Previous studies have provided evidence that dark inorganic carbon fixation is an important process for the functioning of the ocean interior. However, its quantitative relevance and ecological significance in benthic deep-sea ecosystems remain unknown. We investigated the rates of inorganic carbon fixation together with prokaryotic abundance, biomass, assemblage composition, and heterotrophic carbon production in surface sediments of different benthic deep-sea systems along the Iberian margin (northeastern Atlantic Ocean) and in the Mediterranean Sea. Inorganic carbon fixation rates in these surface deep-sea sediments did not show clear depth-related patterns, and, on average, they accounted for 19% of the total heterotrophic biomass production. The incorporation rates of inorganic carbon were significantly related to the abundance of total Archaea (as determined by catalyzed reporter deposition fluorescence in situ hybridization) and completely inhibited using an inhibitor of archaeal metabolism, N1-guanyl-1,7-diaminoheptane. This suggests a major role of the archaeal assemblages in inorganic carbon fixation. We also show that benthic archaeal assemblages contribute approximately 25% of the total 3H-leucine incorporation. Inorganic carbon fixation in surface deep-sea sediments appears to be dependent not only upon chemosynthetic processes but also on heterotrophic/mixotrophic metabolism, as suggested by estimates of the chemolithotrophic energy requirements and the enhanced inorganic carbon fixation due to the increase in the availability of organic trophic resources. Overall, our data suggest that archaeal assemblages of surface deep-sea sediments are responsible for the high rates of inorganic carbon incorporation and thereby sustain the functioning of the food webs as well as influence the carbon cycling of benthic deep-sea ecosystems.
 Deep-sea ecosystems cover about 65% of the Earth's surface, and they play a key role in biomass production and biogeochemical cycles on a global scale [Tyler, 2003]. Except for hydrothermal vents and cold seeps, life in deep-sea ecosystems is dependent upon the supply of organic matter that reaches the ocean interior through vertical fluxes [Jahnke, 1996; Smith et al., 2008] and lateral transport mechanisms from the continental shelves [Canals et al., 2006]. The inputs of organic carbon (OC) that reach the ocean floor sustain the metabolism of the benthic food webs, which are largely dominated by prokaryotic biomass [Turley, 2000; Dell’Anno and Danovaro, 2005; Jørgensen and Boetius, 2007]. Here, despite severe food limitations, the decomposition and utilization processes of OC proceed at relatively high rates compared with shallow ecosystems, which suggests that benthic deep-sea prokaryotes are highly efficient in exploiting the available trophic resources [Dixon and Turley, 2001; Dell’Anno and Danovaro, 2005; Jørgensen and Boetius, 2007; Danovaro et al., 2008]. However, combinations of direct measurements of OC supply versus consumption and various modeling approaches have revealed that the OC input to the ocean interior is not sufficient to fulfill the metabolic requirements of the deep-sea biota [Smith and Kaufmann, 1999; Danovaro et al., 2001; Reinthaler et al., 2010]. This apparent paradox of the imbalance between the OC supply and its consumption has been explained on the basis of potential underestimation of the allochthonous inputs of OC [Arístegui et al., 2009; Burd et al., 2010] and/or that of additional sources of OC that are produced in situ through chemoautotrophic processes [Gasol et al., 2008; Herndl et al., 2008; Reinthaler et al., 2010].
 Nitrification is believed to be one of the main processes responsible for dark CO2 fixation in the ocean interior [Herndl et al., 2005; Wuchter et al., 2006; Middelburg, 2011]. Nitrifying prokaryotes belong to both the Bacteria (i.e., the β-subgroup and γ-subgroup of Proteobacteria [Kowalchuk and Stephen, 2001; Arp and Stein, 2003]) and Archaea [Venter et al., 2004; Könneke et al., 2005; Hallam et al., 2006a, 2006b; Martens-Habbena et al., 2009] domains. Different studies have suggested that chemoautotrophic processes that involve aerobic ammonia oxidation in marine ecosystems are mainly due to archaeal rather than bacterial assemblages [Francis et al., 2005, 2007; Wuchter et al., 2006; Santoro and Casciotti, 2011; Santoro et al., 2011; Yakimov et al., 2011]. Potentially, archaeal inorganic C fixation significantly contributes to the sustaining of the prokaryotic heterotrophic C requirements at bathyal depths [Herndl et al., 2005; Baltar et al., 2010; Reinthaler et al., 2010]. Conversely, Swan et al.  reported that bacterial assemblages might also be responsible for the chemoautotrophic processes that occur in the dark ocean, probably through oxidation of reduced sulfur, CO, and CH4. If the mechanisms and microbial components that are responsible for inorganic C use in deep water masses of the oceans remain to be defined, information on the quantitative importance and ecological role of dark inorganic C fixation in surface deep-sea sediments is practically nonexistent.
