Induction of a neurotoxin in diatoms by iron limitation via cysteine synthase

Significance

Neurodegenerative diseases like Alzheimer’s disease (AD) and amyotrophic lateral sclerosis/parkinsonism dementia complex (ALS/PDC) represent an increasing global health crisis. Understanding the biochemical causes of these diseases thus merits high priority. The neurotoxin β-N-methylamino-L-alanine (BMAA) produced by marine diatoms can potentially cause AD and other neurodegenerative illnesses through the production of abnormal protein aggregates that are cytotoxic. Here, we elucidate the biochemical reactions that lead to the production of BMAA-containing proteins in diatoms and show that the production of BMAA-proteins is stimulated by reduced concentrations of iron (Fe) in the culture medium. Falling Fe levels in many parts of the ocean may enhance BMAA-protein production, potentially increasing the risk of neurodegenerative diseases.

Abstract

The β-N-methylamino-L-alanine (BMAA) is an emerging neurotoxin associated with human neurodegenerative diseases such as Alzheimer’s disease. Here, we report the prevalence of BMAA synthesis in protein forms by marine diatoms and reconstruct its tentative biosynthesis pathway. Remarkably, the BMAA production is strongly induced by iron limitation. Transcriptomic analyses suggest that cysteine synthase (CysK) is involved in BMAA synthesis. This is verified as CRISPR/Cas9-based CysK knockout abolished BMAA production and addition of the recombinant CysK to the mutant restored BMAA synthesis. As diatoms are the most abundant primary producers in ocean, the prevalence of BMAA in diatoms has significant public health implications. The biosynthesis pathway provides biomarkers for further investigation of BMAA production in marine diatoms and insights for understanding the pathological mechanism for human neurodegenerative diseases.

Phytoplankton are the foundation of the aquatic ecosystem and fisheries, responsible for 50% of the total atmospheric CO₂ drawdown and O₂ production. Many species produce various biotoxins, and their food chain transfer and bioaccumulation cause public health risks. Among the diverse phyla of microalgae, diatoms contribute ~20% of global CO₂ drawdown, yet some species of genera such as Pseudo-nitzschia and Nitzschia are notorious producers of domoic acid (DA) (1, 2), a phycotoxin responsible for amnesic shellfish poisoning (ASP) through the contaminated seafood consumption (3). Recently, some diatoms were reported to produce β-N-methylamino-L-alanine (BMAA) (4–8), a neurotoxin initially discovered in terrestrial plant cycads (9) and cyanobacteria (10, 11). This nonproteinogenic amino acid could selectively damage motor neurons and trigger neurodegenerative diseases such as amyotrophic lateral sclerosis/parkinsonism dementia complex (ALS/PDC) in Guam and sporadic Alzheimer’s disease (AD) in other regions (10, 12–14). The health implications of BMAA through consumption of seafood is enormous as organisms may function as an endogenous neurotoxic reservoir (12), where BMAA is bioaccumulated along food chains, enhancing its risk to human health around the world (6, 15). BMAA at environmental concentrations (ng L⁻¹) also exerts adverse influences on organisms in the food chain such as the behavior of crustacean Artemia salina and the development of zebrafish embryos (16, 17). The global number of dementia patients rapidly increases, rising from 20.2 to 43.8 million during the period of 1990-2016 (18).

The dissolved iron (Fe) plays a critical role in regulating marine primary productivity as it is required for chlorophyll biosynthesis, electron transport in chloroplasts and mitochondria, and activity of various key enzymes such as nitrate reductase and superoxide dismutase (19). However, phytoplankton growth is often limited by iron availability due to its extremely low solubility in seawater (20–22). In particular, concentrations of iron are extremely low in the surface seawater (21), as oxidizing conditions reduce iron solubility (23). Global changes further influence the iron availability as they may depress upwelling of dissolved iron from deep sources. The prevalent iron deficiency in seawater suppresses the photosynthesis and growth of phytoplankton, resulting in inadequate nitrogen utilization (24–26), creating high-nutrient low-chlorophyll regions in the global ocean. Consequently, iron limitation decreases atmospheric CO₂ drawdown and indirectly influences the global climate system (27). While nutrient limitation is widely known to influence toxin production in phytoplankton (28, 29), potentially as a defense mechanism (30), iron limitation has been reported to enhance the synthesis of domoic acid (DA) by diatoms (31, 32). Although they are different types of toxins, DA and BMAA are both amino acid derivatives and contain nitrogen atoms. This raises a question whether iron limitation also promotes BMAA production in diatoms.

