of fragmentation, adduct ions, isotopic peaks, and loss of signal at higher masses (3). Further, electrophores can be detected in general with much higher sensitivity than DNA molecules by mass spectrometry (4, 5). Previously, we reported the detection of a mixture of four electrophore-labeled DNA oligomers (6), and, in a separate study (7), of albumin labeled with six different

10.1021/bc010072v CCC: $22.00 © 2002 American Chemical Society Published on Web 07/09/2002

Releasable Electrophore Tags for Mass Spectroscopy

electrophores, by electron capture mass spectrometry (EC-MS). While these results were encouraging, the synthetic strategies that we used to prepare the electrophore tags did not open up a practical way to make a large number of them. Here we report our success in overcoming this limitation by designing and preparing five “branchpoint” electrophore tag precursors, each of which is easy to prepare, and reactive toward phenols as the moieties that provide the variation in mass. EXPERIMENTAL SECTION

Materials and Software. 3-Bromo-2-butanone was purchased from Lancaster Inc, Windham, NH. 2,3,4,5,6Pentafluorobenzyl alcohol, 2,3,5,6-tetrafluoro-4-(pentafluorophenyl)phenol, piperazine, pyridine, various acid chlorides, 2,3-O-isopropylidene-D-erythronolactone, ethyl and methyl isonipecotate, decafluorobiphenyl, phosphomolybdic acid, R-bromoacetic acid, thionyl chloride, dicyclohexylcarbodiimide (DCC), 2,2′-azobisisobutyronitrile, Nhydroxysuccinimide (NHS), N-bromosuccinimide, and all phenols were purchased from Aldrich Chemical Co. unless indicated otherwise. Acetonitrile was distilled over CaH2 before use. All solvents were from Fisher Scientific Co. Most reactions were monitored by silica fluorescenceindicator thin-layer chromatography (TLC) with detection by UV (short wavelength) or phosphomolybdic acid. The 1 H, 19F, and 13C spectra were obtained on a Varian 300 MHz spectrometer with tetramethylsilane as internal reference for the proton spectra and CFCl3 for the fluorine spectra. Cation exchange beads in NH4+ form (AG 50wX8, 100-200 mesh 143-5441,) were from Bio-Rad. Hyperchem software was from Hypercube (Waterloo, Ontario). For flash chromatography, the ratio of compound to Si was 1:100 to 1:200. Mass Spectrometry. GC-EC-MS was performed using a Hewlett-Packard (HP) 5973 mass spectrometer coupled to a HP 6890 gas chromatograph. The instrument was controlled by a HP 5973 MS Chemstation computer for data acquisition in the scan mode. An Ultra-1 HP capillary column (25 m, 0.20 mm i.d., 0.11 µm film thickness) from Hewlett-Packard was used. With helium as carrier gas, the column head pressure was set to 36 psi. The source pressure (methane) was set to 2.0 Torr at 250 °C. Inject (1 µL sample of 100 pg/µL in ethyl acetate) on-column at 50 °C; ramp column oven to 250 C at 30 °C/min after a 4 min solvent delay; then ramp to 300 °C at 50 °C/min and hold for 12 min. MALDI-TOF-MS experiments were done using a Bruker Daltonics Proflex laser desorption TOF-MS instrument in the linear positive mode. A literature procedure (8) was followed with modifications. To 1 µL of oligonucleotide (0.5-1.0 µg/µL water) was added 2 µL of ammonium ion exchange bead suspension. After the suspension was mixed (rocking plate) for 3 min, 3 µL of 70 mg/mL, 3-hydroxypicolinic acid:picolinic acid (9:1 w/w) in 50% acetonitrile was added, and, after brief mixing and settling, 1 µL of the supernatant was placed onto the target and allowed to dry. LI-EC(Ag)-TOF-MS experiments were carried out as described (9). Synthesis. Ethyl N-(4-Hydroxy-2,3-O-isopropylidenebutyryl)isonipecotate (2). A 25 mL dry roundbottomed flask equipped with a N2 inlet and a stir bar was charged with 1.0 g (6.3 mmol) of 2,3-O-isopropylidene-D-erythronolactone (1), followed by 2.0 mL of ethyl isonipecotate. The resulting mixture was allowed to stir at 70 °C (oil bath) until TLC analysis showed no lactone remained (about 12-16 h). The reaction mixture was

