Characteristics of the Biochemical Composition and Bioavailability of Phytoplankton-Derived Particulate Organic Matter in the Chukchi Sea, Arctic

: Analysis of the biochemical composition (carbohydrates, CHO; proteins, PRT; lipids, LIP) of particulate organic matter (POM, mainly phytoplankton) is used to assess trophic states, and the quantity of food material is generally assessed to determine bioavailability; however, bioavailability is reduced or changed by enzymatic hydrolysis. Here, we investigated the current trophic state and bioavailability of phytoplankton in the Chukchi Sea (including the Chukchi Borderland) during the summer of 2017. Based on a cluster analysis, our 12 stations were divided into three groups: the southern, middle, and northern parts of the Chukchi Sea. A principal component analysis (PCA) revealed that relatively nutrient-rich and high-temperature waters in the southern part of the Chukchi Sea enhanced the microphytoplankton biomass, while picophytoplankton were linked to a high contribution of meltwater derived from sea ice melting in the northern part of the sea. The total PRT accounted for 41.8% ( ± 7.5%) of the POM in the southern part of the sea, and this contribution was higher than those in the middle (26.5 ± 7.5%) and northern (26.5 ± 10.6%) parts, whereas the CHO accounted for more than half of the total POM in the northern parts. As determined by enzymatic hydrolysis, LIP were more rapidly mineralized in the southern part of the Chukchi Sea, whereas CHO were largely used as source of energy for higher trophic levels in the northern part of the Chukchi Sea. Speciﬁcally, the bioavailable fraction of POM in the northern part of the Chukchi Sea was higher than it was in the other parts. The ﬁndings indicate that increasing meltwater and a low nutrient supply lead to smaller cell sizes of phytoplankton and their taxa (ﬂagellate and green algae) with more CHO and a negative e ﬀ ect on the total concentration of POM. However, in terms of bioavailability (food utilization), which determines the rate at which digested food is used by consumers, potentially available food could have positive e ﬀ ects on ecosystem functioning. fractions under su ﬃ cient nitrogen conditions and growth stages of phytoplankton, which could lead to a high ( > 1) PRT:CHO ratio [13,15,16]. Fogg and Thake [60] and Hu [61] reported that as prolonged stressful conditions (such as nitrogen limitation) occur, metabolic changes in synthesizing enzyme systems can convert CHO into LIP synthesis. Thus, our results suggest that at least in the southern region, nitrogen limitation was not severe enough to limit phytoplankton growth. The northern part has not been exposed to nitrogen stress for a long time.


Introduction
In terms of bottom-up controls, phytoplankton is key organism that serves as a primary producer and primary food source for organisms at higher trophic levels in the foodwebs of aquatic ecosystems. Climate change enhances the sea ice melting in the Arctic Ocean with increasing concerns about primary production and nutrient cycling. Sea ice loss reduces surface albedo and enhances light penetration, creating irregularities on the timing and the duration of phytoplankton blooms [1,2]. These conditions can create discontinuity between the available food resources and the nutritional demands of higher

Field Sampling and Measurements of the Environmental Variables
This study was carried out at 12 stations in the Chukchi Sea onboard the R/V Araon icebreaker from 7 to 24 August 2017 ( Figure 1A). The potential temperature, salinity, and photosynthetically active radiation (PAR) from the surface to a 100 m depth were measured by a rosette-mounted Sea-Bird conductivity-temperature-depth (CTD) system-1% PAR at the surface light level was defined as the euphotic layer [34] by a Secchi disc using the vertical attenuation coefficient (Kd = 1.7/secchi depth). The meltwater percentage (MW; %) was calculated from the salinity at each sampled depth (S meas ) and the greatest depth (either the bottom depth or 100 m in this study; S deep ), assuming an average sea ice salinity of 6 [35,36] since the mean salinity at a melt pond in the western Arctic Ocean was 5.9 [37]: (1) The water samples used to determine the dissolved inorganic nutrients, chlorophyll a (chl-a), photosynthetic pigments, and POM (from carbon isotope samples at the surface), were obtained from the surface to the euphotic layer (2-5 depths) using a CTD/rosette sampler with 10-L Niskin bottles (Ocean Test Equipment Inc., Fort Lauderdale, FL, USA). The dissolved inorganic nutrients (nitrate + nitrite, ammonium, silicate, and phosphates) were analyzed onboard using a 4-channel QuAAtro Auto Analyzer (Seal Analytical, Norderstedt, Germany). The concentrations of the nutrients were measured using standard colorimetric methods, and the reference material for nutrients in seawater (Lot. No. "BV", Kanso Technos Co., Ltd., Osaka, Japan) were used in addition to standards for every batch of runs to ensure accurate and comparable measurements during the cruise.
