Anal. Chem. 2002, 74, 1355-1359
A Method To Accurately Determine the Extent of Solid-Phase Reactions by Monitoring an Intermediate in a Nondestructive Manner Jeffery Mihaichuk, Christopher Tompkins,* and Wolfgang Pieken
Proligo LLC, 2995 Wilderness Place, Boulder, Colorado 80301
The application of Fourier transform infrared (FT-IR) spectroscopy to the quantitation of effective nucleoside loading on a solid support, via detection of a reaction intermediate, was developed. 4,5-Dicyanoimidazole (DCI) activator is reacted with a known amount of polystyrenebound phosphoramidite to form a diisopropylamine-DCI salt complex that is detected by FT-IR spectroscopy at a unique wavenumber of 1104 cm-1. Based on a linear standard curve, the species’ absorbance at 1104 cm-1 is then directly translated into an appropriate concentration. This concentration characterizes the number of active phosphoramidite sites on the solid support. This method consistently determines the loading extent of nucleosides (deoxy-A/T/G/C) on the solid support to the nearest 0.02 mmol/g. The coupling reaction extent of a deoxy-T derivative with the support-bound phosphoramidite compares well with the nucleoside loading, as determined by the FT-IR-based analytical method. The nucleoside loading was further verified by phosphorus elemental analysis of the 3′-phosphate linker concentration. Solid-phase organic synthesis (SPOS) is a continually expanding practice necessitating better characterization of the support. Various solid-phase analytical techniques are extensively reviewed.1-4 These analytical techniques can be grouped into two general strategies: passive quantitation of species loading by direct observation, and removal of chemical species from the support followed by solution-phase quantitation. Both types of assays have limitations. In the first strategy, nuclear magnetic resonance (NMR) and Fourier transform infrared spectroscopy (FT-IR) techniques have been applied to directly analyze the solid support. NMR techniques include gel phase and magic angle spinning. Gel-phase NMR has had limited utilization for quantification of resin loadings.5-7 Furthermore, this technique is complicated by broad resonances and spectral interference of the * To whom correspondence should be addressed. E-mail:ctompkins@ proligo.com. (1) Hermkens, P. H.; Ottenheijm, H. C.; Rees, D. Tetrahedron 1996, 52 (13), 4527-4554. (2) Crowley, J. I.; Rapoport, H. Solid-Phase Org. Synth. 1976, 9, 135-144. (3) Yan, B. Acc. Chem. Res. 1998, 31 (10), 621-630. (4) Gallop, M. A.; Fitch, W. L. Curr. Opin. Chem. Biol. 1997, 1 (1), 94-100. (5) Svensson, A.; Fex, T.; Kihlberg, J. Tetrahedron Lett. 1996, 37 (42), 76497652. (6) Stones, D.; Miller, D. J.; Beaton, M. W.; Rutherford, T. J.; Gani, D. Tetrahedron Lett. 1998, 39 (27), 4875-4878. 10.1021/ac010884k CCC: $22.00 Published on Web 02/16/2002
© 2002 American Chemical Society
Figure 1. Byproduct liberation from a controlled reaction. Example 1 depicts a scenario in which compound c replaces compound b, a quantifiable chemical compound. Example 2 depicts a scenario in which compound C reacts and cleaves compound b to form compound Cb, a quantifiable chemical compound.