Bioconjugate Chem., Vol. 13, No. 5, 2002 1003

then purified on a silica gel column (1:1 ethyl acetate: hexane, followed by 100% ethyl acetate), and 1.5 g (75%) of desired product was obtained. 13C NMR (CDCl3): 13.78, 24.97, 25.07, 26.78, 26.91, 27.18, 27.33, 27.86, 40.45, 40.41, 41.11, 41.21, 44.22, 44.37, 60.20, 61.33, 61.41, 74.31, 74.46, 77.35, 77.47, 109.10, 109.13, 166.11, 165.99, 173.49 ppm. 1H NMR(CDCl3): 1.23-1.29 (m, 3H), 1.40 (s, 3H), 1.55 (s, 3H), 1.6-2.05 (m, 4H), 2.51-2.64 (m, 1H), 2.71-3.31 (m, 2H), 3.56-3.77 (m, 2H), 3.954.17 (m, 3H), 4.29-4.45 (m, 2H), 4.91 (dd, J ) 3.5 Hz, 1H) ppm. TLC: Rf ) 0.40 (100% ethyl acetate). The methyl derivative (2a, 1.2 g, 60%) was prepared in a similar way. 13C NMR (CDCl3): 25.28, 25.36, 27.12, 27.26, 27.54, 27.65, 28.24, 28.31, 40.39, 40.70, 41.56, 41.67, 44.64, 44.79, 51.79, 61.93, 62.01, 75.01, 75.05, 77.98, 77.04, 109.532, 166.47, 166.60, 174.26 ppm. 1H NMR (CDCl3): 1.34 (s, 3H), 1.48 (s, 3H), 1.54-1.80 (m, 2H), 1.86-1.96 (m, 2H), 2.47-2.60 (m, 1H), 2.71-3.24 (m, 2H), 3.48-3.72 (m, 5H), 3.85-3.95 (m, 1H), 4.204.4 (m, 2H), 4.85 (m, 1H) ppm. TLC: Rf ) 0.40 (100% ethyl acetate). Ethyl N-[2,3-O-Isopropylidene-4-(4′-pentafluorophenyl-2′,3′,5′,6′-tetrafluorophenoxy)butyryl]isonipecotate (I). A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 0.67 g (2 mmol) of 2, followed by 0.7 g (2.2 mmol) of decafluorobiphenyl, 0.5 g (3 mmol) of CsOH monohydrate, 10 mL of CH3CN, and 1.5 g of powdered molecular sieves. The resulting mixture was allowed to stir at room temperature until TLC analysis showed no starting material remained (about 16-24 h). The solid was removed by filtration, and the filtrate was concentrated on rotary evaporator and further purified on a silica gel column (1:5 ethyl acetate:hexanes, 300 mL, followed by 1:2 ethyl acetate:hexanes) and 0.62 g (49%) of desired product was obtained. 13C NMR (CDCl3): 14.13, 14.17, 25.41, 25.57, 27.22, 27.45, 27.57, 27.79, 28.29, 40.44, 41.11, 41.49, 41.68, 44.46, 44.61, 60.69, 73.33, 73.40 (m), 73.80 (m), 74.32, 74.53, 75.74, 76.11, 110.46 (m), 136 (m), 139 (m), 143 (m), 146 (m), 165.20, 165.30, 173.84, 173.87 ppm. 1H NMR (CDCl3): 5.01 (d, J ) 5.7 Hz, 1H), 4.60-4.71 (m, 1H), 3.98-4.56 (m, 6H), 2.48-3.32 (m, 3H), 1.60-2.00 (m, 4H), 1.54 (s, 3H), 1.42 (s, 3H), 1.15-1.28 (m, 3H) ppm. 19F NMR (CDCl ): -137.95 (m, 2F), -139.83 (m, 2F), 3 -151.24 (m, 1F), -155.95 (m, 1F), -156.43 (m, 1F), -161.27 (m, 2F) ppm. TLC: Rf ) 0.50 (1:1 ethyl acetate: hexanes). Ethyl N-{{4-{4′-[4′′-(p-Acetophenyl)-2′′,3′′,5′′,6′′-tetrafluorophenyl]-2′,3′,5′,6′-tetrafluorophenoxy}-2,3O-isopropylidenebutyryl}}isonipecotate (I-447a). A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 0.40 g (0.64 mmol) of I, followed by 0.15 g (1 mmol) of 4-hydroxyacetophenone, 0.7 g K2CO3, and 10 mL of CH3CN. The resulting mixture was allowed to stir at 60 °C oil bath overnight. The solid was removed by filtration and was washed with CH3CN (3 × 10 mL). The combined organic solution was concentrated and purified on a silica gel column (1:5 ethyl acetate:hexanes, 500 mL, followed by 1:2 ethyl acetate: hexanes) to give 0.25 g (41%) product. 13C NMR (CDCl3): 14.12, 14.15, 25.39, 25.55, 26.50, 27.21, 27.43, 27.54, 27.76, 28.28, 29.67, 40.41, 41.08, 41.46, 41.66, 44.45, 44.60, 60.66, 73.40 (m), 73.84 (m), 74.30, 74.54, 75.74, 76.08, 110.37, 110.43, 115.37, 115.42, 130.68, 133.13, 136 (m), 139 (m), 143 (m), 146 (m), 160.01, 165.20, 165.29, 173.86, 173.81, 196.29 ppm. 1H NMR (CDCl3): 8.01 (dm, J ) 8.9 Hz, 2H), 7.10 (dm, J ) 8.8 Hz, 2H), 5.01 (d, J ) 6.8 Hz, 1H), 4.68-4.77 (m, 1H), 4.00-4.59 (m, 6H), 2.523.32 (m, 3H), 2.60 (s, 3H), 1.60-2.03 (m, 4H), 1.56 (s,

1004 Bioconjugate Chem., Vol. 13, No. 5, 2002

3H), 1.44 (s, 3H), 1.26 (td, J ) 7.1 Hz, J ) 3.1 Hz, 3H) ppm. 19F NMR (CDCl3): -137.92 (m, 2F), -139.71 (m, 2F), -153.33 (m, 2F), -155.93 (m, 1F), -156.38 (m, 1F) ppm. TLC: Rf ) 0.45 (1:1 ethyl acetate:hexanes). Ethyl N-{{2,3-O-Isopropylidene-4{4′-[[4′′-[p-(3′′′oxobutyl)phenoxy]-2′′,3′′,5′′,6′′-tetrafluorophenyl]]-2′,3′,5′,6′-tetrafluorophenoxy}butyryl}}isonipecotate (I-475a). Similarly, the reaction of I with 4-(hydroxyphenyl)-2-butanone gave 0.35 g (55%) of product. 1 H NMR (CDCl3): 7.18 (d, J ) 8.5 Hz, 2H), 6.97 (d, J ) 8.6 Hz, 2H), 5.04 (d, J ) 6.1 Hz, 1H), 4.69-4.77 (m, 1H), 4.02-4.59 (m, 4H), 3.69 (m, 3H), 2.52-3.32 (m, 7H), 2.15 (s, 3H), 1.61-2.14 (m, 4H), 1.56 (s, 3H), 1.40 (s, 3H) ppm. 19F NMR (CDCl ): -138.77 (m, 2F),-139.75 (m, 2F), 3 -153.95 (m, 1F), -156.10 (m, 1F), -156.50 (m, 1F) ppm. TLC: Rf ) 0.25 (1:1 ethyl acetate:hexanes). Methyl N-(Bromoacetyl)isonipecotate (4). A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 1.30 g (10 mmol) of bromoacetic acid (3), followed by 1.38 g (12 mmol) of Nhydroxysuccinimide, and the flask was placed in an icebath. After 30 min of stirring, 15 mL of acetonitrile containing 3.68 g (12 mmol) of DCC was added dropwise, and the reaction mixture was allowed to stir overnight before addition of 1.43 g (10 mmol) of methyl isonipecotate. After the reaction mixture was stirred another 24 h, it was filtered and the filtrate was concentrated on an evaporator, yielding residue that was redissolved into 200 mL of ethyl acetate. The resulting solution was washed with 3 × 30 mL of H2O and 30 mL of brine solution and dried over MgSO4. The organic solvent was removed on a rotary evaporator, and the crude product 4 was dried further under vacuum (0.05 mmHg) for 16 h and used without further purification. Methyl N-[r-(4-(Pentafluorophenyl)-2,3,5,6-tetrafluorophenoxy)acetyl]isonipecotate (II). A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 1.1 g of 4-pentafluorophenyl2,3,5,6-tetrafluorophenol, 2.0 g of crude 4, 1.5 g of potassium carbonate, and 20 mL of acetonitrile. The resulting mixture was allowed to stir at 60 °C (oil bath) until TLC analysis showed no phenol remained (about 12-16 h). The K2CO3 was removed by filtration and washed several times with 3 × 10 mL of acetonitrile. The combined organic solution was concentrated on rotary evaporator, and the residue was purified on a silica gel column (1:5 ethyl acetate:hexanes, followed by 1:2 ethyl acetate:hexanes), to give 1.1 g (65%) of product. 13C NMR (CDCl3): 28.05, 27.48, 40.39, 41.24, 43.85, 51.74, 70.38 (m), 136-146 (m), 164.63, 174.08 ppm. 1H NMR (CDCl3): 1.62-1.78 (m, 2H), 1.91-1.97 (m, 2H), 2.512.60 (m, 1H), 2.89 (td, J ) 10.3 Hz, 1H), 3.16 (tm, J ) 11.3 Hz, 1H), 3.63 (s, 3H), 3.63-3.71 (m, 1H), 4.25 (dt, J ) 13.5 Hz, 1H), 5.01 (s, 2H) ppm. 19F NMR (CDCl3): -137.80 (m, 2F), -139.94 (m, 2F), -151.48 (m, 1F), -156.75 (m, 2F), -161.70 (m, 2F) ppm. TLC: Rf ) 0.40 (1:1 hexanes: ethyl acetate). Methyl N-{{{r-{{4-{4′-[p-(3′′-Oxobutyl)phenoxy]2′,3′,5′,6′-tetrafluorophenyl}-2,3,5,6-tetrafluorophenoxy}}acetyl}}}isonipecotate (II-475a). A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 0.12 g (0.23 mmol) of II, followed by 0.10 g of 4-(4-hydroxyphenyl)-2-butanone, 0.26 g K2CO3, and 3 mL of CH3CN. The resulting mixture was allowed to stir at 60 °C overnight. The K2CO3 was removed by filtration and washed with CH3CN. The combined organic solution was concentrated and purified on a silica gel column (1:5 ethyl acetate:hexanes, followed by 1:2 ethyl acetate:hexanes) to yield 0.15 g (96%) of

