1.1. Study site

The sampling sites are located on the Sitka stream, Czech Republic (Fig. 1). The Sitka is an undisturbed, third-order, 35 km long lowland stream originating in the Hruby Jesenik mountains at 650 m above sea level. The catchment area is 118.81km2, geology being composed mainly of Plio-Pleistocene clastic sediments of lake origin covered by quaternary sediments. The mean annual precipitation of the downstream part of the catchment area varies from 500 to 600 mm. Mean annual discharge is 0.81 m3.s-1. The Sitka stream flows in its upper reach till Sternberk through a forested area with a low intensity of anthropogenic effects, while the lower course of the stream naturally meanders through an intensively managed agricultural landscape. Except for short stretches, the Sitka stream is unregulated with well-established riparian vegetation. River bed sediments are composed of gravels in the upper parts of the stream (median grain size 13 mm) while the lower part, several kilometres away from the confluence, is characterised by finer sediment with a median grain size of 2.8 mm. The Sitka stream confluences with the Oskava stream about 5 km north of Olomouc. More detailed characteristics of the geology, gravel bar, longitudinal physicochemical (e. g. temperature, pH, redox, conductivity, O2, CH4, NO3, SO4) patterns in the sediments and a schematic view of the site with sampling point positions have been published previously (Rulik et al. 2000, Rulik & Spacil 2004). Earlier measurements of a relatively high production of methane, as well as potential methanogenesis, confirmed the suitability of the field sites for the study of methane cycling (Rulik et al. 2000, Hlavacova et al. 2005, 2006).

Soil heavy metal contamination has become an increasing problem worldwide. Among the heavy metals, Cu, Zn, Mn, Cd, Pb, Ni and Cr are considered to be the most common toxicity problems causing increasing concern. Growth inhibition and reduced yield are common responses of horticultural crops to nutrient and heavy metal toxicity [2]. Nevertheless, sometimes less common responses happen under metal toxicity conditions. For example, in the case of Pb it has been suggested that inhibition of root growth is one of the primary effects of Pb toxicity through the inhibition of cell division at the root tip [59]. Significant reductions in plant height, as well as in shoot and root dry weight (varying from 3.3% to 54.5%), as compared with that of the controls, were found for Typha angustifolia plants in different Cr treatments [60]. Furthermore, according to Caldelas et al. (2012) [19], not only growth inhibition happened (reached 65% dry weight) under Cr toxicity conditions, but also root/shoot partitioning increased by 80%. Under Cr stress conditions, it was found that root and shoot biomass of Genipa americana L. were significantly reduced [20]. The biomass reduction of Genipa americana trees is ascribed, according to the same authors, to the decreased net photosynthetic rates and to the limitations in stomatal conductance. The disorganization of chloroplast structure and inhibition of electron transport is a possible explanation for the decreased photosynthetic rates of trees exposed to Cr stress [20]. In contrast to the above, Cd and Pb applications induced slight or even significant increase in plant height and biomass. The fact that Cd and Pb addition enhanced Ca and Fe uptake suggests that these two nutrients may play a role in heavy metal detoxification by Typha angustifolia plants; furthermore, increased Zn uptake may also contribute to its hyper Pb tolerance, as recorder in the increased biomass over the control plants [60]. According to the

same authors (Bah et al., 2011), plants have mechanisms that allow them to tolerate relatively high concentrations of Pb in their environment without suffering from toxic effects.