 In the present study, we investigated the rates of dark inorganic C fixation in surface sediments of different benthic deep-sea systems along the Iberian margin (northeastern Atlantic Ocean) and in the Mediterranean Sea. Coupled with the determination of prokaryotic abundance, biomass, assemblage composition, and heterotrophic C production, this analysis was aimed at providing new insights into C cycling and the functioning of the microbial food webs in benthic deep-sea ecosystems.
2 Materials and Methods
2.1 Study Areas and Sample Collection
 The sediment samples were collected between August 2007 and October 2008 at depths ranging from 1207 to 4381 m in the northeastern Atlantic Ocean and from 1200 to 4100 m in the Mediterranean Sea (Figure 1, supporting information, Table S1). In the northeastern Atlantic Ocean, the sediments were collected at 4 stations along the Iberian margin, of which 3 were located along the Galicia Bank and the remaining station was located in the Gulf of Cadiz. In the Mediterranean Sea, the sediments were collected at 10 stations across the entire basin, with 3 located in the western Mediterranean and the others in the central-eastern Mediterranean basin (see supporting information for further details). At all of these stations, the sediments were collected with a NIOZ-type box corer, which allowed the collection of samples that were hermetically sealed. Visual inspection of the overlying waters and sediment surfaces revealed very limited resuspension effects during sampling. Previous studies on deep-sea sediments have demonstrated that microbiological analyses carried out on sediments collected with box corers and that on sediments collected with multiple corers used synoptically provide the same results [Danovaro et al., 1998]. Immediately after sampling, subsamples of the top 1 cm of the sediment surfaces were collected for analysis of grain size, porosity, total OC, total nitrogen, and microbiological parameters. Additional sediment cores were collected with the overlying seawaters for the analysis of inorganic nutrient and oxygen concentrations, which was carried out on board (for details, see supporting information, Supplementary Materials and Methods). For the incorporation of radiolabeled substrate, sediment subsamples were immediately analyzed on board after sampling. For prokaryotic counts, sediment subsamples were added to pre-filtered (0.2 µm pore size) seawater containing formaldehyde (final concentration, 2%) and stored at 4°C until analysis (within 4 weeks from sampling; for details, see supporting information, Supplementary Materials and Methods). For analysis of prokaryotic assemblage composition by catalyzed reporter deposition fluorescence in situ hybridization (CARD-FISH), sediment subsamples were fixed with paraformaldehyde, washed with phosphate-buffered saline (PBS), and stored in ethanol/PBS (1:1) at −20°C until further laboratory analysis (for details, see supporting information, Supplementary Materials and Methods). The incorporation of radiolabeled substrate and microbiological analyses were carried out on replicate cores collected at each site from different deployments.
2.2 Determination of Dark Inorganic Carbon Fixation Rates
 Rates of dark fixation of inorganic C in surface deep-sea sediments were determined using [14C]bicarbonate (specific activity, 58 mCi mmol−1; Amersham) using a modification of the methodology previously reported for measurements of photosynthetic primary production in shallow sediments [MacIntire and Cullen, 1995; Cibic et al., 2007]. To define the most appropriate amount of [14C]bicarbonate to be used to reach saturation, incubations were initially carried out on surface deep-sea sediment samples collected at four stations using three different [14C]bicarbonate dilutions (3, 12, and 30 μCi ml−1). To do this, sediment subsamples (ca. 1 cm3) had 1 ml of 0.2 µm pre-filtered seawater (collected at the sediment-water interface of each station) that contained the [14C]bicarbonate added. These were then incubated in the dark at the in situ temperature (ca. 2°C for the Atlantic Ocean samples and 13°C–14°C for the Mediterranean Sea samples; see supporting information, Table S1) for 6 to 8 h. The incubation time was selected on the basis of time course experiments of [14C]bicarbonate incorporation carried out over 48 h. After the incubations, inorganic C incorporation was stopped with the addition of 0.2 µm pre-filtered formaldehyde solution (final concentration, 2%). The sediment blanks were made by adding 0.2 µm pre-filtered formaldehyde solution immediately before [14C]bicarbonate addition. To remove the excess [14C]bicarbonate in solution and/or adsorbed to detrital particles, all of the samples were subjected to three washes with PBS and centrifugation at 13,500×g for 10 min steps. After removal of the final supernatant, the samples were incubated overnight in 1 N HCl (final concentration). After this incubation, 10 ml of scintillation cocktail was added, the mixed samples were centrifuged at 3000×g for 10 min, and the supernatants were transferred into scintillation vials for the quantification of radioactivity using a liquid scintillation counter (Packard Tri-Carb 300). On the basis of these data, the analysis of inorganic C incorporation in the sediments of all of the other deep-sea sites was carried out using a final [14C]bicarbonate activity of 12 μCi ml−1 and an incubation time of 6 to 8 h. These samples were then processed as described above.
