**Impact of Water Pollution on Trophic Transfer of Fatty Acids in Fish, Microalgae, and Zoobenthos in the Food Web of a Freshwater Ecosystem**

**Shahid Mahboob 1,2,\*, Khalid Abdullah Al-Ghanim 1, Fahad Al-Misned 1, Tehniat Shahid 3, Salma Sultana 2, Tayyaba Sultan 2, Bilal Hussain <sup>2</sup> and Zubair Ahmed <sup>1</sup>**


Received: 26 April 2019; Accepted: 8 June 2019; Published: 14 June 2019

**Abstract:** This research work was carried out to determine the effects of water contamination on the fatty acid (FA) profile of periphyton, zoobenthos, two Chinese carps and a common carp (*Hypophthalmichthys molitrix, Ctenopharygodon idella* and *Cyprinus carpio*), captured from highly polluted (HP), less polluted (LP), and non-polluted (NP) sites of the Indus river. We found that the concentration of heavy metals in the river water from the polluted locations exceeded the permissible limits suggested by the World Health Organization (WHO) and the US Environmental Protection Agency (EPA). Fatty acid profiles in periphyton, zoobenthos, *H. molitrix*, *C. idella*, and *C. carpio* in the food web of river ecosystems with different pollution levels were assessed. Lauric acid and arachidic acids were not detected in the biomass of periphyton and zoobenthos from HP and LP sites compared to NP sites. Alpha-linolenic acid (ALA), eicosadienoic acid and docosapentaenoic acid were not recorded in the biomass samples of periphyton and zoobenthos in both HP and LP sites. Caprylic acid, lauric acid, and arachidic acid were not found in *H. molitrix*, *C. idella*, and *C. carpio* captured from HP. In this study, 6 and 9 omega series FAs were identified in the muscle samples of *H. molitrix*, *C. idella* and *C. carpio* captured from HP and LP sites compared to NP sites, respectively. Less polyunsaturated fatty acids were observed in the muscle samples of *H. molitrix*, *C. idella*, and *C. carpio* collected from HP than from LP. The heavy metals showed significant negative correlations with the total FAs in periphyton, zoobenthos, and fish samples.

**Keywords:** Fatty acid; fish; food web; periphyton; trophic transfer; water pollution; zoobenthos

### **1. Introduction**

The aquatic fauna and flora of river ecosystems comprise a complex assemblage of different communities and are biologically important because of the interlinking between different trophic levels. These aquatic food chains are very feeble and sensitive to contaminants, especially to the toxicity of exogenous chemicals and heavy metals that are discharged into freshwater reservoirs due to various human activities. Heterotrophic aquatic organisms in food chains consume a variety of metalloids and xenobiotic compounds, which usually cause immutable degradation of the planktonic life at higher concentrations [1,2]. The toxic response in freshwater fish species to contaminated environments has been reported on a global scale [3,4]. The uptake of heavy metals into the aquatic food chain can occur either by dietary or non-dietary routes [5]. Therefore, the concentration of heavy metals in fish normally indicates levels present in sediment and water that is specifically in freshwater reservoirs

where the fish is captured from [6], as well as the time of exposure [7]. The concentration of essential metals, if increased above the normal metabolic needs of fish, may become toxic for the fish and for the ultimate consumer, humans [8]. Heavy metals may accumulate in primary producers such as microalgae, where diatoms ultimately transfer them to other trophic levels [9]. Heavy metals are ingested by fish and bio accumulate in the liver, kidneys, and other vital organs through adsorption and absorption [7].

Lipids are considered to be one of the most essential nutrients, which affect the growth and development of fish and other organisms [10], and alleviate immune competence [11]. Essential lipids are nutritionally important for the consumers in the food chain because they promote the growth and development and overall health of aquatic fauna and flora of aquatic communities in freshwater ecosystems [12–14]. Kainz et al. [15] proposed that the trophic movement of fatty acid (FA) in the food chain may be used as a physiological biomarker for monitoring the status of contamination in freshwater ecosystems. Kainz et al. [15] further mentioned that this trophic movement of FA in the food chain may be used as a physiological biomarker for observation of the status of contamination in freshwater ecosystems. Thus, it is necessary to ensure the abundance of microalgae and zoobenthos for trophic transfer into higher levels in the food web to ensure the transfer of FA and polyunsaturated fatty acid (PUFA) to the fish [16]. Currently, there are no comprehensive reports in the literature describing the interlinking trophic movements of PUFA and the impact of water pollution in the river ecosystems. The latter is still poorly understood with reference to FA profiles of periphyton, zoobenthos, and fish, and effect of contaminants. periphyton and zoobenthos can be used as valuable indicators to determine the effect of contamination and the synthesis of FA in freshwater ecosystems [17].

Aquaculture plays an important role in providing good quality animal protein and provides sustainable livelihood opportunities and food security for the ever increasing world population [18,19]. Fish are recognized as an important part of the human diet, owing to its balanced ratio proteins/PUFAs, including omega series FAs [20] which may reduce the risk of heart diseases. Because of the nutritional and pharmaceutical importance of PUFAs, researchers in the discipline of fishery sciences have been paying them increasing attention [21,22].

The bioaccumulation of metals in fish is triggered by the accumulation of these elements in phytoand zooplankton; however, this event ceases to be the most relevant as long as biomagnification takes place. Biomagnification can have serious impacts on the food chain [23]. Fish has the potential to accumulate more metals from food and water [24]. Kainz and Fisk [25] mentioned that most of the FAs and pollutants move trophically through the food chain, ultimately having effects on the final consumer. This situation warrants an understanding of the fate of FAs and the impact of heavy metals contamination on the variability of FA on the food chain in river ecosystems. Variation in FAs dynamics in the food web is linked to increases in environmental stress and habitat destruction due to water pollution within a freshwater ecosystem [26]. Moreover, the disparity in FAs and pollutant trophic movement in the food chain may give insights into ecological functions and their impact on habitat and environmental stress. Keeping this in mind, it is necessary to investigate the interlink and biotransformation of FAs in the food web and the relationship with water pollution. This requires assessing and contrasting the trophic movement of lipids and pollutants in the aquatic food chain. The main aim of research was (i) to assess the fatty acid profiles in periphyton, zoobenthos, *Hyphpthalmichthys molitrix*, *Ctenopharyngodon idella*, and *Cyprinus carpio* in the food web of river ecosystems with different pollution levels; (ii) to assess their flow in aquatic ecosystems; (iii) to explore their potential for evaluating and monitoring the health of aquatic habitats; (iv) and to apply FA profiles as a possible physical biomarker of environmental stress from heavy metal pollution.

### **2. Materials and Methods**

### *2.1. Study Area*

The Indus river is the longest river in Pakistan. The Indus River originates on the Tibetan Plateau, enters into towards Gilgit-Baltistan from Ladakh, and then flows from Punjab Province and joins into

the Arabian Sea. It is the largest river in Pakistan with a total catchment basin of about 1,165,000 km2 (450,000 m2) https://en.wikipedia.org/wiki/Indus\_River).

The Mianwali District is situated in the province of the Punjab and is about 200 m above sea level (Figure 1; 25). The Mianwali is one of the districts in the province of the Punjab and is about 200 m above sea level [27]. This district is rich in minerals, clay, coal, gypsum, limestone, etc., which are excavated for commercial purposes. In this district there is also a nuclear power plant and the Chashma Hydel power plant, which are adding electricity into the national electricity grid. The temperature ranged between –2 ◦C and 51 ◦C with 255 mm of rainfall [28]. The experimental sites were selected in Area 1 (Kalabagh; high pollution (HP) site), Area 2 (Chashma; low pollution (LP) site), and non-polluted (NP) site (Area 3; Attock) along the River, and these sites were 35 km apart from each other.