Zhang et al.

desired product. 13C NMR (CDCl3): 27.60, 28.19, 28.83, 30.13, 40.51, 41.39, 44.05, 45.10, 51.96, 70.73 (t, J ) 4.1 Hz), 115.83, 129.61, 136-146 (m), 155.30, 164.63, 174.15, 207.67 ppm. 1H NMR (CDCl3): 1.69-1.78 (m, 2H), 1.912.03 (m, 2H), 2.15 (s, 3H), 2.60 (m, 1H), 2.73-3.01 (m, 5H), 3.17-3.26 (m, 1H), 3.72-3.79 (m, 1H), 3.72 (s, 3H), 4.34 (dt, J ) 13.9 Hz, 1H), 5.05 (s, 2H), 6.95 (dm, J ) 8.5 Hz, 2H), 7.17 (dm, J ) 8.7 Hz, 2H) ppm. 19F NMR (CDCl3): -138.60 (m, 2F), -139.70 (m, 2F), -154.00 (m, 2F), -156.9 (m, 2F) ppm. TLC: Rf ) 0.40 (1:1 hexanes: ethyl acetate). Electrophore II-861a was Synthesized in a Similar Way (0.1 g, 51% yield). 1H NMR (CDCl3): 7.557.58 (m, 2H), 7.29-7.32 (m, 2H), 7.15-7.18 (m, 2H), 6.93-6.96 (m, 2H), 6.04-6.08 (m, 1H), 4.98 (s, 2H), 4.254.30 (m, H), 3.60-3.71 (m, 6H), 3.49 (s, 2H), 3.11-3.18 (m, H), 2.84-2.93 (m, 3H), 2.51-2.58 (m, H), 2.25-2.30 (m, 4H), 1.90-1.98 (m, 2H), 1.61-1.70 (m, 2H), 1.171.53 (m, 32H), 0.80 (m, 6H). 19F NMR (CFCl3): -138.4 (m, 2F), -139.7 (m, 2F), -153.9 (m, 2F), -156.7 (m, 2F). Electrophore II-981a was Synthesized in a Similar Way (0.11 g, 53%). 1H NMR (CDCl3): 7.72 (m, 2H), 7.48 (m, 2H), 7.31 (m, 2H), 7.24 (m, 2H), 7.01 (m, 2H), 6.93 (m, 2H), 6.21 (t, J ) 5.6 Hz, 1H), 5.12 (s, 2H), 5.05 (s, 2H), 4.34 (m, 1H), 3.73 (m, 6H), 3.21-3.47 (m, 5H), 2.94 (t, J ) 7.0 Hz, 2H), 2.96-3.00 (m, 1H), 2.59-2.62 (m, 1H), 2.03 (m, 2H), 1.25-1.81 (m, 34H), 0.87 (t, J ) 6.1 Hz, 6H). 19F NMR (CFCl3): -138.42 (m, 2F), -139.67 (m, 2F), -153.82 (m, 2F), -156.66 (m, 2F). Acid Form of II-475a. II-475a (145 mg) was dissolved in 3.5 mL of tBuOH/H2O (2:1), and 35 mg of LiOH was added. The resulting solution was stirred at room temperature until no starting material remained by TLC analysis (usually about 10-20 min). The solution was then quenched in 20 mL of ice H2O, acidified to pH 4-5 with 5% HCl and extracted with 5 × 10 mL of ethyl acetate. The combined organic solution was washed with 2 × 10 mL of H2O and 10 mL brine and then dried over MgSO4 or Na2SO4. The organic solvent was removed by the rotary evaporator, giving a residue was dried further at 0.05 mmHg for several hours (135 mg, 95%) before storing in a desiccator for further use. N-Hydroxysuccinimide Ester of II-475a. A 10 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 50 mg of the above acid form of II-475a, followed by 15 mg of NHS, 3 mL of dry CH3CN, and 23 mg of EDC. The resulting mixture was stirred at room-temperature overnight and then was quenched with 40 mL of ice-water. The resulting solution was extracted with 5 × 10 mL of ethyl acetate, and the combined organic solution was washed with 2 × 10 mL of ice-cold water and 10 mL of ice-cold brine and dried over MgSO4. Rotary evaporation gave a residue that was dried at (0.05 mmHg) for several hours, giving 52 mg (95%) of product, Rf ) 0.5 (100% ethyl acetate), that was stored in a desiccator. 3-(4′-(Pentafluorophenyl)-2′,3′,5′,6′-tetrafluorophenoxy)-2-butanone (III). A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 1.0 g (3 mmol) of 2,3,5,6-tetrafluoro-4-(pentafluorophenyl)phenol, 0.55 g of 3-bromo-2-butanone (5), 1.0 g of potassium carbonate, and 20 mL of acetonitrile. The resulting mixture was allowed to stir at 60 °C until TLC showed no phenol remained (about 12-16 h). The K2CO3 was moved via filtration and washed with 3 × 10 mL acetonitrile. The organic solvent was removed by rotary evaporation, and the residue was recrystallized from hexanes/diethyl ether to give 1.08 g (90%) of the desired product. 1H NMR (CDCl3): 4.94 (q, J ) 6.9 Hz, 1H), 2.37

