I recently completed a pilot study with collaborators at James Cook University, Australia to evaluate the feasibility of using RP-HPLC to analyze single foraminifer (Hearty et al., in press). I analyzed foraminifera from seven marine sediment cores taken from the Queensland shelf, northeast Australia. Two chromatograms of RP-HPLC analysis illustrate the typical analytical performance of the technique (Fig. 6). Eight amino acid pairs are separated with baseline resolution. I focused on two amino acids: aspartic acid (Asp) and glutamic acid (Glu). More accurately, the Asp and Glu probably include a small component of their amidated forms, asparagine and glutamine, respectively, which were converted to Asp and Glu during laboratory hydrolysis. These two amino acids are among the most abundant in foraminifer protein, are eluted in the first 20 min of the RP-HPLC analysis, and exhibit the highest analytical reproducibility (Kaufman and Manley, 1998).
The concentrations of amino acids in fossil foraminifer are intermediate between those of ostracodes and gastropods. In contrast to the other two types, foraminifer that differ in age by hundreds of thousands of years contain similar concentrations of amino acids. This suggests that amino acids in Quaternary foraminifer (Pulleniatina) from marine sediment are retained more completely than in ostracodes or gastropods from terrestrial deposits. By retaining amino acids, a foraminifer test comes closer to approximating a closed system, which simplifies and adds confidence to models of racemization kinetics.
I analyzed between 5 and 16 (average = 8) tests from each of the 45 samples (367 total). Some samples comprised tests with D/L ratios that were much different than the others of the group. The higher-than-expected ratios might indicate reworking of older tests into younger sediment; the lower-than-expected ratios might indicate contamination by secondary amino acids. Because RP-HPLC separates multiple D and L amino acids, the covariance of D/L ratios could be used as a cross-check to identify and exclude aberrant results. Asp and Glu D/L ratios strongly co-vary in the foraminifera (Fig. 7), with a few subsamples clearly deviating from this trend. Of the 367 subsamples analyzed in this study, 36 (9.8%) were rejected. Following data screening, the inter-shell variability in Asp and Glu D/L averaged 8 and 14%, respectively, for the 45 samples (the standard error of the means are considerably lower). This is similar to the inter-shell variation reported for single-age populations of mollusks (Miller and Brigham-Grette, 1989) and ostracodes (Kaufman, 2003a, 2003b).
Downcore trends in D/L afford a fundamental test of the integrity of the amino acid data for geochronology. Cores PC10 and PC42 were taken 50 km apart from similar depths and oceanographic settings. The stratigraphy and sedimentation rates of these two cores are similar (Dunbar et al., 2000), and the downcore progressions of D/L ratios are identical (Fig. 8). A higher-order assessment of the geochronological integrity of the amino acid data involves the comparison of D/L ratios with independently derived ages. D/L ratios were determined for 14 samples whose ages are known from previously analyzed 14C ages. In all cases, the D/L ratios are concordant with 14C age. Beyond the range of 14C, previously determined marine oxygen-isotope stage boundaries provide approximate ages of the sediments up to about 500 ka (Dunbar et al., 2000). Again, the D/L ratios vary systematically with isotope-correlated ages. In detail, however, slightly different racemization rates for foraminifers in each core are implied from the 14C comparison. This difference is most apparent for horizons about 15 ka old in cores PC16, PC14, and PC42, and might result from differences in bottomwater temperature history among the core sites.
