Needles in the Blue Sea: Sub-species Specificity by Targeted Metaproteomics of the Vast Oceanic Microbial Metaproteome

Targeted metaproteomics is the application of mass spectrometry-based quantitative protein measurements to natural microbial populations. It has significant potential for providing novel biogeochemical insights, for example by measurement of specific biomarkers indicative of nutrient stress, as well as direct measurements of enzymes that can be used to provide estimates of potential reaction rates. High microbial diversity presents unique challenges to this method. We examined the feasibility of a targeted metaproteomics workflow in Pacific Ocean environments for two cyanobacterial nitrogen regulatory proteins, NtcA and P-II. Genomic analyses using new METATRYP software found the number of shared (redundant) tryptic peptides between different marine bacteria species to typically being 1% or less. Closely related cyanobacteria Prochlorococcus and Synechococcus shared an average of 4.8+1.9% of their tryptic peptides, while shared intraspecies peptides were higher, with 13+15% shared peptides. Measurements of an NtcA peptide in the Pacific Ocean was found to target multiple cyanobacteria species, whereas a P-II peptide showed sub-species specificity to the high-light Prochlorococcus ecotype. Distributions of NtcA and P-II in the Central Pacific Ocean were similar except at the Equator likely due to differential nitrogen stress responses between Prochlorococcus and Synechococcus. The number of unique tryptic peptides coded for within three combined oceanic microbial metagenomes was estimated to be ~4e7, 1000-fold larger an individual microbial proteome and 27-fold larger than the human proteome, yet still 20 orders of magnitude lower than the peptide diversity possible in all protein space, implying that peptide mapping algorithms should be able to withstand the added level of complexity in metaproteomic samples. These oceanic quantitative protein distributions in the oceans demonstrate sub-species resolution is achievable combining in silico and empirical approaches.