INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Viruses are now considered to be an important component of aquatic microbial communities. The reevaluation of the virus role in marine ecosystems is mostly due to discovery of very abundant viral densities in all pelagic systems, densities that generally exceed bacterial densities by 1 to 2 orders of magnitude (30). Due to their high density (10⁹ to 10¹⁰ liter¹) and their high capability of infecting bacteria and phytoplankton, viruses may have profound effects on microbial loop dynamics (14, 29).

An increasing amount of information indicates that viruses might be responsible for ca. 10 to 30% (up to 72%) of bacterial mortality in aquatic systems (18). These processes have cascade effects also on the biogeochemical cycling of organic matter in the marine environment (24). The ability of viruses to destroy bacteria (through infection followed by host cell lysis) is recognized as one of the most relevant mechanisms of DOM release (including dissolved DNA [14]), thus affecting nutrient recycling and the pathways of organic carbon utilization by bacteria.

Previous studies on virus-bacterium interactions in pelagic systems have demonstrated that viral density and capability of infecting bacteria (the so-called "viral-loop" control) increase with increasing trophic conditions, with the highest percentage of infected cells being in highly eutrophic systems (31). However, all studies dealing with virus distributions have been surprisingly restricted to the plankton domain and, excluding one study from Canadian lake sediments (20) and one from interstitial waters (12, 30), no information is available on virus concentration and distribution in the entire sediment matrix.

Epidemiological models predict that viral infection increases with increasing host cell density (32). In this regard, marine sediments could represent the optimal environment for virus development and for testing the hypothesis of a stronger viral loop control with increasing eutrophication. In fact, marine sediments are characterized by (i) high organic matter concentrations, ca. 5 orders of magnitude higher in the sediments than in the water column (ca. 10 to 30 mg of C ml of wet sediment¹ versus 100 to 300 µg of C liter¹ in the water column), (ii) high bacterial densities, ca. 3 orders of magnitude higher than in the water column (generally from 10⁸ to 10⁹ ml of sediment¹ versus 10⁸ to 10⁹ liter¹), and (iii) as a result benthic bacteria display the lowest intercell distance and the probability of virus-bacterium contact is 3 orders of magnitude higher.

Among marine environments, deep-sea sediments, covering about 50% of the world's surface, are responsible for processes controlling organic carbon cycling on a global scale, and knowledge of the mechanisms regulating benthic bacteria is essential for a better understanding of organic matter diagenesis (10). Deep-sea benthic bacteria are assumed to be controlled "bottom-up" (by the flux of organic carbon from the photic layer [11] and by organic matter availability [5]) and/or "top-down" (e.g., by nanoflagellate grazing [7]) and possibly by viral infection, although this latter hypothesis is yet to be tested. In this study we examined deep-sea sediment viral and bacterial densities and the biochemical composition of the sedimentary organic matter (measured as lipids, proteins, carbohydrates, and phytopigments), comparing highly oligotrophic bathyal sediments with deep-sea trenches characterized by benthic hot spots of microbial activity (3). The specific aims of this investigation were (i) to gather information about the relative significance of bacteria and viruses in deep-sea sediments and (ii) to identify their possible interactions and the role of different environmental conditions in controlling the virus-to-bacterium density ratio.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Study area and sampling. Sediment samples were collected in the Aegean Sea (Eastern Mediterranean) between 28 December 1997 and 18 January 1998. The sampling strategy included three areas characterized by different ecological conditions: the Sporades Basin (average depth of sampled area, 1,232 m), the Ierapetra Basin (average depth of sampled area, 4,235 m), and the Cretan Sea (average depth of sampled area, 1,840 m) (Table 1). The Sporades Basin is located in the northwestern Aegean and is characterized by nutrient-rich waters coming from the Black Sea. The Cretan Sea is considered to be one of the most oligotrophic areas of the Mediterranean, due to the extremely low primary productivity (7, 13). Finally, the Ierapetra Basin is located southeast of Crete and is considered a trap for particulate matter and a benthic microbial "hot-spot" area (3). The sampling schedule included three stations in the Sporades Basin, four stations in the Ierapetra Basin, and two stations in the Cretan Sea (Table 1). Undisturbed sediment cores were collected using a multicorer. Immediately after sampling, sediment corers were vertically sectioned into six layers: 0 to 0.3, 0.3 to 1, 1 to 2, 2 to 4, 4 to 6, and 6 to 10 cm in depth. For bacterial and viral analysis, replicate subsamples (n = 3) of ca. 0.5 ml were added to 3 ml of prefiltered (0.02-µm [pore size]) seawater containing 2% formalin and stored at 4°C.

Phytopigment analysis. Chloroplastic pigments (chlorophyll a and pheopigments) were determined according to the method of Lorenzen and Jeffrey (19). Pigments were extracted from ca. 1 g of sediment with 90% acetone overnight. After centrifugation (800 × g), the supernatant was used to determine the chlorophyll a concentration and then acidified with 0.1 N HCl to estimate pheopigments using a Perkin-Elmer fluorometer (model LS50B).