Tzerakis et al. (2012) [2] found that excessively high concentrations of Mn and Zn in the leaves of cucumber (reached 900 and 450 mg/kg d. w., respectively), grown hydroponically under toxic Mn and Zn conditions, reduced the fruit biomass due to decreases in the number of fruits per plants, as well as in the net assimilation rate, stomatal conductance and transpiration rate. However, it was found that significant differences concerning biomass production between different species of the same genus exist under metal toxicity conditions; Melilotus officinalis seems to be more tolerant to Pb than Melilotus alba because no differences in shoot or root length, or number of leaves, were found between control plants and those grown under 200 and 1000 mg/kg Pb [15]. In addition to the above, genotypic differences between cultivars of the same species, concerning biomass production, under metal toxicity conditions may also be observed; Chatzistathis et al. (2012) [13] found that under excess Mn conditions (640 pM), plant growth parameters (shoot elongation, as well as fresh and dry weights of leaves, root and stem) of olive cultivar ‘Picual’ were significantly decreased, compared to those of the control plants (2 pM), something which did not happen in olive cultivar ‘Koroneiki’ (no significant differences were recorder between the two Mn treatments) (Figure 1). According to the same authors, some factors related to the better tolerance of ‘Koroneiki’ not only at whole plant level, but also at tissue and cell level, could take place. Such possible factors could be a better compartmentalization of Mn within cells and/or functionality of Mn detoxification systems [13]. Significant growth reductions of several plant species, grown under Mn toxicity conditions, have been mentioned by several researchers [61-65].

Nickel (Ni) toxicity, which may be a serious problem around industrial areas, can also cause biomass reduction. At high soil Ni levels (>200 mg/kg soil) reduced growth symptoms of Riccinus communis plants were observed [18]. According to Baccouch et al. (1998) [66], the higher concentrations of Ni have been reported to retard cell division, elongation, differentiation, as well as to affect plant growth and development. Excess Cd, which causes direct or indirect inhibition of physiological processes, such as transpiration, photosynthesis, oxidative stress, cell elongation, N metabolism and mineral nutrition may lead in growth retardation, leaf chlorosis and low biomass production [67]. According to the same authors,

Cd stress could induce serious damage in root cells of grey poplar (Populus x canescens). Arsenic (As) toxicity may be another (although less common) problem contributing to soil contamination. Repeated and widespread use of arsenical pesticides has significantly contributed to soil As contamination [4]. According to the same authors, plant growth parameters, such as biomass, shoot height, and root length, decreased with increased As concentrations in all soils.

The physicochemical sediment and interstitial water properties of the investigated sites showed large horizontal and vertical gradients. Sediment grain median size decreased along a longitudinal profile while organic carbon content in a sediment fraction < 1 mm remained unchaged (Table 2). Generally, interstitial water revealed relatively high dissolved oxygen saturation with the exceptions of localities IV and V where concentration of dissolved oxygen sharply decreased with the depth, however, never dropped below ~ 10%. Vice versa, these two localities were characterized by much higher concentrations of ferrous iron and dissolved methane (Table 2) compared to those sites located upstream. Concentration of the ferrous iron reflects anaerobic conditions of the sediment and showed the highest concentration to occur in the deepest sediment layers (40-50cm). Average annual temperatures of interstitial water at localities in downstream part of the Sitka stream were about 2.5 oC higher compared to localities upstream and may probably promote higher methane production occuring here. Precursors of methanogenesis, acetate, propionate and butyrate were found to be present in the interstitial water at all study sites, however, only acetate was measured regularly at higher concentration with maximum concentration reached usually during a summer period.

CH4 can be produced and released into overlying near-bottom water through exchange at sediment-water interface. Methane released from the sediments into the overlying water column can be consumed by methanotrophs. Methanotrophs can oxidize as much as 100 % of methane production (Le Mer & Roger 2001). According to the season, 13-70 % of methane was consumed in a Hudson River water column (de Angelis et Scranton 1993). For the Sitka stream, measurement of benthic fluxes into the overlying surface waters indicates that methane consumption by methanotrophic bacteria is likely a dominant way of a methane loss, nevertheless some methane still supports relatively high average methane concentrations in the surface water and, in turn, high emissions to the atmosphere.

