Fig 1.
Experimental design.
Fig 2.
Visual summary of the HPLC optimisation.
HPLC separation was optimised by modifying the gradient, flow rate and composition of the mobile phases as well as testing different temperatures and stationary phase chemistries of the separation column. The first column lists the various conditions tested and the following columns display the results for each external standard analysed in this study. Base Peak Chromatograms (BPCs) are displayed from 2.5 min to 32.5 min on the x axis (retention time). The same intensity scale was displayed on the y axis for a given standard and parameter. Yellow arrows on the right hand side point to optimum conditions for each parameter tested. aCN, alpha casein; bCN, beta casein; kCN, kappa casein; aLA, alpha lactalbumin; bLG, beta lactoglobulin; BSA, bovine serum albumin.
Fig 3.
UV traces at 214 nm and EICs over time (5–25 min) of external protein standards.
Standards were prepared at the same concentration, run independently and overlaid to illustrate that ionisation efficiency varied from one protein to the other. All external standards purchased from Sigma (aCN, bCN, kCN, aLA, bLG, BSA, and myoglobin) were dissolved in 50% Solution A to a 10 mg/mL concentration. The coloured arrows in between the UV traces and the EICs represent the elution windows of the bovine protein standards.
Table 1.
MS parameters for each milk protein external standards and myoglobin internal standard following mass deconvolution.
Table 2.
Response using EIC peak areas of each external standards over increasing concentrations.
Table 3.
Averaged RTs and response of myoglobin internal standard prepared in Solution A or spiked in milk matrices a cross 2 technical replicates.
Table 4.
Slope, standard error (SE), intercept, Pearson correlation coefficient (R2) values, limits of detection (LOD) and quantitation (LOQ), and working range of myoglobin calibration curve.
Table 5.
Quantitative reproducibility of standards with or without IS across triplicates.
Fig 4.
Method validation using milk samples.
Optimum method was tested on milk samples (3 replicates) with or without internal standard (IS, myoglobin). Panel A, base peak chromatograms (BPCs) and UV trace at 214 nm of the Jersey bulk milk sample spiked with IS and run in triplicates. Panel B, spectra averaged across 5–25 min (see arrow in panel A) of the Jersey BPC and displayed along the whole m/z (600–3000) range. Panel C, BPCs and UV trace at 214 nm of the Holstein bulk milk sample spiked with IS and run in triplicates. Panel D, spectra averaged across 5–25 min (see arrow in panel C) of the Holstein BPC and displayed along the whole m/z (600–3000) range. Panel E, BPCs of Jersey sample, Holstein sample, and IS overlaid. Panel F, extracted ion chromatograms (EICs) of the Jersey sample spiked with IS and run in triplicates. Panel G, averaged spectra of kCN B-1P (see arrow in panel F) along 600–3000 m/z and zoomed in on the most abundant ion (1056.6 m/z) in inset. Panel H, EICs of the Holstein sample spiked with IS and run in triplicates. Panel I, averaged spectra of kCN A-1P (see arrow in panel H) along 600–3000 m/z and zoomed in on the most abundant ion (1058.6 m/z) in inset. Panel J, overlaid EICs of one Jersey sample replicate and one Holstein sample replicate.
Table 6.
Quantitation of protein variants from milk samples.