**Figure 1.** Map of the locations of sampling sites on the Indus River (Kundian Barrage, Kalabagh Barrage, and Chashma Barrage) [25]. Source: OCHA (United Nation Office for the Coordination Humanitarian Affairs).

### *2.2. Collection and Preparation of Fish Samples*

*Hypophthalmichthys molitrix* is planktivorous and consumes the organisms within lower multiple lower trophic levels across a range of habitats. Grass carp (*Ctenopharyngodon idella*) is a large cyprinid and is a voracious feeder. Small grass carp consume planktonic crustaceans, rotifers, and insect larvae, while the adults are completely vegetarian. *Cyprinus carpio* is a popular benthivorous fish that has larger bottom–up effects than other benthivorous fish. The bottom–up effects of *C. carpio* mainly depend on the incorporation of benthos-derived nutrients and the release of nutrients from bottom sediment during grazing on benthos. Twenty-one specimens of *H. molitrix, C. idella*, and *C. carpio* each were captured from HP, LP and NP sites for an evaluation of the fatty acid profiles. A total of 63 fish specimens were procured with the help of fishermen. The average weight ranged from 900 to 1200 g. Fish specimens were transferred live in polyethylene bags to the laboratory. Muscle samples were processed as per the method mentioned by reference [29]. This study was approved by the Ethics and Animal Welfare Committee of the Department of Zoology, GC University, Faisalabad (Ethical code number: GCUF/Zool/EAWC/34).

### *2.3. Analysis of Water Samples*

Water samples were collected in hydrographic bottles of 32 oz capacity at the depth of 30 cm below the surface from the three determined sampling sites for the determination of selected physiochemical parameters and selected heavy metals through an atomic absorption spectrophotometer ("Hitachi polarized Zeeman AAS, 2000 series") by following the procedure as mentioned by reference [30]. The water samples were collected in the morning and these were stored in iceboxes before being taken to the laboratory for analysis. Different dilutions of Hg, Sn, Cr, Pb, Zn, Mn, Cu, and Cd were made to check the accurateness of the equipment during the analysis of samples. The quality control and quality assurance protocol was followed as mentioned in our previous published work [9]. Calibration curves were plotted and validated with their corresponding R<sup>2</sup> values for the detection of each metal. The values of R<sup>2</sup> of the curves were 0.99983, 0.99981, 0.99951, 0.99984, 0.99926, 0.99987, and 0.99982 for Hg, Sn, Cr, Pb, Zn, Mn, Cu, and Cd, respectively.

### *2.4. Periphyton Sampling*

Periphyton samples were obtained from the three experimental locations by following the methodology of references [28,29]. "A 10 × 10 cm steel frame was fixed at the bottom at three points of each location and composite them to collect the periphyton, then the pebble was removed. The periphyton samples was cleaned from the pebble surface using brushes, after which it was washed with river water. Aliquots from this volume were centrifuged at 2500 g for 15 min for the further analysis of metals and fatty acids" [31].

### *2.5. Zoobenthos Sampling*

Zoobenthos samples were obtained from the experimental locations at three points and composite them through a Samples Surber-type kick-bottom sampler as mentioned by reference [31].

### *2.6. Fatty Acid Profiling*

The lipid components were obtained from the fish muscle, periphyton, and zoobenthos samples with help of Soxhlet extractor (Electrothermal EME6 England), as described by reference [9]. "The extracted lipids were transformed to fatty acid methyl esters using methanolic sulfuric acid by an esterification procedure", as described by references [32,33]. The fatty acid profiling was carried out by following the methods of reference [34], through gas chromatograph (Perkin Elmer Model 3920) with flame ionization detector (FID) column 2 m in length and 2 mm in diameter. The chromatograms recorded from all samples were used to observe the retention time of each fatty acid (Fatty acid methyl esters (FAMEs)) and these were compared to the chromatogram of a standard (mixture of pure FAMEs) as described by reference [35].

### *2.7. Statistical Analysis*

The data obtained was processed using Minitab software for analysis of variance (ANOVA) to assess the dissimilarity between various parameters of this study between the three sampling sites. Duncan's multiple range test (DMR test) (*p* < 0.05) was used to compare the means. "Shapiro-Wilk's W test and Levene's test" was used for normality and homogeneity of the data when necessary [36]. Correlation coefficients were calculated to determine the relationship between the concentration of heavy metals and the total FAs profile in fish and planktonic life from three sampling sites.

### **3. Results**

### *3.1. Physico-Chemical Factors and Heavy Metals*

The physico-chemical parameters of the water samples from sampling sites (HP, LP and NP) are presented in Table 1. The level of salinity of HP was about 2%, found to be close to the level of salinity of the open ocean (normally about 3%). pH levels were 12.1 ± 0.36, 8.6 ± 0.12, and 8.1 ± 0.08 in HP, LP, and NP sites, respectively. The pH level was very high at the HP site. The highest biochemical oxygen demand (81.2 ± 1.10 mg/L) and chemical oxygen demand (195.8 ± 1.16 mg/L) were recorded at the HP site. The concentration of total dissolved solids (2445.5 ± 8.41 mg/L) and total suspended solids (329.6 ± 6.41 mg/L) were very high at the HP site. The concentration of phenols and sulfates were highest at the HP site, closely followed by the LP site. The level of phenols at the HP site was 15 times higher than at the NP site (Table 1).

**Table 1.** Mean Physico-Chemical parameters and metal concentrations (± SE) at different sampling locations of Indus River.


BOD: Biological oxygen demand, COD: Chemical oxygen demand. TDS: Total dissolved solids, TSS: Total suspended solids. Different letters (a, b, c) in the same row represent significant (*p* < 0.05) differences. D; Desirable limits. P; Permissible limits. †; Effluent inland surface water quality standards. \*\* No relaxation.

The concentration of studied heavy metals are presented in Table 1. These concentrations exhibited significant variations between the three sites. The level of Sn, Cr, Pb, Mn, Cu, Cd, and Hg at the HP in fish muscle, periphyton, and zoobenthos biomass were highest at HP compared to LP and NP site and was above the upper limits stated by reference [34] (Tables 1 and 2). The highest level of Cu in muscle samples was detected in *C. carpio* and plankton from HP, followed by LP sites. The maximum level of metals was recorded in the muscle samples of *C. carpio* captured from HP, followed by LP and NP sites. The lowest concentration of these metals was recorded in the muscles of *C. idella* (Table 2).

### *3.2. Fatty Acids Profile*

The saturated fatty acids (SFAs) were low in the biomass of periphyton and zoobenthos obtained from HP and LP sites, compared to NP sites (Table 3). Lauric acid and arachidic acids were not detected in the biomass of periphyton and zoobenthos from HP and LP. The Environmental Protection Agency (EPA) value was significantly higher in the biomass sampled from NP, compared to HP and LP. The number of monounsaturated fatty acids (MUFAs) was higher in samples of periphyton and zoobenthos from NP (Table 3). Palmitoleic acid, vaccenic acid, oleic acid, eicosenic acid and erucic acid were not detected in periphyton biomass samples from HP and LP. PUFAs level was greater in periphyton and zoobenthos biomass from NP, compared to HP and LP sites. Alpha-linolenic acid (ALA), eicosadienoic acid, docosapentaenoic acid and docosapentaenoic acid were not detected in the biomass of periphyton and zoobenthos sampled from HP and LP. The percentage of EPA and DHA were higher in the periphyton biomass from HP, compared to NP (Table 3).


