Vector analysis of ecoenzyme activities reveal constraints on coupled C, N and P dynamics

• Single, 2- and 3-way comparisons of enzyme activities reflect hierarchical microbial C, N and P demands.
• Vector length and angle quantifies relative C, N and P enzyme activities.
• Vector length indicates relative C vs. nutrient limitation.
• Vector angle indicates relative P vs. N limitation.
• Single enzyme activity has no clear relationship to vector characteristics.

The primary goal of this study was to devise a quantitative method of interpreting the simultaneous microbial C, N, and P acquisition through the activities of four, key extracellular enzymes, 1,4-β-glucosidase (BG), leucine aminopeptidase (LAP), 1,4-ß-N-acetylglucosaminidase (NAG) and acid/alkaline phosphatases (AP). To this end, the proportional activity of C vs. N acquiring enzymes (BG/[BG + NAG + LAP]) was plotted against C vs. P acquiring enzymes (BG/[BG + AP]). We then calculated the length and angle of the vector created by connecting a line between the plot origin and point represented by these proportions; the length quantifies relative C vs. nutrient limitation and the angle quantifies the relative P vs. N limitation. Analyses of large data sets obtained from soil, freshwater periphyton and aquatic sediments revealed that logarithmic, arcsine, arcsine-square-root and logit transformations did little to improve the statistical distribution of data over raw proportions. More importantly, the vector characteristics of enzyme activity loci are much easier to interpret for raw proportions than when data have been previously transformed. Analyses also revealed the importance of using consistent methods, i.e., omitting NAG assays for an acidic soil led to overestimates of P limitations, and the importance of understanding the nature of the system, i.e., soils of the Antarctic Dry Valleys had low BG activities, likely because there is little to no local production of cellulose. Further, analyses of four sites in Luquillo Forest, Puerto Rico, showed no differences among sites in relative C demand despite differences in BG activities, and higher P demand in the cloud than lower montane forest, despite higher AP activity at the latter site. Finally, analyses of three decomposing litter types over time revealed contrasting patterns of change in relative C, N and P demand over time, reflecting both stoichiometric and C quality effects on decomposer communities. Thus enzyme activity vectors reflect the simultaneous, relative resource demands of the microbial community independent of variations in total enzyme activity, such as may result from variations in total microbial biomass.