Discussion
Biomagnification of BMAA along food chains was synthetically demonstrated in the diatoms-dominated marine ecosystem. The concentration of BMAA accumulated in zooplankton was generally higher than that in phytoplankton (Fig. 2), and only in January the average level of BMAA in zooplankton was slightly lower than that in phytoplankton (Supplementary Fig. 3). We note that the average concentration of BMAA in phytoplankton was much higher in January than that in other seasons. The lowest density of phytoplankton (> 20 µm) and the highest percent of diatoms (99.9%) also occurred in January. Inversely, the lowest percent (87.2%) of diatom in phytoplankton net samples occurred in June, meanwhile the average concentration of BMAA was the lowest in four seasons, which hints that the concentration of BMAA in phytoplankton was positively related to diatoms. The diatom Bacillaria paxillifera was the dominant species only in January, but it has not been reported as a producer of BMAA in the previous studies17-19. Similarly, high average concentrations of BMAA accumulated in zooplankton in January and June, which may be related to the feeding behavior of zooplankton. Only the zooplanktonOikopleura dioica predominated in the zooplankton community as the dominant species in both January and June (Supplementary Table 1). The zooplankton O. dioicacreates a mucus net ”house” for capturing phytoplankton during feeding. Water is pumped through this house and phytoplankton are filtered out of the water and then transferred into the mouth. We suggest that this feeding method would result in an accumulation of BMAA by filtering large amounts of diatoms, while Calanus sinicus and Acartia pacificaare omnivorous, primarily feeding on phytoplankton but also consuming rotifers, ciliates, and their own eggs. Furthermore,Noctiluca scientillans is a heterotroph, primarily engulfing fish eggs, bacteria, in addition to plankton, diatoms and dinoflagellates. Therefore, the feeding habit of zooplankton and the species composition of phytoplankton are important factors for the accumulation of BMAA in zooplankton populations.
The average concentration of BMAA in phytoplankton and zooplankton net samples at all sampling stations in four seasons were 0.14 and 0.30 µg g-1 DW, respectively. Compared across different sampling stations, the highest trophic magnification factor (TMF) between phytoplankton and zooplankton trophic levels are shown in Supplementary Fig. 4. The TMF values of BMAA ranged from 0.45 to 11.82 in most of sampling stations except for C1 station in all seasons, with an average value 4.58. The highest value occurred at the sampling station C3 in September (11.82), and A3, A5, B2, D6 and D7 stations had relatively low TMF values (0~2). Compared across different seasons, the highest TMF values for BMAA occurred in different sampling stations was in September (8.50), followed by June (3.98) and January (1.13). In a previous study, the concentration of BMAA in cyanobacteria-enriched samples collected from Baltic Sea ranged from 0.001 to 0.015 μg g-1 DW, and zooplankton-enriched fraction contained 6-fold higher on average than the cyanobacterial samples20, of which the TMF is in the range of our study from phytoplankton to zooplankton (1.13~8.50). For another case, average BMAA concentrations were 4.12, 3.21, 3.76, and 6.05 µg g-1DW in cyanobacteria, mollusks, crustaceans and various fish species, respectively, and the TMF values were below 1.5 along food chains in Gonghu Bay of Lake Taihu, China46. A relatively low biomagnification of BMAA, trophic magnification factor ~1.20, was also found in six freshwater aquaculture products from Taihu Lake Basin in China21. Comparatively, Jiaozhou Bay showed a relatively high biomagnification along food chains from diatom-dominated phytoplankton.
In order to assess the risk of human exposure to BMAA, marine animals including filter-feeding bivalve mollusks, carnivorous crustacea, and saprophytic gastropod mollusks, were collected together to monitor BMAA accumulation. Results showed that BMAA was universally detected in marine animals collected during the whole year. Average concentrations of BMAA accumulated in bivalve, crustacea and gastropod mollusks were 0.84, 1.19, and 2.08 μg g-1 WW, respectively, corresponding to 4.22, 5.95, and 10.42 μg g-1 DW based on the conversion factor 5: 1 (wet weight: dry weight). The TMF of BMAA from phytoplankton (average 0.14 μg g-1 DW) to these animals were about 30.1, 42.5, and 74.4, respectively (Fig. 6). The production ability of diatom-dominated phytoplankton in this study was higher 1~2 orders of magnitude of that produced by cyanobacteria in Baltic Sea20. The concentrations of invertebrates (mussels and oysters, TMF 7~22) and various vertebrates (fish, TMF up to 200 in some tissues of fish) also demonstrated the magnification of BMAA along food chains20, but a recent study questioned the concentration of BMAA in the benthic fishes in the ecosystem of Baltic Sea45. However, the detection of BMAA with higher concentrations in gastropod mollusks living in the benthic ecosystem was consistent with that of fishes living closer to the bottom of the Baltic Sea which also contained higher BMAA concentrations than pelagic fishes20. It is possible that BMAA settles to benthic environments with the death of cyanobacteria and diatoms and subsequently stored in marine sediments. More research is required to discover the accumulation mechanism within marine animals living in the benthic ecosystem. Compared to another study, although the total average concentration of BMAA detected in freshwater cyanobacterial samples (4.12 µg g-1 DW) from Gonghu Bay, Taihu Lake, China46 is about 30 times that detected in the diatom-dominated phytoplankton from Jiaozhou Bay (0.14 µg g-1 DW), the average concentration of BMAA detected in the high-trophic N. didyma (21.4 μg g-1 DW) in this study is about 5-times higher than that in the freshwater mollusks and crustaceans (3.21 and 3.76 μg g-1 DW) reported in the previous study46. This phenomenon demonstrated that the TMF of BMAA from producer to benthic animals in the marine ecosystem studied here is great higher than that in the freshwater ecosystem of Gonghu Bay.