The methane production (measured as methanogenic potential) was found to be 3 orders of magnitude lower than the oxidation (methanotrophic activity), thus, almost all methane should be oxidized and consumed by methanotrophic bacteria and no methane would occur within the sediments. However, situation seems to be quite different suggesting that namely methanotrophic activity measured in a laboratory could be overestimated. Since oxidation of methane requires both available methane and oxygen, methanotrophic activity is expected to be high at sites where both methane and dissolved oxygen are available. Therefore, high values of the MA were usually found in the upper layers of the sediments (Segers 1998) or at interface between oxic and anoxic zones, respectively. Relatively high methanotrophic activity found in deeper sediments of the localities III-V indicates that methane oxidation is not restricted only to the surface sediments as is common in lakes but it also takes place at greater depths. It seems likely that oxic zone occurs in a vertical profile of the sediments and that methane diffusing from the deeper layer into the sedimentary aerobic zone is being oxidized by methanotrophs here. Increased methanotrophic activity at this hyporheic oxic-anoxic interface is probably evident also from higher abundance of type II methanotrophs in the same depth layer. Similar pathway of methane cycling has been observed by Kuivila et al. (1988) in well oxygenated sediments of Lake Washington, however, methane oxidation within the sediments would be rather normal in river sediments compared to lakes. All the above mentioned findings support our previous suggestions that coexistence of various metabolic processes in hyporheic sediments is common due to vertical and horizontal mixing of the interstitial water and occurrence of microbial biofilm (Hlavacova et al. 2005, 2006).

Five localities alongside stream profile were chosen for sampling sediment and interstitial water samples based on previous investigations (Figure 2, Table 1). Hyporheic sediments were collected with a freeze-core using N2 as a coolant (Bretschko & Klemens 1986) throughout summer period 2009-2011. At each locality, three cores were taken for subsequent analyses. After sampling, surface 0-25 cm sediment layer and layer of 25-50 cm in depth were immediately separated and were stored at a low temperature whilst being transported to the laboratory. Just after thawing, wet sediment of each layer was sieved and only particles < 1 mm were considered for the following microbial measurements and for all microbial activity measurements since most of the biofilm is associated with this fraction (Leichtfried 1988).

A few randomly selected subsamples (1 mL) were used for extraction of bacterial cells and, consequently, for estimations of bacterial numbers; other sub-samples were used for measurement of microbial activity and respiration, organic matter content determination, etc. Sediment organic matter content was determined by oven-drying at 105 oC to constant weight and subsequent combustion at 550 oC for 5 hours to obtain ash-free dry weight (AFDW). Organic matter values were then converted to carbon equivalents assuming 45 % carbon content of organic matter (Meyer et al. 1981). Sediment from another freeze-core was oven-dried at 105 °C and subjected to granulometric analysis. Grain size distribution and descriptive sediment parameters were computed using the database SeDi (Schonbauer & Lewandowski 1999).

Soil pollution represents a risk to human health in various ways including contamination of food, grown in polluted soils, as well as contamination of groundwater surface soils [68].

Classical remediation techniques such as soil washing, excavation, and chelate extraction are all labor-intensive and costly [69].

Phytoremediation of heavy metal contaminated soils is defined as the use of living green plants to transport and concentrate metals from the soil into the aboveground shoots, which are harvested with conventional agricultural methods [70]. The technique is suitable for cultivated land with low to moderate metal contaminated level. According to Jadia and Fulekar (2009) [71], phytoremediation is an environmental friendly technology, which may be useful because it can be carried out in situ at relatively low cost, with no secondary pollution and with the topsoil remaining intact. Furthermore, it is a cost-effective method, with aesthetic advantages and long term applicability. It is also a safe alternate to conventional soil clean up [17]. However, a major drawback of phytoremediation is that a given species typically remediates a very limited number of pollutants [24]. For example, a soil may be contaminated with a number of potentially toxic elements, together with persistent organic pollutants [72]. There are two different strategies to phytoextract metals from soils. The first approach is the use of metal hyperaccumulator species, whose shoots or leaves may contain rather high levels of metals [25]. The important traits for valuable hyperaccumulators are the high bioconcentration factor (root-to-soil metal concentration) and the high translocation factor (shoot to root metal concentration) [73]. Another strategy is to use fast-growing, high biomass crops that accumulate moderate levels of metals in their shoots for metal phytoremediation [25]. Phytoextraction ability of some fast growing plant species leads to the idea of connecting biomass production with soil remediation of contaminated industrial zones and regions. This biomass will contain significant amount of heavy metals and its energetic utilization has to be considered carefully to minimize negative environmental impacts [74].

