Exploring the Cumulative Selectivity of Polycyclic Aromatic Hydrocarbons in Phytoplankton, Water, and Sediment in Typical Urban Water Bodies

Abstract

To understand the interactions among eutrophication, algal bloom, and POPs (persistent organic pollutants) in freshwater ecosystems, the cumulative selectivity of PAHs (polycyclic aromatic hydrocarbons) in phytoplankton, water, and sediment with different eutrophication level waters were identified in a typical plain river network region located in Nanjing City. Results showed that a total of 33 algal species belonging to 27 genera and 4 phyla were identified in 15 sites of urban water bodies, and most of them belonged to the type Cyanobacteria–Bacillariophyta. The eutrophication level of these rivers and lakes led to the sample site specificity of algal composition and abundance. The planktonic algae mainly accumulated the 2-ring and 3-ring PAHs, and the sediment mainly enriched the high-ring PAHs. However, the enrichment capacity of planktonic algae on PAHs was much higher than that of sediment. Cyanophyta and some species of Bacillariophyta and Chlorophyta in mesotrophic (βm) and meso-eutrophic water bodies (ßαm) preferentially accumulated lower-ring PAHs (naphthalene, acenaphthylene, and phenanthrene). Some other specific algae species, such as Euglenophyta, some species of Bacillariophyta, and most Chlorophyta in mesotrophic and moderate eutrophic water bodies, had strong capacities to enrich high-ring PAHs subsuming benzo[a]anthracene, chrysene, and anthracene. The eutrophication level of water bodies affected the cumulative selectivity of PAHs by shaping the site specificity of phytoplankton composition, which may be related to water quality, sediment characteristics, phytoplankton composition, and the algal cell walls.

1. Introduction

Polycyclic aromatic hydrocarbons (PAHs) refer to compounds with two or more benzene rings arranged by linear, angular, and cluster forms, which possess high melting and boiling points, strong hydrophobicity, low vapor pressure, and a large Kow (the octanol/water partition coefficient). They can exist in natural environments persistently and have carcinogenic, teratogenic, and mutant characteristics [1].

Planktonic algae as a good biosorbent have a strong bioenrichment ability to PAHs [2,3,4,5,6,7]. The PAHs accumulated in planktonic algae are affected by a series of factors which are composed of pH value, temperature, ion intensity, molecular structure and initial concentration of organic compounds, and metabolic types of microorganisms [8,9,10,11,12,13,14]. Mailhot et al. [11] indicated a positive correlation between the enrichment coefficient of planktonic algae and the hydrophobicity of PAHs. Koelmans et al. [15] verified the non-linear relations between the enrichment coefficient of planktonic algae and the Kow of organic micropollutants through simulating experiments in the lab. Martinez et al. [16] demonstrated that the difference in PAHs enrichment was mainly caused by the difference in hydrophobic lipids in the cell wall of algae. Yan et al. [2] studied the enrichment of PAHs by Chlorella protothecoides with different metabolic types (autotrophic and heterotrophic), and the results showed that heterotrophic Chlorella protothecoides showed stronger enrichment capacity than the autotrophic ones. Soto et al. [17] found that Chlamydomonas grew in an airtight tube equipped with a substrate containing 30 mg∙L⁻¹ naphthalene, which could enrich naphthalene until up to a high concentration at 7 d. Subsequently, naphthalene could be quickly released from Chlamydomonas after it is transferred to a pure substrate. Lei et al. [18] found that two kinds of freshwater microalgae, Selenastrum capricornutum and Chlorella vulgaris, were able to remove 48~78% of fluoranthene and pyrene in the culture solution. Kris et al. [19] showed that the concentration level of benzopyrene in brown algae, red algae, green algae, and chara algae was species-specific. Hong et al. [20] indicated that two typical PAHs can be enriched simultaneously by Skeletonema costatum (Greville) Cleve and Nitzschia sp., and the accumulation and degradation abilities of Nitzschia sp. were higher than those of S. costatum. The microalgal species also showed comparable or higher efficiency in the removal of the Phe-Fla mixture than Phe or Fla singly, suggesting that the presence of one PAH stimulated the degradation of the other. Chung et al. [21] found that the percentage removals of aqueous Phe by Sargassum for all the investigated factors (e.g., initial pH and ionic strength) were in the range of 91.7–98.4%. Sargassum is an effective sorbent for removing HOCs (hydrophobic organic compounds) in wastewaters and urban runoffs. Chan et al. [22] found that the removal efficiency of phenanthrene, fluoranthene, and pyrene by Selenastrum capricornutum was directly related to the initial cell density, and the optimal density was l × 10⁷ cells/mL. The removal rate of the three pollutants was in the order: pyrene > fluoranthene > phenanthrene. Meanwhile, the removal mechanism included an initial adsorption onto the cell walls of both live and dead cells, and the adsorbed PAHs were then absorbed and degraded in live cells only. García et al. [14] found that sorption is an important phenomenon for BaP removal by Selenastrum capricornutum; however, biodegradation is the principal means of removing BaP in live cells. Further, most of the BaP removal of Selenastrum acutus is due to sorption rather than degradation. Ke et al. [23] found that the presence of metal ions was beneficial to the adsorption of low molecular weight PAHs (Flu and Phe) by algae, which was attributed to metal-induced cell degeneration. However, the presence of metal ions could not affect the absorption of high molecular weight PAHs (Fla, Pyr, and BaP). This conclusion partly confirmed the results of other studies in the same period. Most studies showed that the biosorption efficiency and adsorption capacity of organics decreased due to the shielding effect of heavy metal ions. However, some researchers have found that the presence of heavy metal ions could also improve the adsorption capacity of algae because metal ions played a bridging role between microorganisms and organic matter. Li et al. [24] found that the basal metabolism associated with the growth and proliferation of algal cells could be significantly promoted by bacterial pyrene metabolites, which suggests a close relationship between algae and bacteria in the transformation and ecological effects of toxic contaminants. Luo et al. [25] found that the two light sources (gold and white light irradiation) did not result in significant differences in the biodegradation of the selected PAHs in live algal cells, but white light was more effective in promoting photodegradation than was gold light in dead cells, and dead algal cells most likely acted as a photosensitizer and accelerated the photodegradation of PAHs. Warshawsky et al. [26] showed that algae are important in their ability to degrade PAHs, but the degradation is dependent on the dose of light energy emitted and absorbed, the dose of PAHs to which the algae are exposed, the phototoxicity of PAHs and their metabolite(s), and the species and strain of algae involved. All of these factors are important in assessing the degradation and detoxification pathways of recalcitrant PAHs by algae.