Results showed that BAMA was more widely present than BMAA and other isomers in phytoplankton and zooplankton samples collected from Jiaozhou Bay in this study. Unfortunately, the lack of commercialized BAMA reference material restricted an accurate analysis of this compound. Therefore, BAMA was verified by the precolumn AQC-derivatization method using C18 column for most of samples because it does not have baseline separation from BMAA on the HILIC column. In order to assess the variation trend of BAMA in different trophic organisms, the concentration of BAMA was estimated by the mass spectrometry response of BMAA standard in this study. We note that the ratios of BAMA to BMAA in different organisms roughly tended to decrease with increasing trophic levels. The relationship between BAMA and BMAA requires further studies to be explained in detail in the future.
The phytoplankton community (> 20 µm) was dominated by diatoms in Jiaozhou Bay in all four seasons in 2019 in this study, which was also consistent with some previous studies using 76 µm-net47-49. The density ratio of dinoflagellate to diatom in most months was less than 0.2, except in July, in 201047, and the annual average ratio of dinoflagellate to diatom was less than 0.08 in 201148. Diatoms were the most dominant phytoplankton group throughout the year in 2017, except in spring and winter, whenNoctiluca scintillans was co-dominant49. In this study, only the dinoflagellate Ceratium kofoidii was co-dominant species in June, and diatoms Skeletonema spp.,Chaetoceros spp., Thalassiosira spp. were co-dominant species almost in the whole year. The positive detection of BMAA in phytoplankton net samples suggested that diatoms may be the producer of BMAA in Jiaozhou Bay. The production of BMAA in marine diatoms17-19,50 and freshwater diatoms51 have been reported in some previous studies. Therefore, a total of 56 strains of diatoms pertaining to five genera including Pseudo-nitzschia , Chaetoceros ,Thalassiosira , Planktoniella , and Minidiscus were analyzed for BMAA, in which only Thalassiosira andChaetoceros were reported before to produce BMAA. In the previous study by Jiang et al .17, five genera of diatoms (Achnanthes sp., Navicula pelliculosa , Proboscia inermis , Skeletonema marinoi , Thalassiosira sp.) were cultured and tested to produce BMAA, in which the concentration of protein-bound BMAA in T. sp. was 3.28 ng g-1 DW. Seven species of marine diatoms (Navicula pelliculosa , Chaetoceros socialis ,Coscinodiscus granii , Thalassiosira weissflogii ,Phaeodactylum tricornutum , Skeletonema marinoi ,Ditylum brightwellii ) were cultured and analyzed for BMAA, in which the concentration of protein-bound BMAA in Ch. socialis andT. weissflogii were 0.62 and 2.19 µg g-1 DW, respectively, and varied concentration of free-form BMAA was also detected except for D. brightwellii 50. The highest production of BMAA was observed at the stationary growth phase of four diatom strains includingPh. tricornutum , Ch. sp., Ch. calcitrans , andT. pseudonana , at 1.4, 1.6, 1.8, and 0.28 µg g-1 DW, respectively18. Large discrepancy of BMAA concentration occurred in the laboratory cultures ofThalasososira and Chaetoceros in above different studies. In this study, 37.5% (21/56) of the studied diatom strains contained the protein-bound BMAA which belonged to four genera except forMinidiscus . Only three species of Minidiscus were cultured and analyzed here (Supplementary Table 5), and we could not determine that the diatoms of this genus are non-toxic for BMAA due to variation of BMAA production in the same diatom genus. Interestingly, BMAA was mainly present in the precipitated-bound fraction of the diatom cultures, followed by the total soluble fraction. In the genusPseudo-nitzschia , a total of 9 strains detected BMAA, in whichPs. multiseries (0.41 μg g-1 DW) and Ps. caciantha (0.30 μg g-1 DW) cultures contained total soluble BMAA, and the other 7 strains were only found precipitated bound BMAA ranging with 0.11~3.05 μg g-1 DW. In the genus Thalassiosira , 8 strains detected BMAA in the precipitated bound fraction ranging from 0.33~3.95 μg g-1 DW, in which T. minima , T. alleniiand T. gravida also contained total soluble BMAA. In the genusChaetoceros , 3 strains contained precipitated bound BMAA below 0.3 μg g-1 DW. The precipitated bound BMAA was also detected in a strain of Planktoniella at 0.16 μg g-1 DW. The average production level of BMAA in different diatom genera decreased in order of Thalasosira (2.07 μg g-1 DW) > Pseudo-nitzschia(1.00 μg g-1 