Methane concentrations ranged between 0.18 — 35.47 pg L-1 in surface water and showed no expected trend of gradual increase from upstream localities to those laying downstream. However, significant enhancement of CH4 concentration was found on locality IV and V, respectively. Concentrations of dissolved CH4 inboth surface and interstitial waters peaked usually during summer and autumn period (Hlavacova et al. 2005, Mach et al. in review).

Generally, methane concentrations measured in interstitial water were much higher compared to those from surface stream water and on a long-term basis ranged between 0.19 — 11 698.9 pg L-1. Due to low methane concentrations in interstitial water at localities I and II, vertical distribution of its concentrations was studied only at the downstream located sites III-V. Significant increase of the methane with the sediment depth was observed at the localities IV and V, respectively. Namely locality IV proved to be a methane pool, methane concentrations in a depth of 40 cm were found to be one order of magnitude greater than those from the depth of 20 cm (Tab. 3). Recent data from locality IV show much lower methane concentrations in the upper sediment horizons compared to those from deeper layers (Fig. 3a). Considerable lowering of methane concentration in upper sediment horizons is likely caused by oxidizing activity of methanotrophic bacteria (Fig. 3d). while dissolved oxygen concentration sharply decreased with the sediment depth (Fig. 3b).

Usually, both the surface and interstitial water were found to be supersaturated compared to the atmosphere with locality IV displaying saturation ratio R to be almost 195 000. This high supersaturation greatly promote diffusive fluxes of methane to the atmosphere across air-water interface and is also an important mechanisms for loss of water column CH4.

Stable carbon isotope signature of carbon dioxide (S13C-CO2) measured in the interstitial water ranged from -19.8 %o to -0.8 %o, while carbon isotope signature of methane (S13C — CH4) ranged between -72 % to -19.8 %. This relatively high variation in the methane isotopic values could be caused due to consequential fractionation effects preferring light carbon isotopes like methane oxidation or fractionation through diffusion and through flow of an interstitial water. Contrary, the narrow range of the S13C-CH4 was found in the sediment depth of 40-60 cm where a high methane production has occured. Here, the S13C-CH4 values varied only from -67.9 % to -72 %. Apparent fractionation factor (ac) varied also greatly from 1,004 to 1,076. Usually values of ac > 1.065 and ac < 1.055 are characteristic for environments dominated by hydrogenothropic and acetoclastic methanogenesis, respectively. Our measurements indicate predominant occurrence of a hydrogenothropic methanogenesis in the high methanogenic zones where the most amount of methane is produced and S13C of CO2 values were markedly depleted (i. e. 13C enriched). This could be caused by enhanced carbon dioxide consumption by hydrogenothrophic methanogens, strongly preferring light isotopes. Nevertheless, both acetoclastic and hydrogenotrophic pathways take part in the methanogenesis along the longitudinal profile of the Sitka stream.

Methanogenic potential (дд CH4 kg-1 DW day-1)

Figure 3. Vertical distribution of methane concentration in the interstitial water at study site IV, horizontal bars indicate 1 SE

The presence of relatively rich assemblage of methanogenic archaea in hyporheic river sediments is rather surprising, however it is in accordance with other studies. The number of total different bands (i. e. estimated diversity of the methanoges) observed in the DGGE patterns of the methanogenic archaeal communities was comparable with a number of the DGGE bands found in other studies. For example, Ikenaga et al. (2004) in their study of methanogenic archaeal community in rice roots found 15-19 DGGE bands, while Watanabe et al. (2010) showed 27 bands at different positiosns in the DGGE band pattern obtained from Japanese paddy field soils. Our results from the DGGE analysis are supported by cloning and sequencing of methyl coenzyme M reductase (mcrA) gene which also retrieved relatively rich diversity (25 different mcrA gene clones) of the methanogenic community in the Sitka stream hyporheic sediments. Similar richness in number of clones was also mentioned in a methanogenic community in Zoige wetland, where 21 different clones were found (Zhang et al. 2008a), while 20 clones were described in the methane cycle of a meromictic lake in France (Biderre-Petit et al. 2011). In addition, soils from Ljubljana marsh (Slovenia) showed 17 clones (Jerman et al. 2009), for example. Both DGGE and mcrA gene sequencing results suggest that both hydrogenotrophic and acetoclastic methanogenesis are an integral part of the CH4 — producing pathway in the hyporheic zone and were represented by appropriate methanogenic populations. Further, these methanogenic archaea form important component of a hyporheic microbial community and may substantially affect CH4 cycling in the Sitka stream sediments.