The enrichment of PAHs by algae is a complex biochemical process, which is caused by a combination of multiple mechanisms. The enrichment routines are composed of biosorption and active transportation. Of them, biosorption is the main route of algal bioenrichment, which is fast and has nothing to do with metabolism. Meanwhile, dead cells can also participate in it. Due to cell wall fragmentation, the carboxyl, amino, aldehyde, phosphoryl, and other internal functional groups in the cytoplasm are exposed to environmental media and are easy to bind with organic matter, which makes the dead cells have strong biosorption ability. Research showed that biosorption may promote biodegradation (the bioavailability of PAHs is mainly improved by shortening the distance between pollutants and degrading bacteria, including the attachment of pollutants to planktonic algae biomass and the chemotaxis of algae to pollutants). Compared with biosorption, active transportation contributes little to algal enrichment, and only living cells can participate in it. Active transportation is an active intracellular absorption process that relies on energy transfer, which requires the participation of some specific enzymes, such as glutathione transferase and cytochrome P450, and is also affected by temperature, light, and metabolic inhibitors [27,28,29,30,31].

The enrichment of PAHs by algae is an important route for the physical transfer of poisons, and PAHs could be transferred through food chains at a high level [32]. Meanwhile, PAHs enriched by algae can reduce the direct impact on other organisms, especially humans. However, previous research is concerned more with algae enrichment of PAHs in laboratories rather than in rivers and lakes. In this work, the main purpose was to identify species-specific algae and their cumulative selectivity of PAHs in phytoplankton, water, and sediment in rivers and lakes with different eutrophication levels, so as to provide a technical basis and reference for ecological restoration and management of similar water bodies.

2. Materials and Methods

2.1. Sample Collection and Pretreatment

The samples of water, sediment, and phytoplankton were collected from 15 sampling sites of Qinhuai River and Xuanwu Lake in Nanjing City on 15 October 2021 (Figure 1). The physicochemical environmental parameters, including dissolved oxygen (DO), total nitrogen (TN), ammonia nitrogen (NH₄⁺-N), chemical oxygen demand (permanganate index, COD_Mn), and chlorophyll a (Chl-a), were measured according to the reference of Zhao et al. [33,34]. The PAHs in water samples were extracted with a C18 solid-phase extraction column (Borui Co., Ltd., Tianjin, China) and concentrated with a vacuum rotary evaporator (RE-3000, Yarong Biochemical Instrument Plant, Shanghai, China) for the determination of PAHs concentration with HPLC. The PAHs accumulated by phytoplankton in a 25 L water sample were collected from the concentrated phytoplankton by filtration (passing through a 25# phytoplankton net and a 0.2 μm glass fiber filter, respectively) and ultrasonic extraction method according to the reference of Zhao et al. [33]. The PAHs were analyzed by Agilent1100 HPLC with fluorescence and UV detectors in a gradient program at 0.75 mL/min, and were quantified by using external standard methods according to the references of Zhao et al. [35]. The algae in 1 L water samples were identified and accounted for by the microscopic examination method after being settled out and concentrated to 30 mL. More detailed information can be found in the Supplementary Materials S1.1–S1.6, and Quality Assurance/Quality Control can be found in the Supplementary Materials S1.7.

2.2. Data Analysis

To explore the influence of algae on the partitioning of PAHs among water, sediment, and algae, RDA analyses of phytoplankton species, sampling sites, and residues of PAHs in algae were performed using the CANOCO 5.0 software package (Demo version, Microcomputer Power, Ithaca, NY, USA). The values of phytoplankton abundance/biomass and residual PAHs were normalized by log₁₀ (x + 1)-transformation. Because of the potential interaction between algae and PAHs, the data matrix exchange technique was adopted for algal and PAHs data during the RDA analysis. By comparing the RDA ordination results based on this technique, we identified the dominant algal species and residual PAHs in the different sampling sites and determined the ecological contribution and role of phytoplankton to the bioaccumulation of PAHs in the studied rivers and lakes.

Given the importance of eutrophication to algae enrichment of PAHs, our study applied a comprehensive nutrition index to evaluate the eutrophication level of water bodies in the research area. This method overcomes the one-sidedness evaluation of eutrophication by a single factor, and takes chlorophyll a, total nitrogen (TN), total phosphorus (TP), and the permanganate index (COD_Mn) into account. The comprehensive nutrition status index formula is as follows:

$$TLI(\sum )={{\displaystyle \sum }}_{j=1}^m{W}_i\times TLI\left(j\right)$$

(1)

where TLI(∑) means comprehensive nutritional status index (trophic level index); TLI(j) represents the nutrition status index of the j parameter; and W_j is the relative weight of the nutrition status index of the j parameter.

