Assessing the degradation of ancient milk proteins through site-specific deamidation patterns

Abstract

The origins, prevalence and nature of dairying have been long debated by archaeologists. Within the last decade, new advances in high-resolution mass spectrometry have allowed for the direct detection of milk proteins from archaeological remains, including ceramic residues, dental calculus, and preserved dairy products. Proteins recovered from archaeological remains are susceptible to post-excavation and laboratory contamination, a particular concern for ancient dairying studies as milk proteins such as beta-lactoglobulin (BLG) and caseins are potential laboratory contaminants. Here, we examine how site-specific rates of deamidation (i.e., deamidation occurring in specific positions in the protein chain) can be used to elucidate patterns of peptide degradation, and authenticate ancient milk proteins. First, we characterize site-specific deamidation patterns in modern milk products and experimental samples, confirming that deamidation occurs primarily at low half-time sites. We then compare this to previously published palaeoproteomic data from six studies reporting ancient milk peptides. We confirm that site-specific deamidation rates, on average, are more advanced in BLG recovered from ancient dental calculus and pottery residues. Nevertheless, deamidation rates displayed a high degree of variability, making it challenging to authenticate samples with relatively few milk peptides. We demonstrate that site-specific deamidation is a useful tool for identifying modern contamination but highlight the need for multiple lines of evidence to authenticate ancient protein data.

Figure 1

Deamidation of any casein and BLG proteins in milk powder samples. Numbers at the top of the bars show the number of deamidating amino acids (i.e., asparagine and glutamine combined) represented by this result. The y-axis represents the relative remaining amount of the deamidating amino acid—therefore high values represent less deamidation; whereas low values imply higher levels of deamidation. Outliers are shown as points, while suspected outliers are shown as hollow points.

Figure 2

Site-specific deamidation of milk powder proteins. The y-axis shows the relative remaining portion of the deamidating amino acid, while the x-axis represents the half-time of the sites of deamidation. The size of the points is relative to the intensity of the peptide identified. Half-times as estimated by Robinson and Robinson, plots produced through deamiDATE.

Figure 3

Number of BLG and any casein peptides identified in experimentally contaminated samples (extracted and analyzed in this study) according to MaxQuant’s ‘razor and unique peptides’ output. The number of asparagine and glutamine (respectively) residues in all peptides in each category is shown in the bars.

Figure 4

Bar graph of deamidation of asparagine (left) and glutamine (right) in contamination experiment samples. Error bars represent the standard deviation. The number of residues involved in this calculation is shown in each bar; plots produced through deamiDATE.

Figure 5

Site-specific deamidation in contamination experiment, where the y-axis represents the proportion of remaining deamidating amino acid, and the x-axis shows the half-time of the sites. The size of the points represents the intensity of the peptide; plots produced through deamiDATE.

Figure 6

BLG and casein identifications in samples with at least one peptide in reanalysed datasets from previously published works. The average number of unique BLG and casein peptides is shown per sample in each study for samples that have at least one peptide. The error bars represent standard deviation. The summed total number of asparagine and glutamine residues attributed to any milk peptides for all sample in each study is shown above the error bars.

Figure 7

Average glutamine deamidation per sample, (A) BLG, (B) Caseins. The estimated age of the samples (years before present) is on the x-axis. Plots produced through deamiDATE.

Figure 8

Site-specific deamidation of (A) BLG and (B) Caseins, where the points are coloured by the samples’ age, and the marker represents the paper from which the sample originated. Plots produced through deamiDATE.