Introduction

Algae blooms are serious environmental problems around the world, especially in developing countries (Kagaloua et al. 2008; Pernet-Coundrier et al. 2012). Algae blooms occur in an aquatic environment if too much phosphorus (P) enters the system (Giles et al. 2015). For a eutrophic lake, release of P from dead algae is an important nutrient source that will support continuous algal blooms in lakes (Li et al. 2009; Feng et al. 2016a). Algae are not usually collected by humans because they are not valuable, so that the debris is allowed to decompose in situ. The decomposition of algal residues affects the bio-cycling and release of P, increasing the risk of resurgence of algal blooms (Feng et al. 2016b; Lehman et al. 2017). Decaying algal debris releases both inorganic P (Pi) and Po. The Po constituents need to be hydrolyzed to bioavailable Po by various enzymes (Feng et al. 2016b). Therefore, the forms and concentrations of P in algae shall be evaluated while algae debris is decomposing. However, until now, studies of the species, concentrations, and effects of Po in algae of eutrophic lakes have been limited because of the complexity and limitations of analytical methods (Turner et al. 2005; Bell et al. 2017).

Typical analytical approaches, such as enzymatic reactivity, high performance liquid chromatography (HPLC), and mass spectrometry are based on operational definitions so that they cannot discern P classes at a molecular level (Suzumura 2005; Baldwin 2013; Karl 2014). Phosphorus-31 nuclear magnetic resonance spectroscopy (31P NMR) is a non-destructive, non-invasive technique for identifying chemical forms in various environmental samples. Solid-state 31P NMR is focused on inorganic P compounds, solution 31P NMR is mainly used to determine organic P compounds (Turner et al. 2005; He et al. 2011; Abdi et al. 2014; Sørensen et al. 2014). Several P compounds have been detected by solution 31P NMR, including phosphonates, orthophosphate,monoesters P, diesters P, pyrophosphates and polyphosphates (He et al. 2007; Turner et al. 2012; Zhu et al. 2013). Generally, monoesters P represent a wide range of important Po compounds, such as inositol phosphate and sugar phosphates (He et al. 2007; Doolette et al. 2009; Jarosch et al. 2015). Therefore, it is an ideal technique for analyzing Po species in algae of eutrophic lakes, as it would not only provide important information pertaining to P biogeochemical cycling in lake ecosystems, but also yield abundant insight into identities of specific P compounds.

Chao Lake (31°25′28″–31°43′28″N, 117°16′54″–117°51′46″E) is one of the five largest freshwater lakes in China. It is situated on the flood plains between the Yangtze River and Huai River in the central Anhui Province of eastern China (Zan et al. 2010; Tang et al. 2015). Chao Lake is a typical shallow lake with a mean depth of 3 m and a surface area of 780 km2 and drainage area of 13,350 km2 (Wang et al. 2013). Due to the rapid increase in anthropogenic activities in the lake’s watershed over recent decades, the lake has suffered from serious pollution, eutrophication and algae blooms (Xu et al. 2005). As a matter of fact, since the mid 1980s, algae blooms have occurred each year in Chao Lake (Chen and Liu 2014). In order to assess the environmental risks of algae on the eutrophic lake, this study analyzed the P species in algae by solution 31P NMR, and based on the knowledge obtained, predicted the P bioavailability of the algal biomass in aquatic environments.

Materials and methods

Study sites and sample collection

Samples of algae were collected from six sites in Chao Lake in September 2015 (Fig. 1). These sampling sites were located in different eutrophic areas. Generally, Chao Lake is divided into two parts: the western part (samples C1 and C2) and the eastern part (samples C3, C4, C5 and C6) along the line of Zhongmiao–Mushan–Qitouzui, as shown in Fig. 1. The quality of water was worst in the western part of the lake and gradually became better from west to east (Zhu et al. 2006; Tang et al. 2015). The lake’s annual mean concentrations of total nitrogen (TN) and TP approached 2.85 and 0.26 mg l−1, respectively, and the annual mean chlorophyll-a reached up to 25.6 μg l−1 (Li et al. 2015). Algae were collected by use of a plankton collector (HB403-BWS). Samples were placed in sealed bags and put in ice boxes immediately. These algal samples were freeze-dried, ground, then passed through a 2-mm sieve before being stored at −20 °C (Feng et al. 2016a). The dominant species in Chao Lake was Microcystic aeruginosa with an appearance frequency of 90.9% (Yang et al. 2011).

