Estimates of protein in coastal marsh soils: A case study of the utility of the Bradford assay for quantifying soil protein

•Bradford-based protein estimates correlated with extracted nitrogen and amino acids.•Bradford assay tended to grossly overestimate protein in the soil extracts.•Bradford overestimation was correlated with the carbon:nitrogen ratio of extracts.•Bradford overestimation always increased after the extracts were diluted. The fraction of soil protein extracted by autoclaving soil (AESP) has emerged as a potential indicator of available organic nitrogen (N), aggregate stability, and soil carbon (C) storage potential. Estimates of AESP are often based on colorimetric assays—primarily the Bradford assay—yet the results of these assays can be confounded by interference from polyphenolic contaminants of soil protein extracts. The extent to which polyphenolic interference in the Bradford assay may contribute to over- or underestimation of AESP content and whether a correction can be made for the interference has remained unclear. Here, we present a case study of how Bradford-based estimates of AESP compared to maximum and minimum estimates of AESP across two chronosequences of marsh soils that ranged widely in soil organic matter concentration (20–580 mg·g−1). Our maximum estimates were based on the amount of N in the dialyzed extracts; our minimum estimates were based on the proportion of the extracted N that was accounted for by hydrolysable amino acids. We observed that the Bradford-based estimates usually followed trends that were qualitatively similar to those of the maximum and minimum estimates across different marshes. However, the Bradford-based estimates were almost always greater than the maxima (by up to 450%) and never below the minima. This extent of overestimation was positively correlated with the C:N ratio of the extracts, but the relationships between the overestimates and C:N ratios differed between the two chronosequences. Contrary to what has been previously suggested, the dilution of extracts only exacerbated the extent of overestimation in the Bradford assay. Instead of a continued reliance on the Bradford assay for AESP measurements, we propose that the N content of semi-purified soil extracts serve as a surrogate metric for AESP, primarily because it may be more amenable to inter-study comparisons while still being feasible for most soil laboratories.