**Table 2.** Heavy metal concentrations (mg/kg) in the biomass of periphyton, zoobenthos and in the muscle of fish species from different sampling locations of the Indus River.

Values (Mean ± SE) are averages of five samples analyzed in triplicate. Different letters (a, b, c) in the same row represent significant (*p* < 0.05) differences.


**Table 3.** Fatty acids (% ±SE) in periphyton and zoobenthos from three sampling sites at different pollution levels in the Indus River.


**Table 3.** *Cont.*

SFAs: Saturated fatty acids; MUFAs; Monounsaturated fatty acids; PUFAs: Polyunsaturated fatty acids; Values (Mean ± SE) are averages of five samples for each fish species analyzed in triplicate. Different letters (a, b, c) in the same row represent significant (*p* < 0.05) differences.

The fish captured from HP exhibited lower FAs and SFAs compared to the fish captured from LP (Table 4). The percentage of PUFAs in *H. molitrix, C. idella*, and *C. carpio* captured from HP was 32.32 ± 0.65, 7.19 ± 0.35, and 26.13 ± 0.82%, respectively. The percentage of PUFAs in *H. molitrix, C. idella*, and *C. carpio* captured from NP was 48.65 ± 1.11, 41.55 ± 0.97, and 44.15 ± 1.90%, respectively. The percentage of MUFAs and SFAs in *H. molitrix, C. idella*, and *C. carpio* captured from HP were 6.74 ± 0.29, 5.14 ± 0.17, and 5.46 ± 0.54 and 43.38 ± 2.45, 62.94 ± 3.05, and 74.07 ± 4.14%, respectively. The total MUFAs and SFA profiles in *H. molitrix, C. idella*, and *C. carpio*, captured from LP showed a similar trend of fluctuations to fish from HP (Table 4).

**Table 4.** Fatty acid profile % (±SE) of fish muscle from three sites at different pollution levels.



**Table 4.** *Cont.*


**Table 4.** *Cont.*

SFAs: Saturated fatty acids; MUFAs; Monounsaturated fatty acids; PFAs: Polyunsaturated fatty acids; Values (Mean ± SE) are averages of five samples for each fish species analyzed in triplicate. Different letters (a, b, c) in the same row represent significant (*p* < 0.05) differences.

The maximum percentage of SFAs in *C. carpio* was observed in HP. A decrease in the abundance of *C. carpio* was noticed during the study period (Table 4). Caprylic acid (C8:0), lauric acid (C12:0) and C20:0 arachidic acid were not found in *H. molitrix, C. idella*, and *C. carpio* from HP. A very small amount of lauric acid (C12:0) and C20:0 arachidic acid was recorded in *H. molitrix, C. idella*, and *C. carpio* from LP. Eicosapentaenoic acid was not detected in any of the fish species collected from HP sites. In this study, 6 and 9 omega series FAs were found in muscle samples of *H. molitrix, C. idella*, and *C. carpio* from HP and LP, respectively. Linoleic acid (C18:4(n-3), eicosadienoic acid (C20:2 (n-6), and docosapentaenoic acid (C22:4 (n-6) were not recorded in fish from HP. Eicosapentaenoic acid (C20:5 (n-3) was detected only in the muscle samples of *H. molitrix* from HP. Total 11 omega series FA were recorded in muscle of *H. molitrix, C. idella*, and *C. carpio* from NP sites. Caprylic acid was not detected in *H. molitrix, C. idella*, and *C. carpio* from LP sites (Table 4). Myristic acid (C14:0) was recorded as 0.44 ± 0.05, 1.826 ± 0.21, and 2.651 ± 0.22% in *H. molitrix, C. idella*, and *C. carpio*, respectively, from HP. Myristic acid was determined as 1.54 ± 0.04, 1.37 ± 0.03, and 0.53 ± 0.02 and 0.14 ± 0.01, 4.68 ± 0.55, and 7.79 ± 0.44% in *H. molitrix, C. idella*, and *C. carpio* from LP and NP sites, respectively. Arachidic acid was not found in *H. molitrix* and *C. carpio* from HP sites. Oleic acid (C18:1 (n-9) was not detected in *C. idella* collected from HP sites. C16:1 (n-7) (palmitoleic acid), C16:1 (n-9) (Cis-7 hexadecenoic acid), and C20:1 (N-9) (Eicosenoic acid) were not found in these fish species captured from HP. C18:1 (n-7) (cis-vaccenic acid) was detected only in the muscle sample of *H. molitrix* from HP. C16:1 (n-7) was detected only in the muscle samples of *H. molitrix* from LP sites. The concentration of C16:1 (n-7) was only determined as 0.72 ± 0.04 in *C. carpio* from NP sites (Table 4).

Correlation indices that were calculated among the concentrations of total FAs in periphyton, zoobenthos, fish muscle, and heavy metals in water samples are presented in Table 5. It has been observed that Sn, Cr, Pb, Zn, Mn, Cu, Cd, and Hg indicated significantly negative correlations with total FA profile in periphyton, zoobenthos, and fish samples from HP and LP (Table 5). Highly significant negative correlations were observed among Cr, Zn, Mn, and Cu and the total fatty acid profile samples of periphyton, zoobenthos, and fish collected from the HP site. The variation in FA found positively correlated with the level of contamination of these heavy metals in the food web. The health of the aquatic system was found to be significantly affected by the water quality of the HP and LP sites of the river compared to the NP site, which possibly causes decreases in the abundance of periphyton and fish populations in the aquatic system.


*Biomolecules* **2019** , *9*, 231

### **4. Discussion**

The trophic transfer of FAs from periphyton to the organisms at higher trophic levels is important for their health and growth [14,37]. This movement of important nutrients in the food chain may be affected by different contaminants in the freshwater ecosystem [15]. The metals and metalloids, phenols, and organic contaminants in freshwater ecosystems enter the food of aquatic animals from various sources, including anthropogenic activities, and accumulate in planktonic life and fish. The heavy metals which accumulate can cause physiological stress on FA at different trophic levels in the food chain, and ultimately in humans [16,37].

Fish are used as a bioindicator for different organic and inorganic pollutants in freshwater ecosystems due to their presence in different trophic levels, because of their age, size, and mode of nutrition [15]. Various factors have effects on the distribution of aquatic fauna and flora in freshwater reservoirs [38,39]. Abiotic parameters are considered to mostly affect the pattern of distribution and richness of planktonic life [39,40]. The metals assessed in this study accumulated in fish directly from the water and planktonic life in the Indus River in the study area. In this study, higher concentrations of salinity, sulfates, phenol and heavy metals were the driving force which decreased the abundance of phytoplankton and zooplankton, and their FA profile. The phenol, sulfate, total dissolved solids (TDS), and TS values clearly indicated difference in their concentration at HP and LP sites. The level of salinity at HP was very close to that of brackish waters. The higher concentration of total TDS and TS at HP may be due to high turbidity. The presence of different metals in freshwater ecosystems varied with the physico-chemical factors of the corresponding ecosystem, particularly the pH and redox state. Reference [41] reported that the decrease in pH at high river discharges may release metals from complexes in the river and streams, which may be toxic to the aquatic fauna and flora in the ecosystem.