Releasable Electrophore Tags for Mass Spectroscopy

(s, 3H), 1.60 (d, J ) 7.0 Hz, 3H) ppm. 19F NMR (CDCl3): -137.92 (m, 2F), -139.43 (m, 2F), -151.12 (m, 1F), -155.95 (m, 2F), -161.30 (m, 2F) ppm. Rf ) 0.50 (ethyl acetate:hexanes, 1:9). General Procedure for the Preparation of Phenol Derivatives of Compound (III). A 25 mL dry roundbottomed flask equipped with a N2 inlet and a stir bar was charged with 0.41 g (1 mmol) of 3-[2′,3′,5′,6′-tetrafluoro-4′-(pentafluorophenyl)phenoxy]-2-butanone, 1.1 mmol of a selected phenol (see below), 0.5 g of potassium carbonate, and 20 mL of acetonitrile. The resulting mixture was allowed to stir at 60 °C until TLC analysis showed no starting material remained (about 12-16 h). The K2CO3 was removed via filtration and washed with 3 × 10 mL of acetonitrile. The combined organic solution was concentrated on a rotavapor giving a residue that was further purified on a silica gel column to give the product. 3-{4′-[4′′-(p-Iodophenoxy)-2′′,3′′,5′′,6′′-tetrafluorophenyl]-2′,3′,5′,6′-tetrafluorophenoxy}-2-butanone (III-531a). By the above general procedure, the reaction of 4-iodophenol with III gave 0.51 g (85%) of product. 1H NMR (CDCl3): 7.66 (dm, J ) 9.2 Hz, 2H), 6.84 (dm, J ) 9.3 Hz, 2H), 4.90 (qm, J ) 6.8 Hz, 1H), 2.37 (s, 3H), 1.60 (d, J ) 7.6 Hz, 3H) ppm. 19F NMR (CDCl3): -138.19 (m, 2F), -139.25 (m, 2F), -153.5 (m, 2F), -155.9 (m, 2F) ppm. 3-{{4′-{4′′-[o,p-Bis(r,r-dimethylbenzyl)phenoxy]2′′,3′′,5′′,6′′-tetrafluorophenyl}-2′,3′,5′,6-tetrafluorophenoxy}}-2-butanone (III-641a). By the above general procedure, the reaction of 2,4-bis(R,R-dimethylbenzyl)phenol with III yielded 0.80 g (87%) product. 1H NMR (CDCl3): 7.03-7.36 (m, 11H), 4.47 (d, J ) 2.3 Hz, 1H), 6.56 (d, J ) 8.3 Hz, 1H), 4.88 (q, J ) 6.7 Hz, 1H), 2.37 (s, 3H), 1.76 (s, 6H), 1.74 (s, 6H), 1.56 (d, J ) 7.0 Hz, 3H) ppm. 19F NMR (CDCl3): -139.40 (m, 2F), -139.76 (m, 2F), -154.17 (m, 2F), -156.19 (m, 2F) ppm. N,N-Didecyl 4-{4′-[4′′-(r-Methyl-β-oxopropyloxy)2′′,3′′,5′′,6′′-tetrafluorophenyl]-2′,3′,5′,6′-tetrafluorophenoxy}benzamide (III-728a). By the above general procedure, the reaction of N,N-didecyl 4-hydroxybenzamide with III gave 0.28 g (49%) major product. 1H NMR (CDCl3): 7.39 (dm, J ) 8.4 Hz, 2H), 7.06 (dm, J ) 8.5 Hz, 2H), 4.94 (q, J ) 7.4 Hz, 1H), 3.21-3.47 (m, 4H), 2.37 (s, 3H), 1.61-1.66 (m, 32H), 1.60 (d, J ) 6.9 Hz, 3H), 0.87 (t, J ) 6.6 Hz, 6H) ppm. 19F NMR (CDCl3): -138.27 (m, 2F), -139.29 (m, 2F), -153.39 (m, 2F), -156.02 (m, 2F) ppm. Also a minor product (III-848a) was obtained. 1H NMR (CDCl3): 8.25 (dm, J ) 9.3 Hz, 2H), 7.44 (dm, J ) 8.5 Hz, 2H), 7.25 (dm, J ) 8.8 Hz, 2H), 7.15 (dm, J ) 8.5 Hz, 2H), 4.93 (q, J ) 7.2 Hz, 1H), 3.11-3.48 (m, 4H), 2.38 (s, 3H), 1.61 (d, J ) 6.5 Hz, 3H), 1.16-1.66 (m, 32H), 0.88 (m, 6H) ppm 19F NMR (CDCl3): -137.68 (m, 2F), -139.15 (m, 2F), -153.10 (m, 2F), -155.82 (m, 2F) ppm. 4′-(Bromomethyl)acetophenone (7). A 50 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 30 mL of CCl4, 2.0 g of NBS, 1.34 g of 4-methylacetophenone (6), and a catalytic amount of 2,2′-azobisisobutyronitrile. The resulting solution was refluxed for 8 h and was monitored by TLC during reflux. After the solvent was removed by rotary evaporation, the residue was quenched with 30 mL of ice-water and extracted with 3 × 30 mL of diethyl ether. The combined organic solution was washed with 2 × 30 mL of water and dried over MgSO4. After removal of solvent by rotary evaporation, the residue was purified further by silica column chromatography (methylene chloride:hexanes, 1:3) to give 1.06 g (50%) of product. 1H NMR (CDCl3):

