LETTER doi:10.1038/nature10905 Endospore abundance, microbial growth and necromass turnover in deep sub-seafloor sediment Bente Aa. Lomstein 1 , Alice T. Langerhuus 1 , Steven D’Hondt 2 , Bo B. Jørgensen 3 & Arthur J. Spivack 2 Two decades of scientific ocean drilling have demonstrated wide- spread microbial life in deep sub-seafloor sediment, and surprisingly high microbial-cell numbers. Despite the ubiquity of life in the deep biosphere, the large community sizes and the low energy fluxes in this vast buried ecosystem are not yet understood 1,2 . It is not known whether organisms of the deep biosphere are specifically adapted to extremely low energy fluxes or whether most of the observed cells are in a dormant, spore-like state 3 . Here we apply a new approach—the D:L-amino-acid model—to quantify the distributions and turnover times of living microbial biomass, endospores and microbial necromass, as well as to determine their role in the sub-seafloor carbon budget. The approach combines sensitive analyses of unique bacterial markers (muramic acid and D-amino acids) and the bacterial endospore marker, dipicolinic acid, with racemization dynamics of stereo-isomeric amino acids. Endospores are as abund- ant as vegetative cells and microbial activity is extremely low, lead- ing to microbial biomass turnover times of hundreds to thousands of years. We infer from model calculations that biomass production is sustained by organic carbon deposited from the surface photo- synthetic world millions of years ago and that microbial necromass is recycled over timescales of hundreds of thousands of years. Deep sub-surface sediment material was obtained from the eastern tropical Pacific during the Ocean Drilling Program (ODP) Leg 201 expedition with DS JOIDES Resolution. Coring sites ranged from the continental shelf off the coast of Peru to ocean depths of 5,000 m. The expedition recovered sediment at depths of up to 420 metres below the sea floor (mbsf), and this sediment was found to be up to 35 million years old. We report estimates of high bacterial endospore numbers in the order of 10 7 endospores per cm 3 (Fig. 1) in deep sediment and sediment that is up to 10 million years old. We used two different analyses to quantify endospores: muramic acid, which is a unique building block in cell walls of both bacteria and endospores; and dipicolinic acid, which is uniquely formed by endospores. In the first analysis, muramic acid of endospores is calculated as the total muramic acid (Supplementary Fig. 1) minus muramic acid from vegetative cells. Vegetative cells (and intact but recently dead cells) were quantified by acridine orange direct counts (AODCs) 4 . Endospore muramic acid is converted to endospore numbers by the use of cell-specific conversion factors that are obtained from the literature (Supplementary Information). It remains unclear whether bacteria or archaea predominate in the studied sediment 5,6 . We therefore use two extreme scenarios for our estimates of endospore abundance, either that bacteria completely dominate 6 (bacterial dominance) or that 90% are archaea 5 (archaeal dominance). Archaea do not contain muramic acid in the cell wall. We also assume that 35% of the bacteria are Gram positive and 65% are Gram negative, as was found throughout the sediment column at Site 1229 (ref. 7). Muramic-acid-based endospore numbers are 0.2 3 10 7 to 3 3 10 7 endospores per cm 3 . Archaeal or bacterial dominance did not affect the estimated endospore numbers because muramic acid levels calculated from AODC only marginally contributes to the measured muramic acid concentrations (Fig. 1). At Site 1227 we also estimated endospore numbers from dipicolinic acid concentrations assuming an average dipicolinic-acid content of 2.2 3 10 216 mol spore 21 (ref. 8). Dipicolinic-acid-based endospore numbers are 0.3 3 10 7 to 1.0 3 10 7 endospores per cm 3 and confirmed the muramic-acid-based estimates by a mean deviation factor of 4.6. There was a significant positive correlation between the two endospore estimates from muramic acid and dipicolinic acid as judged from a statistical analysis of their depth trend (regression of muramic acid and dipicolinic acid; P 5 0.0008; R 2 5 0.7297). To our knowledge, this is the first time that such com- parative data are published for endospores in environmental samples. This discovery of high endospore abundances raises the question of why endospores have not been detected previously and whether total cell abundance in the deep biosphere, including endospores, has been globally underestimated. Previous studies of the deep biosphere did not quantify endospores. Endospores are unlikely to be stained by fluor- escent DNA dyes such as acridine orange 9,10 or by ribosomal RNA staining techniques such as catalysed reporter deposition fluorescence in situ hybridization because endospore walls are impermeable 6 . In this study we processed samples with 3 N HCl (for 4 h at 95 uC), which com- pletely extracts muramic acid and dipicolinic acid from endospores 11 . Our results suggest that endospores are as abundant as vegetative prokaryotes in this deep marine biosphere. Studies in other oceano- graphic regions are needed to clarify how the earlier global estimate of 3.5 3 10 30 cells 12 should be adjusted to account for this. 1 Department of Bioscience, Section for Microbiology, Aarhus University, Building 1540, Ny Munkegade 114, DK-8000 Aarhus C, Denmark. 2 Graduate School of Oceanography, University of Rhode Island, 100A, Horn Building, Narragansett, Rhode Island 02882, USA. 3 Center for Geomicrobiology, Department of Bioscience, Aarhus University, Building 1535, Ny Munkegade 114, DK-8000 Aarhus C, Denmark. 5 6 7 8 9 10 0 50 100 150 200 250 300 Cells (log 10 cm -3 ) Cells (log 10 cm -3 ) Cells (log 10 cm -3 ) Depth (mbsf) Site 1227 6 7 8 9 10 Site 1229 6 7 8 9 10 Site 1230 Figure 1 | Profiles of AODCs and estimated endospore numbers on the Peruvian continental shelf (sites 1227 and 1229) and in the Trench in Peru (site 1230). Filled circles, AODC 4 ; coloured open squares, muramic-acid- based estimated number of endospores (bacterial dominance); crosses, muramic-acid-based estimated number of endospores (archaeal dominance); black open squares, dipicolinic-acid-based estimated number of endospores (n 5 6). Grey shading, sulphate–methane transition zones. Error bars, s.d. 5 APRIL 2012 | VOL 484 | NATURE | 101 Macmillan Publishers Limited. All rights reserved ©2012