Furthermore, how prevalent is BMAA production in diatoms, what conditions are conducive to its biosynthesis, and what biosynthesis pathway is responsible for BMAA synthesis are poorly understood and underexplored. We conducted a survey of diatoms for the prevalence of BMAA production in this lineage, growth condition manipulation to find inducing factors, transcriptomic analysis to identify potential metabolic pathways involved, and CRISPR/Cas9 gene knockout to verify the responsible genes. Our data reveal that BMAA production is strongly induced by iron limitation and involves cysteine synthase (CysK). This study unveils the biosynthesis pathway of protein-bound BMAA, revealing that cysteine residue and methylamine are catalyzed by cysteine synthase to produce BMAA in marine diatoms.

Results

Iron Limitation Is a Strong Inducer of Protein-Bound BMAA in Marine Diatoms.

The analytical results of BMAA in 12 diatom species showed that BMAA was detected in all of these tested strains, including three strains of Chaetoceros, seven strains of Thalassiosira, one strain of Planktoniella, and one strain of Phaeodactylum (SI Appendix, Table S1). Only protein-bound BMAA was detected in these diatoms, with no free form found (SI Appendix, Table S1). In the batch cultures of Thalassiosira minima under various concentrations of iron, when the addition of Fe was reduced to 1/3 of the control group (f/2 medium), the Fe concentration in the growth media declined faster and intracellular concentration was lower (Fig. 1). All the 12 BMAA-producing diatom species exhibited growth suppression in the batch cultures under Fe-limited condition (SI Appendix, Fig. S1), with significantly lower specific growth rates in most of the diatom species examined (SI Appendix, Fig. S2). Another Fe-stress symptom was the reduction in chlorophyll a concentration (SI Appendix, Fig. S3). However, the cellular contents of BMAA-containing proteins increased by the Fe-limited condition in eight of the 12 BMAA-producing species, by 2.26- to 7.47-folds comparing to the Fe-replete control groups (SI Appendix, Fig. S4). The most prominent promoting effect of Fe stress on BMAA production was found in T. minima and Tephritis sinica.

Fig. 1.

Growth status, variation of iron concentration in media and cells, and cell quotas of BMAA in triplicated cultures of the diatom T. minima grown under different concentrations of iron (Fe). The addition of normal level of Fe following the f/2 recipe is labeled as 1 × Fe; other treatments included the addition of threefold elevated Fe (3 × Fe), threefold decreased Fe (1/3 × Fe), and sixfold decreased Fe (1/6 × Fe). (A) Changes in diatom cell density over time. (B) Changes in iron concentrations in diatom cultures. (C) Iron concentration in the collected cells at day 14. (D) Changes in cell quotas of BMAA in different Fe-treatment groups. Shown are means and SD (arrows) of the triplicate cultures. Asterisks depict significant differences between the iron limitation group and the control group (one-way ANOVA or Welch’s t test, *0.01 < P < 0.05, **0.001 < P < 0.01, ***P < 0.001).

Due to its pronounced BMAA production response to Fe limitation, T. minima was selected for further investigation of the Fe dose–response. Four different concentrations of Fe (3 ×, 1 ×, 1/3 ×, and 1/6 × of f/2 medium concentration) were used, each in triplicate cultures. Population growth increased with Fe concentration (Fig. 1). The production of BMAA-containing proteins significantly increased when Fe decreased (1/3 × and 1/6 × Fe), but no significant changes of BMAA content in diatom under Fe “overdose” (3 × Fe) (Fig. 1). Notably, BMAA production under Fe limitation condition (1/3 × Fe) was 7.7-fold higher than that under normal condition (1 × Fe).

BMAA Production Involves a Cysteine-Based Biosynthesis Pathway.

We generated transcriptome data on T. minima cultured under 1/3 ×Fe limitation and common f/2 medium conditions (available at NCBI under the accession number PRJNA946367). De novo assembly of the combined RNA-seq data from all samples yielded a reference transcriptome (RefT) of 111,388 unigenes. Based on the RefT, a total of 26,476 differentially expressed genes (DEGs) were identified between the Fe-limited and control groups (Fig. 2), of which 12,279 and 14,197 genes were up- and down-regulated (|fold change| ≥1.5), respectively. The top 20 KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment pathways with P-adjust < 0.05 are shown in Fig. 2. Twelve KEGG pathways were mainly related to energy metabolism, protein biosynthesis, and carbon fixation functions, of which four pathways are involved in protein biosynthesis, including Proteasome (map03050; P-adjust 1.33 × 10⁻⁹) and Ribosome (map03010; P-adjust 3.49 × 10⁻⁹), Protein processing in endoplasmic reticulum (map04141; P-adjust 0.013), and SNARE interactions in vesicular transport (map04130; Padjust 0.022). The biosynthesis and accumulation of BMAA-containing proteins are potentially related to these four KEGG pathways (SI Appendix, Fig. S5). Notably, a sequence annotated as CysK was among the most dramatically upregulated genes (SI Appendix, Fig. S6). Interestingly, a positive correlation was found between the quotas of BMAA per cell and the expression levels of cysK gene in these tested BMAA-producing diatoms (Fig. 3).