With Chl-a as the reference parameter, the normalized relative weight formula of the j parameter is as follows:

$$W_j=\frac{r_{ij}^2}{{{\displaystyle \sum }}_{j=1}^m{r}_{ij}^2}$$

(2)

where r_ij is the correlation coefficient between the j parameter and the benchmark parameter Chl-a, and m is the number of evaluation parameters. The m value in our study is 4. The correlation r_ij between Chl-a, TP, TN, and COD_Mn in Chinese river and lake water bodies was 0.84, 0.82, and 0.83, and $r_{ij}^2$ was 0.7056, 0.6724, and 0.6889, respectively. The comprehensive nutrient index and nutrient status classification of water bodies at different sampling sites were shown in

Table 1. List of algae and their indication function of eutrophication levels.

Code	Genus Name	Eutrophication Level	Algae Abundance	Code	Genus Name	Eutrophication Level	Algae Abundance
Cyanophyta				Euglenophyta
Cy1	Oscillatoria	αm	++++	E1	Euglena	ps, αm, βm	++
Cy2	Phormidium	ps	++++	E2	Phacus	βm, os	+
Cy3	Merismopedia	αm	+++	Chlorophyta
Bacillariophyta				C1	Chlorella	ps, αm	++
B1	Navicula	βαm	+	C2	Ankistrodesmus	αm, βm	++
B2	Melosira	βαm	+++	C3	Pandorina	βm, os	++
B3	Cyclotella	βαm	+	C4	Coelastrum	βm, os	+
B4	Stephanodiscus	αm	++	C5	Actinastrum	βm, os	++
B5	Tabellaria	os	+	C6	Oocystis	βm, os	+
B6	Fragilaria	βm	+	C7	Scenedesmus	αm, βm	++
B7	Synedra	βαm	+	C8	Characium	os, βm	++
B10	Gyrosigma	βαm	+	C9	Closterium	αm, βm	+
B11	Gomphonema	βαm	+	C10	Pediastrum	βm, os	+
B12	Cymatopleura	αm, βm	+	C11	Schroederia	ps	+
B13	Stauroneis	os	+

Note: eutrophication level: os: oligotrophication; ßm: mesotrophic; ßαm: mesotrophic-eutrophication; αm: moderate eutrophic; ps: hypertrophic type [37]. Algae abundance: +, 0–10; ++, 10–50; +++, 50–100; ++++, more than 100; unit, 10⁴ cells per liter.

. The classification criteria of eutrophication level, TLI(∑) < 30: oligotrophication (os), refers to a water body that lacks nutrients and contains a large number of species but a small number of aquatic organisms; 30 ≤ TLI(∑) ≤ 50: mesotrophication (ßm); 50 < TLI(∑) ≤60: Meso-eutrophication (ßαm); 60 < TLI(∑) ≤70: moderate eutrophication (αm); 70 < TLI(∑): hypereutrophics (α-ps and ß-ps) [36].

3. Results

3.1. Characteristics of Phytoplankton Community Composition in Autumn

A total of 33 species, 27 genera, and 4 phyla of phytoplankton were detected in rivers and lakes in the studied area, which were mainly composed of Cyanophyta (3 species, 3 genera), Bacillariophyta (15 species, 11 genera), Euglenophyta (3 species, 2 genera), and Chlorophyta (12 species, 11 genera). Chondria dominated the Cyanophyta, accounting for 74.61% of the total number of Cyanophyta. Further, 68.54% of the Bacillariophyta was Melosira. The number of phytoplankton communities in the study area ranged from 44.76 × 10⁴ to 527.14 × 10⁴ cells/L, and the mean value was 221.51 × 10⁴ cells/L. The abundance of phytoplankton was in the order: Cyanophyta (92.0 × 10⁴ cells/L, 41.5%) > Bacillariophyta (81.20 × 10⁴ cells/L, 36.7%) > Euglenophyta (27.0 × 10⁴ cells/L, 12.2%) > Chlorophyta (21.20 × 10⁴ cells/L, 9.6%). Cyanophyta and Bacillariophyta occupied 78.19% of total phytoplankton. Thus, the study area belonged to the type Cyanobacteria–Bacillariophyta in autumn. The codes of phytoplankton and their relative abundance and the indication function of eutrophication levels are listed in Table 1.

The relatively large number of phytoplankton observed in the sample sites included: Outside the Sancha Estuary (527.14 × 10⁴ cells/L), Hanzhongmen Bridge (420.32 × 10⁴ cells/L), Xiafu Bridge (292.72 × 10⁴ cells/L), Yuhua Bridge (277.48 × 10⁴ cells/L), and Inside the Sancha Estuary (264.53 × 10⁴ cells/L). Meanwhile, a relatively small abundance of phytoplankton was detected in three sample sites—Yixian Bridge (44.76 × 10⁴ cells/L), the upper reaches of Qinhuai River (55.99 × 10⁴ cells/L), and Xuanwu Lake (76.99 × 10⁴ cells/L)—which was much lower than the mean value. Thus, more phytoplankton was distributed in the External Qinhuai River and in the south of the Internal Qinhuai River. Meanwhile, the abundance of phytoplankton in the External Qinhuai River increased from the upper reaches to the lower reaches (Figure 2). The spatial distribution difference of algal abundance and composition may be related to the water flow velocity and eutrophication degree at different sampling points. Due to the sluice controlled stagnant flow in the study area, the water flow speed of rivers and lakes is very slow (about 0.02 m/s) in Xuanwu Lake and the Internal Qinhuai River, and the flow rates of the External Qinhuai River (sites 9#–13#) and Yangze River (14# site: Outside the Sancha Estuary) are more than those of Xuanwu Lake and Internal Qinhuai River. According to the report of Zhang et al. [38], the flow rate significantly inhibited the growth of phytoplankton and dramatically shifted the phytoplankton composition and enhanced the inter-relationships among environmental variables, and the effect of the flow rate on phytoplankton is interlinked with many other environmental variables. Similar results have been obtained by Li et al. [39].

However, the algal biomass showed different spatial distribution characteristics, which may be related to the different weights of individual algal bodies of different algal species.