After prefiltration through a 200 μm mesh net to remove large zooplankton, the water samples used to determine total chlorophyll a (chl-a) and accessory pigments were filtered onto GF/F filters (precombusted at 450 °C for 4 h; Whatman, Port Washington, NY, USA) immediately after collection. The filters were stored at −80 °C until the analyses were performed. Size-fractionated chl-a was determined from samples passed sequentially through 20 μm (>20 μm; microphytoplankton), 2 μm (2-20 μm; nanophytoplankton) and Whatman GF/F filters (0.7-2 μm; picophytoplankton). All the chla concentrations were calculated by the methods described by Parsons et al. [39] using a Trilogy fluorometer (Turner Designs, San Jose, CA, USA). The phytoplankton community composition was determined with photosynthetic pigments measured by high performance liquid chromatography (HPLC; Agilent 1260 Infinity LC, Agilent Technologies Inc, Santa Clara, CA, USA)-CHEMTAX analyses. For the stable carbon isotope composition of POM, seawater was filtered onto precombusted (450 °C for 4 h) 25 mm GF/F (Whatman, 0.7 μm pore) filters. The filters were immediately stored at −80 °C until further analysis. Stable carbon isotope composition was The water samples used to determine the dissolved inorganic nutrients, chlorophyll a (chl-a), photosynthetic pigments, and POM (from carbon isotope samples at the surface), were obtained from the surface to the euphotic layer (2-5 depths) using a CTD/rosette sampler with 10-L Niskin bottles (Ocean Test Equipment Inc., Fort Lauderdale, FL, USA). The dissolved inorganic nutrients (nitrate + nitrite, ammonium, silicate, and phosphates) were analyzed onboard using a 4-channel QuAAtro Auto Analyzer (Seal Analytical, Norderstedt, Germany). The concentrations of the nutrients were measured using standard colorimetric methods, and the reference material for nutrients in seawater (Lot. No. "BV", Kanso Technos Co., Ltd., Osaka, Japan) were used in addition to standards for every batch of runs to ensure accurate and comparable measurements during the cruise.
After prefiltration through a 200 µm mesh net to remove large zooplankton, the water samples used to determine total chlorophyll a (chl-a) and accessory pigments were filtered onto GF/F filters (precombusted at 450 • C for 4 h; Whatman, Port Washington, NY, USA) immediately after collection. The filters were stored at −80 • C until the analyses were performed. Size-fractionated chl-a was determined from samples passed sequentially through 20 µm (>20 µm; microphytoplankton), 2 µm (2-20 µm; nanophytoplankton) and Whatman GF/F filters (0.7-2 µm; picophytoplankton). All the chl-a concentrations were calculated by the methods described by Parsons et al. [39] using a Trilogy fluorometer (Turner Designs, San Jose, CA, USA). The phytoplankton community composition was determined with photosynthetic pigments measured by high performance liquid chromatography (HPLC; Agilent 1260 Infinity LC, Agilent Technologies Inc, Santa Clara, CA, USA)-CHEMTAX analyses. For the stable carbon isotope composition of POM, seawater was filtered onto precombusted (450 • C for 4 h) 25 mm GF/F (Whatman, 0.7 µm pore) filters. The filters were immediately stored at −80 • C until further analysis. Stable carbon isotope composition was determined using isotope ratio mass spectrometry (IRMS; visION, Elementar UK, Manchester, UK) in the stable isotope laboratory at the University of Hanyang, Ansan, Korea, after HCl fuming overnight to remove the carbonate. The carbon isotope fractionation, δ 13 C (% ), was calculated using the following equation: where the standard for δ 13 C is IAEA-CH-3 [40].

