1

Supplementary Material to JASMS: January 26th, 2015

A Novel and Intuitive Method of Displaying and Interacting with Mass Difference Information: Application to Oligonucleotide Drug Impurities

Stilianos G. Roussis

Isis Pharmaceuticals

Carlsbad, California 92008

Isis Pharmaceuticals,

2282 Faraday Avenue, Carlsbad, CA 92008

Tel. 760-603-3579

SUPPLEMENTARY METHODS

Reagents and Samples. Triethylamine (TEA), tributylamine (TBuA) and ethylenediaminetetraacetic (EDTA) free acid (Fluka BioUltra) were obtained from Sigma-Aldrich (St. Louis, MO). Glacial acetic acid and HPLC grade acetonitrile (ACN) were purchased from J. T. Baker (Phillipsburg, NJ). Distilled, deionized water of low metal ion content (Mediatech Cellgro Sterile WFI) was used for the preparation of the mobile phases. Deionized water (HPLC grade, J. T. Baker or Mediatech Cellgro Sterile WFI) was used for sample preparation. Oligonucleotide samples were obtained from Isis Pharmaceuticals Inc. (Carlsbad, CA). Residual ammonium hydroxide potentially present in crude samples was removed by using a vacuum centrifuge in a well-ventilated fume hood. Samples were prepared at approximately 0.2 mg/mL concentrations in dilute (e.g., 0.01% v/v) aqueous TEA.

The Agilent Technologies ESI-L low concentration tuning mixture (40:1 dilution ratio) was used to tune and calibrate the instrument in the negative ionization mode. A dedicated calibrant delivery system was used to introduce additional internal reference compounds to increase the mass accuracy of the experimental data at a second calibration step. The reference compounds were continuously co-introduced with the LC effluent stream using a second nebulizer integrated into the ionization source and controlled by software. Agilent’s purine (m/z 119.036320) and HP-0921 (acetate adduct, m/z 980.016375) compounds from the API-TOF reference mass solution kit (G1969-85001), and HP-1821 (acetate adduct, m/z 1879.958890) and HP-2421(acetate adduct, m/z 2479.920567) from the ES-TOF Biopolymer analysis reference mass standards kit (G1969-85003) were used as internal mass references for the secondary mass correction.

Liquid Chromatography-Mass Spectrometry.Ion-pair high performance liquid chromatography (IP-HPLC) coupled with a time-of-flight (TOF) mass spectrometer was used for the analysis of the samples. The HPLC system was an Agilent 1100 series instrument equipped with a degasser (G1322A), a binary pump (G1312A), a heated column chamber (G1316A), an autosampler (G1313A), and a variable wavelength UV detector (DAD G1315B). For the analysis of the purified (DMT-on) and detritylated (DMT-off) drug substance samples the HPLC column was an XBridge OST C18 3.5 m particle size, 2.1mm i.d. x 150 mm column (Waters, Milford, MA). Mobile phase A consisted of 10% ACN, 5mM TBuAA, 1M EDTA in water. Mobile phase B consisted of 80% ACN, 5mM TBuAA, 1M EDTA in water. The gradient profile was: 0% B from 0 to 1.0 min, then 0-55% B from 1.0 to 3.0 min, then 55-65% B from 3.0 to 20.0 min, then 65-95% B from 20.0 to 25.0 min, then kept constant at 95% B until 29.8 min, and finally 5-0% B from 29.8 to 30.0min. The post-run time was 20 min for a total run time of 50 min. The flow rate was 0.25 mL/min. For the analysis of the crude sample set prepared under excess and no-capping reaction conditions, respectively, the column was an Acquity UPLC OST C18 1.7 m particle size, 2.1mm i.d. x 50 mm column (Waters, Milford, MA).The gradient profile was: 0% B from 0 to 0.3 min, then 0-55% B from 0.3 to 1.0 min, then 55-65% B from 1.0 to 7.0 min, then 65-95% B from 7.0 to 10.2 min, then kept constant at 95% B until 13.5 min, and finally 5-0% B from 13.5 to 13.7min. The post-run time was 3.3 min for a total run time of 17 min. The flow rate was 0.30 mL/min. The column temperature was 50 °C for both sets of experiments.Samples were introduced using the autosampler. The sample volume injected was 25 L.

The time-of-flight (TOF) mass spectrometer was an Agilent 6224 instrument, equipped with a dual nebulizer electrospray ionization (ESI) source (G3251B) to accommodate the simultaneous introduction of the reference compounds used in the secondary mass correction step to increase the mass accuracy of the data. Nitrogen was used as both drying and nebulizing gas. The drying gas temperature was set to 340 °C and the gas flow to 12 L/min. The nebulizer pressure was 30 psig. The dual ESI Vcap (voltage applied on the capillary) was 4000 V. The fragmentor voltage was 150 V and the skimmer voltage 65 V. The RF peak-to-peak voltage on octopole 1 was 250 V. Data were acquired in the negative ionization mode and stored in the profile format. The absolute threshold of acquisition was set to 50 counts. The instrument was operated in the high-resolution mode (4 GHz). The acquisition mass range was set to 80-2500 m/z and the acquisition rate to 1 spectrum/s.9904 transients are acquired per spectrum.

AbundanceThreshold.Generally a low abundance threshold value is used in the digital acquisition of the spectra (e.g., 50 counts) to ensure the detection of the lowest abundance components in the sample. This can also be achieved by the injection of higher amounts of sample, up to a certain point, as this may lead to unwanted side effects such as reduced chromatographic separation (overlapping peaks) and potentially ion signal suppression in the ionization source caused by the increased sample concentration. We prefer to control the abundance threshold in the spectra with the MassHunter data system as it provides the ability to experiment with different values to determine the optimal value for the determination of the lowest abundance sample components. We have found that the averaging of spectra has a particularly significant benefit as it improves the S/N ratio thus greatly improving the ability to define peaks of low abundance. As the number of data points acquired in a high-resolution experiment is considerably higher than in a low-resolution experiment the actual number of data points in the averaged mass spectrum significantly affects the computational time of the mass difference program. We typically run the mass difference program starting with a relatively high abundance threshold (e.g., 10-20% of the most abundant peak) and progressively reduce the threshold to lower than 5% or 1% to determine the presence of weak sample components. Running the program at the low abundance threshold values significantly lengthens the computational time. We have found by experience that it is possible to determine the optimal abundance threshold by initially running the program for one or two spectra, acquired under the same conditions, at the lowest abundance threshold values and then, once the optimal threshold is obtained, subsequent spectra are treated at the determined optimal abundance threshold value.