Bioconjugate Chem., Vol. 13, No. 5, 2002 1005

7.93-7.96 (m, 2H), 7.48-7.5 (m, 2H), 4.51 (s, 2H), 2.61 (s, 3H) ppm. p-(4′-(Pentafluorophenyl)-2′,3′,5′,6′-tetrafluorophenoxy)methylacetophenone (IV). A 100 mL dry roundbottomed flask equipped with a N2 inlet and a stir bar was charged with 2.73 g of 2,3,5,6-tetrafluoro-4-(pentafluorophenyl)phenol, 1.75 g of 7, 2.0 g of potassium carbonate, 30 mL of THF: acetonitrile (1:1), and 0.1 g of potassium iodide. The resulting mixture was allowed to stir at 60 °C until TLC analysis showed no phenol remained (about 12-16 h). The K2CO3 was removed by filtration and washed with acetonitrile 3 × 10 mL. The combined organic solution was concentrated on a rotary evaporator, and the residue was recrystallized from hexanes/diethyl ether to give 2.93 g (77%) of product. (Further recrystallization could raise the combined yield to 90%.) 13C NMR (CDCl3): 197.53, 146.27 (m), 142.92 (m), 140.37, 140.36 (m), 138.60 (m), 137.32, 128.75, 127.80, 75.54 (t, J ) 4.0 Hz), 26.51 ppm. 1H NMR (CDCl3): 8.00-8.03 (m, 2H), 7.56-7.59 (m, 2H), 5.42 (s, 2H), 2.63 (s, 3H) ppm. 19F NMR (CDCl3): -137.99 (m, 2F), -139.55 (m, 2F), -151.01 (m, 1F), -155.77 (m, 2F), -161.18 (m, 2F) ppm. p-[4′-(4′′-Piperidinyl-2′′,3′′,5′′,6′′-tetrafluorophenyl)2′,3′,5′6′-tetrafluorophenoxy]methylacetophenone (IV-396a). A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 0.93 g of IV, 0.30 g of piperidine, 1.0 g of potassium carbonate, and 15 mL of acetonitrile. The resulting mixture was allowed to stir at 60 °C until TLC analysis showed no starting material (about 24 to 48 h). The K2CO3 was removed filtrated and washed several times with 3 × 10 mL acetonitrile. The combined concentration on a rotary evaporation, and the residue was purified further on a silica gel column to give 0.53 g (50%) of product. 1H NMR (CDCl3): 7.99-8.02 (m, 2H), 7.56-7.59 (m, 2H), 5.39 (s, 2H), 3.27 (m, 4H), 2.63 (s, 3H), 1.68 (m, 6H) ppm. 19F NMR (CDCl3): -139.80 (m, 2F), -141.50 (m, 2F), -151.55 (m, 2F), -156.50 (m, 2F) ppm. p-[4′-(4′′-Piperazinyl-2′′,3′′,5′′,6′′-tetrafluorophenyl)2′,3′,5′6′-tetrafluorophenoxy] methyl acetophenone (IV-397a). A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 0.93 g of IV, 0.30 g of anhydrous piperazine,1.0 g of potassium carbonate, and 15 mL of acetonitrile. The resulting mixture was allowed to stir at 60 °C until TLC analysis showed no starting material left (about 24-48 h). After filtration, washing, and concentration, the residue was redissolved in 200 mL of ethyl acetate, and this solution was washed with 3 × 30 mL of water and 30 mL of brine and dried over MgSO4. The solvent was removed by rotary evaporation, and the crude product was further dried under vacuum for several hours prior to the following reaction. (Note: this key intermediate solution could be stored at 4 °C for 7 days without significant decomposition.) p-{4′-[4′′-(4′′′-Acetyl-1′′′-piperazinyl)-2′′,3′′,5′′,6′′-tetrafluorophenyl]-2′,3′,5′,6′-tetrafluorophenoxy}methylacetophenone (IV-439a). A 25 mL dry round-bottomed flask equipped with a N2 inlet and a stir bar was charged with 1.0 g of above crude material and 15 mL of dry THF. The solution was cooled to in an ice-bath for 0.5 h prior to addition of 1.0 mL of acetic chloride dropwise, followed by 1.0 mL of pyridine. The resulting mixture with stirring was allowed to warm to roomtemperature overnight. The reaction mixture was quenched with 30 mL of ice-water and neutralized with 0.5% HCl. This mixture was extracted with ethyl acetate (5 × 20 mL). The combined organic solution was washed with 3

1006 Bioconjugate Chem., Vol. 13, No. 5, 2002

Zhang et al.

Figure 1. Synthesis of gateway electrophores I-V. a: Ethyl isonipecotate, 70-80 °C oil bath, 24 h, 80% yield. b: Decafluorobiphenyl, CsOH, molecular sieves, CH3CN or DMF, room temperature, 48 h, 49% yield. c: N-Hydroxysuccinimide, DCC, CH3CN, room temperature, overnight; then add methyl isonipecotate, 24 h, then evaporate and use crude material in next reaction. d: 2,3,5,6Tetrafluoro-4-(pentafluorophenyl)phenol, K2CO3, CH3CN, 60 °C, 24 h, 90% yield. e: 2,3,5,6-Tetrafluoro-4-(pentafluorophenyl)phenol, K2CO3, CH3CN 60°, 24 h, 90% yield. f: NBS, CCl4, reflux for 16 h, 50% yield. g: 2,3,5,6-Tetrafluoro-4-(pentafluorophenyl)phenol, K2CO3,CH3CN, 60 °C, 24 h, 90% yield. h: Ac2O, pyridine, CH2Cl2, reflux for 3 h, 91% yield. i: 4′-Hydroxyacetophenone, K2CO3, toluene, 18-crown ether (cat.), reflux for 16 h, 81% yield. j: 20% NaOH (aq), methanol, room-temperature overnight, 95% yield. k: SOCl2, chloroform, reflux for 6 h, 92% yield.

Figure 2. General scheme for conversion of electrophores I-V to releasable electrophore mass tags.

× 30 mL of water and 30 mL of brine solution and dried over MgSO4. After removal of solvent by rotary evaporation, the residue was further purified on a silica gel chromatography column to give 0.72 g (63%) of product. 13C NMR (CDCl ): 197.86, 169.40, 140.80 (m), 137.49 (m), 3 128.92 (m), 128.1, 75.75 (m), 50.98 (m), 47.04, 42.13, 26.93, 21.61 ppm. 1H NMR (CDCl3): 7.99-8.02 (m, 2H), 7.56-7.59 (m, 2H), 5.40 (s, 2H), 3.75-3.79 (m, 2H), 3.603.64 (m, 2H), 3.31-3.36 (m, 4H), 2.63 (s, 3H), 2.16 (s,