Fig. 2.

Transcriptional analysis of diatoms cultured with normal f/2 medium (CK groups) and iron-limited groups (1/3 × Fe of f/2 medium) in four replicates. (A) Volcano map of DEGs in the iron limitation groups (1/3 × Fe) and CK groups (1 × Fe). (B) Cluster analysis of DEGs in eight samples of diatom in the iron-limited group (1/3 × Fe, Iron_1 to Iron_4) compared to the CK group (1 × Fe, CK_1 to CK_4). (C) Scatterplot of the top 20 KEGG pathways with P-adjust for the DEGs in diatoms from the iron-limited group (1/3 × Fe) vs. the CK group (1 × Fe). The P-adjust values of the KEGG pathways in the red dashed box are less than 0.05.

Fig. 3.

Positive correlations between BMAA production and transcriptional expression level (A) or concentration (B) of the CysK enzyme in diverse diatom strains under different culture conditions.

To prove that cysK is directly involved in BMAA biosynthesis, we conducted functional genetic and biochemical experiments. When the cysK gene was knocked out using CRISPR/Cas genome editing tool (Fig. 4A), which achieved biallelic disruption (Fig. 4B), the ability to produce BMAA in both mutant strains was abolished (Fig. 4C). A knockout control (KOC) was also created for comparison by subjecting the cells to the same gene editing procedure as for cysK knockout but omitting Cas9, BMAA production in KOC cultures was comparable to the wild type (Fig. 4C). When 0.03 mg mL⁻¹ of the CysK enzyme was externally added to both knockout strains, BMAA production was restored in both mutant strains, with a quota at 5 × 10⁻⁴ fg cell⁻¹ after incubation for 72 h, which increased to 8 × 10⁻⁴ and 7 × 10⁻⁴ fg cell⁻¹ in KO1 and KO2, respectively, after incubation for 96 h (Fig. 5).

Fig. 4.

Mutation generation of cysK and characters of mutants. (A) Schematic of cysK showing sequence and location information of knockout target. SMART software predicts one PALP domain in the protein sequence. (B) Alignment of partial cysK sequences in the CRISPR/Cas9-generated mutants showing indels compared to wild type. Highlighted dotted boxes represent the knocked out bases in both KO1 and KO2. (C) Changes in BMAA concentration in P. tricornutum showing the abolishment of BMAA production in cysK knockout mutants (n = 3).

Fig. 5.

The cysK knockout strain of P. tricornutum produces BMAA upon exogenous addition of CysK enzyme. (A) BMAA production over time with the addition of CysK enzyme (0.01 or 0.03 mg mL⁻¹) in the first cysK knockout strain of P. tricornutum (KO1). A1, A2, and A3 represents the mass chromatograms for different ion transmission m/z 119 -> 44, 88, and 102, respectively, obtained in the mutant strain KO1 spiked with 0.03 mg mL⁻¹ of CysK enzyme after incubation for 48 h, 72 h, and 96 h. (B) BMAA production over time with the addition of CysK (0.01 or 0.03 mg mL⁻¹) in the second cysK knockout strain of P. tricornutum (KO2). B1, B2, and B3 represent the mass chromatograms for different ion transmissions m/z 119 -> 44, 88, and 102, respectively, obtained in the mutant strain KO2 spiked with 0.03 mg mL⁻¹ of CysK enzyme after incubation for 48 h, 72 h and 96 h.

In another biochemical experiment, the recombinant CysK enzyme expressed and purified from Escherichia coli (SI Appendix, Text Materials and Methods 1.10) was used to induce non-BMAA-producing diatoms, Planktoniella blanda, Thalassiosira nordenskioeldii, and Thalassiosira delicatula collected from the Chinese coast, to produce BMAA. We detected BMAA in the CysK-amended cultures of P. blanda after addition of 0.1 mM methylamine and 0.02 mg mL⁻¹ CysK enzyme at 30 min, with a BMAA cellular content of approximately 8 × 10⁻⁴ fg cell⁻¹ after 60 min. T. nordenskioeldii also showed BMAA production, detectable at 30 min with a gradual increase at 60 and 120 min, reaching about 1.5 × 10⁻³ fg cell⁻¹ after 180 min.