 The [14C] radioactivity (as disintegration per minute, DPM) in the samples was converted into the amount of inorganic C incorporated per gram of dry sediment per hour (for details, see supporting information, Supplementary Materials and Methods). The nonspecific binding of [14C]bicarbonate to the sediments was taken into account by analyzing the [14C] radioactivity of replicate sediment subsamples treated with formaldehyde or pre-combusted in a muffle furnace (450°C for 4 h) before the [14C]bicarbonate addition. These samples were used as the controls, and they showed that the [14C] radioactivity bound to the sediment particles was very low, with values always significantly lower than those measured in the untreated sediments (by up to 20-fold; p < 0.001).
2.3 Bacterial Versus Archaeal Incorporation of [14C]Bicarbonate
 To identify the relative involvement of Bacteria versus Archaea in inorganic C uptake, specific experiments were conducted on the sediment samples collected at all of the benthic deep-sea sites using N1-guanyl-1,7-diaminoheptane (GC7) to inhibit archaeal metabolism [Jansson et al., 2000]. GC7 blocks protein synthesis by inhibition of deoxyhypusine synthase (DHS) [Jansson et al., 2000; Lee et al., 2002]. DHS catalyzes the first step of hypusination, the final product of which is a modified amino acid (i.e., hypusine) in initiation factor 5A (IF5A) of eukaryotes and Archaea [Bartig et al., 1990; Park et al., 1997]. Bacteria do not have IF5A and/or deoxyhypusine and hypusine [Bartig et al., 1990; Park et al., 1997; Kyrpides and Woese, 1998]. Although some bacterial taxa contain genes related to the DHS-like homospermidine synthase genes, these appear not to have functional importance [Brochier et al., 2004]. Replicate sediment samples collected at all of the sites were added with 1 mM GC7. Additional replicate samples collected at selected sites were also added with 0.1 and 2 mM GC7. All of these samples were incubated for 2 h in the dark at the in situ temperature (as above; see supporting information, Table S1). Other replicate sediment samples without GC7 addition were maintained under the same conditions and used as the controls. All of the samples were added with [14C]bicarbonate (final activity, 12 μCi ml−1) and processed as described above following incubation for 6 and 60 h at the in situ temperature (as above; see supporting information, Table S1).
2.4 [14C]Bicarbonate Incorporation in the Presence of Leucine
 To provide insight into the ecological significance of the inorganic C incorporation, additional experiments were carried out to mimic the potential input of labile organic matter to the seafloor using leucine as substrate. The rationale for these experiments was to determine whether mixotrophic/heterotrophic metabolism is also responsible for the inorganic C assimilation, as strictly chemoautotrophic processes should be directly influenced to a minor extent by organic matter availability. Triplicate sediment samples and duplicate formalin-fixed blanks with added [14C]bicarbonate (final activity, 12 μCi ml−1) and in the presence of leucine (final concentration, 1 μM) were incubated for 6 h in the dark at the in situ temperature (as above; see supporting information, Table S1) and then processed as described above. Other sediment samples used as controls only had the [14C]bicarbonate added.
2.5 [3H]Leucine Incorporation
 Sediment subsamples with the addition of an aqueous solution of [3H]leucine (specific activity, 68 Ci mmol−1; final concentration, 0.5–1.0 μM) were incubated for 1 h in the dark at the in situ temperatures (as above; see supporting information, Table S1). Time course experiments over 6 h and concentration-dependent incorporation experiments (seven concentrations, from 0.05 to 5.0 μM leucine) were also carried out to define the linearity of the [3H]leucine incorporation and to estimate the leucine saturation level, respectively. To investigate the contribution of Archaea to the heterotrophic C production, additional sediment subsamples that were collected at different deep-sea sites were added with GC7 (final concentrations, 0.1, 1.0, and 2.0 mM). After the incubation, all of the samples were added with ethanol (final concentration, 80%) and processed according to Danovaro et al.  before scintillation counting. Sediment blanks were made by adding ethanol immediately after the [3H]leucine addition. [3H]Leucine incorporation was converted into C produced by the heterotrophic prokaryotes according to Simon and Azam . Additional details are reported in the supporting information.