5. Conclusion

To our knowledge this study is the first analysis of the composition of active methanogenic/methanotrophic communities in river hyporheic sediments. By use of various molecular methods we have shown that both methanogenic archaea and aerobic methanotrophs can be quantitatively dominant components of hyporheic biofilm community and may affect CH4 cycling in river sediments. Their distribution within hyporheic sediments, however, only partly reflects potential methane production and consumption rates of the sediments. Rather surprising is the detection of methanotrophs in the deep sediment layer 25-50 cm, indicating that suitable conditions for methane oxidation occur here. In addition, this work constitutes the first estimation of sources, sinks and fluxes of CH4 in the Sitka stream and in 3rd order stream environment. Fluxes of CH4 from supersaturated interstitial sediments appear to be a main CH4 source toward the water column. Compared with CH4 production rates, the diffusive fluxes are very low due to efficient aerobic oxidation by methanotrophic bacteria, especially during higher flow discharges. Although fluxes to the atmosphere from the Sitka stream seems to be insignificant, they are comparable or higher in comparison with fluxes from other aquatic ecosystems, especially those measured in running waters. Finally, our results suggest that the Sitka Stream is a source of methane into the atmosphere, and loss of carbon via the fluxes of this greenhouse gas out into the ecosystem can participate significantly in river self-purification.

Author details

Martin Rulik, Adam Bednarik, Vaclav Mach, Lenka Brablcova, Iva Buriankova,

Pavlina Badurova and Kristyna Gratzova

Department of Ecology and Environmental Sciences, Laboratory of Aquatic Microbial Ecology, Faculty of Science, Palacky University in Olomouc, Czech Republic

Surface water was collected from running water at a depth of 10 cm below the surface level in autumn 2009 at each study site. Interstitial water samples were collected using a set of 5-6 minipiezometers (Trulleyova et al. 2003) placed at a depth of about 20-50 cm randomly in sediments at each study site. The initial 50-100 mL of water was used as a rinse and discarded. As usual, two subsamples of interstitial water from each minipiezometer were collected from a continuous column of water with a 100 mL polypropylene syringe connected to a hard PVC tube, drawn from a minipiezometer and injected into sterile, clear vials (40 mL) with screw-tops, covered by a polypropylene cap with PTFE silicone septa (for analysis of dissolved gasses) and stored before returning to the laboratory. All samples were taken in the morning between 9 a. m. and 12 noon. All measurements were done during the normal discharge levels (i. e. no spates or high flood levels were included). Interstitial water temperature, dissolved oxygen (mg L-1 and percent saturation) and conductivity were measured in the field with a portable Hanna HI 9828 pH/ORP/EC/DO meter. Dissolved organic carbon (DOC) was measured by Pt-catalysed high temperature combustion on a TOC FORMACSHT analyser. Long term observation of interstitial water temperature was carried out using temperature dataloggers Minikin (EMS Brno, Czech Republic) buried in the sediment depth of 25-30 cm for a period of one year. Dissolved ferrous iron (Fe2+) concentration was measured using absorption spectrophotometry after reaction with 1,10- phenanthroline. Concentrations of organic acids were meausred using capilary electrophoresis equipped with diode array detector HP 3D CE Agilent (Waldbron, Germany). Limits of detection for particular organic acids were set as following: LOD (acetate) = 6,2 pmol L-1; LOD (propionate) = 4,8 pmol L-1; LOD (butyrate) = 2,9 pmol L-1; LOD 32 (valerate) = 1,8 pmol L-1.