support.8 31P Cross-polarization magic angle spinning (CPMAS) NMR has been successfully used to characterize oligonucleotide phosphorus chemistries.9 However, experimental conditions did not allow for quantification. Recently, two 13C MAS NMR methods were developed to quantify loading of functionalized Wang and trityl resins.10 Although the researchers were able to minimize spinning sideband overlap, the primary method was limited to highly swelled resins. The second method was developed for lowswelling resins but cannot be used for highly loaded support or in cases where there is potential for side products or incomplete loading reactions. As was the case of gel phase, MAS NMR interpretation can also be complicated by interference of the support structure. FT-IR techniques that have been utilized to directly analyze solid support include conventional, photoacoustic, and transmission technologies. Conventional FT-IR spectroscopy has been applied to monitor solid-phase surface chemistries.11-13 Mechanical contact with the sample is required and can influence spectral accuracy from matrix inhomogeneities and signal overlap. A technique of photoacoustic (PA) FT-IR was developed that circumvented surface sensitivities but did not provide sufficient data for loading quantification.14 Finally, transmission FT-IR has been used to study etched surfaces of porous Si but is limited to nontortuous pore structures.15 The second analytical strategy, in which a species is released from the porous support and then detected by MS, TLC, HPLC, (7) Look, G. C.; Holmes, C. P.; Chinn, J. P.; Gallop, M. A. J. Org. Chem. 1994, 59 (25), 7588-7590. (8) Russell, K.; Cole, D. C.; McLaren, F. M.; Pivonka, D. E. J. Am. Chem. Soc. 1996, 118, 7941-7945. (9) Macdonald, P. M.; Damha, M. J.; Ganeshan, K.; Braich, R.; Zabarylo, S. V. Nucleic Acids Res. 1996, 24 (15), 2868-2876. (10) Hany, R.; Rentsch, D.; Dhanapal, B.; Obrecht, D. J. Comb. Chem. 2001, 3, 85-89. (11) Yan, B.; Kumaravel, G.; Anjara, H.; Wu, A.; Petter, R. C.; Jewell, C. F., Jr.; Wareing, J. R. J. Org. Chem. 1995, 60, 5736-5738. (12) Chan, T. Y.; Chen, R.; Sofia, M. J. Tetrahedron Lett. 1997, 38 (16), 28212824. (13) Pivonka, D. E.; Russell, K.; Gero, T. Appl. Spectrosc. 1996, 50, 1471-1478. (14) Gosselin, F. J. Org. Chem.1996, 61 (23), 7980-7981.
Analytical Chemistry, Vol. 74, No. 6, March 15, 2002 1355
Figure 2. DIPA assay chemistry. The DIPA byproduct is denoted in boldface.
Figure 3. Fingerprint region: 0.300 mmol of DIPA + excess DCI in 1 mL of ACN to form 300 mM DIPA-DCI complex.
or GC, requires destructive irreversible cleavage of the product from the support followed by additional sample preparation. When applicable, an alternative to these two analytical strategies is to quantitate, in solution phase, a byproduct of a reaction involving the immobilized species (Figure 1). This technique is most useful when the liberating reaction is part of the standard chemical process that utilizes the solid support. When this is true, the assay will quantify active sites rather than total sites and also does not require development of any new chemistries. Specifically, we describe a unique method to determine the effective load of solid supports for use in a novel oligonucleotide synthesis process, developed by Proligo LLC, termed as “dimethoxytrityl resin product anchored sequential synthesis” (DMT PASS).16 This process differs from traditional oligonucleotide synthesis in that the phosphoramidite is immobilized on polystyrene solid support; the immobilized amidite is then sequentially coupled with a growing oligonucleotide fragment. The active site loading of the immobilized phosphoramidite is determined by first reacting the amidite with 4,5-dicyanoimidazole (15) Yau, S. L.; Arendt, M.; Bard, A. J.; Evans, B.; Tsai, C.; Sarathy, J.; Campbell, J. C. J. Electrochem. Soc. 1994, 141 (2), 402-409. (16) Mihaichuk, J. C.; Hurley, T. B.; Vagle, K. E.; Smith, R. S.; Yegge, J. A.; Pratt, G. M.; Tompkins, C. J.; Sebesta, D. P.; Pieken, W. A. Org. Process Res. Dev. 2000, 4 (3), 214-224.
1356
Analytical Chemistry, Vol. 74, No. 6, March 15, 2002
(DCI).17 This reaction generates a solution-phase byproduct, N,Ndiisopropylamine (DIPA), which forms a complex with excess DCI (Figure 2). The DIPA-DCI complex contains a highly selective IR chromophore, which can be monitored by Fourier transform infrared spectroscopy. This powerful assay has broad applications to almost any standard nucleoside support that utilizes phosphoramidite chemistry.18-21 Since the reaction generating the DIPADCI complex is identical to the standard activation used in the oligonucleotide synthesis process, this assay accurately quantifies the active resin-bound phosphoramidite. Additionally, this FT-IRbased assay offers a convenient, fast, and highly selective methodology that has capabilities of monitoring the reaction in situ. EXPERIMENTAL SECTION The immobilized phosphoramidite was prepared from 5′-DMT polystyrene via standard methods using synthesis reagents from (17) Vargeese, C.; Carter, J.; Yegge, J.; Krivjansky, S.; Settle, A.; Kropp, E.; Peterson, K.; Pieken, W. Nucleic Acids Res. 1998, 26 (4), 1046-1050. (18) Caruthers, M. H. Acc. Chem. Res. 1991, 24, 278-284. (19) Beaucage, S. L.; Iyer, R. P. Tetrahedron 1992, 48 (12), 2223-2311. (20) Sinha, N. D.; Biernat, J.; McManus, J.; Koster, H. Nucleic Acids Res. 1984, 12 (11), 4539-4557. (21) Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22, 1859-1862.