3H) ppm. 19F NMR (CDCl3): -139.77 (m, 2F), -140.31 (m, 2F), -150.90 (m, 2F), -156.22 (m, 2F) ppm. p-{4′-(4′′-(4′′′-Benzoyl-1′′′-piperazinyl)-2′′,3′′,5′′,6′′tetrafluorophenyl]-2′,3′,5′6′-tetrafluorophenoxy}methylacetophenone (IV-501a). Similarly, the crude product from the reaction of 0.5 g of with 0.2 g of piperazine was reacted with 0.6 mL of benzoyl chloride to produce 0.28 g (53%) desired product. 1H NMR (CDCl3): 7.99-8.02 (m, 2H), 7.56-7.58 (m, 2H), 7.45 (s, 5H), 5.39 (s, 2H), 3.31-3.94 (m, 8H), 2.63 (s, 3H) ppm; 19F NMR (CDCl ): -139.79 (m, 2F), -140.28 (m, 2F), 3 -150.90 (m, 2F), -156.20 (m, 2F) ppm. p-{{4′-{4′′-[[4′′′-[3′′′′,5′′′′-Di(trifluoromethyl)benzyl]1′′′-piperazinyl]]-2′′,3′′,5′′,6′′-tetrafluorophenyl}2′,3′,5′6′-tetrafluorophenoxy}}methylacetophenone (IV-623a). The crude product of the reaction of 0.5 g of IV with 0.2 g of piperazine was reacted with 0.2 g of 3,5-ditrifluoromethylbenzyl bromide in the presence of catalytic amount of KI for 24 h to give the product after preparative TLC purification. 1H NMR (CDCl3): 7.998.02 (m, 2H), 7.85 (s, 2H), 7.79 (s, 1H), 7.56-7.59 (m, 2H), 5.39 (s, 2H), 3.68 (s, 2H), 3.40 (m, 4H), 2.60-2.64 (m, 7H) ppm. 19F NMR (CDCl3): -63.27 (s, 6F), -139.81 (m, 2F), -140.86 (m, 2F), -151.28 (m, 2F), -156.35 (m, 2F) ppm. 2,3,4,5,6-Pentafluorobenzyl Acetate (9). To a mixture of 1 g (5 mmol) of 2,3,4,5,6-pentafluorobenzyl alcohol (8) and 1 mL of pyridine in 50 mL of anhydrous methylene chloride was added 0.6 g (5.8 mmol) of acetic anhydride. After being refluxed for 3 h, the mixture was cooled to room temperature, and then 10 mL of water

Releasable Electrophore Tags for Mass Spectroscopy

Bioconjugate Chem., Vol. 13, No. 5, 2002 1007

Figure 3. Mass spectra of electrophores by GC-EC-MS. Sample injection: 1 µL injection, 100 pg/µL in ethyl acetate.

was added. The aqueous phase was extracted with methylene chloride three times, and the combined organic solution was washed with 10 mL of brine and then dried over magnesium sulfate. After removal of the solvent on a rotavapor, 1.1 g (91%) of product was obtained as a colorless oil. 4′-(4′′-Acetoxymethyl-2′′,3′′,5′′,6′′-tetrafluorophenoxy)acetophenone (10). A mixture of 1 g (4.16 mmol) of 9, 0.57 g (4.18 mmol) of 4′-hydroxyacetophenone, 5 g of potassium carbonate, and 50 mg of 18-crown-ether in 50 mL of anhydrous toluene was refluxed for 16 h. After cooling to room temperature, the reaction mixture was

filtered through a Celite bed. The filtrate was concentrated, and the residual solid was loaded on a silica gel column and eluted with 5:1 hexanes/ethyl acetate to afford 1.2 g (81%) of product as a white solid. 4′-(4′′-Hydroxymethyl-2′′,3′′,5′′,6′′-tetrafluorophenoxy)acetophenone (11). To a solution of 1 g (2.8 mmol) of 10 in 50 mL of methanol was added dropwise 10 mL of 20% aqueous sodium hydroxide. The mixture was then stirred at room-temperature overnight, neutralized to pH 7, and concentrated to about 20 mL to remove most of methanol. The aqueous solution was extracted with ethyl acetate (3 × 50 mL). The combined organic solution was

1008 Bioconjugate Chem., Vol. 13, No. 5, 2002

Zhang et al.

Figure 4. Detection of electrophores by LI-EC(Ag)-TOF-MS: 100 pmol on the target, 10 laser shots.

washed with water and brine and then dried over magnesium sulfate. After removal of the solvent, 0.84 (95%) of product was obtained as a white solid. 4′-(4′′-Chloromethyl-2′′,3′′,5′′,6′′-tetrafluorophenoxy)acetophenone (V). A mixture of 0.85 g (2.7 mmol) of 11 and 5 mL of thionyl chloride in 30 mL of chloroform was refluxed for 6 h. The solvent was removed on a rotavapor, and the crude product was further purified by chromatography on silica gel to afford 0.83 g (92%) of product. 4′-(4′′-Phenoxymethyl-2′′,3′′,5′′,6′′-tetrafluorophenoxy)acetophenone (V-93a). A mixture of 0.83 g (2.5 mmol) of V, 0.3 g (3.2 mmol) of phenol, and 2 g of potassium carbonate in 30 mL of acetone was stirred under reflux for 5 h. After being cooled to room temperature, the mixture was filtered through a Celite bed. The filtrate was concentrated and subjected to chromatography on silica gel with 6:1 hexanes/ethyl acetate to afford 0.95 g (97%) of product as a white solid. 1H NMR (CDCl3): 7.97-8.00 (m, 2H), 7.32-7.37 (m, 2H), 7.007.06 (m, 5H), 5.18 (s, 2H), 2.58 (s, 3H) ppm. 19F NMR (CDCl3): -142.66 (m, 2F), -154.10 (m, 2F) ppm.

RESULTS AND DISCUSSION

Shown in Figure 1 are five branchpoint precursors (I-V) of electrophore tags, along with their synthetic pathways. Each compound represents a point in a branching synthetic pathway that can lead to many releasable electrophore mass tags. Nucleophilic reactions with phenols can be employed at this branch point in all cases, involving substitution of a p-fluorine atom of I-IV, and of the chlorine atom of V, as summarized in Figure 2. Overall this can provide, as explained in more detail later, to two signal libraries, one shared in common by I-IV, and a second library from V. As seen in Figure 1, precursor tag I is established by a synthetic pathway that consists of just two reactions, and all of the reactants are commercially available and inexpensive, including the starting compound, 2,3-Oisopropylidine D-erythronolactone. The isolated yield of I from compound 2 (reaction b) is 49%, and the overall yield of I from 1 is 37%. Considerable effort went into the development of reaction conditions b (CsOH monohydrate, molecular sieves, acetonitrile). Many other conditions, involving a variety of bases, solvents, and temperatures, gave a lower

Releasable Electrophore Tags for Mass Spectroscopy

Bioconjugate Chem., Vol. 13, No. 5, 2002 1009

Figure 5. A: Detection of a mixture of six electrophores by LI-EC(Ag)-TOF-MS: 100 fmol of each on the target, 10 laser shots. B: LI-EC(Ag)-TOF-MS of blank silver surface under same condition as in A.

yield of I. Since the reactants in this sequence are commercially available and inexpensive, and I is easily purified from the side products, the synthetic procedure to form I is attractive. Compound I can be converted into electrophore mass tags by a general reaction sequence consisting of the following series of steps: replace the para fluorine with a phenol; acid-hydrolyze the isopropylidine moiety to a corresponding glycol; base-hydrolyze the ester to a corresponding acid; and convert the latter to an N-hydroxy-

succcinimide (NHS) active ester (see Figure 2). Thus, the ensuing reactivity, release, and signal groups in the final release tags derived from I are NHS ester, β-hydroxy amide, and a para-substituted octafluorobiphenylate (signal ion form), respectively. The β-hydroxy amide moiety functions as a release group by undergoing a retro thermal-aldol type of reaction at an elevated temperature, as described before (6). Relative to our prior work (6) this pathway is more practical for the synthesis of many electrophore mass tags.