Biochemical composition of sediment organic matter. Total sedimentary carbohydrates were analyzed according to the method of Gerchacov and Hatcher (15). Glucose solutions were used as a standard. Proteins were determined according to the method of Hartree (17) as modified by Rice (26) to compensate for phenol interference. Albumin solutions were used as a standard. For each analysis ca. 0.5 g of sediment was used. Lipids were extracted from dried sediment samples by direct elution with chloroform and methanol. Analyses were carried out using the methods of Bligh and Dyer (2) and Marsh and Weinstein (21). All analyses were carried out on ca. 0.5 g of sediment previously sonicated for 3 h in Milli-Q water to increase the extraction efficiency. Sediments treated in the muffle furnace (450°C, 2 h) were used as blanks for all analyses, and these analyses were performed in three to six replicates per sediment layer. Biopolymeric carbon (BPC) concentrations were calculated by converting protein, carbohydrate, and lipid concentrations into carbon equivalents assuming the conversion factors (4) of 0.49, 0.40, and 0.75 µg of C µg¹, respectively.

Viral and bacterial abundance. Bacterial and viral counts were made from the same sample using 0.02-µm-pore-size filters. Each sample was treated with pyrophosphate (0.01 M) for 15 min (6, 20), which was efficient for both bacteria and viruses, and sonicated three times (Branson Sonifier 2200; 60 W for 1 min in an ice bath) to optimize extraction. The concentration of pyrophosphate was selected after testing different concentrations down to 0.001 M. Subsamples were diluted 100 to 500 times. Aliquots of the subsamples were stained with SYBR-1 (22) and filtered on Anodisc Al₂O₃ filters (0.02-µm pore size). The filters were analyzed by epifluorescence microscopy using a Zeiss Axioplan microscope equipped with a 50-W lamp. From 10 to 50 fields were viewed, and a minimum of 400 cells were counted for both viruses and bacteria. Detection limits for viral density were calculated on the basis of no cells encountered in more than 50 fields (equivalent to a density lower than 4 × 10³ viruses g of sediment¹). Viruses were also analyzed by transmission electron microscopy (TEM) to gather information on their morphology and characteristics. Using the method of Maranger and Bird (20), ca. 3 ml was centrifuged at 3,000 × g for 30 min. A subaliquot of 1 ml was diluted 1- to 100-fold and centrifuged at 100,000 × g directly for 30 min onto a 400-mesh Formvar-coated Cu and stained with uranyl acetate. Counts were done using electron microscopy at magnifications of ×50,000 to ×80,000. Methodological aspects of these experiments will be discussed elsewhere (R. Danovaro, A. Dell'Anno, A. Trucco, M. Serresi, and S. Vanucci, manuscript in preparation).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Environmental parameters. No clear differences were observed between sampling areas in terms of water content and porosity (Table 1). The water content of the superficial layer (0 to 0.3 cm) for the three investigated areas ranged from 32.7% (in the Ierapetra Basin) to 34.4% (in the Cretan Sea). Differences in average depth (4,235 versus 1,840 m, for the Ierapetra Trench and the Cretan Sea) did not influence sediment compactness. Porosity varied with water content, ranging from 0.56 in the Ierapetra Basin to 0.58 in the Cretan Sea.

Phytopigments and biochemical composition of sedimentary organic matter. Chloroplastic pigment concentrations in the sediments are illustrated in Fig. 1. The chlorophyll a concentration was, on average, 0.04 µg g¹ in all of the investigated areas. At all stations, total pheopigment concentrations were significantly higher than those for chlorophyll a. Organic matter biochemical composition is illustrated in Fig. 2. Carbohydrates were the dominant biochemical class of organic compounds and, in the top 3 mm of the sediment, displayed concentrations more than double for the Ierapetra Trench with respect to the Cretan Sea (on average, 2,406 and 1,092 µg g¹, respectively). The second main biochemical class was represented by proteins that followed a gradient of decreasing productivity ranging from 479 to 760 µg g¹ (at the Ierapetra Trench and the Sporades Basin, respectively). Finally, lipids displayed the same trend reported for proteins with the highest values at the Sporades Basin (on average, 279 µg g¹) and the lowest values at the Ierapetra Trench (on average, 163 µg g¹). Soluble compounds generally accounted for a minor fraction of carbohydrates and lipids (5 to 20% [data not shown]) but represented the main protein fraction, accounting always for more than 80% of the total protein concentrations.

View larger version (25K): [in this window] [in a new window]   FIG. 1.   Vertical distribution of phytopigment concentrations determined fluorometrically at the three sampling areas. Shown are results for chlorophyll a and pheopigments at the Sporades, Ierapetra, and Cretan locations. Average values for each basin are indicated (n = 3 to 6). Data are expressed as micrograms per gram of sediment (dry weight).