Discussion

This study expands the list of diatom taxa that produce public health-relevant toxins. As diatoms account for about 40% of marine primary productivity (20% of global primary productivity) and are globally distributed, our finding that all of the 12 diatom species surveyed in this study produce BMAA, along with that in other species previously reported (6, 33), implicates a broad potential health risk of BMAA. The BMAA-producing diatoms such as T. minima, Thalassiosira gravida, Chaetoceros hirtisetus, and Phaeodactylum tricornutum, are often dominant species in the coastal environments (6, 8), where fishing and shellfish harvest or aquaculture take place. In contrast with cyanobacteria, which produce BMAA both in free and protein-bound forms (11), the majority of BMAA is protein-bound, which agrees with the recent finding in P. tricornutum, Thalassiosira spp., Chaetoceros spp., Pseudo-nitzschia spp., and Navicula pelliculosa (5, 6, 33). Our recent study also showed that BMAA-containing proteins are mainly located in the Golgi apparatus and endoplasmic reticulum of diatom cells (34).

The inducibility of BMAA production by iron limitation is striking and carries potentially significant health risks for marine organisms and humans who consume seafood. When exposed to BMAA at environmental concentrations (ng L⁻¹), the crustacean Artemia salina and the zebrafish embryos exhibited abnormal behavior and development, respectively (16, 17). Increasing production of BMAA in diatoms under iron limitation can aggravate human health risks through accumulation and biomagnification along food chains (10, 12–15). Meanwhile, Pseudo-nitzschia multiseries and other Pseudo-nitzschia species produce protein-bound BMAA (6) as well as DA, which increased under iron limitation (31, 32). If iron limitation promotes BMAA production in these species as diatoms examined in the present study, bloom outbreaks of Pseudo-nitzschia in iron-limited waters would exacerbate the dual risk. Iron limitation occurs not only in oceanic areas but also in coastal waters (35). The ecological and public health implications of toxigenic Pseudo-nitzschia spp. blooms should be considered in future studies, particularly in Fe-limited regions for both BMAA and DA production.

Understanding the biosynthesis pathway of BMAA is crucial for future research to unravel evolutionary trajectory and metabolic costs of BMAA production as well as for the development of molecular markers for monitoring and managing BMAA outbreaks. For plants, Brenner et al. (36) hypothesized a simple two-step pathway for the biosynthesis of BMAA in cycads, starting with a β-substituted alanine involving a cysteine synthase-like enzyme and a methyltransferase. Based on this hypothesis and research on cyanobacterial BMAA (37), we posit that BMAA biosynthesis can occur in several potential pathways, involving phosphoserine, cysteine, or O-acetylserine and CysK to substitute the beta-positioned functional group with methylamine (SI Appendix, Fig. S7). If phosphoserine is the precursor, serine synthase would be involved. Our transcriptomic data showed no significant upregulation of serine synthase under Fe limitation, when BMAA remarkably increased compared to Fe-replete conditions. Two alternative pathways are possible, either by cysteine modification to BMAA by replacing the sulfhydryl with methylamine, or by O-acetylserine conversion to BMAA by replacing the O-acetyl group with methylamine. The former would involve the biosynthesis of cysteine by CysK acting on O-acetylserine, whereas the latter would rely on a CysK-like enzyme that can transfer methylamine instead of O-acetyl group.

Our experimental results unequivocally demonstrate the essential role of the CysK enzyme in BMAA biosynthesis in diatoms. First, the positive correlation between the cell quotas of BMAA and the expression levels of cysK in the tested BMAA-producing diatoms (Fig. 3) provides supportive evidence. Consistent with this, the expression of cysK in the BMAA-producing wild strains of T. minima is higher than that in a BMAA-absent natural mutant strain (38). More importantly, when the CysK synthesis was inhibited by cysK gene knockout, BMAA production was abolished. Conversely, when heterologously synthesized CysK enzyme was added to the cysK-disrupted mutants, their ability to produce BMAA was restored. Furthermore, the provision of this enzyme induced the production of BMAA in the natural non-BMAA-producing diatoms P. blanda and T. nordenskioeldii. The exception observed in T. delicatula (SI Appendix, Table S2) is likely due to the failure of the enzyme to enter its cells. The BMAA-inducing effect of CysK showed dosage dependency, further bolstering the enzyme’s crucial role. Consistent with this role, the expression of the cysK gene was significantly upregulated in diatoms under Fe limitation, which also promoted BMAA production (SI Appendix, Fig. S6). Based on the Ocean Gene Atlas data, in sampled global oceans, the mRNA abundance of cysK homologs derived from Bacillariophyta or Dinophyceae shows a negative correlation with iron concentrations in surface seawater (SI Appendix, Fig. S8), further supporting that the expression of CysK in natural assemblages of diatoms and dinoflagellates is promoted by Fe limitation in seawater.

