Ongoing Modification of Mediterranean Pleistocene Sapropels Mediated by Prokaryotes

See allHide authors and affiliations

Science  28 Jun 2002:
Vol. 296, Issue 5577, pp. 2407-2410
DOI: 10.1126/science.1071893


Late Pleistocene organic-rich sediments (sapropels) from the eastern Mediterranean Sea harbor unknown, metabolically active chemoorganotrophic prokaryotes. As compared to the carbon-lean intermediate layers, sapropels exhibit elevated cell numbers, increased activities of hydrolytic exoenzymes, and increased anaerobic glucose degradation rates, suggesting that microbial carbon substrates originate from sapropel layers up to 217,000 years old. 16Sribosomal RNA gene analyses revealed that as-yet-uncultured green nonsulfur bacteria constitute up to 70% of the total microbial biomass. Crenarchaeota constitute a smaller fraction (on average, 16%). A slow but significant turnover of glucose could be detected. Apparently, sapropels are still altered by the metabolic activity of green nonsulfur bacteria and crenarchaeota.

Deep-sea sediments of the eastern Mediterranean are characterized by the cyclic occurrence of dark sediment layers, called sapropels. Sapropels differ from most other subsurface environments in that they contain high concentrations of total organic carbon (2 to 30.5% dry weight) (1), consisting mainly of dark brown amorphous, highly refractory kerogen (2). They are embedded in hemipelagic carbonate oozes that are poor in organic carbon (<0.5 weight percent) (3) and are likely to represent Pleistocene analogs of the widespread Mesozoic black shales (4). The presence of iso and anteiso fatty acids and of β-hydroxy fatty acids in the sapropels points toward a bacterial contribution to the organic matter (5). Because large numbers of bacteria and dividing cells have been detected in sapropels (1), the organic matter could still provide a source of carbon and energy for microbial life despite its age, and as a consequence, sapropels may still be subject to microbial alteration. To gain further insight into the fate of sapropel organic matter during diagenesis, we analyzed the composition and the physiological state of natural bacterial communities present in four sapropels (S1, S6, S7, and S8; Fig. 1, A and B), which are 8000, 172,000, 195,000, and 217,000 years old, respectively (6).

Figure 1

(A) Longitudinal section of core 69-2SL. Sapropels (S1 and S6 to S8) and intermediate layers (Z0, Z1, Z6, and Z7), which were sampled for microbiological analyses, are indicated. al denotes ash layer. (B) Detail of the longitudinal section of sapropel S8 revealing the fine lamination after the outer layer was frozen and lifted off (to the left of the break indicated by the arrow) (6). The top of the sapropel is oriented toward the left. (C) Epifluorescence photomicrograph of a bacterial microcolony as detected in sapropel S6 after acridine orange staining.

Acridine orange counting revealed that all four sapropels harbored large numbers of prokaryotes, which in some cases were present in microcolonies (Figs. 1C and 2A). Total cell counts (6 × 107 to 13 × 107cells cm−3) were significantly enhanced as compared to those in neighboring carbon-lean sediment layers (Fig. 2A). The frequency of dividing and divided cells in the different depths ranged between 6.7 and 13.9%. Elevated bacterial numbers and dividing cells have also been reported for older sapropels (1) and are a first indication of the presence of active microbial communities.

Figure 2

(A) Total bacterial cell counts (horizontal gray bars) and frequency of dividing cells (•) at eight consecutive depths. Bars indicate 1 SD. (B to D) Physiological activities of microorganisms in sapropels and carbon-poor intermediate layers of core 69-2SL. Exoenzymatic activities of (B) β-glucosidase (▴), (C) aminopeptidase (•) and alkaline phosphatase (○), and (D) respiration of D-[U-14C]-glucose (▾) are shown. Note the different scales of the abscissa in (B) and (C). Bars indicate 1 SD. Enzyme activities represent potential rates, because substrate analogs were used at saturating concentrations (250 μM). (E) Total organic carbon (▪) and total genomic DNA extracted (□). The positions of sapropels and sampling depths are denoted on the right.