Biochemical Composition and Enzyme-Hydrolysable Experiments Related to the POM
The water samples (n = 51) used to assess the biochemical composition of the POM were obtained from two to six different depths at each site within the euphotic layer, and for each macromolecule (PRT, CHO, and LIP), 0.5-1 L of the seawater sample went through a precombusted 25 mm GF/F filter (at 450 • C for 4 h). The filter was immediately stored at −80 • C until analysis. Analysis of the PRT and CHO was performed using the methods described by Lowry et al. [41] and Dubois et al. [42], respectively. For the total PRT extraction, we added deionized water to a filter and, alkaline copper solution and Folin-Ciocalteu phenol regent to the sample tube. The CHO content was measured by a phenol-sulfuric acid reaction. The LIP were extracted from the filter with chloroform and methanol (1:2; v:v) [43], followed by sulfuric acid at 200 • C [44]. The absorbance of the samples, blanks and standards was determined at wavelengths of 750, 490, and 360 nm for the PRT, CHO, and LIP, respectively, using a spectrophotometer (Hitachi, Tokyo, Japan). The concentrations of the macromolecules were determined by comparison to the standard curve created with blank filters (procedural control filters, Whatman GF/F filter). The standard solutions for the PRT, CHO, and LIP were used a protein standard (2 mg mL For enzyme-hydrolysable experiments, sampling was conducted by randomly selected samples of 35. Three enzymes were used in the enzyme-hydrolysable experiments: proteinase K derived from Tritirachium album (CAS No. 39450-01-6), β-glucosidase from almonds (CAS No. 9001-22-3), and lipase from Rhizopus oryzae (CAS No. 9001-62-1) (Sigma-Aldrich). Since these enzymes have hydrolytic activities similar to those of natural marine organisms, including autotrophs and heterotrophs [45], proteinase K, β-glucosidase, and lipase were chosen for the hydrolysis of PRT, CHO, and LIP, respectively [22,24,[46][47][48][49]. The sample filters and blank filters were placed in enzyme solutions (100 mg L −1 in 0.1 M sodium phosphate buffer) to react for 2 h (proteinase K), 2 h (β-glucosidase), and 30 min (lipase). After hydrolysis, each filter was rinsed with buffer and deionized water and the concentrations of PRT, CHO, and LIP were determined as previously described. The concentration of the hydrolyzed biochemical fractions was calculated by the difference between before and after treatment of enzyme for each fraction.

Statistical Analysis
The statistical analyses (t-test, Pearson's correlation, and principal component analysis (PCA)) were performed with SPSS statistical software (version 12.0; SPSS Inc., Chicago, IL, USA) and R software (version 3.4). Cluster analysis was performed by using a hierarchical clustering algorithm with Ward's method to identify the groups of sampling stations. A t-test evaluates whether the means of two independent groups are significantly different from each other. The relationships between the depth, nutrients, chl-a, and biochemical components were tested using Pearson's correlation. PCA was used to evaluate the differences in the biochemical components and identify the significance of the environmental factors (e.g., salinity, temperature, density, phytoplankton size, MW (%), and major inorganic nutrient concentrations) among the groups and at each station. The average value of each variable within the euphotic layer was used for PCA (Table S1). We adopted the principle that an eigenvalue >1.0 can be used to determine the number of principal components.

Physicochemical and Biological Characteristics During the Sampling Period
The potential temperature and salinity diagram reveal different hydrodynamic conditions during the sampling periods ( Figure 1B). Based on Gong and Pickart's work [50], the summertime water mass properties of Stations 2, 3, and 6 were mainly composed of Alaskan coastal water (potential temperature (T) ≥ 3 • C and salinity (S) ≥ 0). Chukchi summer water (−1 • C < T < 3 • C and 30 < S < 33.6), and Pacific winter water (T < −1 • C and S > 31.5) were found in the other stations ( Figure 1B). Since sea-surface temperature and salinity are strongly affected by sea ice and related meltwaters, brine rejection, continental runoff, and the heat flux in the Arctic Ocean [51], we assumed that the temperature and salinity at the surface were representative of the ambient water conditions. As a result, the cluster analysis of the surface potential temperature and salinity data allowed the identification of the three types of regions in the Chukchi Sea: cluster 1 (hereafter, the southern part; Stations 2, 3, and 6) was located at a latitude of approximately 66-70 • N; cluster 2 (hereafter, the northern part; Stations 15, 17, 20, 23, 31, and 33) was located at a latitude of 74.7-78 • N and included the Chukchi Borderland; cluster 3 (hereafter, the middle part; Stations 12, 14, and 35) was located between two areas of the Chukchi Sea (the southern and northern parts) ( Figure 1C and Table 1). The potential temperature at the surface was approximately 8 • C in the southern part, while in the northern part, it fell further, to below 0 • C, ranging from −1.6 to −0.6 • C. The salinity at the surface in the southern Chukchi Sea (shallow continental shelf) was above 31.9, with the maximum value (32.5) recorded at Station 3, while the salinity in the northern part was below 30.3, with the minimum value (27.2) recorded at Station 33 ( Figure 1B). Overall, the northern part of the Chukchi Sea is characterized by a relatively cold temperature and low salinity, while we found higher temperatures and salinities in the surface water in the southern part ( Figure 1B). Hydrodynamic characteristics are subject to the considerable influence of sea ice. The meltwater percentage (MW; %) in the euphotic layer of the study area ranged from 0 to 21.1%, with large spatial variations. Such a situation is specific to the northern part, with an average MW (%) ranging from 4.6 to 18.4 and a mean of 12.8% (SD = ± 3.6). Based on the sea ice extent, the MW (%) accounted for <15% of the surface water at the inner stations (Stations 17, 20, and 23) while at the outer stations (Stations 15, 31, 33, and 35), the MW accounted for more than 15% of the surface water (t-test, p < 0.05; Figure 1D). This result suggests that the salinity was greatly influenced by the regional melting of sea ice.
