Structural Prediction of Impurities in Drugs using MSn Data

Significance of Topic

Identification of trace impurities in macrolide antibiotics such as erythromycin is essential to guarantee product purity and patient safety and to meet stringent regulatory requirements.

Objectives and Study Overview

This work demonstrates a strategy combining multi-stage mass spectrometry (MSn) and formula prediction software to assign molecular formulas and structural features of unknown impurities in an erythromycin A oxime sample.

Methodology and Instrumentation

• Sample preparation: Erythromycin A oxime (1 mg/mL in methanol) was injected onto a Phenomenex Gemini C18 column (150×2 mm, 5 μm) at 40 °C.
• Chromatography: Isocratic elution (40% 0.1% NH4OH/60% acetonitrile) at 0.2 mL/min monitored by UV at 200 nm.
• Mass spectrometry: Shimadzu LCMS-IT-TOF operated in negative electrospray, acquiring full-scan MS and MS2 spectra with high mass accuracy.
• Data processing: Enhanced formula prediction software used MSn fragmentation data, isotopic distribution, DBE and H/C constraints, and nitrogen rule filtering.

Main Results and Discussion

• An impurity at m/z 783.4421 exhibited fragment ions corresponding to the core lactone region (Area C) and loss of a sugar moiety (Area B); a mass shift of 35.9841 Da in Area A suggested a modified deoxy sugar.
• A second impurity at m/z 733.4439 showed a similar fragmentation pattern with a 14.0157 Da deficit in the sugar region (Area B), corresponding to a demethylated derivative (C36H66N2O13).
• A third impurity at m/z 763.4581 matched an addition of an oxygen atom in the sugar region based on a 15.9949 Da shift, assigning formula C37H68N2O14.
• An ion at m/z 761.4375 lacked characteristic Area C fragments and sugar losses, indicating an unrelated contaminant.
• Mass accuracy for all identified formulas was within low ppm, confirming high-confidence assignments.

Benefits and Practical Applications

The combined MSn and predictive algorithm approach accelerates impurity profiling in complex drug samples, reducing manual interpretation and improving structural elucidation workflows in QA/QC and R&D laboratories.

Future Trends and Potential Applications

• Integration of high-throughput automated MSn analysis with machine-learning prediction tools.
• Application to a broader range of biopharmaceuticals and small-molecule drugs.
• Development of real-time impurity monitoring during drug manufacturing.

Conclusion

The study demonstrates that high-resolution MSn data coupled with advanced formula prediction rapidly identifies and characterizes erythromycin impurities, highlighting the method’s utility in pharmaceutical analysis.

References

• Shackman H.M., Ginter J.M., Fox J.P., Nishimura M. Structural Prediction of Impurities in Drugs Using MSn Data; Technical Report vol.3, Shimadzu Scientific Instruments.