The levels of heavy metals in the water samples collected from HP passed the upper limits recommended by reference [37]. The heavy metals level in the water samples and in the muscles of *H. molitrix, C. idella*, and *C. carpio*, and planktonic biomass collected from HP. The bioaccumulation of heavy metals is known to influence the FA profile of fish. Reference [42,43] mentioned that metals stimulate cellular synthesis and metabolism of FA through β-oxidation, while pharmaceutical products act as peroxisomal proliferators [43–45]. Very limited information is available about the influence of heavy metals on Proliferation Activate Receptors (PPARs) expression and the transcription factors of FA metabolism in fish [46]. Elements in these fish species captured from HP and LP sites were accumulated by bio-concentration, and through food and water [16,37]. The increased concentration of heavy metals along with salinity and phenols at HP and LP sites compared to N P site probably are major factors which caused physiological variation in the food web and disturb the biosynthesis of FAs in *H. molitrix, C. idella*, and *C. carpio* [16,38]. The concentration of many heavy metals decreased in higher trophic levels in the food web [45]. In this study, similar results were obtained for most of the heavy metals, except for Hg [47,48]. This was particularly so in the higher trophic levels, and ultimately affected terrestrial ecosystems through fish [16,38].

PUFAs enter at the first trophic level of the food chain via FA synthesis in periphyton. Reference [49] has mentioned that light causes multiple effects on periphyton lipid metabolism and FA profiles. In general, higher light intensity normally causes oxidative damage to PUFA. In addition to the contamination, low light intensity and poor water quality at HP and LP sites influenced the abundance of periphyton producing high quality FA, thereby affecting PUFAs. The movement of FAs from periphyton to the fish level was found to be increasing with the pH at HP and LP sites. Thus, alkaline pH stress promoted an accumulation of TAG (Triacylglycerols) and a proportionally decrease in membrane lipids [50] In this research work, the changes in physio-chemical factors influenced the production of lipids in the planktonic life at HP and LP sites. The current findings seem to agree with the results of reference [51]. They had mentioned that phytoplanktonic abundance and their diversity were affected by eutrophication, which influence the FAs production due to interspecific variations in periphyton FA levels. However, there remains very little information on the molecular mechanisms involved in these abiotic environmental stressors.

The concentration of EPA, DHA and PUFAs was greater in the microalgae at LP compared to HP, which may due to the higher biomass of microalgae. The increase in microalgae growth is promoted by the higher concentration of nutrients, which might have promoted the synthesis of EPA [31,52]. Reference [53] reported that fluctuations in nutrient availability in the food chain affect on FAs profiles of periphyton. The fluctuations in FA profiles in the trophic levels of the aquatic food chain are probably due variations in the periphyton community composition. Our results of increases in the percentage of EPA were not in line with the results of reference [54]. Total PUFA and PUFA:SAFA ratios were reduced in periphyton and zoobenthos with increase in pollution at the HP site compared to the NP site. The level of PUFA in zoobenthos relies on various biotic and abiotic factors [50] such as food types and levels of contamination [31]. Reference [55] reported that increases in the concentration of Cd decreased the production FA profiles in *Chlorella vulgaris*. However, more Cd accumulated under N stress, which reduced the production of triglycerols in algae. DHA is necessary for the good growth of these freshwater fish species in aquatic ecosystems. The low level of DHA in planktonic food may affect the growth and development of different organs in freshwater fish species [31,56]. Here the reduction in the level of DHA was detrimental to the fatty acid profiles of *H. molitrix, C. idella*, and *C. carpio* from HP sites. The accumulation of PUFA in zoobenthos depends on various biotic and abiotic factors [52,57], food types [58] and pollution levels [37].

Fish are considered to be the best source of animal protein, globally. However, deterioration in their quality and losses in FAs cannot be recouped. Differences were non-significant for the FA profile in *C. carpio* sampled from HP and LP sites compared to NP sites, which exhibit an identical response to the chemical pollutants. The maximum percentage of SFAs in *C. carpio* was recorded in the fish procured from HP. Fish with a high concentration of SFAs need more energy for their movement and to search for food [59]. The SFAs C8:0, C12:0 and C20:0 were not recorded for *H. molitrix, C. idella*, and *C. carpio* captured from HP sites. The higher levels of heavy metals at HP and LP sites in the river adversely affected the synthesis of FAs in the three fish species. The higher SFA levels are probably due to de novo synthesis within these fish species. The heavy metals accumulate towards the bottom of the river, and *C. carpio* feeds on a variety of benthic organisms and macrophytes, thereby exposing it to high proportions of heavy metals [60]. The total MUFA concentrations recorded were supported by the findings of reference [61].

A significant lower percentage of PUFAs was noticed in *H. molitrix, C. idella*, and *C. carpio* from HP and LP. However, H. molitrix exhibited higher levels of ω-3 FAs and a large loss of ω-6 fatty acids. Eicosapentaenoic acid (C20:5n3) was not recorded in *C. idella* and *C. carpio* from HP. The concentrations of EP and C20: 5n3 were lower than those of menhaden oil. Identical results were reported by reference [62]. Linoleic acid, eicosadienoic acid and docosapentaenoic acid were not detected in fish procured from HP. Eicosapentaenoic acid was detected only in the muscle samples of H. molitrix from HP. The reduction in the production of PUFAs in *H. molitrix*, *C. idella*, and *C. carpio* from HP and LP may be due to increased levels of metals in the water at these locations in the river [63]. The zooplankton is a source of EPA and DHA for fish in the aquatic ecosystem [26]. The alterations in the food web, linked with an increase in environmental stress in freshwater ecosystems, invasive species, and habitat deterioration, may cause a significant variation in pollutant and lipid trophic transfer [26,64]. *C. idella* captured from HP and LP sites surprisingly responded to the general environment for FA profiles, although they feed on aquatic vegetation. We did not work on the FA profile of aquatic vegetation, and suspect that the alteration in the FA profile was due to an increased water pollution. The higher metal concentration might have affected the FA profile of the aquatic vegetation. This aspect may be verified in future studies. The alterations in the food web, linked with an increase in environmental stress in freshwater ecosystems, invasive species, and habitat deterioration, may cause a significant variation in pollutant and lipid trophic transfer [16]. The variations in FA and heavy metals trophic transfers in the food chain can provide insights into ecological functioning and the fallout of environmental stressors on the FA profile of different organisms in freshwater food webs.

### **5. Conclusions**

Lipids play a significant role in the bioaccumulation of lipophilic pollutants in freshwater fish. The increase of heavy metals in the waters of the Indus River has produced trophic transfers to periphyton, zoobenthos, and fish in highly polluted (HP) and less polluted (LP) sites. Polyunsaturated Fatty Acids (PUFAs) level was greater in periphyton and zoobenthos biomass from non-polluted (NP) sites, compared to HP and LP sites. Fatty acids in the fish muscles were affected by the level of contamination due to the alterations in the food web, linked with an increase in environmental stress, invasive species, and habitat deterioration. It has been inferred that abiotic factors and chemical pollutants induced the trophic transfer in the food, and ultimately the loss of essential fatty acids (FAs) in fish meat. The variations in FA and heavy metals trophic transfers in the food chain can provide insights into ecological functioning and the fallout of environmental stressors on the FA profile of different organisms in freshwater food webs.