The concentrations of the dissolved inorganic nitrate + nitrite + ammonium (DIN), silicate (DSi) and phosphate (DIP) are shown in Figure 2. In the sampling period, the DSi and DIP concentrations from the surface to the euphotic layer ranged from 1.9 to 29.0 µM and 0.2 to 1.7 µM with means of 7.8 (SD = ±6.4 µM) and 0.8 (SD = ±0.3 µM), respectively ( Figure 2A). The concentration of DIN, which was generally depleted (<1 µM) at the surface layer throughout our study area, was in the range of 0-13.2 µM, with an average of 1.5 µM (SD = ±3.0 µM) ( Figure 2B). All the mean nutrient concentrations decreased from the southern to the northern parts of this region.  The average total chl-a concentration of phytoplankton from the surface to the euphotic depth ranged from 0.04 to 5.3 μg L −1 with a mean of 0.8 μg L −1 (SD = ±1.3 μg L −1 ) at all stations, decreasing northward ( Figure 3A). The phytoplankton community was dominated by picophytoplankton, which accounted for 46.2% (SD = ±15.0%) of the total chl-a concentration, followed by nanophytoplankton (mean ± SD = 27.8 ± 10.0%) and microphytoplankton (mean ± SD = 26.0 ± 17.3%) in the northern part of the Chukchi Sea ( Figure 3B). In the southern and middle parts, microphytoplankton were dominant (mean ± SD = 80.1 ± 5.9% for the southern part and mean ± SD = 35.0 ± 34.5% for the middle part) within the euphotic layer ( Figure 3B). The average total chl-a concentration of phytoplankton from the surface to the euphotic depth ranged from 0.04 to 5.3 µg L −1 with a mean of 0.8 µg L −1 (SD = ±1.3 µg L −1 ) at all stations, decreasing northward ( Figure 3A). The phytoplankton community was dominated by picophytoplankton, which accounted for 46.2% (SD = ±15.0%) of the total chl-a concentration, followed by nanophytoplankton (mean ± SD = 27.8 ± 10.0%) and microphytoplankton (mean ± SD = 26.0 ± 17.3%) in the northern part of the Chukchi Sea ( Figure 3B). In the southern and middle parts, microphytoplankton were dominant (mean ± SD = 80.1 ± 5.9% for the southern part and mean ± SD = 35.0 ± 34.5% for the middle part) within the euphotic layer ( Figure 3B).  The average total chl-a concentration of phytoplankton from the surface to the euphotic depth ranged from 0.04 to 5.3 μg L −1 with a mean of 0.8 μg L −1 (SD = ±1.3 μg L −1 ) at all stations, decreasing northward ( Figure 3A). The phytoplankton community was dominated by picophytoplankton, which accounted for 46.2% (SD = ±15.0%) of the total chl-a concentration, followed by nanophytoplankton (mean ± SD = 27.8 ± 10.0%) and microphytoplankton (mean ± SD = 26.0 ± 17.3%) in the northern part of the Chukchi Sea ( Figure 3B). In the southern and middle parts, microphytoplankton were dominant (mean ± SD = 80.1 ± 5.9% for the southern part and mean ± SD = 35.0 ± 34.5% for the middle part) within the euphotic layer ( Figure 3B).