3.2. Residual Characteristics of PAHs in Phytoplankton, Water, and Sediment

The mean residual concentration of ∑PAHs in phytoplankton was 16,373.6 ng/g. Among these sample sites, the maximum concentration of ∑PAHs in Yixian Bridge was 35,171.9 ng/g, followed by the Inside the Sancha Estuary, Fengtai Bridge, Wetland Park, Yuhua Bridge, Xuanwu Lake, the Outside the Sancha Estuary, Wenjin Bridge, and Xiafu Bridge. The minimum concentration of ∑PAHs in Zhenzhu Bridge was only 1467.2 ng/g, which was approximately 1/24 of the maximum concentration. The PAHs residual components of phytoplankton samples in this region were mainly 2- and 3-ring PAHs, accounting for 99.6% of the total residual amount on average, while the residual concentration of 4-ring PAHs was very low, accounting for only 0.36%. Even 5- or 6-ring PAHs were not detected in phytoplankton (Figure 3A). The eutrophication level of these rivers and lakes led to the sample site specificity of algal composition and abundance. The planktonic algae mainly accumulated PAHs with two and three rings, and the sediment mainly enriched the high-ring PAHs. However, the enrichment capacity of planktonic algae on PAHs was much higher than that of sediment.

From Figure 3B, we found that there was a very poor correlation between the concentrations of ∑PAHs in water and phytoplankton samples (R² = 0.033, p = 0.5166). This may be because there are many factors influencing the adsorption of PAHs by phytoplankton, such as the cell structure of algae, the type and structure of PAHs, etc. [8,9,10,11,12,13,14]. In addition to the physical and chemical properties and hydrological conditions of each sampling site were different, and the composition and abundance of the algal community were also different, so the concentrations of ∑PAHs between the water and phytoplankton had no obvious correlation. However, there was a significant negative correlation between the concentrations of PAHs in sediments and that of algae (R² = 0.4149, p = 0.0175), indicating a direct competitive relationship with PAH accumulation (Figure 3B).

3.3. Relationship Analysis between Phytoplankton Community and Residual PAHs

3.3.1. Correlation Analysis

Figure 4 shows the relationship between PAHs and phytoplankton biomass based on Pearson correlation analysis. The results showed that there was an extremely significant positive correlation between Nap, Ace, and Phe (p < 0.01). Meanwhile, Phe was positively correlated with Ant, BaA, and Chr (p < 0.01). Nap and Ace showed a significant negative correlation with E1 (p < 0.05), while Ace showed a significant negative correlation with E1 and C3 and a significant positive correlation with C4 (p < 0.05). In addition, Cy3, B2, B11, C7, and C10 also showed a high positive correlation with Nap and Ace. For Phe, B1 was significantly positively correlated with Phe (p < 0.05), and Cy1, B2, B3, B7, and B10 were also high positively correlated with Phe. For Ant, BaA, and Chr, B1 and B7 showed significant and extremely significant positive correlations with them (p < 0.05 or <0.01). The relationship between phytoplankton abundance and PAHs showed similar results, which are shown in Figure S1 in the Supplementary Materials. These results indicated that these algae played an important role in the absorption and degradation of PAHs.

3.3.2. Redundancy Analysis (RDA)

The structure of the cell walls and the surface characteristics of phytoplankton determine their enrichment efficiency and selectivity on pollutants, and the physicochemical properties of water bodies also affect the phytoplankton composition and their enrichment ability. In our study, CANOCO 5.0 was applied to explore the effect of phytoplankton on the occurrence of PAHs in water. For the first RDA ordination (algae vs. PAHs, Figure 5A), phytoplankton abundance data represented the species matrix and the residues of PAHs in algae represented the environmental matrix. The second RDA ordination reversed these data matrices (PAHs vs. algae, Figure 5B). The codes for each sampling site and for phytoplankton were exhibited in Figure 1 and Table 1.

When phytoplankton abundance data acted as the species matrix and the residues of PAHs as the environmental matrix, the eigenvalues of the first two species axes were 0.210 and 0.118, and the total eigenvalue was 1.000 (

Table 2. Eigenvalues for RDA axis and the cumulative percentage variance.

Species Matrix vs. Environmental Matrix	Ordination Axes	AX1	AX2	AX3	AX4	Total Variance
Algae vs. PAHs	Eigenvalues	0.210	0.118	0.220	0.129	1.000
	Cumulative percentage variance of fitted response data	45.70	71.30	71.80	73.60
PAHs vs. algae	Eigenvalues	0.608	0.3259	0.0521	0.0122	1.000
	Cumulative percentage variance of fitted response data	60.8	93.39	98.6	99.83

). The first two species axes explained 71.30% of the total information amount. The correlation coefficients between the first two species axes (phytoplankton abundance) and the first two environmental axes (PAHs) were 0.8212 and 0.8858, respectively. The correlation coefficient of the first two species ordination axes and of the first two environmental axes (PAHs) were 0.1298 and 0.000 (indicating little correlation), respectively. When we reversed these data matrices (Figure 5B), the residues of PAHs acted as the species matrix and the phytoplankton abundance data as the environmental matrix. The eigenvalues of the first two species axes were 0.6080 and 0.3259 (Table 2), respectively, explained 93.39% of the total information amount. The correlation coefficients between the first two species axes (PAHs) and the first two environmental axes (phytoplankton abundance) were all 1.000, respectively. The correlation coefficients of the first two species ordination axes and of the first two environmental axes (PAHs) were all 0.000. These RDA results suggested that the ordination could reflect the exact relationship between phytoplankton species and the residues of PAHs.

Table 3. Correlation coefficients among PAHs, algae, and RDA ordination axes.