Fig. 1
figure 1

Major inlets and outlets of Chao Lake with algal sampling sites

Full size image

Extraction of P and chemical analysis

Samples of algae were extracted by use of optimized NaOH–EDTA extracting agent (mixtures of 0.5 mol l−1NaOH and 25 mmol l−1EDTA) with a ratio of 150:1 (ml g−1), and the mixtures were shaken for 18 h at room temperature (Cade-Menun and Preston 1996; Feng et al. 2016b). The extracting solutions were centrifuged (8000 × g) for 30 min, and filtered through 0.45-μm glass-fiber filters (Whatman GF/C). Extractable total phosphorus (ETP) after digestion and free molybdate reactive phosphorus (MRP) were measured using the molybdenum blue method (He and Honeycutt 2005). Extractable organic phosphorus (EOP) was calculated by the difference between ETP and MRP. The remaining extracts were freeze-dried for solution 31P NRM spectroscopy analysis.

Percentages of carbon (C) and nitrogen (N) in algae were determined by use of an elemental analyzer (Elementarvario macro EL, Germany). Total phosphorus (TP) and inorganic phosphorus (Pi) were determined by the SMT method described by Ruban et al. (2001). Organic phosphorus (Po) in algae samples was calculated by the difference between TP and Pi. There were three replications for extraction of P and chemical analysis. Data were checked for deviations from normality and homogeneity of variance before performing statistical analyses.

31P NMR analysis

A 100-mg sample of freeze-dried algal extracts was ground, and then redissolved in 1 ml 1 mol l−1 NaOH + 0.1 mol l−1 EDTA and 0.2 ml D2O. After ultrasonication for 30 min and equilibration for 5 min, 2% (v/v) of bicarbonate buffered dithionite (0.11 mol l−1 NaHCO3 + 0.11 mol l−1 Na2S2O4) was added to the extract to reduce interference from paramagnetic ions, such as Fe and Mn (He et al. 2009; Giles and Cade-Menun 2014). The pH of the supernatant solution was adjusted using 10 mol l−1 NaOH to ensure a pH > 12. The supernatant solution was centrifuged (8000 × g) for 30 min and transferred to a 5-mm NMR tube. Solution 31P NMR spectra were acquired at 24 °C on a Bruker AVANCE 400 MHz spectrometer at a 31P frequency of 161.98 MHz, using a 90° pulse, a 5 s relaxation delay and a 0.21 s acquisition time, similar to the parameters used in Feng et al. (2016b). The scan time for each sample was more than 15 h. Peak areas were calculated by integration and completed using MestrelabMNova v.10.

Spiking experiments

The peak of specific monoesters P forms (i.e., glucose 6-phosphate, riboncleotides, α-glycerophosphate, β-glycerophosphate, myo-inositol hexaphosphate) needs to be confirmed with spiking experiments (Fig. 4). Methods for identifying specific P forms in NMR spectra of soil and other environmental samples are well-established and have been used for many years (Smernik and Dougherty 2007; He et al. 2011; McDowell and Hill 2015), combined with P compound libraries developed by Turner et al. (2003); Doolette et al. (2009) and Cade-Menun (2015). Standard samples of monoesters P were purchased from Sigma-Aldrich. Spiked samples were analyzed by 31P NMR as described above. Monoesters P compounds were identified by their chemical shifts, with the orthophosphate peak in all spectra standardized to 6.0 ppm. Spectral processing was done using MestReNove software version 9.0.1 (MestReNove Research SL).