Biochemical Composition (PRT, LIP, and CHO) of POM
The LIP and PRT concentrations in the POM ranged from 5.4 to 169.1 µg L −1 (mean ± SD = 32.4 ± 32.8 µg L −1 ) and 9.7 to 573.8 µg L −1 (mean ± SD = 61.6 ± 101.0 µg L −1 ) within the euphotic layer, respectively (Figure 4). The CHO concentration ranged from 29.9 to 406.4 µg L −1 with a mean of 86.6 µg L −1 (SD = ± 67.9 µg L −1 ) ( Figure 4A). The vertical distribution of the LIP, PRT, and CHO concentrations did not show a specific trend (p > 0.05) but was characterized by significant spatial changes ( Figure 4A). In the southern part of the Chukchi Sea, the average PRT concentration (198.8 µg L −1 ) was approximately 5.1 and 6.9 times higher than those of the stations in the middle (39.2 µg L −1 ) and northern parts (28.7 µg L −1 ) (t-test, p < 0.05), respectively. Similarly, at the southern stations, the average LIP (80.3 µg L −1 ) concentration was approximately 3.4 and 3.8 times higher than the average LIP concentrations in the middle and northern parts, respectively, while the average CHO (162.7 µg L −1 ) concentration was approximately 1.8 and 2.7 times higher than the average CHO concentrations in the middle and northern parts ( Figure 4A).
Water 2020, 12, 2355 7 of 17 μg L −1 (SD = ± 67.9 μg L −1 ) ( Figure 4A). The vertical distribution of the LIP, PRT, and CHO concentrations did not show a specific trend (p > 0.05) but was characterized by significant spatial changes ( Figure 4A). In the southern part of the Chukchi Sea, the average PRT concentration (198.8 μg L −1 ) was approximately 5.1 and 6.9 times higher than those of the stations in the middle (39.2 μg L −1 ) and northern parts (   The food material (FM) is represented by the sum of PRT, CHO, and LIP concentrations in POM ( [13] and reference therein) and concentration of each biochemical constituent (PRT, CHO, and LIP) covaried with the FM, as expected. The FM ranged from 53.7 to 1074.2 µg L −1 , with an average of 180.5 µg L −1 (SD = ± 195.3 µg L −1 ), and the FM decreased northward in this study.
In this study, the bioavailable fraction of POM (BFM, as the sum of HPRT, HLIP, and HCHO concentrations) can be considered the actual nutritional constituents and/or potentially available food for consumers that are able to be digested. In FM, the remaining values (excluding BFM) are expressed as a non-bioavailable form (N-BFM). In our study, similar to FM, the BFM concentration was much higher (mean ± SD = 214.4 ± 194.5 µg L −1 ) in the southern than in the middle (mean ± SD = 97.8 ± 52.6 µg L −1 ) and northern (mean ± SD = 84.1 ± 36.7 µg L −1 ) parts. Similarly, the average N-BFM (256.0 µg L −1 ) at the southern stations was approximately 4.3 and 8.4 times greater than that at the middle and northern stations, respectively. These results show that the positive effect of a large amount of FM is influenced by the quantity of BFM and that the majority of POM is not actually composed of bioavailable PRT, CHO, and LIP.

Multivariate Statistical Analysis
PCA was performed to determine the similarity among the environmental variables between stations. The PCA ordination of the sampled stations according to the measured environmental parameters is plotted in Figure 5 with eigenvalues presented in Table 2. The first two principal components (PC1 and PC2) accounted for 60.2% and 25.0% of the total variability, respectively. The temperature, salinity, density, and microphytoplankton (%) (eigenvectors of 0.931, 0.882, 0.785, and 0.891, respectively) were differentiated from the MW (%) and picophytoplankton (%) (eigenvectors of −0.907 and −0.880, respectively) by PC1, while PC2 was positively correlated with the major inorganic nutrient variables (eigenvectors ≥ 0.8). The analysis indicated general latitudinal groupings of stations in terms of their physical, chemical, and biological characteristics. The southern part was distinguished from the northern part by relatively high nutrient concentrations, temperatures, salinity, densities and relative contribution (%) of microphytoplankton. The northern part was characterized by a high MW (%) and relative contribution (%) of picophytoplankton values. The diagonal trajectory of the stations in the middle part within the ordination indicated that they represented a combination of PC1 and PC2.