2.6 Prokaryotic Abundance, Biomass, and Assemblage Composition
 The prokaryotic abundance and biomass in the sediments were determined by epifluorescence microscopy [Danovaro et al., 2002; Danovaro et al., 2008]. The composition of the benthic prokaryotic assemblages was determined using CARD-FISH analysis [Molari and Manini, 2012]. Methodological details are reported in the supporting information, Supplementary Materials and Methods.
2.7 Estimates of Benthic Respiration Rates
 The benthic respiration rates were estimated through two different approaches: (i) using the estimates of oxygen fluxes that were derived from the measurements of oxygen concentrations at the sediment-water interface and in pore waters along the vertical profiles of the sediments and (ii) using the algorithm of Wenzhöfer and Glud , which is based on the net photosynthetic primary production (extracted from the Ocean Productivity database at http://www.science.oregonstate.edu/ocean.productivity/index.php and derived from a C-based productivity model algorithm [Behrenfeld et al., 2005]). The values of the photosynthetic net primary production used in this study referred to the same sampling periods when the sediments were collected (supporting information, Table S1). Additional details are reported in the supporting information.
 Denitrification rates were estimated using the meta-model developed by Middelburg et al. , which is based on OC fluxes (as derived using the equation of Martin et al. ) as well as oxygen and nitrate concentrations at the sediment-water interface.
2.8 Statistical Analysis
 Analysis of variance (ANOVA) was carried to test the effects of incubation time and [14C]bicarbonate concentrations on the incorporation rates of inorganic C. When significant differences were encountered, a Student-Newman-Keuls post hoc comparison test (α = 0.05) was also carried out. This statistical analysis was also used to identify differences for all of the other experiments and variables investigated.
3.1 Sediment Characteristics
 The deep-sea sediments investigated in the present study were dominated by the silt-clay fraction as well as characterized by total OC from 3.5 to 6.8 mg C g−1 and total N from 0.4 to 1.1 mg N g−1 (supporting information, Table S2). All of the inorganic nutrients analyzed were lower at the water-sediment interface than in the pore waters of the top 1 cm of the sediments. Nitrate concentrations in sediment pore waters ranged from 8.7 to 27.6 μM, while nitrite concentrations varied from 0.14 to 0.74 μM (supporting information, Table S2). The surface sediments were fully oxygenated, as indicated by the high oxygen concentrations at the water-sediment interface (range, 238–263 μM) and by the oxygen penetration depth into the sediments (always below 4 cm).
3.2 Incorporation of Radiolabeled Compounds
 Incubations were carried out with the surface deep-sea sediments collected at the selected sites using increasing amounts of [14C]bicarbonate. In all of the sediments analyzed, the uptake of [14C]bicarbonate was significantly greater when it was used at 12 μCi ml−1 than at 3 μCi ml−1 (p < 0.05), but it was not significantly increased further at 30 μCi ml−1 (Figure 2a). Therefore, independent of the sampling sites and water column depths, 12 μCi ml−1 [14C]bicarbonate appeared to be the most suitable amount to provide accurate estimates of inorganic C incorporation rates. The incorporation of the [14C]bicarbonate increased with time (Figure 2b), although the [14C]bicarbonate incorporation normalized per unit time was significantly greater after 6 to 8 h than with longer incubation times (p < 0.01). The [14C]bicarbonate uptake in the deep-sea sediments incubated with GC7 was negligible, as shown by the [14C]bicarbonate incorporation, which was very close to, and not statistically different from, the rates measured in the formalin-treated sediments. This effect was evident even at the lowest concentration of GC7 (i.e., 0.1 mM; Figure 3a). The incubation of deep-sea sediments with GC7 showed a significant decrease in the [3H]leucine incorporation rates (on average, by 26%, compared to untreated sediments; p <0.05), and this decrease was similar independent of the concentration of GC7 used (data not shown) and the different benthic deep-sea sites investigated (range, 24%–29%; Figure 3b). The addition of unlabeled leucine to the deep-sea sediment samples increased the [14C]bicarbonate incorporation rates significantly (by up to approximately fourfold; p < 0.01) compared to the control sediment samples (i.e., without unlabeled leucine; Figure 3c). At the end of the incubations (i.e., after 6–8 h), the dissolved O2 concentrations in the samples ranged from 186 to 234 μM. Thus, any potential influence of anoxic and/or hypoxic conditions on the inorganic C incorporation rates can be ruled out. This applied also to the sediment samples with added leucine, where the oxygen concentration was, on average, 13% lower than that in the untreated samples (i.e., those without leucine addition).