To gain more insight into the potential modification of sapropel organic matter by microorganisms, we focused on the initial steps of the anaerobic food chain, choosing hydrolytic exoenzymes as indicators of physiologically active prokaryotes. Extracellular enzymatic hydrolysis of biopolymers is the rate-limiting step for the use of organic matter in surface aquatic environments (7). Some exoenzymes, such as alkaline phosphatase and β-glucosidase, are subject to substrate induction and catabolite repression, so that short-term measurements of cell-specific hydrolysis rates also provide information on the actual availability of bacterial substrates (8, 9). We determined the activities of the three exoenzymes alkaline phosphatase, β-glucosidase, and leucine aminopeptidase, using a method recently established for sapropels (6). As an additional indicator of physiological activity, we used the degradation of radiolabeled glucose to14CO2 (6). Exoenzymatic activities in sapropels are associated with intact cells (10).

Of the three exoenzymes, the activity of β-glucosidase was very low throughout the sediment core (Fig. 2B). The cell-specific β-glucosidase activities [0.01 to 2.3 attomoles (amol) cell−1 hour−1] fall well within the range observed for other marine sediments and pelagic water samples (10). Aminopeptidase activity and glucose degradation rates reached their highest values at the sediment surface. However, the activities of alkaline phosphatase and leucine aminopeptidase, as well as the glucose degradation rates, were consistently and significantly elevated within the sapropels as compared to intermediate layers (Fig. 2, C to D). In contrast to the sapropels, exoenzyme activities decrease rapidly with depth in other marine sediments (11–13), following gradients of easily degradable substances. The distinct maxima in the sapropels are therefore unexpected and cannot be explained by downward diffusion of easily degradable organic carbon. Instead, our data strongly indicate that carbon provided by the sapropel layers themselves must support bacterial metabolism in situ. This conclusion is further supported by the values of cell-specific activities of leucine aminopeptidase, which are significantly increased by a factor of 36 in sapropels S7 and S8 (10.4 to 12.0 amol cell−1 hour−1) over those in the intermediate layers (0.2 to 0.4 amol cell−1h−1). These data indicate higher physiological activity of the cells in the sapropels (10).

Kerogens are assumed to be highly resistant to microbial attack under anoxic conditions (14), and adsorption can preserve intrinsically labile molecules such as amino acids and sugars against microbial degradation (15). The degradation of14C-glucose and the β-glucosidase activity detected suggest that glucose is one of the substrates for microbial growth in sapropels. Therefore, we determined the turnover of glucose under in situ conditions by adding radiolabeled glucose at different concentrations to anoxic sediment slurries prepared from sapropel S6 (6). The turnover time of glucose at in situ concentrations (21.6 μM) as calculated from this experiment was 180 days (corresponding to a first-order rate constant of degradation ofk = 0.0056 day−1). This value is considerably shorter than that for bulk organic matter in other marine sediments [turnover time, 537 years (15)], thus indicating that reactive organic carbon compounds are still present in the kerogen and still subject to microbial degradation.

To identify the microorganisms potentially involved in transformations of sapropel organic matter, the dominant 16S ribosomal RNA (rRNA) gene sequences were analyzed. The only sediments containing old organic carbon for which bacterial communities have been characterized are Pacific deep subsurface sediments. So far, numerous barophilic sulfate-reducing bacteria and a few Proteobacteria and Crenarchaeota have been identified by molecular or cultivation-based methods (16, 17).

The total amounts of DNA extracted (6) from the four sapropels exceeded those in the intermediate layers 5- to 250-fold (Fig. 2E). Even the 217,000-year-old Mediterranean sapropel S8 contained DNA fragments up to 30 kb long. Its high molecular weight indicates that the majority of the extracted DNA originates from extant and not subfossil prokaryotes, because hydrated DNA found in other aquatic sediments is split into short (100 to 400 base pairs long) fragments within a few thousand years (18).