It is proposed that FAs may be used to evaluate trophic relationships among water, planktonic life forms, and fish in the food web in order to provide information to consumers about the safety of fish meat. Thus, the variation in FA profiles may be used as a biomarker to assess the status of the health of the ecosystem, and possibly to identify the causes of decreases in the abundance of fish populations.

**Author Contributions:** Conceptualization, S.M. S.S., T.S. (Tayyaba Sultan) and B.H.; methodology; T.S. (Tehniat Shahid) and B.H.; software, B.H.; validation, K.A.A.-G., T.S. (Tehniat Shahid) and F.A.-M.; formal analysis, S.M.; investigation, B.H.; resources, K.A.A.-G.; data curation, B.H.; writing—original draft preparation, S.M.; writing—review and editing, S.S., T.S. (Tayyaba Sultan); and F.A.-M.; visualization, Z.A.; supervision, S.M.; project administration, F.A.-M.; funding acquisition, K.A.A.-G.

**Funding:** We (S.M. and K.A.A.-G.) would like to express our sincere appreciation to the Deanship of Scientific Research at the King Saud University for the funding of this research through the Research Group Project No. RG-1435-012. We also thank the Deanship of Scientific Research at King Saud University for the technical support.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Origin of Carbon and Essential Fatty Acids in Higher Trophic Level Fish in Headwater Stream Food Webs**

**Megumu Fujibayashi 1,2,\*, Yoshie Miura 1, Reina Suganuma 1, Shinji Takahashi 3, Takashi Sakamaki 4, Naoyuki Miyata <sup>1</sup>**


Received: 30 June 2019; Accepted: 12 September 2019; Published: 13 September 2019

**Abstract:** Dietary carbon sources in headwater stream food webs are divided into allochthonous and autochthonous organic matters. We hypothesized that: 1) the dietary allochthonous contribution for fish in headwater stream food webs positively relate with canopy cover; and 2) essential fatty acids originate from autochthonous organic matter regardless of canopy covers, because essential fatty acids, such as 20:5ω3 and 22:6ω3, are normally absent in allochthonous organic matters. We investigated predatory fish *Salvelinus leucomaenis* stomach contents in four headwater stream systems, which are located in subarctic region in northern Japan. In addition, stable carbon and nitrogen isotope ratios, fatty acid profile, and stable carbon isotope ratios of essential fatty acids were analyzed. Bulk stable carbon analysis showed the major contribution of autochthonous sources to assimilated carbon in *S. leucomaenis.* Surface baits in the stomach had intermediate stable carbon isotope ratios between autochthonous and allochthonous organic matter, indicating aquatic carbon was partly assimilated by surface baits. Stable carbon isotope ratios of essential fatty acids showed a positive relationship between autochthonous sources and *S. leucomaenis* across four study sites. This study demonstrated that the main supplier of dietary carbon and essential fatty acids was autochthonous organic matter even in headwater stream ecosystems under high canopy cover.

**Keywords:** fatty acids; dietary sources; allochthonous; *Salvelinus leucomaenis*

### **1. Introduction**

Aquatic animals are supported by two basal organic carbon sources, autochthonous (aquatic primary producers) and allochthonous sources (fallen leaf litter and insects from surrounding terrestrial ecosystems) [1]. Contributions of these basal organic carbon sources to aquatic food webs depend on the proportion of microalgal and phytoplankton productivity, and the number of terrestrial subsidies [2]. For lower order headwater streams, as high canopy cover promotes abundant inputs of litter falls [3] and limited productivity of attached algae due to shading effects [4], the main carbon sources for aquatic consumers are predicted as allochthonous by the river continuum concept (RCC) [5]. However, food web studies in headwater streams have demonstrated that the dominant dietary contribution for macroinvertebrates is both autochthonous [6,7] and allochthonous [8,9], indicating that the predominant carbon sources for headwater stream food webs are unclear.

Previous studies on the dietary contribution of these basal organic carbon sources have mainly focused on quantitative contribution; however, studies focusing on dietary quality are relatively rare in headwater stream food webs [10,11]. For instance, essential fatty acids are known to be important nutrition for fish health [12–14]. In particular, the roles of 20:5ω3 and 22:6ω3 in fish growth, survival, and reproduction have been studied in many species [15–17]. These studies have demonstrated that dietary essential fatty acids could improve fish condition. Freshwater fish can synthesize 20:5ω3 and 22:6ω3 if 18:3ω3, which is precursor of these two essential fatty acids, is available from dietary sources [13]. However, the conversion efficiency in aquatic animals is generally very low. Consequently, direct intake of these fatty acids from dietary sources is required [18]. However, terrestrial organic matter contains only 18:3ω3 but not 20:5ω3 and 22:6ω3, indicating that terrestrial organic matter is nutritionally poor [18]. 20:5ω3 is present at high levels in diatoms [19], which are sometimes the dominant algae attached to the surface of substrates in headstream ecosystems [20]. This implies that autochthonous organic carbon (i.e., attached algae) may be the main essential fatty acid source for consumers, although other carbon components are derived from allochthonous inputs in headwater streams.

Stable isotope ratios of bulk carbon in consumers reflect those of assimilated diet with only minor fractionation (<1%), which enable us to infer its dietary carbon sources [21]. In addition, stable carbon isotope ratios of autochthonous organic sources (e.g., attached algae) and allochthonous organic sources are distinguishable in many cases [8,22]. Accordingly, the contribution of allochthonous organic matter to stream food webs has been evaluated by bulk carbon stable isotope ratios [8]. Stable isotope ratios of bulk nitrogen (δ15N) have also been used in food web studies, because of its usefulness to evaluate the trophic position of animals owing to the substantial enrichment of approximately 3% relative to that of assimilated diet [23–25].

For tracing the dietary essential fatty acid origin, although their carbon stable isotope ratios would be helpful, information about isotopic fractionation of essential fatty acids between diet and consumers is limited. For instance, Budge et al. [26] demonstrated that there was no isotopic fractionation in several essential fatty acids between the diet and serum of fish in a feeding experiment with Atlantic Pollock. Moreover, an almost isotopically unchanged transfer of 18:2ω6 and 18:3ω3 between diet and zebrafish *Danio rerio* was observed in a 100-day feeding experiment with constant dietary sources [27]. Several aquatic food web studies have already assumed no fractionation of isotopic value of essential fatty acids [28,29]. On the contrary Gradyshev et al. [30] found gradual depletion of stable carbon isotope ratios of essential fatty acids, including 18:2ω6 and 18:3ω3, through higher trophic levels in a stream food chain, suggesting that fractionation was negative. Depleted fractionation in 18:3ω3 was also reported in zooplankton in a feeding experiment [31]. Depleted fractionation in other essential fatty acids have been also reported by a feeding experiments with Daphnia [32]. This can be explained by lighter compounds being assimilated preferentially. As above, the information on fractionation of fatty acids are not sufficient and they were conflicting (i.e., no or small fractionation and negative fractionation). Nielsen et al. [33] pointed out more information on the fractionation of fatty acids are required for diet tracing study. Thus, we did not apply a constant value of isotopic fractionation of essential fatty acids in a food chain in this study. We applied a correlation analysis of stable carbon isotope ratios of essential fatty acids between diet and consumer among the study sites. If organic sources consistently contribute to consumers and have wider isotopic differences, one would expect to detect a significant and positive relationship in stable carbon isotope ratios of essential fatty acids between consumers and assimilated food source regardless of isotopic fractionation [34–36].