	Algae vs. PAHs		PAHs vs. Algae
Axes	SPEC AX1	SPEC AX2	SPEC AX1	SPEC AX2
Nap	0.1235	−0.4714	−0.7713	0.2783
Ace	0.2914	−0.5094	−0.6067	0.7643
Phe	−0.1766	−0.1804	−0.6816	−0.7247
Ant	0.0052	−0.0169	0.9831	−0.0828
BaA	−0.007	−0.0548	0.9832	−0.0362
Chr	0.0011	−0.0192	0.9559	−0.0792
Cy1	−0.6637	−0.0902	−0.1693	−0.3668
Cy2	−0.5175	−0.1398	−0.074	−0.1767
Cy3	0.6662	−0.4762	−0.1661	0.5866
B1	0.0056	0.0086	−0.0798	−0.4991
B2	−0.134	−0.7793	−0.4556	−0.0083
B3	−0.4524	−0.255	−0.1951	−0.3489
B4	0.187	0.0642	−0.1098	−0.0116
B5	0.3154	0.0465	0.1103	−0.1785
B6	−0.1734	−0.2423	−0.0702	0.0765
B7	0.1868	0.0579	−0.0999	−0.3387
B10	−0.2584	−0.1233	−0.1745	−0.3535
B11	0.5094	−0.4376	−0.1695	0.4963
B12	0.2422	−0.126	0.3625	0.5375
B13	−0.3717	0.0951	0.1176	−0.0955
E1	0.4241	0.6874	0.6355	−0.2645
E2	0.0978	0.1568	0.4065	0.1085
C1	0.0897	0.2804	0.1076	−0.0033
C2	0.1405	0.6918	0.5239	−0.2825
C3	−0.166	0.5891	0.4964	−0.2784
C4	0.1423	0.0922	−0.3705	0.1283
C5	−0.222	−0.1693	−0.1219	−0.0209
C6	0.2121	0.5879	0.5077	−0.1591
C7	0.3507	−0.5079	−0.2782	0.2434
C8	0.23	0.7712	0.5496	−0.2223
C9	−0.3358	−0.3092	−0.1357	0.052
C10	0.1354	−0.3871	−0.2099	−0.0264
C11	0.2808	−0.1685	−0.0217	0.2809

Note: (1) “Algae vs. PAHs” shows the RDA results in which algae data act as the species variable and PAHs data act as the environment variable. “PAHs vs. algae” shows the RDA results in which PAHs data act as the species variable and algae data act as the environment variable. (2) The data marked with blue color font show the algae have a high correlation with Nap and Ace, the green color font shows the algae have a high correlation with Phe, and the red color font shows the algae have a high correlation with Ant, BaA, and Chr.

shows that in the first RDA ordination (algae vs. PAHs), Nap and Ace represented a positive correlation with the first species ordination axis, and the algae including Cy3 (correlation coefficient r = 0.6662), B11 (0.5094), B12 (0.2422), C4 (0.1423), C7 (0.3507), C10 (0.1354), and C11 (0.2808) were also positively correlated with the first ordination axis. This showed that Nap and Ace affected these Algae more. Phe represented a negative correlation with the first species ordination axis, and the corresponding algae including Cy1 (−0.6637), Cy2 (−0.5175), B3 (−0.4524), B6 (−0.1734), B10 (−0.2584), B13 (−0.3717), C5 (−0.2220), and C9 (−0.3358) also showed a negative correlation, which indicated that these algae may have an effect on the accumulation of Phe. BaA, Chr, and Ant were not significantly correlated with the first two species axes in the first kind of RDA analysis. In the second kind of RDA analysis (PAHs vs. algae), these PAHs were positively correlated with the first species ordination axis (r = 0.9559–0.9831), and the corresponding algae consisting of B5 (0.1103), B12 (0.3625), B13 (0.1176), E1 (0.6355), E2 (0.4065), C1 (0.1076), C2 (0.5239), C3 (0.4964), C6 (0.5077), and C8 (0.5496) were also positively correlated with the first ordination axis. This showed that these algae had important effects on the adsorption and accumulation of these three types of PAHs. In the second RDA analysis, Nap and Ace were significantly positively correlated with the second species ordination axis, and the algae that could have a significant impact on PAHs accumulation was similar to the results of the first RDA analysis. Phe was highly negatively correlated with the first two species axes in the second RDA analysis. The other two algae, B1 and B7, also presented significant effects, but B6, B13, and C9 were excluded. The above results showed that when the matrix exchange technique was used to discuss the interaction of the two kinds of species (phytoplankton and PAHs) that interacted with each other, the results of the two RDA analyses could be mutually confirmed and supplemented.

The algae, PAHs, and sampling sites could be well divided into three groups (dotted circles in Figure 5), and the components included in groups 1, 2, and 3 (Figure 5A) and in groups 1′, 2′, and 3′ (Figure 5B) were basically consistent, and these results were also showed in Table 3. The predominant algal species and the dominant residual PAHs at different sampling sites were determined from the length of the arrows in Figure 5. The relationship between residual PAHs, indicator algae, and sampling sites in RDA analysis was summarized in

Table 4. Summary about the relationship among residual PAHs, indicator algae, and sampling sites in RDA.

Matrix	Group	PAHs	Algae type	Sample Sites
Algae vs. PAHs	1	Nap, Ace	Cy3, B2, B11, B12, C7,C10, C11, B12	4A, 5A, 10A
	2	Phe	Cy1, Cy2, B1, B3, B6, B10, B13, C5, C9	2A,3A, 8A, 11A, 13A, 14A
	3	BaA, Chr, Ant	B1, B5, B7, E1, E2, C1, C2, C3, C4, C6, C8, B4	1A, 6A, 7A, 9A, 12A, 15A
PAHs vs. algae	1	Nap, Ace	Cy3, B2, B6, B11, C4,C7,C9,C10, C11, B4	4A, 5A, 10A,12A, 15A
	2	Phe	Cy1, Cy2, B1, B3, B7, B10, C5	2A, 3A, 6A, 8A, 11A, 13A,14A
	3	BaA, Chr, Ant	B1, B5, B12, B13, E1, E2, C1, C2, C3, C6, C8	1A, 7A, 9A

Note: Different eutrophication levels indicated by different colors, os: oligotrophication; ßm: mesotrophic; ßαm: meso-eutrophic; αm: moderate eutrophic; ps: hypertrophic type.