Another part of the BMAA biosynthesis pathway involves acquiring the methylamino group in BMAA, which could be derived from either the addition of methylamine or serial addition of an amine followed by methylation. If protein-bound BMAA is produced through modification of cysteine residue in proteins or peptides, methylamine addition might be the most parsimonious mechanism. In addition, the absence of free-form BMAA in all monocultures of experimental diatoms in laboratory conditions also suggests that the synthesis of BMAA is more likely initiated from cysteine residues in proteins than free cysteine molecules. Methylamine is one of the common low-molecular-weight organic nitrogen compounds in the ocean, as it is a metabolite produced by marine animals and bacteria (39), as well as by diatoms (40). Thus, there are both external and internal sources of methylamine available to diatoms for BMAA synthesis. Our methylamine addition experiment provides supportive evidence for this pathway. When methylamine was added to the growth medium at 0.12 mM, T. minima exhibited suppression for growth but showed a 2.1-fold increase in BMAA production (SI Appendix, Fig. S9). Based on these findings, we conclude that the cysteine residues and methylamine are catalyzed by cysteine synthase to produce BMAA in diatoms (Fig. 6).

Fig. 6.

The hypothesized biosynthesis pathway for protein-bound BMAA in marine diatoms. Cysteine residue and methylamine are combined under the catalysis of CysK to form BMAA structure in peptides or proteins.

Our identification of the cysteine-based biosynthesis pathway represents a pioneering effort to elucidate the mechanism of BMAA biosynthesis in marine diatoms. Elucidating that this pathway is not only significant for further understanding metabolic shifts that lead to BMAA production in response to environmental changes but is also instrumental to the development of a molecular marker for predicting toxin risks in the coastal marine ecosystems. The neurotoxin BMAA has been detected in phytoplankton, zooplankton, and mollusk samples collected from several important mariculture regions in China’s Yellow Sea and Bohai Sea (41), highlighting the need to monitor its risk within marine food chains. Most diatoms that form toxic outbreaks do not cause water discoloration, making it hard to visually detect and issue warnings about the risks of the phycotoxins. A molecular marker will be instrumental to diagnostic or forecast efforts.

Much remains to be studied about this pathway and its implications. For instance, whether all harmful algal blooms (HABs)-causing, DA-producing species of diatoms also produce BMAA remains unknown. Besides, dinoflagellates are major producers of HABs-related microalgal toxins, yet although some studies have reported BMAA production by marine dinoflagellates (42, 43), research on this topic remains limited. In addition, as cysteine plays a central role in sulfur metabolism, protein folding, heavy metal detoxification, and antioxidative response (44–46), understanding what role BMAA plays within cells and why iron limitation induces its production are critical questions that remain to be addressed.

Materials and Methods

A total of 12 strains of diatoms isolated from Chinese coastal waters were used to test the universal effects of Fe-limitation on the biosynthesis of BMAA in diatoms. The cell status and growth rates of the diatoms were recorded using flow cytometry (BD Accuri C6 plus, Biosciences, NJ). Contents of BMAA were analyzed with a Thermo Ultimate 3000 HPLC (Thermo Fisher Scientific, Bremen, Germany) coupled to an AB-Sciex Qtrap 4500 mass spectrometer (AB Sciex Pte. Ltd, Singapore). Transcriptome analysis of T. minima cultured under Fe limitation was performed by Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd (Shanghai, China). Prokaryotic expression and purification of CysK protein were carried out by Sangon Biotech (Shanghai, China). The purified CysK protein was used in in vitro biochemical experiments to verify its catalytic function in protein-bound BMAA production involving cysteine residues and methylamine. To generate cysK knockout mutants (cysK-KO1 or KO2) and corresponding controls (KOC) in the diatom P. tricornutum, the CRISPR/Cas9 genome-editing system was employed in this study (47–49). The detailed information for materials and methods is available in SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

The transcriptome raw data have been deposited at NCBI under the accession number PRJNA946367 (50). All other data are included in the article and/or SI Appendix.