Origin and Quantity of POM
Our POM samples were collected by filtration and consisted of a variety of complex mixtures of compounds. Many studies have reported that chemical markers, such as chl-a, natural abundance of the stable isotopes of carbon (δ 13 C), and the C:N ratio, can be used to distinguish phytoplankton, as live components, from POM [9,12,30,52]. In our samples, the respective concentrations of PRT, LIP, and CHO in the POM had a linear relationship to the chl-a concentration (r = 0.689, 0.714, 0.724, n = 47, p < 0.01 for PRT, LIP, and CHO, respectively), which was used as a proxy for phytoplankton biomass. The δ 13 C value of the POM ranged from −28.5 to −22.1% (mean ± SD = −26.2 ± 2.8% ); our values were within the range previously reported in phytoplankton samples (Kim et al. [30]; Ahn et al. [33]; reference therein). Kim et al. [30] and Ahn et al. [33] reported that POM is mainly derived from phytoplankton during summer in the Arctic Ocean based on δ 13 C and the C:N ratio. In addition, during the sampling period, the low DIN concentration (<1 µM) and salinity distribution in the surface water suggest that the POM was greatly influenced by regional sea ice rather than a riverine source (terrigenous). Therefore, in our study, the POM was considered to have mainly come from a marine phytoplankton origin.
A field study has shown a large spatial variability in the concentration of FM in the Chukchi Sea. Kim et al. [30] reported that FM concentrations ranged from 80.5 to 698.8 µg L −1 , with an average of 294.4 µg L −1 (SD = ± 228.1 µg L −1 ) in the euphotic layer of the Chukchi Sea, a value that was approximately 1.6 times higher than that in this study (mean ± SD = 180.5 ± 195.3 µg L −1 ). Yun et al. [31] also found concentrations of FM similar to results from a previous study in this area in this area that similar to results from a previous study in this area, ranged from 89.7 to 362.4 µg L −1 with an average of 156.4 µg L −1 in the euphotic layer during summer. These variations are thought to result from spatial and temporal variations in the biomass, composition, and productivity of phytoplankton which are common in the Arctic Ocean [53].
Early studies indicated that primary production is higher in the southern Chukchi Sea than in the northern Chukchi Sea, which is consistent with chl-a abundance [27,[54][55][56][57]. Based on 13 C uptake in the southern Chukchi Sea, Lee et al. [54] estimated a daily production of 0.6 g C m −2 day −1 (0.1 to 1.5 g C m −2 day −1 ). In comparison, the estimated averages of the daily primary production rates for the southern Chukchi Sea are 1.6 g C m −2 day −1 and 1.7 g C m −2 day −1 from Zeeman [58] and Korsak [27], respectively. The mean daily production in the northern Chukchi Sea measured by Yun et al. [56] was somewhat lower (mean ± SD = 0.14 ± 0.10 g C m −2 day −1 ) than the rate (0.66 ± 0.62 g C m −2 day −1 ) in the southern region, which is consistent with the findings from Lee et al. [54] (0.16 ± 0.16 g C m −2 day −1 ) and Lee et al. [55] (mean ± SD = 0.18 ± 0.07 g C m −2 day −1 ). Similarly, the mean chl-a concentration (2.0 µg L −1 ) in the southern part of the Chukchi Sea during the summer of 2017 was approximately one order of magnitude higher than the average value (0.2 µg L −1 ) in the northern part ( Figure 3). These results suggest that the regional differences in quantitative POM may have resulted from the different levels of phytoplankton biomass in the Chukchi Sea.

Biochemical Composition in Relation to Environmental Parameters
Overall, CHO accounted for 53.3% of the POM for all the survey stations, followed by PRT (29.2%) and LIP (17.5%) ( Figure 6A), which led to a low PRT:CHO ratio (0.6). Consistent with this observation, the DIN:DIP (mean ± SD = 1.3 ± 2.0) molar ratio within the euphotic layer was also low compared with the N:P Redfield ratio of 16 [59], indicating substantial nitrogen limitation in this region ( Figure 2B). However, interestingly, the PCA revealed that there were significant differences in the compounds among the groups ( Figure 5). More specifically, the biochemical composition of the POM was dominated by PRT (41.8%); in the southern part, there was PRT:CHO ratio of 1.2 despite a low DIN:DIP ratio (mean ± SD = 2.7 ± 3.0), while a CHO-dominant (>50%) system was found in the northern Chukchi Sea with a PRT:CHO ratio of 0.5. In general, the PRT fraction was greater than the CHO and LIP fractions under sufficient nitrogen conditions and growth stages of phytoplankton, which could lead to a high (>1) PRT:CHO ratio [13,15,16]. Fogg and Thake [60] and Hu [61] reported that as prolonged stressful conditions (such as nitrogen limitation) occur, metabolic changes in synthesizing enzyme systems can convert CHO into LIP synthesis. Thus, our results suggest that at least in the southern region, nitrogen limitation was not severe enough to limit phytoplankton growth. The northern part has not been exposed to nitrogen stress for a long time.