.

Group 1/1′ was associated with the algae species with high accumulation capacity or strong influence for/on Nap and Ace, and its typical sites were Wenzheng Bridge (Site 4), Xiafu Bridge (Site 5), and Fengtai Bridge (Site 10), which all belonged to meso-eutrophic water bodies. Group1/1′ consisted primarily of Cyanophyta of Merismopedia (Cy3, αm), Bacillariophyta of Melosira (B2, βαm), Fragilaria (B6, βm), Gomphonema (B11, βαm), and Cymatopleura (B12, αm, βm), Chlorophyta of Scenedesmus (C7, αm, βm), Pediastrum (C10, βm, os), and Schroederia (C11, ps). In which, Cy3, B6, B12, C4, C7, and C10 are the indicator species from mesotrophic to moderate eutrophic water bodies (ßm, ßαm, αm), and C11 is the dominant species of hypertrophic water bodies. This group indicated that these algae, which are indicator species from mesotrophic to moderate eutrophic water bodies [39], had a strong ability to bioaccumulate or strong influence for/on Nap and Ace.

Group 2/2′ was associated with the algae species with a high accumulation capacity or strong influence for/on Phe, and its typical sites were Yixian Bridge (site 2), Wenjin Bridge (site 3), Wetland Park (site 8), Hanzhongmen Bridge (site 11), the Inside the Sancha Estuary (site 13), and the Outside the Sancha Estuary (site 14), which belonged to meso-eutrophic water bodies (ßαm), except for Yixian Bridge (site 2, ßm). Group 2/2′ mainly included Cyanophyta of Oscillatoria (Cy1, αm) and Phormidium (Cy2, ps), Bacillariophyta of Navicula (B1, βαm), Cyclotella (B3, βαm), Fragilaria (B6, βm), Synedra (B7, βαm), and Gyrosigma (B10, βαm), and Chlorophyta of Actinastrum (C5, βm, os) and Closterium (C9, αm, βm). Cy1 and Cy2 are the indicator species of moderate eutrophic and hypertrophic water bodies (αm and ps), B1, B3, B7, and B10 are the indicator species of meso-eutrophic water bodies (ßαm), B6, C5, C9, and B13 are the indicator species of oligotrophic and mesotrophic water bodies (os and ßm, ßαm, and αm), and C11 is the dominant species of hypertrophic water bodies [39]. This grouping suggested that these algae, which are the indicator species from oligotrophic to hypertrophic water bodies, had a strong ability to enrich or strongly influence Phe.

Group 3/3′ was associated with the algae species with high accumulation capacity or strong influence for/on BaA, Chr, and Ant, and the typical sites were Zhenzhu Bridge (site 1, ßm), Yunliang River (site 7, αm), and the External Qinhuai River (site 9, ßm). Group 3/3′ mainly included Bacillariophyta of Navicula (B1, βαm) and Tabellaria (B5, os), Euglenophyta of Euglena (E1, ps, αm, βm) and Phacus (E2, βm, os), Chlorophyta of Chlorella (C1, ps, αm), Ankistrodesmus (C2, αm, βm), Pandorina (C3, βm, os), Oocystis (C6, βm, os), and Characium (C8, os, βm). B1, B7, C1, C2, and E1 are the indicator species of meso-eutrophic (βαm) and moderate eutrophic water bodies (αm), E2, C3, C4, C6, and C8 are the indicator species of mesotrophic water bodies (ßm), and B5 is the indicator species of oligotrophic water bodies (os). This grouping suggested that these algae, which are indicator species from oligotrophic to moderate eutrophic water bodies, had a strong ability to enrich or strong influence on BaA, Chr, and Ant.

3.3.3. Stepwise Multiple Regression Analysis

In order to explore the relationship between PAHs content in algae, algae abundance, and biomass, we carried out the stepwise regression analysis based on the data of algae abundance, algae biomass, and PAHs content in different sample sites (

Table 5. Results of stepwise multiple regression between algae abundance/biomass and PAHs content.