Figure 4. Baseline subtraction of ACN and DCI. A measurable absorbance at 1104 cm-1 is unique to the DIPA-DCI complex.
Figure 5. Determination of a quantifiable wavenumber: (A) 0.058 mmol of 5′-DMT DIPA phosphoramidite resin + excess DCI in 1 mL of ACN to form 58 mM DIPA-DCI complex, (B) 0.046 mmol of free phosphoramidite + excess DCI in 1 mL of ACN to form 46 mM DIPADCI complex, (C) 60 mM DIPA in ACN, (D) 120 mM DCI in ACN, (E) 46 mM free phosphoramidite in ACN, and (F) nonfunctionalized solid support + 120 mM DCI in ACN.
Figure 6. Concentration profile. Absorbance increases linearly with DIPA concentration in a 120 mM DCI solution. Data points (from left to right): (A) 56 mM DIPA-DCI complex and (B) 30, (C) 17, (D) 5, and (E) 0 mM DIPA.
Chemgenes Corp., GFS Chemicals, Fluka, and Palmer Sciences.22 Both the deoxyoligonucleotides and 4,5-dicyanoimidazole were obtained from Proligo Biochemie GmbH (Hamburg, Germany). The immobilization was performed on an Applied Biosystems 390Z DNA/RNA synthesizer. ACS-grade acetonitrile (ACN) and meth(22) Gait, M. J. Oligonucleotide Synthesis: A Practical Approach; Oxford University Press: Oxford, U.K., 1990; p 27.
Figure 7. FT-IR monitoring of DIPA-DCI complex. The data are the average of three different trials that varied from one another within 1.5% standard deviation.
ylene chloride (DCM) were obtained from EM Science. Analysis of all assay samples was performed on a ReactIR1000 FT-IR system from ASI using Diamond DuraDisk sampler technology (DurasamplIR). Standardization. To account for daily fluctuations of the FTIR spectrometer calibration, it was necessary to design an assay method that incorporated a daily manual standardization. A single and independent wavenumber was sought that would accurately quantitate the amount of DIPA-DCI in solution. The wavenumber region 1300-625 cm-1, better known as the fingerprint region, was searched for a pattern of peaks specific to the DIPA-DCI interaction. The fingerprint region of a concentrated DIPA-DCI solution in ACN is shown in Figure 3. A unique signal at 1104 cm-1 (9.06-µm wavelength) was found following the subtraction of both standard ACN and DCI profiles (Figure 4). Furthermore, as demonstrated in Figure 5, the peak at 1104 cm-1 also appears when excess DCI and a DIPA-protected phosphoramidite species are in solution together. Above an absorbance of 0.011, there is no spectral interference with the DIPA-DCI complex at 1104 cm-1 from ACN, DCI, DIPA, phosphoramidite, or solid support. The DIPA concentration relationship is linear with absorbance at 1104 cm-1 (Figure 6). Five different FT-IR readings at each concentration demonstrated an average standard deviation of (0.001 absorbance units. This fluctuation in absorbance is indicative of the amount of baseline drift. As indicated by the regression Analytical Chemistry, Vol. 74, No. 6, March 15, 2002
1357
Figure 8. Verification of DIPA assay by coupling with deoxy-T nucleoside. Using the DIPA assay results as a basis, the reaction was run with 2 equiv of nucleoside and DCI. Reactions were complete in 30 min. Table 1. Confirmation of DIPA Loading Results DIPA assay
a
elemental analysis
nucleoside
load (mmol/g)
Std Dev (mmol/g)
coupling load (mmol/g)
% phosphorus
load (mmol/g)
dC dC dC dG dG dT dT dT dT dA dA
0.36 0.39 0.15 0.30 0.35 0.24 0.31 0.43 0.48 0.21 0.23
( 0.02 ( 0.05 ( 0.01 ( 0.01 ( 0.03 ( 0.04 ( 0.01 ( 0.03 ( 0.02 ( 0.02 ( 0.02
0.29 0.33 0.15 0.25 naa 0.19 0.26 0.39 0.41 0.20 0.22
1.260 1.185 0.485 0.895 1.160 0.765 0.760 1.255 1.420 0.490 na
0.41 0.38 0.16 0.29 0.37 0.25 0.25 0.41 0.46 0.16 na
na, not available.
