Search

We prepared a series of planar titanium microfluidic (μLC) columns, each 100 mm long, with 0.15, 0.3 and 0.5 mm i.d.'s. The microfluidic columns were packed with 1.8 μm C18 sorbent and tested under isocratic and gradient conditions. The efficiency and peak capacity of these devices were monitored using a micro LC instrument with minimal extra column dispersion. Columns with serpentine channels were shown to perform worse than those with straight channels. The loss of efficiency and peak capacity was more prominent for wider i.d. columns, presumably due to on-column band broadening imparted by the so-called "race-track" effect. The loss of chromatographic performance was partially mitigated by tapering the turns (reduction in i.d. through the curved region). While good performance was obtained for 0.15 mm i.d. devices even without turn tapering, the performance of 0.3 mm i.d. columns could be brought on par with capillary LC devices by tapering down to 2/3 of the nominal channel width in the turn regions. The loss of performance was not fully compensated for in 0.5 mm devices even when tapering was employed; 30% loss in efficiency and 10% loss in peak capacity was observed. The experimental data for various devices were compared using the expected theoretical relationship between peak capacity P c and efficiency N; (P c -1) = N 0.5  × const. While straight μLC columns showed the expected behavior, the devices with serpentine channels did not adhere to the plot. The results suggest that the loss of efficiency due to the turns is more pronounced than the corresponding loss of peak capacity., (Copyright © 2017 Elsevier B.V. All rights reserved.)

We demonstrate an integrated microfluidic LC device coupled to a QTOF capable of improving sensitivity and linearity for intact protein analysis while also tuning the charge state distributions (CSD) of whole antibodies. The mechanism for sensitivity improvement using microflow ESI is demonstrated by shifting of the CSD to higher charge state, and narrowing of the overall CSD. Both of these aspects serve to improve ion current of the most abundant charge state of antibodies and lead to improvement in sensitivity over high flow ESI by a factor of 15×. Current limits of detection are 0.1 ng (on-column) (n = 100, %RSD = 17.5) using IgG glycosylated antibody, as compared to 5 ng (on-column) (n = 10, %RSD = 15) for high flow LC-ESI-MS. In addition to improvements in sensitivity we also observe improvements in linear dynamic range for microflow ESI that results from a combination of lower limits of detection and narrower CSD. An improvement of linear dynamic range of 1.5 orders of magnitude was observed over conventional high flow LC-MS. In cases where the complexity of the antibody limited both sensitivity and spectral charge state resolution, we employed supercharging and decharging mechanisms to further improve sensitivity and charge state spacing resolution. We demonstrate an 89% increase in sensitivity using glycerol that was added post column, with retention of the glycoform resolution. Since large proteins reside in a relatively low noise region of the mass spectra it is possible to realize effects of supercharging for intact proteins, specifically antibodies of 150 kDa, that are less pronounced for peptide supercharging. We also demonstrate a 51% increase in charge state resolution as imidazole was used to generate lower charge states for high-mass ions. The increase in charge state resolution enables more complex antibodies, or antibody mixtures that coelute in the LC, to be deconvoluted more efficiently. In summary, we demonstrate an analytical technique that yields improved sensitivity and quantitative linear dynamic range for intact protein analysis over conventional LC-MS, and yields ease of use for more complex experimentation such as supercharging and decharging experiments.