Group	PAHs	Algae	Algae Entered	Regression Equation	R	F	p
Group I	Nap	Abu (1)	E1,	CNap = 40.797 − 0.172 × E1	0.533	5.174	0.041
			B12	CNap = 43.890 − 0.202 × E1 − 6.778 × B12	0.724	6.609	0.012
			C4	CNap = 42.579 − 0.209 × E1 − 6.244 × B12 + 5.974 × C4	0.832	8.250	0.004
			C11	CNap = 42.645 − 0.212 × E1 − 11.023 × B12 + 5.980 × C4 + 7.447 × C11	0.913	12.571	0.001
			B11	CNap = 41.541 − 0.208 × E1 − 10.604 × B12 + 6.272 × C4 + 7.442 × C11 + 4.620 × B11	0.946	15.374	0.000
		Bio (2)	E1, B12, Cy3, C4, B5	CNap = 40.068 − 2.123 × E1 − 595.594 × B12 + 751.959 × Cy3 + 13,544.526 × C4 +19.661 × B5	0.964	23.525	0.000
	Ace	Abu	C4,	CAce = 10.048 + 3.248 × C4	0.583	6.685	0.023
			E1,	CAce = 11.195 + 3.464 × C4 − 0.070 × E1	0.809	11.346	0.002
			B11	CAce = 11.195 + 3.634 × C4 − 0.068 × E1 + 3.136 × B11	0.920	20.232	0.000
		Bio	C4, E1, B11	CAce = 11.195 + 7268.369 × C4 − 0.685 × E1 + 156.789 × B11	0.920	20.232	0.000
Group II	Phe	Abu	B1	CPhe = 2.607 + 0.179 × B1	0.575	6.426	0.025
		Bio	B1, B10, C4, C3	CPhe = 1.710 + 12.401 × B1 + 41.121 × B10 + 2379.072 × C4 − 36.288 × C3	0.913	12.472	0.001
Group III	Ant	Abu	B1	CAnt = 0.027 + 0.015 × B1	0.657	9.857	0.008
		Bio	B1	CAnt = 0.025 + 0.738 × B1	0.647	9.345	0.009
	BaA	Abu	B1	CBaA = 0.005 + 0.013 × B1	0.678	11.070	0.005
			B10	CBaA = −0.032 + 0.015 × B1 + 0.080 × B10	0.815	11.860	0.001
			Cy2	CBaA = −0.036 + 0.015 × B1 + 0.164 × B10 − 0.003 × Cy2	0.961	44.868	0.000
			C6	CBaA = −0.025 + 0.016 × B1 + 0.160 × B10 − 0.003 × Cy2 − 0.008 × C6	0.979	57.138	0.000
			Cy3	CBaA = −0.015 + 0.015 × B1 + 0.156 × B10 − 0.003 × Cy2 − 0.008 × C6 − 0.001 × Cy3	0.987	70.048	0.000
			B13	CBaA = −0.002 + 0.016 × B1 + 0.151 × B10 − 0.002 × Cy2 − 0.008 × C6 − 0.001 × Cy3 − 0.011 × B13	0.993	93.543	0.000
			B4	CBaA = 0.015 + 0.016 × B1 + 0.146 × B10 − 0.002 × Cy2 − 0.007 × C6 − 0.001 × Cy3 − 0.014 × B13 − 0.002 × B4	0.996	132.037	0.000
		Bio	B1, B10, Cy2, C6	CBaA = −0.028 + 0.783 × B1 + 4.023 × B10 − 3.826 × Cy2 − 0.412 × C6	0.978	54.740	0.000
	Chr	Abu	B1	CChr = −0.018 + 0.023 × B1	0.862	37.670	0.000
			B7	CChr = −0.015 + 0.015 × B1 + 0.007 × B7	0.924	35.079	0.000
			B3	CChr = −0.012 − 0.003 × B1 + 0.016 × B7 − 0.034 × B3	0.954	36.843	0.000
			B7,B3	CChr = −0.014 + 0.015 × B7 + 0.030 × B3	0.953	59.420	0.000
			C4	CChr = −0.028 + 0.015 × B7 + 0.031 × B3 + 0.048 × C4	0.970	58.397	0.000
		Bio	B1, B7, C1, B5	CChr = −0.029 + 0.766 × B1 + 0.349 × B7 − 11.653 × C1 − 0.406 × B5	0.959	28.535	0.000

Note: ⁽¹⁾: Abu refers to “Abundance”; ⁽²⁾: Bio refers to “Biomass”.

).

The PAHs accumulated by algae can also be divided into three groups, which was consistent with the results of RDA analysis. When phytoplankton abundance data acted as the independent variable and PAHs content in algae as the dependent variable, the algae species, which can enter the regression equation of PAHs accumulation amount, suggested their important contribution for PAHs accumulation in algae. From the results of regression equation parameters, the R value was generally greater than 0.533, most of them were 0.9 (p = 0.041–0.000) and F value was greater than 5.174, suggesting that the fits of the regression equation were good.

Group 1 consisted primarily of Nap, Ace, and Bacillariophyta of B11 (Gomphonema), B12 (Cymatopleura), Chlorophyta of C4 (Coelastrum), C11 (Schroederia), and Euglenophyta of E1 (Euglena) based on the regression analysis of algae abundance, and also included Cyanophyta of Cy3 (Merismopedia) and Bacillariophyta of B5 based on the regression analysis of algae biomass.

Group 2 mainly included Phe and Bacillariophyta of B1 (Navicula) according to the regression analysis of algae abundance, and also included Bacillariophyta of B10 (Gyrosigma), Chlorophyta of C4 (Coelastrum) and C3 (Pandorina) based on the regression analysis of algae biomass.

Group 3 consisted primarily of Ant, BaA, Chr, and Cy2 (Phormidium) and Cy3 (Merismopedia), Bacillariophyta of B1 (Navicula), B3 (Cyclotella), B4 (Stephanodiscus), B7 (Synedra), B10 (Gyrosigma), and B13 (Stauroneis), and Chlorophyta of C4 (Coelastrum) and C6 (Oocystis) according to the regression analysis of algae abundance, and also included Chlorophyta of C1 (Chlorella) and Bacillariophyta of B5 (Tabellaria) based on the regression analysis of algae biomass.

4. Discussion

The research on the coupling relation between eutrophication and hydrophobic organic contaminants (HOCs) in rivers and lakes with different trophic statuses has become a research hotspot. The biological pump, equilibrium partitioning, and the indirect influence of eutrophication (e.g., high pH induced by phytoplankton, and phytoplankton life cycling) play critical roles in the occurrence and fate of HOCs (e.g., PAHs) in different eutrophic level waters [40,41]. Eutrophication-induced regime shifts from macrophytes to phytoplankton dominance had a stronger influence on the burial ability of PAHs with higher hydrophobicity, varying with the dominant primary producers associated with the trophic level index of the water column [42]. Moreover, the rising surface air temperature, the declining wind speed, and the reducing days with precipitation had weakened the ability of Chinese lakes to bury 16 PAHs by 69.2 ± 9.4% − 85.7 ± 3.6% from 1951 to 2017. The relative contributions of the climatic variables to the reduced burial ability depended on the properties of the PAHs and lakes [43]. Our study also observed the significant negative correlation between the concentration of PAHs in sediments and that of algae (Figure 3B), which implied that phytoplankton had a stronger influence on the burial ability of PAHs in different trophic status waters.