coefficient (R2) value, the linear model accounts for 99.95% of variation in absorbance at any concentration. The linear correlation of absorbance with species concentration at 1104 cm-1 follows Beer’s law (eq 1), where A is the absorbance at a particular
and the effective loading calculated using eq 2, where L is the
A ) bc
phosphoramidite resin loading (mmol/g), Ci is the deoxy-T initial concentration (mmol/mL), Cf is the deoxy-T final concentration (mmol/mL), V is the reaction volume (mL), and M is the mass of the support (g). Phosphorus Elemental Analysis Verification Test. The immobilized phosphoramidite loading was also verified by phosphorus elemental analysis of solid support samples. Robertson Microlit Laboratories, Inc, performed this service in duplicate trials.
(1)
wavelength, is the molar adsorptivity coefficient, b is the path length, and c is the solute molar concentration, and indicates that absorbance at this wavenumber is directly proportional to DIPADCI complex concentration. DIPA Loading Assay Protocol. The protocol of the DIPA assay is as follows: A stock solution of 120 mM DCI was prepared in anhydrous ACN. To 1.0 mL of this solution was added 0.1 g of immobilized phosphoramidite support at time 0. The samples were then shaken for a reaction time of 60 min. The phosphoramidite loading was then determined by comparing the absorbance of the DIPA-DCI complex at 1104 cm-1 to a series of standards. The reaction time of the DIPA assay was chosen based on experimental FT-IR data. The DIPA-DCI complex was monitored actively by FT-IR at 1104 cm-1 as the reaction of solid-bound amidite and DCI progressed. By 50 min, the concentration had reached equilibrium (Figure 7). Coupling Extent Verification Test. To verify the results of the DIPA assay, coupling reactions were performed using dT, dC, dG, and dA phosphoramidite resins (Figure 8). In all cases, the loaded resins were coupled to an excess of deoxy-T nucleoside. The amount of consumed nucleoside was determined by HPLC 1358
Analytical Chemistry, Vol. 74, No. 6, March 15, 2002
L ) (Ci - Cf)V/M
(2)
RESULTS AND DISCUSSION The DIPA loading assay results of various dA, dC, dG, and dT immobilized phosphoramidite support samples can be found in Table 1. Also shown is the verification of loading by the coupling extent assay and phosphorus elemental analysis. The DIPA loading assay and the coupling assay were performed in triplicate trials and averaged. The DIPA-assayed samples demonstrated an average standard deviation of (0.02 mmol/g and average variance of 0.0005 mmol/g over the trials. The DIPA assay loading results compared to the coupling extent test within (12% and compared to the elemental analysis within (9%. The coupling extent loading results were consistently lower than interpreted by the DIPA assay. This is most likely due to the inaccessibility of the bulky nucleoside to some of the active sites of the support. The disadvantage of using the coupling extent
test to determine support loading is that it requires more labor and more expensive reagents. Additionally, the resin is destroyed in the process. Phosphorus elemental analysis is costly and is usually not performed in-house where results can be immediately available. Additionally, phosphorus elemental analysis determines the total number of sites rather than the number of active sites. CONCLUSIONS The effective loading of an immobilized phosphoramidite can be easily determined by FT-IR analysis of a DCI activator reaction with the phosphoramidite using ACN as the solvent. The DIPADCI association is linear, as monitored at 1104 cm-1 wavenumber, and can accurately be measured based on a manual standardization of the FT-IR to account for calibration variations. This
methodology is an example of how an intermediate can be monitored in a controlled solution-phase reaction in order to accurately measure the loading of a solid support. The method is particularly useful for optimization of coupling efficiency to maximize yields and minimize reagent usage. Furthermore, this analytical strategy is nondestructive, simple to perform, inexpensive, and indicative of how the support will truly function based on accessible active sites.
Received for review August 7, 2001. Accepted December 20, 2001. AC010884K
Analytical Chemistry, Vol. 74, No. 6, March 15, 2002
1359