1010 Bioconjugate Chem., Vol. 13, No. 5, 2002

Analogous reaction pathways are employed to convert branchpoint compounds II-V into many releasable electrophore mass tags, as also summarized in Figure 2, where an example is shown (II-475a NHS ester). The latter tag, upon electron capture, furnishes an ion (RO-) at m/z 475, due to dissociative cleavage of the O-CH2 bond. For the electrophore tags derived from V, the mass signals can begin at m/z 93, since this value corresponds to a phenolate anion, which ultimately forms when the chlorine atom of V is replaced by phenol in the synthetic pathway. In contrast, the signal libraries derived from I-IV begin at about m/z 343, due to replacing the p-fluorine atom of this compound with methoxy. This latter product eventually forms a 4-[(4′-methoxy)tetrafluorophenyl]tetrafluorophenolate ion (m/z 343). Thus, a main advantage of V over I-IV is that it extends the library of electrophore tags to much lower masses. For all five of the pathways, stable isotope phenols can be employed along with corresponding, ordinary phenols, as needed, to provide additional mass signals or to enhance quantitation by acting as an internal standard. While the use of stable isotopes would be necessary to fill in the mass range immediately above m/z 93, a large library of mass tags can be prepared without resorting to stable isotopes, because many phenols are readily available, a cost advantage. In Figure 3 is shown a mass spectrum by GC-EC-MS of IV (designated as IV-331a in this figure), along with that of four corresponding electrophore mass tags derived from IV, and of a ketone electrophore derived from V. (Compound IV is designated as IV-331a here since it gives rise to a major ion at m/z 331, and is the first compound or “a”, derived from IV that provides a signal at this mass.) It is meaningful to detect these ketones as synthetic intermediates since they are also the very signal compounds that are released for electron capture in the last step of our analytical technique. Of special interest is the degree to which the compounds of Figure 3 form single ions in the EC-MS, as seen. This important property not only contributes to their sensitivity, but also helps to simplify the overall mass spectrum that a library of such compounds would yield. We consider that all of these derivatives are successful in this respect, since the electron capture mass spectrum of each one is dominated by a single ion (aside from M + 1). A minor, secondary ion arising from loss of a fluorine is seen in some of the spectra. While a 10-fold variation in response is observed for the major ion among these compounds, this result may have no meaning for our purposes. This is because the variation in response may be due to different recoveries of the compounds in the GC injection and separation steps, and these steps will be omitted in our ultimate detection scheme (which will involve laser desorption or laser-induced electron capture desorption). Precursor tag I and the representative, corresponding derivative, I-474a, can be measured directly by laserinduced electron capture time-of-flight mass spectrometry on a silver surface (LI-EC[Ag]-TOF-MS), as shown in Figure 4, relying on the conditions that we introduced recently for the latter technique (9). We did not comparatively inject these compounds into the GC-EC-MS, because we would expect an irreproducible, low response for compounds of this type (high in mass and containing isopropylidine, amide, and ester groups). Whether the potentially labile isopropylidine moiety of I and I-475a remains intact throughout the electron capture process that leads to the observed ions is unknown.

Zhang et al.

Figure 6. A: Detection by positive ion MALDI-MS of oligonucleotide P107 (5′-[aminoalkyl]-ACCGGCCATGCTCGCGACCA) in trace a; and of P107, after labeling with mass tag II475a, in trace b. The oligomer was labeled with an NHS ester of II-475a in aqueous dimethyl sulfoxide containing 5% triethylamine using the conditions described previously (10). B: Detection by LI-EC(Ag)-TOF-MS of the latter, electrophore-labeled oligonucleotide, 500 fmol on the target, six laser shots.

Why does compound I-475a, but not compound I, lose fluorine in the LI-EC(Ag)-TOF-MS experiments of Figure 4? At least part of the reason is the variation in the experimental conditions. Arbitrarily, we used a lower power density for the laser in the detection of I (Figure 4B) vs I-475a (Figure 4A). Note the noisier baseline in Figure 4B both in terms of thickness and waviness, associated with a less abundant signal. This was confirmed by observing that fluorine loss from I could be increased by raising the laser intensity (data not shown). In Figure 5 are shown the structures, fragmentation, and simultaneous detection by LC-EC(Ag)-TOF-MS, of a mixture of intermediates from the synthetic schemes presented in Figure 1. Importantly, all of the compounds are detected with high sensitivity: the amount of each compound applied to the silver target is 100 fmol, and only about 0.25% of each sample spot is in the laser beam. As seen, a relatively clean baseline is observed for a corresponding blank sample in Figure 5B. Also of note is the success of compound II-475a in this experiment. Compound II-475a possesses a new release group relative to prior literature. The preparation and testing of this latter compound was motivated by the behavior of I-475a (Figure 4A). Extension of our detection strategy to a DNA oligomer labeled with an electrophore mass tag, for eventual use as a sequencing primer or hybridization probe, is shown in Figure 6. A 5′ amino linked DNA oligomer was labeled

Releasable Electrophore Tags for Mass Spectroscopy

Bioconjugate Chem., Vol. 13, No. 5, 2002 1011

Figure 7. Detection of electrophores by LI-EC(Ag)-TOF-MS: A. 100 pmol of II-861a on the target, 17 laser shots. B. 1200 pmol of II-981a on the target, nine laser shots.

with a N-hydroxysuccinimide (NHS) ester of release tag II-475a. Analysis of the starting and electrophore-labeled product oligomers (a and b, respectively) by conventional MALDI-TOF-MS is shown in Figure 6A, and detection of the product oligomer (actually of the electrophore tag on this oligomer) by LI-EC(Ag)-TOF-MS is shown in Figure 6B. While we did not strive to demonstrate high sensitivity in this experiment, the observation that the simple conditions of LI-EC(Ag)-TOF-MS enable direct oligomer detection with a clean baseline (especially above 450 units of an electrophore-labeled DNA oligomer) is important in this technology. Toward a goal of expanding the degree of multiplexing available from electrophore release tags, we have begun to explore higher-mass versions of these reagents. Initial success in this endeavor is shown by the data in Figure 7A, where compound II-861a, despite its more complicated structure, is detected by LI-EC(Ag)-TOF-MS as nearly a single ion (m/z 861). In work preceding this experiment, compound II-981a was prepared and found to yield six major ions (Figure 7B). The secondary cleavage at the benzyl ether site in II-981a is not surprising, since it forms a favorable amidophenolate species (the m/z 416 ion). The ion at m/z 496 in both