Water 2020, 12, 2355 11 of 17 which could lead to a high (>1) PRT:CHO ratio [13,15,16]. Fogg and Thake [60] and Hu [61] reported that as prolonged stressful conditions (such as nitrogen limitation) occur, metabolic changes in synthesizing enzyme systems can convert CHO into LIP synthesis. Thus, our results suggest that at least in the southern region, nitrogen limitation was not severe enough to limit phytoplankton growth. The northern part has not been exposed to nitrogen stress for a long time. In addition, the results of the biplots ( Figure 5) based on PCA revealed that microphytoplankton were influenced by relatively nutrient-replete conditions and had maximum chl-a and PRT values, while the picophytoplankton were more sensitive to nutrient deficiency and the MW (%), and were characterized by a high CHO contribution. These conditions were situated between two distinct areas (the southern and the northern parts) ( Figure 5). Similar phenomena have been described by Jin et al. [62] and Li et al. [63] in relation to dominant species, nutrient depletion, and ice cover conditions in In addition, the results of the biplots ( Figure 5) based on PCA revealed that microphytoplankton were influenced by relatively nutrient-replete conditions and had maximum chl-a and PRT values, while the picophytoplankton were more sensitive to nutrient deficiency and the MW (%), and were characterized by a high CHO contribution. These conditions were situated between two distinct areas (the southern and the northern parts) ( Figure 5). Similar phenomena have been described by Jin et al. [62] and Li et al. [63] in relation to dominant species, nutrient depletion, and ice cover conditions in the western Arctic Ocean. Li et al. [63] suggested that small cells (<2 µm) thrive as a result of low nitrate availability and a strong stratification since pico-sized cells have a large surface area to volume ratio compared to that of larger cells, which allows effective nutrient and photon acquisition. CHEMTAX pigment analysis revealed that changes in temperature (caused by the timing of sea ice retreat) influence phytoplankton community structure [64]. Thus, it seems that the variation in biochemical compounds discovered among the two different groups (i.e., the southern and northern groups) could be the result of environmentally (such as the level of nutrients and freshwater) induced differences in the size classes and communities of phytoplankton in the Chukchi Sea.
Generally, the analysis of photosynthetic marker pigments (e.g., fucoxanthin, diadinoxanthin and diatoxanthin for diatoms, zeaxanthin for cyanobacteria, chlorophyll b and prasinoxanthin for green algae, 19 butanoyloxy fucoxanthin and 19 hexanoyloxy fucoxanthin for flagellates) can be useful biomarkers for phytoplankton biomass and species [65]. In our study, thirteen pigments except chl-a were identified through the (HPLC)-CHEMTAX analyses ( Figure S1). As shown by the abundance of specific phytoplankton groups based on their corresponding biomarker pigments, the southern part was dominated by diatoms (88%), whereas pigments associated with flagellates and green algae made up approximately 44% of the total accessory pigment concentration and diatoms (53%) were observed in the northern Chukchi Sea ( Figure S1). Several studies of phytoplankton have documented that species-specific characteristics, such as the cell wall structure and functional characteristics, play a significant role in the variation in biochemical components of phytoplankton [10,33,[66][67][68][69]. Haug et al. [66] found that in diatoms, the concentration of PRT was generally much higher than that of CHO and LIP, whereas dinoflagellates have abundant CHO within their cell walls. Yun et al. [69] also reported that there were significantly close relationships between flagellates and the LIP fraction and diatoms and the PRT proportion in the Chukchi Sea. According to Shifrin and Chisholm [67], green algae and diatoms contained an average of 17.1% and 24.5% LIP (% of total dry weight), respectively, during log-phase growth in 30 cultured phytoplankton species. Therefore, the distributions and the relative contribution of these different dominant species and/or taxa of phytoplankton might also largely affect changes in the biochemical composition in the region.

Bioavailability of POM
Even now, the FM concentration, is used to represent the quantity of food in POM in relation to indicators of energy and material transfer to higher trophic levels [8,13,14,30]. However, FM is ideal when POM is made only of bioavailable constituents. In reality, POM contains bioavailable and non-bioavailable (refractory or less labile) fractions.