Here, we tested the following two hypotheses that: 1) the dietary allochthonous contribution for fish in headwater stream food webs positively relates with canopy cover; and 2) essential fatty acids originate from autochthonous organic matter regardless of canopy covers. To test these hypotheses, we analyzed fatty acid compositions and bulk carbon and compound-specific isotope ratios in fish from four headstream ecosystems.

### **2. Materials and Methods**

### *2.1. Sampling*

We conducted field surveys in four headwater streams, located in subarctic area in the northern part of Japan from July to September 2016 (Table 1). Canopy cover was calculated from a hemispherical photography taken from the center of each stream using CanopOn2 program [37].


**Table 1.** Description of study sites in this study.

*Salvelinus leucomaenis*is a dominant predatory fish in these four study sites. The main dietary sources of *S. leucomaenis* are larvae and adults of aquatic insects and terrestrial insects [38,39]. *S. leucomaenis* was sampled by fishing. The total length and whole wet weight were measured and a muscle near the pelvic fins was dissected for further analyses. The stomach was preserved in 90% ethanol. For autochthonous organic sources analyses, epilithic biofilm, which was mainly composed of attached algae, was removed using a brush from several randomly selected stones. The collected epilithic biofilm was placed in a plastic sampling bottle with distilled water. The bottle containing algae was filtered through two glass filters (GFF; Whatman, Little Chalfont, UK) in the laboratory for fatty acid and bulk stable carbon and nitrogen isotope analyses. For allochthonous organic sources analyses, decomposed immersed leaf litter was sampled into a plastic bag. Both autochthonous and allochthonous organic sources were sampled in triplicate. The larvae of aquatic insects including Ephemeroptera, Trichoptera, and Plecoptera, which are potentially a direct dietary source for *S. leucomaenis*, were collected using D-frame nets (250 μm mesh) and sorted in the laboratory. Heptageniidae and Ephemerellidae were used for further analysis as they were dominant and commonly detected across all four study sites. All collected samples were transported to the laboratory in a cooler box. All samples were placed in a plastic bag separately and stored in a freezer at −20 ◦C until further analysis.

Additional sampling for analyses of bulk stable isotope ratios of *S. leucomaenis*, its stomach contents, epilithic biofilms, and leaf litter were conducted in the same location of Babame in July 2018. The same sampling procedure was applied except for stomach contents. The stomach of *S. leucomaenis* was placed in a plastic bottle with distilled water and moved to laboratory in a cooler. The stomach contents were identified and separated. *S. leucomaenis*, epilithic biofilms, and leaf litter were treated following the method mentioned above. All samples were preserved in a plastic bag and stored in a freezer at −20 ◦C until further analysis.

### *2.2. Analyses*

The stomach contents of each *S. leucomaenis* individual were divided into four groups based on the morphological characteristic using a stereoscopic microscope: water baits (larvae of aquatic insects) and surface baits (adults of aquatic insects and terrestrial insects) according to the definition of Tsuda [38], terrestrial plants, and unknown. Each group was weighted and the contribution of each group was calculated.

Freeze-dried samples of *S. leucomaenis*, both organic sources, and aquatic insects were used for the 'one-step method' [40] for fatty acid analysis. For aquatic insects, two or three individuals were pooled as one sample, and prepared three samples in each study sites. Freeze-dried samples were moved to a 10 mL glass tube. For *S. leucomaenis,* aquatic insects, and leaf litter, approximately 50 mg of homogenized sample was used. For epilithic biofilms, one sheet of GFF was used in the analysis. One milliliter of an internal standard (0.1 mg of tricosanoic acid per 1 mL of hexane), 1 mL of hexane,

and 0.8 mL of 14% boron trifluoride methanol were added to the 10 mL glass tube. Nitrogen gas was then used to fill the head space. The glass tube was placed in a 100 ◦C dry bath for 2 h, followed by cooling to room temperature, and 0.5 mL of hexane and 1 mL of ultrapure water were added. The glass tube was vigorously shaken manually and centrifuged for 3 min at 2,500 rpm. The upper layer of hexane, containing fatty acid methyl esters (FAMEs), was transferred to a 1.5 mL gas chromatography (GC) vial. Solid residues of *S. leucomaenis* were used for bulk carbon and nitrogen isotope ratio analysis.

One microliter of FAME solution was injected in a gas chromatograph (Trace GC, Thermo Fisher Scientific, Bremen, Germany) equipped with a capillary column (Select FAME, 100 m × 0.25 mm i.d.; Agilent Technologies, Santa Clara, CA, USA). The GC analysis was carried out under the analytical conditions described by Fujibayashi et al. [28]. Each fatty acid peak was identified by comparing their retention times with those of commercial authentic standard mixtures (Supelco, Inc., Bellefonte, PA, USA). The peak area was used for calculating the contribution of each fatty acid to total fatty acids.

After fatty acid analysis by GC, the remaining hexane sample was used to analyze the essential fatty acids stable carbon isotope ratio. FAMEs in the hexane solution were injected into a GC–isotope ratio mass spectroscopy instrument (Trace GC Ultra/Delta-V Advantage; Thermo Fisher Scientific, Bremen, Germany), which was equipped with a capillary column (SP2560, 100 m × 0.25 mm i.d.; Supelco, Inc., Bellefonte, PA, USA). The operating conditions were as described by Fujibayashi et al. [29]. Each essential fatty acid peak was identified as described above for the GC-FID analysis. Stable carbon isotope ratios of fatty acids were determined using the following formula:

$$\text{S}^{13}\text{C or N} \left(\%\_{00}\right) = \left(\text{R}\_{\text{sample}} / \text{R}\_{\text{standard}} - 1\right) \times 1000 \tag{1}$$

where Rsample is the 13C/ 12C ratio of the sample, and Rstandard is the 13C/ 12C ratio of the international isotopic standard (i.e., Vienna Pee Dee Belemnite).

Correction for the effect of additional carbon from boron trifluoride methanol on δ13C was conducted according to Fujibayashi et al. [29]. The stable carbon isotopes of fatty acids in *S. leucomaenis*, epilithic biofilms, and terrestrial litter samples were analyzed.

Dried solid residues of *S. leucomaenis*, subsamples of freeze-dried terrestrial litter, GFFs (epilithic biofilms), and aquatic insects were used for bulk stable carbon and nitrogen isotope ratio analysis. Utilization of solid residues after a one-step method potentially changes the isotopic value. Therefore, the relationship between the stable isotope ratios of carbon and nitrogen in original samples and those in the corresponding dried solid residue after the one-step method were checked in advance with freshwater fish muscle samples (Supplementary file Figure S1). For nitrogen, while a significant positive relationship was detected, variation was relatively high. Thus, estimation of trophic position of *S. leucomaenis* using solid resides may include some extents of error. However, stable carbon isotope rations of solid residues well reflected that of the original samples, and the utilization of solid resides for stable carbon isotope analysis was applied in this study. For the sampling of aquatic insects, one individual was used for one sample, and we prepared three samples for each study site. All samples were weighed in microcapsules and injected into an elemental analyzer (Flash EA; Thermo Fisher Scientific, Bremen, Germany) linked to a mass spectrometer (Delta-V Advantage; Thermo Fisher Scientific, Bremen, Germany). Stable isotope ratios of bulk carbon and nitrogen were expressed as Equation (1); where Rsample is the 13C/ 12C or 15N/ 14N ratio of the sample, and Rstandard is the 13C/ 12C and the 15N/ 14N ratio of the international isotopic standard (Vienna Pee Dee Belemnite, and atmospheric N2, respectively).