View larger version (23K): [in this window] [in a new window]   FIG. 2.   Vertical distribution of the main biochemical classes of organic compounds in the sediments of the deep Eastern Mediterranean. Shown are results for proteins, carbohydrates, and lipids. Average values for each basin are indicated (n = 3 to 6). Data are expressed as micrograms per gram of sediment (dry weight).

Benthic bacteria and viruses. Data on bacterial and viral densities in the sediments of the Eastern Mediterranean are reported in Fig. 3. Bacterial densities at the Sporades Basin were ca. twice those at the Ierapetra Trench, on average (11.0 × 10⁸ versus 4.9 × 10⁸ cells g of sediment¹, respectively), and almost threefold higher than in the Cretan Sea (on average, 4.0 × 10⁸ cells g¹). Similar patterns were observed for the bacterial biomass (data not shown). The viral densities displayed very similar values in the Sporades Basin and the Cretan Sea (23.75 × 10⁸ and 20.2 × 10⁸ cells g of sediment¹, respectively), with densities about double that at the Ierapetra Trench (on average, 12.1 × 10⁸ cells g¹). As a result, the virus-to-bacterium density ratio was highest in the Cretan Sea (on average, 5.3), with values approximately twice that in the Ierapetra Trench and the Sporades Basin (on average, 2.2 and 2.6, respectively). Vertical patterns indicated a clear decrease of both bacterial and viral densities with depth in the sediment, but the viral-abundance decrease was much sharper, leading to a strong decrease of the virus-to-bacterium ratio with increasing sediment depth. At one of the four Ierapetra Trench stations, the virus density under the 4-cm depth was below detection limits.

View larger version (24K): [in this window] [in a new window]   FIG. 3.   Vertical distribution of microbial parameters in the deep Eastern Mediterranean. Shown are the bacterial densities (cells per gram of sediment [dry weight]) and viral densities (viruses per gram of sediment [dry weight]). Average values for each basin are indicated. The data are expressed as viruses or bacteria (108) per gram of sediment (dry weight).

Statistical analysis. A Spearman rank correlation analysis was carried out to test the relationships among bacteria, viruses, and environmental parameters. The bacterial density (log transformed) was significantly correlated with viral density (n = 9, r = 0.805, P < 0.01), soluble carbohydrates (r = 0.734, P < 0.01), and total (r = 0.786, P < 0.01) and soluble (r = 0.662, P < 0.05) proteins. Virus density values were significantly correlated with total protein concentrations (r = 0.679, P < 0.05). Differences among stations were tested by using one-way analysis of variance (ANOVA) at the 5% level, after testing of the homogeneity of variance.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Trophic conditions in deep-sea sediments of the Eastern Mediterranean. It has been recently proposed that viral density and bacterial mortality due to viral infection increase with the degree of eutrophication (31). In this study we compared deep areas characterized by evident differences in trophic conditions. The three sites were different both in organic matter content (as indicated by changes in biopolymeric carbon concentrations) and in quality (as highlighted by changes in the relative significance of protein and carbohydrate pools). A clear gradient of increasing trophic conditions was observed along the axis of the North to the South Aegean (i.e., moving from the bathyal Sporades Basin to the abyssal Ierapetra Trench). This gradient reflected differences in primary productivity and particle fluxes previously reported for the North and South Aegean (9). Organic matter concentrations (expressed as protein, carbohydrate, and lipid content of the sediment) were generally low compared with equally deep sediments in temperate coastal areas and similar to those reported in productive environments (5). A specific feature of sedimentary organic matter composition in oligotrophic deep-sea environments is the dominance of carbohydrates over other biochemical components, resulting in a very low protein-to-carbohydrate ratio (5).

Virus and bacterium abundance and distribution in deep-sea sediments. It is generally recognized that bacterial distribution is largely dependent upon the amount of utilizable organic matter in the sediments, which in turn is largely controlled by the sedimentation and degradation rates in the water column (11). The data reported here are consistent with this rule, and the bacterial density is significantly correlated with the amounts of labile organic compounds (5, 7, 9, 10). However, our data from deep-sea environments contrast with most studies on bacterial distribution in marine sediments, which have reported quite-conservative bacterial densities in marine sediments, when densities are expressed with respect to fluid volume (data not shown) instead of to dry sediment mass (27). Bacterial densities were significantly higher at the Sporades Basin than in other sampling areas (ANOVA, P < 0.01) and were correlated to organic-matter content and viral density (Fig. 4). Similar relationships have been observed in lake sediments (20) and in pelagic systems (31).

View larger version (12K): [in this window] [in a new window]   FIG. 4.   Relationship between bacterial and viral densities in deep-sea sediments of the Eastern Mediterranean.

1

1