Bioavailability is a pivotal term for nutritional effectiveness, and the contribution of BFM to FM (%; nutritional efficiency) was used to assess POM bioavailability in this study. The nutritional efficiency ranged between 33.1 and 89.7%, with an average of 64.1% in the Chukchi Sea. More interestingly, the nutritional efficiency in the northern Chukchi Sea (74.0%) was approximately 1.2 times higher (60.0%) than it was in the middle part, while a lower mean value (42.7%) was observed in the southern part ( Figure 6B). These results may have contributed to the different hydrolysis rates among the components, for which a greater presence is also an important factor. For example, the POM in the southern Chukchi Sea had a high contribution from PRT (41.8%) but a low level of HPRT (approximately 31.0% of their total pool), whereas a high level of HLIP (approximately 79% of their total pool) were observed despite a low contribution of LIP to the POM (20.5%). In the northern part, a large contribution from the HCHO (>80% of their total pool) was observed, with CHO accounting, for more than 50%, on average, of the POM in the northern part. In the middle part, HCHO accounted for 72.1% of the total CHO pool, followed by HPRT (55.4% of their total pool) and HLIP (40% of their total pool).
However, our findings, except for the middle part, are contradictory to the conclusions of Handa and Tominaga [70] and to the results obtained by Dawson and Liebezeit [71], Christian and Karl [72], and Fabiano et al. [8]. These reports suggested that cellular and proteinous amino acids were lost more rapidly than extractable sugars and particulate CHO. Such contrasting results suggest that different sizes [62] and species [73] of phytoplankton likely influence bioavailability. In our study, we found that the bioavailable contribution was negatively correlated with the relative amount of microphytoplankton (r = −0.652, p < 0.05, n = 20) and positively correlated with the relative amount of picophytoplankton (r = 0.668, p < 0.05, n = 20) ( Figure S2); this result is consistent with the results from Jin et al. [62], who reported that picophytoplankton is more likely to mineralize and degrade in the upper ocean layers. In addition, CHO and amino acids are more enriched in intracellular materials than in cell wall materials [74]. Liebezeit [73] showed somewhat lower CHO degradation (38%) at stations dominated by diatoms than at stations dominated by Haptophyceae (86%) in the upper 100 m of the water column in the Bransfield Strait. Diatoms are characterized by silica shells (frustules) that are resistant to acid conditions (reviewed in DeNicolar [75]) and crushing forces [76]. In this sense, inherent structural differences in phytoplankton might also affect enzymatic hydrolysis, because phytoplankton was the major source of organic matter in our study. Taken together, although these results cannot be explained simply, different enzymatically hydrolysable efficiencies among the three different regions in the Chukchi Sea resulted from a selective loss of labile compounds and different communities of phytoplankton. Therefore, the higher POM bioavailability in the northern part of the Chukchi Sea could be caused by the different biochemical structures of the dominant picophytoplankton community compared to those of the microphytoplankton and diatom dominated community in the southern part of the Chukchi Sea. Clearly, a higher POM bioavailability provides more effective food materials for potential consumers in the northern part of the Chukchi Sea despite their lower biomass and lower primary productivity.

Conclusions
The biochemical composition of POM in the regions of the Chukchi Sea studied was due to differences in both environmental variables and the structure of the phytoplankton community. We also expect the observed results of the biochemical composition of POM to influence the nutritional quality of the available food. For instance, changes in the size, quantity and bioavailability of prey (phytoplankton) could affect the feeding, growth, reproduction and survival of predators [1,77,78]. In particular, in the Arctic Ocean, recent studies have indicated warming and decreased salinity of the water, with concomitant small phytoplankton sizes and decreased primary production [56,63]. If the sea ice continues to melt, then the quantity, quality, and labile level of POM will change, and consequently, the ecosystem structure, such as the trophic chain and microbial loop efficiency, will change in Arctic ecosystems. Therefore, further studies are needed to better understand the recent potential food materials under rapidly changing environmental conditions in the Arctic Ocean and picophytoplankton trophic roles in the microbial foodweb process.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4441/12/9/2355/s1, Figure S1: Relative contribution of accessory pigments to total accessory pigment (wt:wt) in euphotic layer of southern, middle, and northern part of the Chukchi Sea, Figure S2: The relationship between relative contribution of micro (red dot) and picophytoplankton (green dot) fraction to total phytoplankton biomass (chl-a) and POM bioavailability. Solid lines indicate the fitted regression lines of the raw data points, Table S1: Average environmental parameters (± SD) within euphotic layer at each station in the Chukchi Sea.