### **3. Results**

Although canopy cover was high in both the first order rivers, Babame and Kurikoma, with 93.5% and 91.1%, respectively; the third-order rivers, Hayakuchi and Naruse, had relatively open canopy with 67.6% and 63.5%, respectively.

Ten and 11 individuals of *S*. *leucomaenis* were caught in Babame and Kurikoma by fishing; however, just one individual was caught in Hayakuchi and Naruse. There was a relatively high proportion, ranging from 31% to 68%, of unknown components in the stomach contents that could not be identified because of decomposition (Figure 1). Terrestrial plants were almost not detected in the stomach contents, while water and surface baits were dominant in the stomach contents in *S*. *leucomaenis*. There was no obvious relationship between water and surface bait contribution and canopy cover.

**Figure 1.** Stomach contents (wet weight %) of *S. leucomaenis* from four study sites.

Although stable carbon and nitrogen isotope ratios of *S*. *leucomaenis* varied among study sites, the trophic position of *S*. *leucomaenis* was always the highest (Figure 2). Aquatic insects were generally at a lower position than *S. leucomaenis* with similar carbon isotopic values. Leaf litter showed the most depleted isotopic value for both carbon and nitrogen in all study sites. The range of stable carbon isotope ratios of leaf litter was relatively narrow, from −31.5% in Hayakuchi to −29.8% in Kurikoma. For epilithic biofilms, stable carbon isotope ratios were more enriched than terrestrial litter and showed a wider range, from −27.2% in Kurikoma to −23.8% in Hayakuchi.

All essential fatty acids were detected in *S. leucomaenis* (Figure 3). The major essential fatty acid was 22:6ω3. The average contribution of 22:6ω3 was the highest in *S. leucomaenis* from Kurikoma. With respect to other essential fatty acids, the other omega-3 fatty acids, such as 18:3ω3 and 20:5ω3, presented a higher contribution than that of omega-6 fatty acids. In both aquatic insects, essential fatty acid distribution was similar, with no 22:6ω3. The major fatty acids in both aquatic insects were 18:3ω3 and 20:5ω3. This essential fatty acid pattern was similar among study sites. Epilithic biofilms mainly consisted of 18:3ω3 and 20:5ω3. The contribution of 20:5ω3 was relatively constant in all study sites, while 18:3ω3 contribution varied widely among study sites. Only small amounts of 22:6ω3 were detected from epilithic biofilms. Terrestrial litter only contained 18:2ω6 and 18:3ω3. Other C20 essential fatty acids were only detected at low percentages.

The stable carbon isotope ratios of 18:2ω6, 20:4ω6, and 20:5ω3 in *S. leucomaenis* and epilithic biofilms were almost the same across study sites (Supplementary file Figure S2); consequently, a significant or marginally positive relationship was detected between them (correlation analysis: 18:2ω6, n = 4, r = 0.98, *p* < 0.01; 20:5ω3, n = 4, r = 0.99, *p* < 0.001; 20:4ω6, n = 4, r = 0.94, *p* = 0.063) (Figure 4). Although there was no statistical significance, a positive trend was detected between the stable carbon isotope ratios of 18:3ω3 in epilithic biofilms and that in *S. leucomaenis*. The stable carbon isotope ratios of 18:3ω3 were generally lower in the epilithic biofilms than in *S. leucomaenis*

(Supplementary file Figure S2). Contrarily, the stable carbon isotope ratios of 18:2ω6 and 18:3ω3 in leaf litter were not positively correlated with those of in *S. leucomaenis*.

**Figure 2.** Stable isotope ratios biplot for bulk carbon and nitrogen in basal organic carbon sources and consumers in four study sites. Error bars represent standard deviation.

**Figure 3.** Contribution of essential fatty acids in basal organic carbon sources and consumers in the four study sites. Error bars represent standard deviation.

The bulk carbon and nitrogen isotope ratios of leaf litter in Babame in July 2018 showed values similar to those in July 2016, −29.3% for carbon, and −2.1% for nitrogen (Figure 5). The carbon and nitrogen stable isotope ratios of epilithic biofilms were −25.1% and 4.3%, respectively. Terrestrial insects in the stomach of *S. leucomaenis* were between leaf litters and epilithic biofilms for carbon and nitrogen stable isotope ratios, with a mean value of −26.7% for carbon and 2.2% for nitrogen. The bulk stable carbon and nitrogen isotope ratios of *S. leucomaenis* were the most enriched among all samples, and close to those of epilithic biofilms.

**Figure 4.** Relationship between stable isotope ratios of essential fatty acid in *S. leucomaenis* and basal organic sources from the four study sites. The black triangle and open circle represent autochthonous organic sources (epilithic biofilms) and allochthonous organic sources (leaf litters), respectively. Error bars represent standard deviation.

**Figure 5.** Biplot for stable isotope ratios of bulk carbon and nitrogen of basal organic carbon sources, *S. leucomaenis,* and surface baits from *S. leucomaenis* stomachs in Babame, 2018. Error bars represent standard deviation.

### **4. Discussion**

### *4.1. Origin of Organic Sources*

Canopy cover is an important factor for biogeochemical and biological processes in headwater streams [41]. High leaf litter input and limited primary production are expected in headwater streams. Therefore, a positive relationship between allochthonous contribution and canopy cover was expected [42]. For instance, dietary inputs of surface baits, such as emerged aquatic insects and terrestrial insects, can be expected to increase as the canopy cover increased. However, the stomach contents did not show the expected patterns. Furthermore, our results of bulk carbon stable isotope ratios demonstrated that allochthonous contribution was very rare in *S. leucomaenis*, even in the Babame and Kurikoma study sites, where the canopy cover was very high (>90%). The isotopic positions of *S. leucomaenis* were relatively close to aquatic insects for carbon and higher for nitrogen, indicating that aquatic insects were diet items for *S. leucomaenis*. There were isotopic differences of 2–3% in carbon between epilithic biofilms and aquatic insects. The differences seem to be relatively high if a dietary relationship was assumed between epilithic biofilms and aquatic insects, because 0–1% fractionations have generally been assumed [21]. These relatively high differences may be explained under some assumptions. First, epilithic biofilms are the mixture of various organic sources with not only algal species, but also terrestrial organic matter [43]. According to our fatty acid analysis of epilithic biofilms, 24:0, which is a fatty acid biomarker of higher plants [44], was detected at 1–2% (data not shown), indicating that the analyzed epilithic biofilms contained terrestrial organic matter. Therefore, the stable carbon isotope ratios are the average of all contained organic matter [43]. The stable carbon isotope ratios of leaf litter showed generally low values in our study sites, indicating pure attached algae stable isotope values were likely more enriched than the analyzed values. If aquatic grazers can selectively utilize specific preferred carbon sources from periphyton [45], the observed wider fractionation between aquatic insects and epilithic biofilms is explainable. This wider fractionation could also be explained by temporal variation of stable carbon isotope ratios in epilithic biofilms. Although the leaf litter stable carbon isotope ratios were relatively constant among the study sites, those of epilithic biofilms widely varied, even in the same study sites between 2016 and 2018 in Babame. It is known that algal stable carbon ratios varied along the gradient of some environmental factors such as isotopic value of dissolved inorganic carbon [8,46] and growth stage [46,47]. This potentially high variability made it difficult to infer the algae dietary contribution for consumers under one-time sampling of stable carbon isotope. Contrastingly, stable carbon isotope ratios of animals were considered to integrate relatively long times [22]. The stable carbon isotope ratios of aquatic insects were generally higher than those of leaf litter and relatively similar to those of attached algae. Thus, the main organic source for *S. leucomaenis* seems to be autochthonous, that is, epilithic biofilms transferred through aquatic insects, regardless of canopy cover in the study sites.