3.3 Carbon Fixation Rates and Prokaryotic Assemblages in Deep-Sea Sediments
 The inorganic C incorporation rates in the deep-sea sediments investigated showed significant spatial variability, with a range from 1.8 ± 0.1 nmol C g−1 d−1 at 1200 m depth in the eastern Mediterranean to 19.1 ± 0.5 nmol C g−1 d−1 at 2688 m depth in the western Mediterranean (p < 0.01; Figure 4a). The deep-sea sediments in the western Mediterranean were characterized by inorganic C incorporation rates (mean, 13.8 ± 3.7 nmol C g−1 d−1) that were approximately fourfold higher than those of the benthic systems in the central Mediterranean (mean, 3.6 ± 0.3 nmol C g−1 d−1) and eastern Mediterranean (mean, 3.7 ± 0.9 nmol C g−1 d−1; p < 0.01). These differences were evident even when the benthic systems at similar depths were compared. Also, the incorporation rates of [3H]leucine showed significant changes across the deep-sea sites investigated (p < 0.01), although with different spatial patterns compared to those for inorganic C uptake (Figure 4b). Higher [3H]leucine incorporation rates were indeed observed in the deep-sea sediments of the eastern Mediterranean (mean, 65.4 ± 40.7 nmol C g−1 d−1).
 In the deep-sea sediments of the northeastern Atlantic Ocean, the total prokaryotic abundance ranged from 0.11 ± 0.05 × 108 cells g−1 at 1207 m depth to 7.0 ± 0.73 × 108 cells g−1 at 4381 m depth. In the benthic deep-sea ecosystems of the Mediterranean basin, the prokaryotic abundance ranged from 0.26 ± 0.09 × 108 cells g−1 at 2700 m depth in the central Mediterranean to 10.5 ± 1.5 × 108 cells g−1 at 2688 m depth in the western Mediterranean (Figure 5a). The total benthic prokaryotic abundance decreased significantly from the western sector (mean, 6.8 ± 2.1 × 108 cells g−1) to the eastern sector of the deep Mediterranean basin (mean, 1.4 ± 0.4 × 108 cells g−1; p < 0.01). CARD-FISH analysis revealed that the abundance of Bacteria and Archaea in the investigated deep-sea sediments together accounted for approximately 80% of the total prokaryotic assemblages, as determined using SYBR Green I (range, 63%–91%). On average, Bacteria accounted for 52% of the total prokaryotic counts (range, 39%–59%; Figure 5b). The abundance of total Archaea in the surface deep-sea sediments was significantly related to the inorganic C fixation rates (n = 13, R2 = 0.839, p < 0.01; supporting information, Figure S1).
3.4 Benthic Respiration Rates
 The benthic respiration rates that were estimated from the two different approaches are reported in Table 1. The two independent estimates provided consistent data, as highlighted by the significant relationships (n = 6, R2 = 0.734, p < 0.05) and the slight quantitative differences between the estimates (estimates from measurements of oxygen concentrations were, on average, approximately 25% higher than those derived from the algorithm). Benthic deep-sea systems located in the northeastern Atlantic Ocean and in the western Mediterranean were characterized by higher rates of OC consumption (mean, 1.04 ± 0.20 and 0.99 ± 0.24 mmol C m−2 d−1, respectively) than those seen for the systems located in the central Mediterranean (mean, 0.47 ± 0.03 mmol C m−2 d−1) and eastern Mediterranean (mean, 0.61 ± 0.07 mmol C m−2 d−1).
Table 1. Estimated Net Photosynthetic Primary Productivity (NPP) in the Same Sampling Periods When the Sediments Were Collecteda and Benthic Respiration Ratesb
NPP (mmol C m−2-−1)
A (mmol C m−2 d−1)
B (mmol C m−2 d−1)
For details, see also supporting information, Table S1.
The data under the column marked “A” were obtained using the algorithm of Wenzhöfer and Glud , whereas those under the column marked “B” were obtained using estimates of oxygen fluxes derived from measurements of oxygen concentrations at the sediment-water interface and in pore waters along the vertical profiles of the sediments in the different sampling areas investigated. na, not available.