Results and analysis

Nutrients (C, N and P) in debris of algae

Composition of C, N and their ratios in debris of algae are shown in Table 1. Percentages of C ranged from 31.5 to 50.3% with a mean value of 41.9% in the algae samples from Chao Lake. Content of N was 5.2–7.8% with a mean value of 6.4%. Both contents of C and N of algae was greater than those of aquatic macrophytes, which has been widely reported previously (Zhong et al. 2012; Qu et al. 2013, Feng et al. 2016a). The ratio of C:N was a good predictor of degradation in algae and aquatic macrophytes with a lower ratio of C:N for material more readily degradation (Reitzel et al. 2006). In this study, the ratios of C:N in algae from the western lake (7.1 ± 0.3) were higher than those of the eastern lake (6.4 ± 0.3), suggesting higher lability of the algal debris of the eastern lake. The ratio of C:N of aquatic macrophytes (average 12.9) (Feng et al. 2016a) was greater than that of algae in Chao Lake (average 6.7) (Table 1). Therefore, the algae decomposed more easily than aquatic macrophytes in the same lake. This was also consistent with the results of previous studies (Liu et al. 2016).

Table 1 Contents of C, N, and P in algae and their NaOH–EDTA extraction efficiency in Chao Lake
Full size table

Contents of TP in debris of algae ranged from 2671 to 5385 mg kg−1 dry mass (dm) with a mean value of 3936 mg kg−1. The greatest concentration of TP in debris of algae was observed in sample C6. This may be because of two major inflowing rivers (e.g., the Zhe gao river and Shuang qiao river) with heavy nutrition inputs (Tang et al. 2015). Contents of TP in the surface sediments from Chao Lake only ranged from 420 to 1090 mg kg−1 with a mean value of 687 mg kg−1 (Zhang and Xing 2013). Content of TP in debris of algae from Chao Lake was approximately 5 times higher than TP in surface sediments. Therefore, dead algae-derived P might be an important source of bioavailable P for repeated algal blooming in eutrophic lakes such as Chao Lake. Concentration of Po in these algal samples ranged from 2107 to 4136 mg kg−1with a mean value of 2783 mg kg−1. The greatest concentration of Po was also observed in the heavily polluted region (i.e., sample C6). The proportion of Po in algae of the western lake (78.3% ± 2.6%) was higher than that in the eastern lake (64.7% ± 1.4%). The mean ratio of Po/TP of the six algal samples was 69.2% in Chao Lake (Table 1). This value was greater than that of algae from Tai Lake (mean ratio of Po/TP 57.8%) (Feng et al. 2016b). The previous studies reported that Po could be converted to bioavailable P (e.g., HPO42−) for algae blooming through a series of redox-driven solubilization reactions and phosphatase-mediated hydrolytic processes (Wang and Pant 2010; Zhu et al. 2015). Thus, we believed that Po in debris of algae from Chao Lake possessed larger bioavailability potential than algae from Tai Lake per their difference in Po/TP ratios.

NaOH–EDTA extractable P from algae

In this study, contents of NaOH–EDTA extractable TP ranged from 1292 to 2774 mg kg−1, with an extraction efficiency of 40.5–78.3%, and contents of extractable Po from algae ranged from 351 to 1141 mg kg−1, with an extraction efficiency of 13.8–54.2% (Table 1). The recoveries of Po in debris of algae from Chao Lake were lower than those with pure algae such as Microcystis, Chlorella vulgaris, and Sprilinaplatensis (Feng et al. 2016a). However, the extraction efficiencies of Po of algae from Chao Lake were similar to those of particulate P from Tai Lake (23–56%) (Bai et al. 2017) and of sediments from Haihe River (30–73%) (Zhang et al. 2017). The extraction efficiency of TP from algae from Chao Lake was similar to that of soils and sediments (49–83%) (Xu et al. 2005; Tang et al. 2015). Multiple-step extractions (e.g., additional or sequential HCl extraction) (He et al. 2008; Cade-Menun 2015; Zhu et al. 2016; Liu et al. 2017) seem needed to increase the P recovery from these algal samples.

