Continued from “Soy Milk, Good or Bad? HPLC can tell – Part 1”
In their experiments, the scientists used a number of soy drinks from local stores:
They also analyzed regular cow’s milk with added soy isoflavones. Sample preparation included 23 hour hydrolysis with concentrated HCl (hydrochloric acid ) to break down the proteins, followed by a solid-phase extraction to isolate furosine.
The HPLC method used a C8 column and the potassium chloride in aqueous acetic acid as a mobile phase with a linear gradient. The retention time was a bit over 20 min and the identity of the peak was confirmed as furosine by MS (mass spectrometry).
In the CZE method, furosine was detected by UV at 280 nm and had a migration time of 1.76 min in a separation capillary of effective length 40 cm.
When the soy beverages were analyzed, about half had no detectable furosine, showing little protein degradation. Of the rest, HPLC and CZE gave very similar values, which ranged from 25.2 to 168.71 mg/100 g protein, depending on the beverage. The highest furosine content was in the instant powdered soy milk, which is produced by severe heat treatment.
Based on these results, the researches determined the protein damage to be equivalent to 0.09-7.18% of the lysine residues blocked by the Maillard reaction. The recoveries and repeatabilities of the HPLC and CZE methods were comparable but the HPLC LOD (detection limit) was better at 1.30 ng/100 g protein compared with 5.30 for CZE.
However, despite this better sensitivity, the team recommended the use of CZE due to its faster speed, cheaper instrumentation and lower solvent usage. It is unfortunate that they never explored the option of “Lab on a Chip”. Most of the comparisons studies of CZE to Nano HPLC showed significant improvement in sensitivity as well as resolution without the major pain that CZE methods present which the Spanish scientists did not mention in their article.
Reference: “Chromatographic and electrophoretic approaches for the analysis of protein quality of soy beverages”
Miryam Amigo-Benavent, Mar Villamiel, M. Dolores del Castillo
Digital Object Identifier (DOI): 10.1002/jssc.20060043