Planktonic algae are considered a good biosorbent and have a strong bioenrichment ability to PAHs [3,5,6,7]. The accumulation and removal of PAHs by planktonic algae are affected by a series of factors, such as pH value, temperature, ion intensity, molecular structure and initial concentration of organic compounds, metabolic types of microorganisms, the lipid content and composition of the plankton cells, and eutrophication level [9,11,12,13,14,44,45,46,47,48,49]. The species composition and abundance of phytoplankton vary with the change in environmental conditions and trophic status of rivers and lakes [50]. Cytoderm structures and surface characteristics of algae determine their selectivity and enrichment efficiency for specific contaminants [51]. Tao et al. reported that the cyanobacteria life cycle affected the seasonal occurrence of PAHs in water and sediments, and a high pH induced indirectly by eutrophication decreased the seasonal air–water exchange fluxes of PAHs, and reduced the accumulation of PAHs in surface sediments and the aromaticity of surface sediments [52]. The trophic transfer of the PAHs from phytoplankton to zooplankton increased with phytoplankton biomass and temperature, but decreased with the lake trophic state index. Biomagnification only occurred during phytoplankton bloom periods [40]. The cyanobacterial biomass affected by the trophic status dominated the occurrence of the PAHs in the surface sediments of these lakes. The biomass dilution and biological pump affected the accumulation of the PAHs in phytoplankton and zooplankton, and had a bigger influence on the PAHs with higher hydrophobicity. Both the bioconcentration factors and bioaccumulation factors of the PAHs decreased with the TSI [53].

At the same time, the increasing pollutants’ concentrations (e.g., PAHs) negatively affected the phytoplankton abundance, composition, and physiology, including photosynthetic inhibition, DNA damage, reduced gene expression, and oxidative stress [44,54,55,56]. Othman et al. observed dramatic changes in the phytoplankton community composition and size classes, and smaller cells were more resistant to the highest PAH concentrations. [54]. Sreejith et al. reported that pyrene was found to be more toxic to phytoplankton, picoeukaryotes exhibited higher sensitivity to PAHs than Synechococcus, and the populations in the highly oligotrophic northern region of the Red Sea were more tolerant to PAHs [55]. In our research, we also observed the sample site specificity and algal species specificity for the PAHs accumulation in algae due to the difference in eutrophication levels (Figure 2 and Figure 3), and the significant negative correlation between phytoplankton abundance and PAHs residue, which implied that the PAHs had induced adverse effects on these algae.

As shown in Figure 2 and Figure 3, the predominant algal species and dominant residual PAHs in water, sediment, and algae varied significantly across the study area (p < 0.01). These differences may have been related to water chemical characteristics and phytoplankton composition (Table 1, Figure 1B and Figure 2). Thus, the residual characteristics of PAHs in water, sediment, and algal samples indicated sample site and algal species specificity. The capacity of algae to accumulate PAHs is influenced by many factors, including the physicochemical properties of water, the type and structure of PAHs, and the algal cell characteristics (e.g., lipid content) and community composition [57]. Due to their lipophilic nature, PAHs tend to accumulate in phytoplankton cells. For example, Thalassiosira sp. showed the highest proportion of PAH-sorbed cells (29% and 97% of total abundance for phenanthrene and pyrene, respectively), which may be attributed to its relatively high total lipid content (33.87% dry weight) [44]. Moreover, the changes in the nitrogen status of algae can increase algal lipid content and affect the bioconcentration of hydrophobic organic compounds (HOCs). Bioconcentration factors (BCFs) for several HOCs increased up to nine-fold when the total lipid content of the green algae Selenastrum capricornutum increased from 17% to 44% of algal dry weight under nitrogen starvation [58]. It can be seen that the eutrophication level of river and lake waters could indeed affect the distribution characteristics of PAHs in water, sediments, and algae. In our study, the bioconcentration factors (BCF) of PAHs from water to phytoplankton in the study area were calculated, and the linear equation between logBCF and logKow of PAHs showed that the logBCF of different PAHs were basically linearly positively correlated with logKow of PAHs (R² = 0.3866); however, the BCF of Phe and Chr obviously deviated from the linear trend (Figure 6).

In this study, using RDA based on the data matrix exchange technique, we found that the eutrophication level of water bodies did affect phytoplankton composition, bioaccumulation of PAHs by phytoplankton, and led to sample site and algal species specificity. Some specific algae species, such as Cyanophyta and some species of Bacillariophyta and Chlorophyta, in mesotrophic (βm) and meso-eutrophic water bodies (ßαm), preferentially accumulated lower-ring PAHs (Nap, Ace, and Phe), and some other specific algae species, such as Euglenophyta, some species of Bacillariophyta, and most of Chlorophyta, in mesotrophic and moderate eutrophic water bodies preferentially accumulated higher-ring PAHs (BaA, Chr, and Ant).

5. Conclusions

The eutrophication level of water bodies did affect phytoplankton composition, bioaccumulation of PAHs by phytoplankton, and led to sample site and algal species specificity. The planktonic algae mainly accumulated the PAHs with two and three rings. Cyanophyta and some species of Bacillariophyta and Chlorophyta in mesotrophic (βm) and meso-eutrophic water bodies (ßαm) preferentially accumulated lower-ring PAHs (Nap, Ace and Phe with 2- and 3-ring PAHs). Some other specific algae species, such as Euglenophyta, some species of Bacillariophyta, and most of Chlorophyta in mesotrophic and moderate eutrophic water bodies, had strong capacities to enrich the BaA, Chr, and Ant with high-ring PAHs. The ecological contribution of phytoplankton on PAHs accumulation may be affected by the eutrophication level of water bodies, phytoplankton composition, and the structure of the PAHs and the algal cell walls.