Solution 31P NMR spectra of NaOH–EDTA extracts of algae

Eight main P species including inorganic P species (orthophosphate and pyrophosphate) and Po species (five monoesters P and diesters P) were identified in the NaOH–EDTA extracts of the six algal samples by solution 31P NMR (Fig. 2; Table 2). The peak of orthophosphate was at 6.00 ppm in the 31P NMR spectra, monoesters P was at 3.33–5.49 ppm, diesters P was at −0.69 to −0.31 ppm, and pyrophosphate was at −4.21 to −4.12 ppm (Fig. 2). The sum of orthophosphate and monoesters P in ETP accounted for more than 93% of ETP (Table 1; Fig. 3b). With the peak of orthophosphate as the largest signal in these 31P NMR spectra. The content of orthophosphate was between 615.1 and 1331.7 mg kg−1, and accounted for 41.5–54.0% of ETP (Table 1; Fig. 3). Polyphosphates were not detected in any algae of Chao Lake.

Fig. 2
figure 2

Solution 31P NMR spectra of NaOH–EDTA extracts of algae in Chao Lake

Full size image
Table 2 Concentrations of P compounds in NaOH–EDTA extracts of the algae determined by solution 31P NMR
Full size table
Fig. 3
figure 3

Content and percentage of major P types (a, b) and monoesters P compound forms (c, d) in algae collected from six sites (C1–C6) of Chao Lake

Full size image

Monoesters P comprised the largest Po fraction with NaOH–EDTA extracts, and accounted for 48.4% (average) in ETP (Table 2; Fig. 3b). Through the spiking experiments (Fig. 4), the peak at 4.88 ± 0.02 ppm was assigned to α-glycerophosphates and the peak at 4.50 ± 0.02 ppm was assigned to β-glycerophosphates, based on Turner and Richardson (2004) and Doolette et al. (2009); the percentages of α- and β-glycerophosphates in ETP were 3.4% (65.5 mg kg−1) and 22.6% (432.7 mg kg−1), respectively. β-glycerophosphates were the largest component of monoesters P. The peak at 4.32 ± 0.01 ppm was assigned to ribonucleotides (He et al. 2011) and the percentage of ribonucleotides in ETP was 10.6% (218.2 mg kg−1). The peak at 5.12 ± 0.01 ppm was assigned to glucose 6-phosphate (Cade-Menun 2015), the percentage of which in ETP was 1.9% (38.7 mg kg−1). In addition, a few of the peaks in the monoesters P region were unidentified, because chemical shifts were strongly influenced by subtle differences among samples for viscosity, pH, salts and paramagnetic ions (Young et al. 2013; Abdi et al. 2014; He et al. 2011; Giles et al. 2015). Unidentified monoesters P are defined as ‘other monoesters P’ in this study. These other monoesters P accounted for 9.9% of ETP in algae of Chao Lake (Table 2).

Fig. 4
figure 4

Solution 31P NMR spectra of monoesters P standard compounds (a–e) and C1 algal sample in Chao Lake (f)

Full size image

Pyrophosphate was detected in most samples except the sample C1, which is consistent with a number of other studies (Bedrock et al. 1995; Mahieu et al. 2000; He et al. 2011). In other literature, polyphosphates and pyrophosphate were detected in some samples, but not necessarily (Busato et al. 2005; Feng et al. 2016b).

Discussion

Identification of inositol hexaphosphate (IHP) in algae of Chao Lake

The inositol hexaphosphate (IHP) stereoisomers (scyllo-, myo-, chiro-, neo-IHP) were important monoesters P components in many environmental samples (Turner et al. 2012; Cade-Menun 2015). Each of these compounds contains six phosphates, and the conformation of those phosphate groups causes them to have multiple peaks in a single spectrum, in an arrangement specific to each compound. The only exception is scyllo-IHP, which has one peak for the six phosphates; the peak at 3.55 ± 0.02 ppm was assigned to scyllo-IHP, based on Turner and Richardson (2004) and Doolette et al. (2009). None of the spectra shown in Fig. 2 have a peak at 3.55 ppm, so the scyllo-IHP was not present in algae of Chao Lake. Myo-IHP has four peaks in a 1:2:2:1 arrangement (with respect to peak areas), with peaks at 5.79 ± 0.01, 4.88 ± 0.01, 4.55 ± 0.01 and 4.42 ± 0.01 ppm (in Fig. 4e). Three of these peaks were not present in algae samples in Chao Lake (Fig. 2), so myo-IHP was not present in algae of Chao Lake. The fact that myo-IHP had not been observed in algae in a previous study is consistent with the observation of this study (Feng et al. 2016a). For chiro-IHP [in either the 4 equatorial/2 axial (4e/2a) conformation or the 2e/4a conformation], three peaks must be clearly visible, in a 2:2:2 arrangement, and the diagnostic peaks for each are at 6.2–6.5 ppm (Cade-Menun 2015); however, none of the spectra shown in Fig. 2 have any peaks between 6.0 and 7.0 ppm, so these compounds were not present in algae of Chao Lake. In addition, neo-IHP requires two peaks to be present, in a 4:2 arrangement, so that the peak at 6.4 ± 0.01 ppm is twice as large as the one at 4.3 ± 0.01 ppm. Given that there were no peaks between 6.0 and 7.0 ppm, it can be assumed that neo-IHP was not present in algae of Chao Lake.