DW) > Planktoniella(0.16 μg g-1 DW) > Chaetoceros(0.12 μg g-1 DW). Among the positively detected strains of diatoms, 40-fold difference of BMAA concentration between the highest (T. minima , 3.95 μg g-1 DW) and the lowest (Ps. sp., 0.11 μg g-1 DW) level. An additional batch culture of some BMAA-producing strains of diatoms was conducted to confirm the BMAA-producing ability of diatoms was steady and specific. The results showed that diverse diatom strains isolated from Chinese coast could produce BMAA dominated in the precipitated bound fraction. The distribution form of BMAA in the cultures of isolated diatom strains was largely different from that in the phytoplankton collected from Jiaozhou Bay. Additionally, the isomer BAMA was not found in the laboratory cultures although it was ubiquitous in the field diatom-dominated phytoplankton samples. This discrepancy hints that the production of BMAA in diatoms is affected by environmental factors such as nutrient, temperature, predator pressure, etc. What is the real source of BAMA in the phytoplankton samples collected from Jiaozhou Bay? How do the environmental factors regulate the production of BMAA by diatoms? These problems should be deeply explored in further studies.
Misincorporation of BMAA into proteins in the place of L -serine has been considered as a bioaccumulation and pathogenic mechanism for neurodegenerative diseases33,34,52, but some studies also demonstrated that BMAA was not be miscoded into proteins in higher organisms53,54 which was reviewed and proposed as a result of differing methodologies55. In this study, BMAA in diatoms was the highest in the protein-bound form. So far the biosynthesis of BMAA is still unknown. However, the analytical results of BMAA in the cultures of diatom strains demonstrated that the production of BMAA in diatoms was species- or strain-specific, and affected by nutrient conditions19. For example, the production of BMAA varied in the strains MC4451, MC4185, and MC4423 ofPs. multiseries isolated from Taiwan Strait Y13, Taiwan Strait E6, and Taiwan Strait Y23, respectively. The strain MC4451 produced total soluble BMAA, while MC4185 produced precipitated bound BMAA and MC4423 did not produce BMAA. The strains MC6394 and MC6056 ofPlanktoniella blanda were isolated from Taiwan Strait Y11 and Hong Kong, respectively, in which the former produced precipitated bound BMAA but the latter did not. The strains MC6475 (from Shantou Nan’ao Treasure Island) and MC6447 (Tsing Yi, Hong Kong) of Thalassiosira minima have the same pattern. As it has been pointed above that the BMAA form in the field phytoplankton collected from Jiaozhou Bay was different from that in the cultures of isolated diatom strains under laboratory conditions. We hypothesize that environmental stress or ecological predatory pressure from zooplankton induce diatoms to release a free form BMAA from the protein-bound fraction through the nitrogenous metabolism pathways.
In the present study, the benthic and carnivorous gastropod N. didyma was identified again as a typical vector for BMAA in Jiaozhou Bay ecosystem, of which the average concentration of BMAA reached 21.4 μg g-1 DW corresponding to the TMF of about 150 from phytoplankton. In order to explore the production ability of BMAA by the symbiotic bacteria in the gut of N. didyma , a total of 7 species of symbiotic bacteria were cultured and identified for toxin analysis. Results showed that none of these bacteria produced BMAA, but DAB was detected in all samples, which supports the hypothesis that the neurotoxin BMAA detected in N. didyma was accumulated from food chains. In addition, DAB, as an isomer of BMAA, has been detected before in diverse organisms such as cyanobacteria, phytoplankton, zooplankton, mollusks, etc.19,24,25,30,38,40, but we hypothesize that the marine bacteria isolated from the gastropod gut are possible source of DAB detected in aquatic organisms. Interestingly, some unknown compounds were also detected in these bacteria with the same transition ions as DAB, which may be precursor compounds relative to DAB. If true, DAB could also be produced by some heterotrophic prokaryotes in the environment, but no significant biomagnification phenomenon has been reported for DAB in aquatic ecosystems including this study. However, BMAA is known to be produced only by autotrophic prokaryotes (cyanobacteria) or eukaryotes (diatom or dinoflagellate) in aquatic environments. More research is needed to determine the biosynthesis of BMAA, BAMA and DAB in marine environments.