Northeastern Atlantic Ocean
Gulf of Cadiz
 Dark inorganic C assimilation has traditionally been considered to be negligible in the photic layer of marine ecosystems, due to the overwhelming dominance of photosynthetic processes. However, increasing evidence has indicated that as well as in specific deep-sea ecosystems, such as hydrothermal vents, cold seeps, and anoxic basins [Jørgensen et al., 1991; Mandernack and Tebo, 1999; Taylor et al., 2001; Yakimov et al., 2007; Jost et al., 2008; La Cono et al., 2011], dark inorganic C fixation occurs also in the oxygenated deep water masses of the oceans [Herndl et al., 2005; Agogué et al., 2008; Hansman et al., 2009; Tamburini et al., 2009; Baltar et al., 2010; Yakimov et al., 2011; Swan et al., 2011]. The extent and pathways of inorganic C use in the ocean interior still need to be fully clarified [Agogué et al., 2008; Hansman et al., 2009; Swan et al., 2011; Yakimov et al., 2011], and information on the quantitative importance and ecological role of dark CO2 fixation in surface deep-sea sediments is practically nonexistent.
 The inorganic C incorporation rates in the deep-sea sediments investigated here showed significant spatial variability, although no clear depth-related patterns were observed. These rates fall within the range of those reported for sediments collected at greater depths in the Japan Trench (>9000 m [Seki and ZoBell, 1967]) and the Indian Ocean (>5000 m [Das et al., 2011]). However, when normalized to the same unit of volume, they were up to 3 orders of magnitude greater than those reported for the water masses of the Atlantic Ocean and the Mediterranean Sea at comparable depths [Herndl et al., 2005; Tamburini et al., 2009; Reinthaler et al., 2010; Yakimov et al., 2011]. Caution should be taken when comparing data obtained using single incubation times, as the incorporation rates of radiolabeled substrates can show deviations from linearity over time, which can lead to significant uncertainties in the estimates of the prokaryotic metabolic rates in deep-sea environments [Dixon and Turley, 2001; Luna et al., 2012]. The time course experiments carried out in the present study provide evidence indeed that inorganic C fixation deviates from linearity after a few hours of incubation. The measured rates might also be influenced by decompression, although both positive and negative effects have been reported in previous studies conducted in the deep sea. Previous studies indeed have shown that during the stratification period, decompression can induce underestimation of the measured rates, as opposed to those measured in situ and under the given temperature conditions [Tamburini et al., 2002, and references therein]. Conversely, other studies have shown stimulation by decompression during the mixed-water period [Bianchi and Garcin, 1994] and no changes between decompressed and nondecompressed samples collected at the sediment-water interface [Danovaro et al., 2008] or between sediment samples incubated at 1 and 500 atm [Das et al., 2011].
 In the present study, we estimated that, on average, the inorganic C fixation rates in the surface deep-sea sediments amounted to 25% of the total prokaryotic heterotrophic C production determined synoptically. Such a comparison should be viewed with caution, as the estimates of prokaryotic biomass production assessed through the use of [14C]bicarbonate and [3H]leucine are based on different assumptions and methodological approaches [Nagata et al., 2010; Yamada and Suzumura, 2011]. For instance, [3H]leucine is assumed to be used for protein synthesis by heterotrophic prokaryotes [Simon and Azam, 1989], whereas [14C]bicarbonate can also be used for the biosynthesis of other organic compounds (e.g., lipids [Wuchter et al., 2003]). As only two thirds of dark CO2 fixation is used for microbial protein synthesis [Yamada and Suzumura, 2011], the contribution of inorganic C fixation rates to the total heterotrophic C production might be lower. We found indeed that, on average, inorganic C fixation rates amounted to 19% of the total heterotrophic biomass production, which was estimated on the basis of the C consumption rates of the deep-sea sediments investigated (assuming a C conversion efficiency of 35% [Dunne et al., 2007]). Altogether, these findings suggest that prokaryotic biomass production through inorganic C fixation contributes to sustain the functioning of the benthic food webs in the deep sea.