Analyses of 16S rRNA genes (6) by polymerase chain reaction (PCR), denaturing gradient gel electrophoresis (fig. S1), and sequencing yielded 12 partial sequences that clustered with the environmental clone T78 group of the green nonsulfur bacteria division and were distantly related to as-yet-uncultured bacteria found in deep subsurface habitats (fig. S2). Only a single sequence clustered within the δ group of the Proteobacteria, showing the greatest sequence homology (89.5%) toGeobacter sulfurreducens (American Type Culture Collection number 51573T) (fig. S2). Amplification of archaeal 16S rRNA gene fragments yielded a total of 41 different phylotypes in the sapropels and intermediate layers. Most (that is, 36) of these sequences were affiliated to the group of nonthermophilic marine Crenarchaeota (fig. S3). Most notably, all but one of the 16S rRNA gene sequences from the sapropel layers were affiliated to this group. Archaeal DNA constituted an average of 15.9% of the total genomic DNA (Fig. 3). The abundance of archaea in the eastern Mediterranean sediment thus is higher than in other deep-sea sediments, for which values between 2.5 and 8% have been reported (19). By comparison, green nonsulfur bacteria constituted up to 69% of the total microbial community in sapropel S6 and significantly surpassed the abundance of archaea in five out of the eight layers investigated. On average, green sulfur bacteria and archaea together accounted for 66% of the total microbial community in sapropels and reached a maximum of 84% in sapropel S6 (Fig. 3).

Figure 3

Relative abundance of bacteria of the clone T78 group (▪) and archaea (□) as determined by real-time PCR. Bars indicate 1 SD. The dotted line gives the sum of the abundances of green sulfur bacteria and archaea.

With only two exceptions, none of the numerous members of the clone T78 group has ever been cultured. One isolate of the clone T78 group isDehalococcoides ethenogenes strain 195, which exhibits very restricted physiological capacities and grows exclusively by oxidation of H2, with concomitant reduction of chlorinated hydrocarbons (20). Recently, a thermophilic filamentous member of the clone T78 group was described, which ferments glucose or sucrose in the presence of yeast extract (21). Therefore, no information is available on the biogeochemical significance of nonthermophilic members of the clone T78 group (22). In the Mediterranean sediments, the considerable cell-specific exoenzyme activity, together with the glucose degradation, point toward a fermentative growth mode of these bacteria. So far, members of the clone T78 group have been detected in hydrothermal springs (such as the OPB clones), soil, wastewater, and subsurface environments such as paleosols (the H and TO clones, fig. S2) (22), but their 16S rRNA gene sequences usually constitute only a minor fraction (2.5%) of environmental clone libraries (23). The dominance of the T78 group in the Mediterranean deep-sea sediments is thus unprecedented and suggests an adaptation of these bacteria to the specific conditions in sediments containing kerogen. This view is reinforced by the fact that bacteria of the clone T78 group are much less abundant at the sediment surface than in sapropels and intermediate layers (Fig. 3).

A first insight into the mechanisms of adaptation can be obtained from a calculation of the theoretical energy demand of cells in sapropels. The lowest maintenance energy requirement of anaerobic bacteria has been determined for Acetobacterium woodii and is equivalent to 0.01 kJ [grams of carbon (gC) hour]−1(24). At a mean cell number of 9.3 × 107 cm−3 in the sapropels (Fig. 2A) and a mean cell diameter of 0.7 μm, the concentration of microbial biomass amounts to 3.4 × 10−6 gC cm−3(25) and its maintenance energy demand to 2.9 × 10−4 kJ cm−3 year−1. Because kerogen is composed mainly of long chains of polymethylenic carbon (14), we used the free energy change determined for the methanogenic degradation of hexadecane of –1.95 kJ gCdegraded −1 (26) (the corresponding value for degradation under conditions of sulfate reduction is –3.33 kJ gCdegraded −1). If the prokaryotes present had a maintenance energy requirement similar to that of known laboratory strains, sapropel organic carbon would thus be degraded at a rate of (0.88 to 1.50) × 10−4 gC cm−3 year−1. As a consequence, even sapropels consisting entirely of organic carbon (∼0.9 gC cm−3) would be completely degraded within 10,000 years. Organic carbon compounds in sapropels have been preserved over much longer time intervals, although a high fraction of microbial cells in the sapropels are physiologically active and continue to use organic carbon originating from the sapropels. The above comparison thus indicates that prokaryotes in sapropels have significantly lower maintenance energy requirements than any of the pure cultures investigated to date.

Mediterranean sapropels harbor large populations of previously unknown members of the green nonsulfur bacteria and crenarchaeota. Our cumulative evidence suggests that these prokaryotes are physiologically active, are specifically adapted to the specific conditions as they prevail in sediments with large amounts of subfossil kerogen, and are capable of altering the organic matter in situ even 217,000 years after its deposition.

  • * To whom correspondence should be addressed. E-mail: j.overmann{at}


View Abstract

Navigate This Article