spectra, assuming that we have assigned it correctly, would not be seen if the tag was attached to an oligomer, since the same cleavage would generate an oligomer ion. Our ability to avoid significant secondary cleavages in an ion having a mass as high as 861 units suggests that a considerable mass range can be anticipated for release tag electrophores, creating an opportunity for highly multiplexed assays that utilize them as tags. Considering our need to design electrophores with favorable fragmentation properties upon electron capture (e.g., compound II-861a vs II-981a in Figure 7) we explored, in a preliminary way, whether the semiempirical method AM1 (using Hyperchem software) might be useful as a tool to understand some aspects of electron capture for our compounds. Arbitrarily, we selected a simple electrophore, III, and considered that the energy of the incoming electron was 0 eV for this calculation. For example, we wanted to find out whether the calculation could rationalize the emergence of g50 kcal/mol of excess, internal energy upon electron capture, to account for the observed cleavage of the C-O bond (bond energy 50-67 kcal/mol), aside from kinetic factors. According to AM1, compound III prefers to exist with its two polyfluoro aromatic rings essentially perpendicu-

1012 Bioconjugate Chem., Vol. 13, No. 5, 2002

lar to each other, and this conformation provides a minimum value for the internal energy of -3789 kcal/ mol. The corresponding anion radical of III in the same conformation is calculated to have an energy which is only 18 kcal/mol lower (-3807 kcal/mol), yielding an insufficient excess of internal energy to break the C-O bond. (What is being assumed is that the difference in energy between these two species is present as excess, internal energy in the incipient anion radical intermediate). Nevertheless, calculating the optimum geometry for the anion radical yields a structure in which the two polyfluoroaromatic rings have rotated to become more planar relative to each other. This furnishes another 35 kcal/mol of additional energy (to -3842 kcal/mol), sufficient in total (53 kcal/mol) to break the C-O bond. Much or most of the added electron density from electron capture, according to AM1, resides on the fluorine atoms of the outer perfluorphenyl ring in both the perpendicular and more planar conformers of the anion radical and then largely shifts to the adjacent polyfluorophenylate moiety that arises when the C-O bond breaks. Perhaps this shift in electron density is easier for the more planar conformer, having more conjugation between the rings. The calculated energies for the observed anion and radical dissociation products in an optimum geometry from electron capture are -2704 and -1130 kcal/mol, respectively. The sum of these latter values is nearly the same (within 8 kcal/mol) of the energy calculated for the parent anion radical (the calculation assumes no excess internal energy is present) after aromatic ring twisting. From this argument it follows that the excess internal energy from ring rotation may not truly be needed for bond cleavage, since the emerging, extra stability of the favorable fragmentation products, due in part to resonance, could provide the energy necessary to drive (concurrently) the dissociation, assuming that this energy emerges early in the reaction pathway. Some of this resonance is due to the planarity of the polyfluorobiphenylate product, which creates a barrier for the parent, more perpendicular anion radical (-3789 kcal/mol) to access this alternative energy source. Thus, overall the postulated ring rotation may contribute to the observed dissociative electron capture both by increasing the amount of excess internal energy in the starting anion radical reactant and by enabling a shift in electron density from the outer to the inner polyfluorophenyl ring along the reaction pathway as a second source of excess internal energy, thereby breaking the bond. CONCLUSION

Electrophore mass tags offer the advantages of low cost, physicochemical stability, high multiplicity, high sensitivity, and high throughput detection by mass spectrometry. Some of this promise has been demon-

Zhang et al.

strated in this project. Practical synthetic strategies have been developed to make a large number of these tags which release under both gas phase and surface electron capture conditions, yielding gas phase signal groups that ionize with minimal secondary fragmentation to provide, in turn, nearly mono-ion spectra for high sensitivity. Representative tags with these properties furnishing ions from m/z 93 to 861 have been prepared, opening up a wide range of multiplicity to enhance multiplexing. This success guides and encourages the further development of these reagents. ACKNOWLEDGMENT

ThisworkwassupportedbyNISTAward70NANB5H1038 and, via George Church, DOE Grant DE-FG02-87ER60565 Contribution No. 809 from the Barnett Institute. The authors thank Max Deinzer for reading this manuscript. LITERATURE CITED (1) Salas-Solano, O., Carrilho, E., Lotler, L., Miller, A. W., Goetzinger, W., Sosic, Z., and Karger, B. L. (1998) Routine DNA Sequencing of 1000 Bases in Less Than One Hour by Capillary Electrophoresis with Replaceable Linear Plyacrylamide Solutions. Anal. Chem. 70, 3996. (2) Pennisi, E. (1998) A Closer Look at SNPs Suggests Difficulties. Science 281, 1787. (3) Guo, B. (1999) Mass Spectrometry in DNA Analysis. Anal. Chem. 71, 333R. (4) Zhang, L.-K., and Gross, M. L. (2000) Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry Methods for Oligodeoxynucleotides: Improvements in Matrix, Detection Limits, Quantification, and Sequencing. J. Am. Soc. Mass Spectrom. 11, 854. (5) Giese, R. W. (2000) Electron-capture mass spectrometry: recent advances. J. Chromatogr. 892, 329. (6) Xu, L., Bian, N., Wang, Z., Abdel-Baky, S., Pillai, S., Magiera, D., Murugaiah, V., Giese, R. W., Wang, P., O’Keeffe, T., Abushamaa, H., Kutney, L., Church, G., Carson, S., Smith, D., Park, M., Wronka, J., and Laukien, F. (1997) Electrophore Mass Tag Dideoxy DNA Sequencing. Anal. Chem. 69, 35953602. (7) Bian, N., Wang, P., Wang, Z., Xu, L., Church, G., and Giese, R. W. (1997) Detection Via Laser Desorption and Mass Spectrometry of Multiplex Electrophore-labeled Albumin. Rapid Commun. Mass Spectrom. 11, 1781-1784. (8) Wang, B., and Biemann, K. (1994) Matrix-Assisted Laser Desorption/Ionization Time-of Flight Mass Spectrometry of Chemically Modified Oligonucleotides. Anal. Chem. 66, 19181924. (9) Wang, P., and Giese, R. W. (2000) Laser-Induced Electron Capture Mass Spectrometry. Anal. Chem. 72, 772-776. (10) Wang, Z., Bian, X., Qian, X., and Giese, R. W. (1998) N-Hydroxysuccinimide Ester Labeling 5′-aminoalkyl DNA Oligomers: Reaction Conditions and Purification, J. Chromatogr. 806, 93-95.

BC010072V