 We showed a significant relationship between the inorganic C fixation rates and the abundance of total Archaea, which explained a large fraction of the total variance (84%). To assess the potential contribution of Archaea to inorganic C incorporation, we also used an inhibitor of archaeal metabolism, GC7. This revealed complete inhibition of inorganic C uptake in all of the benthic deep-sea systems investigated. Recent studies reported a major drop in dark inorganic C fixation (up to 76%) when water samples from the Chilean continental shelf were incubated with GC7 [Farías et al., 2009]. Moreover, investigations conducted both on pure cultures [Jansson et al., 2000; Levipan et al., 2007a] and on natural marine assemblages [Levipan et al., 2007a, 2007b; Farías et al., 2009; Loescher et al., 2012] have revealed that GC7 can selectively inhibit archaeal metabolism. Furthermore, in silico analysis that we carried out on the gene sequences of the different marine Archaea (deposited in GenBank) revealed the presence of genes that code for DHS, which is the target enzyme of GC7 (supporting information, Table S3). Finally, as the eukaryotic biomass represents <10% of the prokaryotic biomass in deep-sea sediments and provides a minor contribution to the benthic metabolism [Dell’Anno and Danovaro, 2005], the contribution of eukaryotes to inorganic C incorporation can be assumed to be negligible. Altogether, these findings suggest the major role of archaeal assemblages in inorganic C fixation in benthic deep-sea ecosystems.
 Archaeal assemblages that inhabit surface deep-sea sediments showed a high potential for the use of inorganic C, as revealed by the high cell-specific uptake rates. The inorganic C incorporation rates normalized per archaeal cell (range, 0.45–15.4 fg C cell−1 d−1) were indeed much higher than those calculated on the abundance of Crenarchaeota that were previously reported for the deep water masses of the North Atlantic (0.024–1.2 fg C cell−1 d−1 [Varela et al., 2011]). As Crenarchaeota represent only a fraction of the total archaeal assemblages, this comparison indicates that our cell-specific uptake rates would be even higher if they referred to the abundance of Crenarchaeota. These high cell-specific uptake rates of inorganic C were also associated with rapid turnover times of the archaeal assemblages. Based on our estimates of the C content of benthic deep-sea prokaryotes (range, 20–34 fg C cell−1; mean, 28 fg C cell−1) and assuming that the archaeal biomass production is sustained only by inorganic C, we estimated a mean turnover time of archaeal assemblages in these surface deep-sea sediments of 10 days (range, <2–67 days). These turnover times are more rapid than those previously reported for planktonic Archaea of the deep water masses (20–100 days [Herndl et al., 2005; Varela et al., 2011]), which suggests that the Archaea that inhabit surface deep-sea sediments are a highly dynamic component of the benthic food webs.
 We also used [3H]leucine and GC7 to investigate the potential role of archaeal assemblages inhabiting surface deep-sea sediments in the incorporation of organic compounds. The role of archaeal assemblages in OC use has been repeatedly claimed for pelagic marine environments using different approaches [Ouverney and Fuhrman, 2000; Herndl et al., 2005; Ingalls et al., 2006; Teira et al., 2006; Levipan et al., 2007a; Hansman et al., 2009] and was also highlighted recently for benthic deep-sea ecosystems [Takano et al., 2010]. Our results show that, on average, approximately 25% of the total 3H-leucine uptake may be due to benthic archaeal assemblages. These findings suggest that Archaea in surface deep-sea sediments can be also involved in the exploitation of organic compounds at rates similar to those of Bacteria, as revealed by their similar cell-specific leucine incorporation rates (mean, 1.3 × 10−18 and 1.9 × 10−18 mol cell−1 d−1 for Archaea and Bacteria, respectively).
 The metabolic strategies of marine archaeal assemblages are still subject to debate, although recent studies have provided evidence of their ability to incorporate inorganic C through the oxidation of ammonia [Francis et al., 2005; Könneke et al., 2005; Wuchter et al., 2006; Hallam et al., 2006a, 2006b; Martens-Habbena et al., 2009; Walker et al., 2010; Santoro et al., 2011]. In oxygenated deep-sea sediments such as those investigated in the present study, ammonia is the dominant reduced inorganic compound that is produced by organic matter mineralization processes [Schulz and Zabel, 2006]. Thus, chemoautotrophy is expected to be mainly sustained by ammonia oxidation, whereas the reoxidation of other reduced compounds that are produced during anaerobic degradation of organic matter (primarily sulfides) has a minor role [Middelburg, 2011]. The dominance of aerobic respiration in the deep-sea sediments investigated here is also confirmed by the low contribution of the denitrification rates to C mineralization (<2%), which is typically encountered in open ocean benthic systems [Middelburg et al., 1996].