Major dietary contribution of autochthonous sources for headstream consumers has been reported for tropical [6], subtropical [48], and temperate regions [7]. For instance, Lewis et al. [49] showed a major contribution of autochthonous dietary input even under the dominant input of allochthonous organic matters. These contrary observations against the predictions of RCC can be attributed to poor food quality of terrestrial organic matters and high quality of algae [18,50]. However, as observed from the stomach contents analysis, there was a substantial contribution of fallen insects to *S. leucomaenis* diet, in accordance to previous studies on *S. leucomaenis* stomach content [51]. In this study, we could not identify each species. However, according to the stable carbon isotope ratio analysis of surface baits in stomach contents, the carbon sources of these surface baits must have been derived partly from aquatic algae. Some riparian insects spend larval life in aquatic ecosystems and have been known to play an important role transferring highly unsaturated fatty acids from aquatic to terrestrial ecosystems by emerging [52]. This indicated that parts of surface baits can also be a vector of autochthonous carbon going back to aquatic ecosystems. Several researchers pointed out that dietary utilization of riparian insects is one of the pathways to acquire autochthonous organic sources [18]. Therefore, to evaluate organic matter origin, assimilation-based methods (e.g., stable isotope and fatty acids) are required.

### *4.2. Origin of Essential Fatty Acids*

It is well known that lipids have more negative δ13C values than that of other biochemical compounds because lighter carbons tend to be used for conversion of pyruvate to acetyl coenzyme A in lipid synthesis [53]. Therefore, several studies have reported more depleted isotopic values in essential fatty acids than that of bulk carbon [27,30]. Similarly, stable isotope ratios of essential fatty acids were substantially depleted compared with that of bulk carbon in the current study. Furthermore, spatial difference was also observed in both bulk and essential fatty acids isotope. Enriched isotopic value was observed in Hayakuchi and Naruse Rivers where canopy cover was relatively low, indicating high availability of sunlight for photosynthesis. It is known that high growth rate makes algal isotopic value enriched due to the increase of contribution of heavy CO2 [46]. The observed wider difference in isotopic ratios of epilithic biofilm can be reflected in the photosynthetic activity in each study site.

*S. leucomaenis* contained all essential fatty acids. 18:2ω6 and 18:3ω3 were detected in both organic carbon sources, leaf litters and epilithic biofilms. These fatty acids are not synthesized by animals [54], indicating the origin of these fatty acids in *S. leucomaenis* was either or both of them. We observed a significant positive relationship of stable carbon isotope ratios of 18:2ω6 between epilithic biofilms and *S. leucomaenis* indicating that this essential fatty acids in *S. leucomaenis* was mainly of autochthonous origin. For 18:3ω3, although a positive tendency was observed between epilithic biofilms and *S. leucomaenis*, this relationship was not statistically significant. However, since 18:3ω3 cannot be synthesized by *S. leucomaenis*, 18:3ω3 must come from either epilithic biofilms or leaf litter. Epilithic biofilms seem to be a probable candidate for the origin in 18:3ω3 for *S. leucomaenis* because isotopic values of leaf litter were highly depleted compared to that of *S. leucomaenis*, which cannot be explained by the previously reported isotopic fractionation, small [27] or depleted and inconsistent [32].

The origin of 20:5ω3 and 20:4ω6 seems to be epilithic biofilms or biosynthesis from their corresponding precursors, namely 18:3ω3 and 18:2ω6, respectively [13]. The stable carbon isotope ratios of both essential fatty acids showed positive relationships between *S. leucomaenis* and epilithic biofilm, indicating that these essential fatty acids were also of autochthonous origin. If we assume the origin of essential fatty acids to be epilithic biofilm, isotopic fractionation through two trophic levels, namely epilithic biofilm, aquatic insects, and *S. leucomaenis* for 18:2ω6, 20:5ω3, and 20:4ω6 was−0.9–0.4%, 0–1.5%, and−1.0–3.3%, respectively. On the contrary, for 18:3ω3, the expected isotopic fractionation of epilithic biofilm to *S. leucomaenis* via aquatic insects was 1.6–8.0%. Fujibayashi et al. [29] found no significant difference of isotopic value of 18:3ω3 between freshwater fish and blooming cyanobacteria. However, the mechanism of this inconsistent and variable fractionation for 18:3ω3 was not explainable in this study. Further research is required for isotopic fractionation of essential fatty acids in food chains.

While 22:6ω3 was the most abundant essential fatty in *S. leucomaenis*, both organic sources and aquatic insects did not contain 22:6ω3. However, we only analyzed the fatty acid content in two ephemeral groups. Moreover, the absence or very small contribution of 22:6ω3 in aquatic insects has been reported for a wide range of aquatic insect taxa [55–57]. As 22:6ω3 was less available from dietary sources, 22:6ω3 detected in *S. leucomaenis* must be biosynthesized from its precursor [11]. During elongation from 20:5ω3 to 22:6ω3, lighter carbon may be preferentially added. Consequently, 22:6ω3 isotopic value depleted compared with that of 20:5ω3 [27]. However, we found that almost the same or slightly enriched isotopic values in 22:6ω3 compared with that of 20:5ω3 (Supplementary file Figure S2). The same tendency was observed in several aquatic consumers, including fish, in Yenisei River [30]. Gladyshev et al. [30] pointed out the possibility that the acetate pool, which is required for elongation of fatty acid, is significantly enriched in 13C compared with fatty acids. While further study is needed to comprehensively understand essential fatty acid dynamics in aquatic ecosystems,

our results demonstrated that the main source of essential fatty acids in headstream food webs was autochthonous organic matter.

### **5. Conclusions**

Dietary origin of total organic carbon and essential fatty acids for the predatory fish *S. leucomaenis* was investigated in the four headwater streams with the two hypotheses: (1) the dietary allochthonous contribution for fish in headwater stream food webs positively relate with canopy cover; and (2) essential fatty acids originate from autochthonous organic matter regardless of canopy cover. Our results indicated that autochthonous organic matters were the main dietary origin of not only essential fatty acids, but also total organic carbon regardless of canopy cover.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2218-273X/9/9/487/s1. Figure S1 Relationship between stable isotope ratios of bulk and lipid free fish muscle samples (n = 137). Figure S2 Isotopic value of essential fatty acids in collected samples from each study stream. Error bars represent standard deviation.

**Author Contributions:** Conceptualization, M.F., Y.M. and R.S.; Formal analysis, M.F.; Funding acquisition, M.F. and S.K.; Investigation, M.F., Y.M., R.S. and S.T.; Methodology, M.F., Y.M. and R.S.; Validation, M.F., Y. M. and R. S.; Visualization, M.F., Y. M. and R. S.; Writing – original draft, M.F.; Writing – review & editing, M.F., Y.M., R.S., S.T., T.S., N.M. and S.K.; Supervision, M.F.

**Funding:** This study was funded by JSPS KAKENHI (Grant numbers 16H0236).

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