Concentrations of dissolved methane in the stream and interstitial water were measured directly using a headspace equilibration technique. Dissolved methane was extracted from the water by replacing 10 mL of water with N2 and then vigorously shaking the vials for 15 seconds (to release the supersaturated gas from the water to facilitate equilibration between the water and gas phases). All samples were equilibrated with air at laboratory temperature. Methane was analysed from the headspace of the vials by injecting 2ml of air sub-sample with a gas-tight syringe into a CHROM 5 gas chromatograph, equipped with the flame ionization detector (CH4 detection limit = 1pg L-1) and with the 1.2m PORAPAK Q column (i. d. 3 mm), with nitrogen as a carrier gas. Gas concentration in water was calculated using

Henry’s law. The saturation ratio (R) was calculated as the measured concentration of gas divided by the concentration in equilibrium with the atmosphere at the temperature of the water sample using the solubility data of Wiesenburg & Guinasso (1979).

Many species have been used (either as hyperaccumulators, or as fast growing-high biomass crops) to accumulate metals, thus for their phytoremediation ability. Hyperaccumulators are these plant species, which are able to tolerate high metal concentrations in soils and to accumulate much more metal in their shoots than in their roots. By successive harvests of the aerial parts of the hyperaccumulator species, the heavy metals concentration in the soil can be reduced [23]. According to Chaney et al. (1997) [21], in order a plant species to serve the phytoextraction purpose, it should have strong capacities of uptake and accumulation of the heavy metals when it occurs in soil solution. For example, Sedum plumbizincicola is an hyperaccumulator that has been shown to have a remarkable capacity to extract Zn and Cd from contaminated soils [75]. In addition, a very good also hyperaccumulator for Zn and Cd phytoextraction is Thlaspi caerulescens [23]. Iris pseudacorus L. is an ornamental macrophyte of great potential for phytoremediation, to tolerate and accumulate Cr and Zn [19]. Furthermore, many species of Brassica are suitable for cultivation under Cu and Zn toxicity conditions and may be used for phytoremediation [29]. Phragmites australis, which is a species of Poaceae family, may tolerate extremely high concentrations of Zn, Cu, Pb and Cd, thus can be used as heavy metal phytoremediator [76].

Santana et al. (2012) [20] refer that Genipa americana L. is a tree species that tolerates high levels of Cr3+, therefore it can be used in recomposition of ciliary forests at Cr-polluted watersheds. According to the same authors, this woody species demonstrates a relevant capacity for phytoremediation of Cr. Elsholtzia splendens is regarded as a Cu tolerant and accumulating plant species [77]. Peng et al. (2012) [78] refer that Eucalyptus urophylla X E. grandis is a fast growing economic species that contributes to habitat restoration of degraded environments, such as the Pb contaminated ones. On the other hand, concerning Cd phytoextraction ability, only a few plant species have been accepted as Cd hyperaccumulators, including Brassica juncea, Thlaspi caerulescens and Solanum nigrum. Poplar (Populus L.), which is an easy to propagate and establish species and it has also the advantages of rapid growth, high biomass production, as well as the ability to accumulate high heavy metal concentrations, could be used as a Cd-hypaeraccumulator for phytoremediation [27-28,67]. According to Wang et al. (2012) [28], the increase in total Cd uptake by poplar genotypes in Cd contaminated soils is the result of enhanced biomass production under elevated CO2 conditions. Furthermore, Amaranthus hypochondriacus is a high biomass, fast growing and easily cultivated potential Cd hyperaccumulator [25]. Another species was found to be a good phytoremediator concerning its phytoaccumulation and tolerance to Ni stress is Riccinus communis L. [18]. Finally, Justicia gendarussa, which was proved to be able to tolerate and accumulate high concentration of heavy metals (and especially that of Al), could be used as a potential phytoremediator.

Differences between species, or genotypes of the same species, concerning heavy metal accumulation have been found by many researchers. According to Dheri et al. (2007) [17], the overall mean uptake of Cr in shoot was almost four times and in root was about two times greater in rays, compared to fenugreek. These findings, according to the same authors, indicated that family Cruciferae (raya) was most tolerant to Cr toxicity, followed by Chenopodiaceae (spinach) and Leguminosae (fenugreek). Peng et al. (2012) [78] found that cultivar ST-9 of Eucalyptus urophylla X E. grandis was shown to accumulate more Pb than others of the same species, like ST-2, or ST-29.