 In the present study, we estimated that, on average, benthic respiration rates can provide no more than 20% of the energy required for the sustaining of inorganic C fixation rates through ammonia oxidation (assuming 1 mol of CO2 is fixed per 10 mol of ammonia oxidized [Arístegui et al., 2009; Middelburg, 2011]). Such an imbalance, as previously reported for deep water masses of the Atlantic Ocean [Reinthaler et al., 2010], suggests that ammonia oxidation is unlikely to be the main process that sustains the inorganic C fixation rates in benthic deep-sea ecosystems. Consistently, water samples collected in the Mediterranean Sea at bathyal depth that were poor in ammonia did not show any significant changes in inorganic C fixation rates following ammonia addition [Yakimov et al., 2011]. Other reduced compounds might support chemoautotrophy in the dark ocean, including sulfides, CO, CH4 [Swan et al., 2011], and metals [Bach and Edwards, 2003; Das et al., 2011; Orcutt et al., 2011]. However, these compounds are unlikely to be key energy sources for inorganic C fixation in the deep-sea sediments investigated here, as they are dominated by aerobic processes and not affected by gas emissions or geothermal flow. These findings allow us to hypothesize that a large fraction of the inorganic C incorporation in surface deep-sea sediments is not dependent upon strictly chemolithoautotrophic metabolism.
 This hypothesis is also supported by the significant increase in inorganic C fixation rates that we observed after leucine addition. We estimated indeed that even if all of the added leucine is catabolized to ammonia through heterotrophic activities, ammonia oxidizers would contribute no more than 22% of the increase in the incorporated inorganic C. Previous laboratory experiments have provided evidence that labile organic substrates can significantly enhance CO2 fixation of heterotrophic prokaryotes [Roslev et al., 2004; Hesselsoe et al., 2005] through several of the carboxylation reactions that are involved in central metabolism and peripheral metabolism [Doronia and Trotsenko, 1985; Alonso-Sáez et al., 2010] as well as significantly increase the growth rates of ammonia-oxidizing archaea [Tourna et al., 2011]. Thus, our experimental findings suggest that the availability of labile organic matter in the sediment can increase the inorganic C fixation rates, mainly due to enhanced heterotrophic and/or mixotrophic activities, rather than strictly chemolithoautotrophic metabolism. Consistent with this, we observed a significant relationship between the availability of organic trophic resources (as organic N concentrations) and inorganic C fixation rates in the deep-sea sediments investigated (n = 12, r = 0.816, p < 0.01).
 Mixotrophy might be an important metabolic strategy of Archaea in the oceans, as is suggested by the radiocarbon signature of marine archaeal assemblages [Ingalls et al., 2006] and the presence in ammonia-oxidizing archaea of genes encoding enzymes of the oxidative tricarboxylic acid cycle and oligopeptide transporters [Hallam et al., 2006a, 2006b; Martín-Cuadrado et al., 2008; Walker et al., 2010]. As we found that benthic Archaea can incorporate inorganic C and OC at comparable rates (mean, 8.9 × 10−18 and 7.5 × 10−18 mol C cell−1 d−1 determined on the basis of 14C-bicarbonate uptake and 3H-leucine uptake, respectively), our results suggest that in surface deep-sea sediments, archaeal assemblages might efficiently sustain their metabolic requirements by simultaneous exploitation of the available inorganic C and OC sources. Although further studies are needed to clarify the mechanisms of inorganic C use in benthic deep-sea ecosystems, overall our findings contribute to the rejection of the assumption that only strictly chemoautotrophic processes are responsible for inorganic C assimilation in surface deep-sea sediments, as they suggest that other metabolic strategies are also involved.
 In conclusion, our results suggest that archaeal assemblages of surface deep-sea sediments are responsible for the high rates of inorganic C incorporation and contribute to organic matter consumption, thereby sustaining the functioning of the food webs and influencing the C cycling of the benthic deep-sea ecosystems. The need to include chemolithotrophy in the ocean C budget has been recently reported [Middelburg, 2011]. Findings provided in the present study suggest that the amount of inorganic C incorporated into microbial biomass can be more than fivefold greater than that predicted from chemolithotrophic production estimates derived from diagenetic models, thus providing new clues for our understanding of the functioning of benthic deep-sea ecosystems.
 This work was carried out in the framework of the EU project BIOFUN (Eurocores EuroDEEP) and the Italian national project VECTOR. M.M. thanks the captains and crew of the R/V Urania and R/V Pelagia for their kind assistance during the sampling activities.