Degradation behaviors of diesters P in algae of Chao Lake

Apart from monoesters P, other important Po compounds, diesters P, were detected in some algae samples. The concentration of diesters P was generally low (mean 0.63% of ETP), compared to other Po fractions, and only detected in samples C4, C5 and C6 (9.2–61.6 mg kg−1).

It is well established that some diesters P such as phospholipids and RNA can degrade to monoesters P [e.g., α-,β-glycerophosphates (phospholipids) and various monophosphates (e.g., nucleotides) when analyzed at the high pH required for good peak separation in 31P NMR spectra](Turner et al. 2003; Doolette et al. 2009; He et al. 2011; Schneider et al. 2016). The degree of degradation will vary depending on the length of NMR experiment and other factors (Cade-Menun and Liu 2014; Cade-Menun 2015; Feng et al. 2018). It was essential that these degradation peaks were identified and quantified, in order to determine the correct concentrations of monoesters P and diesters P (Young et al. 2013; Vincent et al. 2013). Therefore, the corrected total monoesters P and corrected total diesters P were those corrected by moving the percentages of α- and β-glycerophosphates and nucleotides from monoesters P to diesters P. When uncorrected, the total monoesters P were significantly higher than the total diesters P, but the reverse was true for the corrected values.

Eutrophication and algal blooming versus biogeochemical cycling of algae-derived P in Chao Lake

Based on the results in this study, we were able to think further about the biogeochemical cycling of P driven by algal blooming in Chao Lake. The TP of algae-derived biomass loading was 10.94 × 103 kg in Chao Lake (Li et al. 2015). With the Po content in algae determined in this study (Table 1), we estimated the Po biomass of algae to be approximately 7.57 × 103 kg in Chao Lake. In previous research (Feng et al. 2016b) with Tai Lake samples, we estimated that approximately 32.7–41.3% of extractable Po from algae has the potential for phosphatase hydrolysis to soluble orthophosphate which can be released into the water body. Thus, in the case of Chao Lake, this bioavailable P would be 2475–3126 kg in algae and would be released into the water and promote repeated algal blooms in Chao Lake if not appropriately removed naturally or artificially. This conclusion indicated that decomposition of algal debris would be a key factor in regeneration of bioavailable P for life in eutrophic lakes, even when external P is excluded. It is therefore necessary to remove algae debris from eutrophic lakes to control the release of P from internal P cycling and the phenomenon of eutrophication of lakes.

Conclusion

This research used solution 31P NMR to provide insights into the P species and distribution of P in algae of the heavily polluted Chao Lake. Data derived from this study showed that the eutrophic lake algae have accumulated remarkable amounts of Pi and Po. The proportion of Po in algae ranged from 55.3 to 80.8% with a mean of 69.2% in Chao Lake.

Eight compounds P were detected in the NaOH–EDTA extracts of algal samples by 31P NMR. The sum of orthophosphate and monoesters P in ETP was greater than 93% of algal P. Our observations implied that the release of P induced by the decomposition of algae debris could be a potential source of bioavailable P in aquatic systems of Chao Lake even without any more external P input. Thus, recycling of the potential bioavailable P in algae might be the mechanism of repeated algae blooming in eutrophic Chao Lake. Remediation of the lake requires a strategy to remove the algal biomass P effectively.